/

Division of Fithe^

y. •. NaflcMl mud

\

BULLETIN

OF THE

UNITED STATES FISH COMMISSION.

VOL. XXII,

FOR

19 0 2.

GEORGE M. BOWERS, Commissioner.

WASHINGTON:

GOVERNMENT PRINTING OFFICE,

19 0 4.

mtmmm ristiy ution

MA m

CONTENTS.

Beardsley, A. E. The Destruction of Trout Fry by Hydra. (Issued March 30, 1903)

Clark, Hubert Lyman. The Echinoderms of the Woods Hole Region. (Issued J line 25, 1904) .

Dean, Bashford. Japanese Oyster Culture.. (Issued February 17, 1903)

Eigenmann, Carl H. The Fresh-water Fishes of Western Cuba. (Issued July 20, 1903)

Evermann, Barton Warren, and Jordan, David Starr. Descriptions of New Genera and

Species of Fishes from the Hawaiian Islands. (Issued April 11, 1903) .<

— . Descriptions of a New Genus and Two New Species of Fishes from the Hawaiian

r Islands. (Issued July 9, 1903)

Harron, L. G.,, and Smith, Hugh M. Breeding Habits of the Yellow Cat-fish. (Issued March

''“3071903-)-/.

Herrick, C. Judson. The Organ and Sense of Taste in Fishes. (Issued July 20, 1903)

Jenkins, Oliver P. Report on Collections of Fishes made in the Hawaiian Islands, with

Descriptions of New Species. (Issued July 23, 1903)

Jennings, H. S. Rotatoria of the United States. II. A Monograph of the Rattulidse. (Issued

July 23, 1903).

Jordan, David Starr. Notes on Fishes collected in the Tortugas Archipelago. (Issued

January 19, 1904)

, and Evermann, Barton Warren. Descriptions of New Genera and Species of Fishes

from the Hawaiian Islands. (Issued April 11, 1903)

. Descriptions of a New Genus and Two New Species of Fishes from the

Hawaiian Islands. (Issued July 9, 1903) -

, and Starks, Edwin Chapin. List of Fishes dredged by the Steamer Albatross off the

Coast of Japan in the Summer of 1900, with Descriptions of New Species and a Review of

the Japanese Macrouridse. (Issued August 13, 1904)

Kendall, William Converse. Habits of Some of the Commercial Cat- fishes. (Issued August 9,

1903)

. Notes on Some Fresh-water Fishes from Maine. (Issued August 14, 1903)

Lydell, Dwight. The Habits and Culture of the Black Bass. (Issued February 26, 1903)

Marsh, M. C. A More Complete Description of Bacterium Truttas. (Issued September 11,

1903)

Moenkhaus, William J. Description of a New Species of Darter from Tippecanoe Lake.

(Issued August 8, 1903)

Parker, G. H. Hearing and Allied Senses in Fishes. (Issued February 26, 1903)

Rutter, Cloudsley. ^^amigmw^he ^^at^almon^ jlss^d^IjrcJj 3£ 19031„. ...
Smith, Hugh M. Observations on the Herring Fisheries of England, Scotland, and Holland.

(Issued February 4, 1903) :

, and Harron, L. G. Breeding Habits of the Yellow Cat-fish. (Issued March 30, 1903).

Snow, Julia W. The Plankton Algae of Lake Erie, with Special Reference to the Chlo-

rophycete. (Issued August 4, 1903)

Snyder, John Otterbein. A Catalogue of the Shore Fishes collected by the Steamer Albatross

about the Hawaiian Islands in 1902. (Issued January 19, 1904)

Starks, Edwin Chapin, and Jordan, David Starr. List of Fishes dredged by the Steamer
Albatross off the coast of Japan in the Summer of 1900, with Descriptions of New Species
and a Review of the Japanese Macrouridse. (Issued August 13, 1904)

Page.

155-160

545-576

13-37

211-236

161-208

209-210

149-154

237-272

417-511

273-352

539-544

161-208

209-210

577-630

399-409

353-368

39-44

411-415

395-398

45-64

65-141

1-16

149-154

369-394

513-538

577-630

HI

ILLUSTRATIONS

Herring Fisheries of England, Scotland, and Holland: Facing page

Plate 1. Views of the herring industry of Yarmouth, England . 1

2. Herring vessels and herring-packing establishments at Vlaardingen, Holland 16

Japanese Oyster Culture:

Plate 3. Oyster park at Tanna. General view showing bamboo collectors arranged in parallel lines 24

4. Oyster park near Nihojima. Typical living ground, or ike-ba, with hedge of shibi 28

5. Oyster park near Nihojima. General view showing newly arranged toya 32

6. Bamboo oyster collectors, or shibi, after having been in use about one month, A; six months, B;

eighteen months, C. Detached oysters shown at D 36

7. Map of the oyster and seaweed concessions in one of the estuaries of Nihojima to show how com-

pletely the cultural area is developed 38

Habits and Culture of the Black Bass:

Plate 8. Black bass pond, fry retainer, and bass bed 40

Hearing and Allied Senses in Fishes:

Plate 9. (1) Side view of aquarium, showing sounding apparatus at right-hand end and suspended glass cage
in which the fish were confined. (2) End view of aquarium showing sounding apparatus. (3)
Dorsal view of brain of Fundulus heteroclitus, dissected to show positions of the roots of the fifth
and seventh nerves (V), the roots of ninth and tenth nerves (X), and the internal ear as indicated
by its otolith. (4) Dorsal view of head of Fundulus heteroclitus, to show region where the following
nerves were cut: The fifth and seventh (V), the eighth (VIII), and the lateral-line nerve (X). (5)

Side view of Fundulus heteroclitus, showing the region where the lateral-line nerve was cut (X) 48

Natural History of the Quinnat Salmon:

Plate 10. (1) Lower McCloud River Falls, which prevent the ascent of salmon. (2) Sacramento River at Duns-

muir. (3) Pit River Falls 67

11. (1) Sacramento River near Sims, ‘'Pool B," referred to on page 102. (2) Sacramento River in the

vicinity of Princeton 72

12. Diagram showing number and size of young salmon taken at Walnut Grove, 1899 94

13. (1) Adult and grilse forms of male salmon, with genital organs mature. (2) Lamprey scar on opercle

of salmon. (3) Died from gill parasites, last of summer run, September, 1900 120

14. Diagram showing the passage of two runs of salmon from Vallejo to Sacramento, 1898 124

15. ( A ) Female that had spawned all but about 500 ova, showing that the injuries are received while
spawning the last few ova or after all have been spawned. (B) Male, apparently exhausted from long
residence in fresh water, but not from being on spawning beds; typical condition of late summer-
run males, Battle Creek, September 15, 1900. (C) Female, with ova but half developed, Battle

Creek, September 15, 1900; died from long residence in fresh water. (D, E) Two males, grilse and

adult, showing extreme cases of fungous growth, October 22, 1900 138

16. Tails of female salmon from spawning beds. (A, B) Tails of spa wned-out salmon. (C) Tail of branded

specimen No. 91, the tail being perfect eight days before photograph was taken. (D) Tail of salmon
with about 500 ova yet remaining 140

17. Sacramento River between Redding and Tehama. Spawning beds of fall salmon indicated by clusters

of dots 142

18. Chart of Sacramento River. Observation stations indicated thus, O 142

Fresh Water Fishes of Western Cuba:

19. (1) San Juan River, looking upstream from the first bend above the United Habana Railroad bridge.

(2) San Juan River from above the United Habana Railroad bridge 213

20. (1) Rio San Diego at Paso Real, looking upstream from below Western Railroad bridge. (2) Rio del

Pinar, looking upstream from bridge 216

21. (1) Stygicola dentatus (Poey). (2) Lucifuga subterraneus Poey. (2) Lucifuga, a blind fish contain-

ing unborn young with well-developed eyes 236

Rotatoria of the United States:

Plate I. (1-6) Diurella tigrisMiiller. (7-10) Diurella tenuior Gosse. (11-14) Diurella weberi, new species 352

II. (15-18) Diurella insignis Herrick. (19-23) Diurella porcellus Gosse. (24-26) Diurella sulcata Jennings.. 352

III. (27-31) Diurella stylata Eyferth. (32-34) Diurella brachyura Gosse. (35, 36) Diurella cavia Gosse 352

IV. (37-39) Diurella rousseleti Voigt. (40-44) Diurella dixon-nuttalli, new species 352

V. (45-49) Rattulus gracilis Tessin. (50-52) Rattulus scipio Gosse. (53, 64) Rattulus macerus Gosse 352

VI. (55-58) Rattulus multicrinis Kellicott. (59-61) Rattulus capucinus Wierz & Zach 352

VII. (62-64) Rattulus cylindricus Imhof. (65, 66) Rattulus latus Jennings 352

IV

ILLUSTRATIONS

y

Rotatoria op the United States — Continued. Facing page

Plate VIII. (67-72) Rattulus longiseta Schrank. (73-76) Rattulus bicuspes Pell 352

IX. (77-80) Rattulus bicristatus Gosse. (81-85) Rattulus pusillus Lauterborn 352

X. (86-91) Rattulus mucosus Stokes. (92-94) Rattulus stylatus Gosse 352

XI. (95-97) Rattulus carinatus Lamarck. (98, 99) Rattulus lophoessus Gosse. (100,101) Rattulus rattus

Muller 352

XII. (102-107) Rattulus elongatus Gosse 352

XIII. (108-110) Diurella intermedia Stenroos. (Ill, 112) Rattulus scipio Gosse. (113) Diurella sulcata

Jennings. (114, 115) Diurella brachyura Gosse. (116, 117) Diurella weberi, new species. (118,

119) Diurella sulcata Jennings : 352

XIV. (120,121) Diurella sejunctipes Gosse. (122) Diurella helminthodes Gosse. (123-126) Diurella marina

Daday. (127) Diurella collaris Rousselet. (128) Diurella brevidactyla Daday. (129) Rattulus
curvatus Levander. (130) Rattulus braehydactylus Glasscott. (131, 132) “Rattulus lunaris"

Ehrenberg. (133) Rattulus dubius Lauterborn. (134) Distemma setigerum Ehrenberg 352

XV. (135) Rattulus unidens Stenroos. (136) Rattulus cuspidatus Stenroos. (137) Rattulus roseus Sten-
roos. (138) “ Rattulus cimolius” Gosse. (139) “ Rattulus calyptus ” Gosse. (140-143) Elosa wor-
rallii Lord. (144) “Ccelopus (?) minutus” Gosse. (145) Bothriocerca affinis Eichwald. (146)

Bothriocerca longicauda Daday 352

Plankton Algie of Lake Erie:

Plate I. (I) Clilamydomonas gracilis Snow. (II) Chlamydomonas communis Snow. (Ill) Chlamydomonas

globosaSnow. (IV) Scenedesmus bijugatus var. flexuosus Lemm. ( V) Staurogenia apiculata 394

II. (VI) Fusola yiridis Snow. (VII) Oocystis borgei. (VIII) Chodatella citriformis Snow. (IX) Pleuro-

coccus regularis Artari 394

III. (X) Pleurococcus aquaticus Snow. (XI) Chlorococcum natans Snow. (XII) Botrydiopsis eriensis

Snow. (XIII) Botrydiopsis oleacea Snow 394

IV. (XIV) Chlorosphaera lacustris Snow. (XV) Chlorosphaera parvula Snow. (XVI) Mesocarpus spec.

(XVII) Ccelosphserium roseum Snow. (XVIII) Chroococcus purpureus Snow 394

A More Complete Description of Bacterium truttas:

Plate I. Bacterium truttae Marsh 411

II. Bacterium truttae Marsh. Pigment produced in agar cultures 415

Fishes of Hawaiian Islands:

Plate I. (1) Dasyatis hawaiiensis Jenkins, new species. (2) Dasyatis sciera Jenkins, new species 420

II. Gymnothorax thalassopterus Jenkins, new species 426

III. Cypsilurus atrisignis Jenkins, new species 436

IV. Deeapterus canonoides Jenkins, new species 442

Shore Fishes of Hawaiian Islands:

Plate I. (1) Carcharias insularum Snyder, new species. (2) Carcharias nesiotes Snyder, new species 538

II. (3) Veternio verrens Snyder, new species. (4) Sphagebranchus flavicaudus Snyder, new species 538

III. (5) Callechelys luteus Snyder, new species. (6) Aphthalmichthys hawaiiensis Snyder, new species. . . 538

IV. (7) Gymnothorax nuttingi Snyder, new species. (8) Gymnothorax berndti Snyder, new species 538

V. (9) Gymnothorax mucifer Snyder, new species. (10) Gymnothorax xanthostomus Snyder, new

species 538

VI. (11) Gymnothorax waialuae Snyder, new species. (12) Uropterygius leucurus Snyder, new species 538

VII. (13) Exonautes gilberti Snyder, new species 538

VIII. (14) Carangus cheilio Snyder, new species. (15) Carangoides ajax Snyder, new species 538

IX. (16) Collybus drachme Snyder, new species. (17) Apogon erythrinus Snyder, new species 538

X. (18) Cirrhilabrus jordani Snyder, new species. (19) Hemipteronotus jenkinsi Snyder, new species . . . 538

XI. (20) Chaetodon corallicola Snyder, new species. (21) Holacanthus fisheri Snyder, new species 538

XII. (22) Stephanolepis pricei Snyder, new species 538

XIII. (23) Antennarius nexilis Snyder, new species. (24) Antennarius duescus Snyder, new species 538

Fishes Collected in the Tortugas:

Plate I. (1) Ctenogobius tortugse Jordan, new species. (2) Gnatholepis thompsoni Jordan, new species 544

II. (2) Elacatinus oceanops Jordan, new species. (3) Ericteis kalisherae Jordan, new species 544

Echinoderms of Woods Hole Region:

Plate 1. (1,2) Asterias forbesi (Desor). (3,4) Asterias vulgaris Verrill 552

2. (5-7) Asterias tenera Stimpson. (8, 9) Asterias austera Verrill 554

3. (10, 11) Cribrella sanguinolenta (O. F. Muller). (12, 13) Solaster endeca (Retzius) 556

4. (14,15) Asterias forbesi (Desor). (16,17) Asterias vulgaris Verrill. (18,19) Asterias austera Verrill.

(20,21) Asterias tenera Stimpson. (22) Cribrella sanguinolenta (O. F. Muller). (23) Solaster endeca
(Retzius) 556

5. (24-27) Ophiopholis aculeata (Linnaeus). (28-30) Ophiura brevispina Say 558

6. (31,32) Ophioglypha robusta (Ayres). (33,34) Amphipholis squamata (Delle Chiaje). (35,36) Gorgo-

nocephalus agassizii (Stimpson) 56o

7. (37, 38) Ophiura brevispina Say. (39,40) Ophioglypha robusta (Ayres). (41, 42) Ophiopholis aculeata

(Linnaeus). (43,44) Amphipholis squamata (Delle Chiaje). (45-47) Gorgonocephalus agassizii

(Stimpson) 560

(48-52) Arbacia punctulata (Lamarck) 553

9. (53-57) Strongylocentrotus drobachiensis (O. F. Muller) 564

10. (58-62) Echinarachnius parma (Lamarck). (63, 64) Meilita pentapora (Gmelin) 564

VI BULLETIN OF THE UNITED STATES FISH COMMISSION.

Echinoderms of Woods Hoi.e Region — Continued. Facing page

Plate 11. (65, 66) Cucumaria frondosa (Gunnerus). (67) Thyone briareus (Lesueur). (68, 69) Thyone unisemita
(Stimpson). (70) Cucumaria pulcherrima (Ayres). (71) Thyone scabra Verrill. (72) Trochostoma
ooliticum (Pourtales). (73) Caudina arenata (Gould). (74) Synapta inheerens (O. F. Muller). (75)

Synapta roseola (Verrill) 566

12. (76-80) Cucumaria frondosa (Gunnerus). (81-85) Cucumaria pulcherrima (Ayres). (86-90) Thyone

unisemita (Stimpson) 566

13. (91-94) Thyone scabra Verrill. (95-102) Thyone briareus (Lesueur). (103,104) Caudina arenata

(Gould). (105-108) Trochostoma ooliticum (Pourtalds) 568

14. (109-112) Synapta inhaerens (O.T. Muller). (113-116) Synapta roseola (Verrill) 571

Japanese Fishes Collected by the Albatross:

Plate 1. (1) Chlorophthalmus albatrossis Jordan & Starks, new species. (2) Chauliodus emmelas Jordan

& Starks, new species 630

2. (1, 2) Neoscopelus alcocki Jordan & Starks, new species. Polyipnus stereope Jordan & Starks, new

species 630

3. (1, 2) Peristedion amiscus Jordan & Starks, new species. (3) Watasea si vicola Jordan & Snyder 630

4. (1) Melanobranchus antrodeg Jordan & Gilbert, new species. (2) Nezumia condylura Jordan &

Gilbert, new species 630

5. (1) Atheresthes evermanni Jordan & Starks, new species. (2) Alaeops plinthus Jordan & Starks, new

species 630

6. (1) Dexistes rikuzenius Jordan & Starks, new species. (2) Araias ariommus Jordan & Starks, new

species 630

7. (1) Versequa achne Jordan & Starks, new species. (2) Microstomus kitaharae Jordan & Starks, new

species 630

8. (1) Engyprosopon iijimse Jordan & Starks, new species. (2) Scaeops grandisquama (Schlegel) 630

TEXT CUTS.

Japanese Oyster Culture: Page.

Map of region of oyster culture on north shore of inland sea near Hiroshima (Sea of Aki) 20

Fisherwoman opening oysters 21

Hand pick for making sockets in gravelly bottom 22

Bamboo collectors (shibi) forming boundary hedge 22

Arrangement of branched collectors 23

Diagrams of oyster farms 24, 25, 26, 29, 30, 34, 35

Oyster hook 27

Oyster rakes 27

Mitten used to hold oyster-bearing shibi while separating oysters 28

Basket for collecting and storing marketable oysters 28

Bamboo collectors as arranged in Kusatsu 31

Bamboo collectors as arranged in Kaida Bay 32

Ground plan of a mound toya of collectors 33

Natural History of the Quinnat Salmon:

Stomach, pyloric appendages and part of the intestine 126, 127, 128

Heads of salmon, showing changes in fresh water •. 130, 131

Genital organs of hermaphrodite salmon 132

Spawned-out female 138

Fishes from Northeastern California:

Pantosteus lahonton Rutter, new species 148

Agosia robusta Rutter, new species 148

Fresh-water Fishes of Western Cuba:

Fundulus cubensis Eigenmann, new species 223

Glaridichthys falcatus Eigenmann, new species 224

Glaridichthys torralbasi Eigenmann, new species 225

Girardinus garmani Eigenmann, new species : 226

Toxus riddlei Eigenmann, new species 227

Heterandria cubensis Eigenmann, new species 229

Atherina evermanni Eigenmann, new species 229

Eucinostomus meeki Eigenmann, new species 229

Heros tetraeanthus torralbasi Eigenmann, new subspecies 230

Heros tetraeanthus tetraeanthus Cuvier and Valenciennes 231

Heros tetraeanthus griseus Eigenmann, new subspecies 232

Heros tetraeanthus latus Eigenmann, new subspecies 233

Heros tetraeanthus cinctus Eigenmann, new subspecies 234

Heros nigricans Eigenmann, new species < 235

ILLUSTRATIONS.

VII

The Organ and Sense op Taste in Fishes: Page.

Brain of yellow cat-fish 242

Section through skin of top of head of Ameiurus melas 248'

Projection of cutaneous branches of communis root of facial nerve in Ameiurus melas 249

Rotatoria op the United States:

Dorsal views of toes in Rattulidae 284,285

Spiral path followed by Diurella tigris Muller 296

Diagram of a reaction to a stimulus in Diurella tigris Muller 297

Notes on Some Fresh-water Fishes prom Maine:

Leuciscus carletoni Kendall, new species 358

Pimephales anuli Kendall, new species 360

Coregonus lab radoricus Richardson 364

Coregonus quadrilateralis 365

Coregonus stanleyi Kendall, new species 367

Description op a New Species of Darter from Tippecanoe Lake:

Hadropterus evermanni. Moenkhaus, new species 398

Report on Collections of Fishes made in the Hawaiian Islands:

Congrelius bowersi J enkins, new species 422

Microdonophis maegregori Jenkins, new species 423

Muraena lampra Jenkins, new species 423

Muriena kauila Jenkins, new spe,cies 424

Gymnothorax leucostictus Jenkins, new species 425

Gymnothorax gracilicauda Jenkins, new species 426

Gymnothorax leucacme Jenkins, new species 1 427

Gymnothorax ercodes Jenkins, new species 428

Echidna leihala Jenkins, new species . 429

Echidna vincta Jenkins, new species 430

Echidna obscura Jenkins, new species. . . 430

Echidna psalion Jenkins, new species 431

Myripristis sealei Jenkins, new species 439

Seriola sparna Jenkins, new species 442

Carangus hippoides Jenkins, new species 443

Carangus rhabdotus Jenkins, new species 445

Carangus politus Jenkins, new species 446

Fowleria brachygrammus J enkins, new species . 448

Apogon menesemus Jenkins, new species 449

Priacanthus meeki J enkins, new species 451

Eteliscus marshi Jenkins, new species 452

Psendupeneus porphyreus Jenkins, new species 455

Chromis elaphrus Jenkins, new species 457

Calotomus cyclurus Jenkins, new species 466

Calotomus snyderi Jenkins, new species 467

Scaridea zonarcha Jenkins, new species 468

Scaridea balia Jenkins, new species 469

Teuthis leucopareius Jenkins, new species 476

Teuthis guntheri Jenkins, new species 478

Acanthurus incipiens Jenkins, new species 480

Callicanthus metoposophron Jenkins, new species 482

Tropidichtbys oahuensis Jenkins, new species 485

Tropidichthys epilamprus Jenkins, new species 486

Lactoria galeodon Jenkins, new species 488

Diodon nudifrons Jenkins, new species 488

Cirrhitoidea bimacula Jenkins, new species 490

Sebastopsis kelloggi Jenkins, new species 493

Sebastapistes corallicola Jenkins, new species 494

Sebastapistes coniorta Jenkins, new species 496

Sebastapistes galactacma Jenkins, new species 497

Dendrochirus chloreus Jenkins, new species 498

Eviota epiphanes Jenkins, new species . 501

Chlamydes laticeps Jenkins, new species 503

Gobionellus lonchotus Jenkins, new species 504

Enypnias oligolepis Jenkins, new species 504

Tripterigion atriceps Jenkins, new species 505

Salarias cypho Jenkins, new species 507

Salarias saltans Jenkins, new species 508

Salarias rutilus Jenkins, new species 509

Aspidontus brunneolus Jenkins, new species 510

VIII

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Japanese Fishes Dredged by Albatross: Page.

Myxine garmani Jordan & Snyder 577

Centroscyllium ritteri Jordan & Fowler - 578

Raja tengu Jordan & Fowler 578

Diaphus watasei Jordan & Starks, new species 581

Synaphobranchus jenkinsi Jordan & Snyder 582

Congrellus megastomus (Gunther) 582

Sphagebranchus moseri Jordan & Snyder 582

Macrorhampliosus sagifue Jordan & Starks -. 583

Hippocampus sindonis Jordan & Snyder 583

Paratrachichthys prosthemius Jordan & Fowler 584

Zen itea Jordan & Fowler 584

Apogon lineatus Schlegel 585

Antigonia rubescens (Gunther) 586

Thysanichthys crossotus Jordan & Starks 587

Lythrichthys eulabes Jordan & Starks 588

Scorpaena izensis Jordan & Starks 689

Ocosia vespa Jordan & Starks 589

Stlengis osensis Jordan & Starks .' 590

Schmidtina misakia Jordan & Starks 590

Daruma sagamia Jordan & Starks 590

Ricuzenius pinetorum Jordan & Starks 591

Pseudoblennius totomius Jordan & Starks t 591

Cottiusculus schmidti Jordan & Starks 592

Cottunculus brephocephalus Jordan & Starks 592

Crystallias rnatsushimae Jordan & Snyder 592

Peristedion orientale Schlegel 593

Lepidotrigla abyssalis Jordan & Starks, new species 595

Lepidotrigla japonica (Bleeker) 596

Suruga fundieola Jordan & Snyder 597

Chaeturiclithys seiistius Jordan & Snyder 597

Trypauchen wakse Jordan & Snyder 597

Callionymus flagris Jordan <& Fowler 598

Draconetta xenica Jordan & Fowler 598

Pteropsaron evolans Jordan & Snyder 699

Osopsaron verecundus (Jordan & Snyder) 600

Eulophias tanneri Smith COO

Lycenchelys pcecilimon Jordan & Fowler COO

Bothrocara zesta Jordan & Fowler 601

Porogadus guntheri Jordan & Fowler 601

Gadomus colletti Jordan & Gilbert, new species 604

Regania nipponica Jordan & Gilbert 605

Coryphaenoides awse Jordan & Gilbert, new species 609

Coryphaenoides garmani Jordan & Gilbert, new species 610

Coryphaenoides misakius Jordan & Gilbert, new species 611

Hymenoeephalus striatissimus Jordan & Gilbert, new species 613

Hymenocephalus papyraceus Jordan & Gilbert, new species 614

Hymenoeephalus lethonemus Jordan & Gilbert, new species 615

Ccelorhynchus kishinouyei Jordan & Snyder 618

Ccelorhynchus anatirostris Jordan & Gilbert, new species 619

Cleisthenes pinetorum Jordan & Starks, new species 622

Lophius litulon (Jordan) 627

Malthopsis tiarella Jordan 628

i

I

i

!

Bull. U. S. F. C. 1902. (To face page 1 .)

Plate 1.

VIEWS OF THE HERRING INDUSTRY OF YARMOUTH, ENGLAND.

OBSERVATIONS ON THE HERRING FISHERIES OF ENGLAND, SCOTLAND,

AND HOLLAND.

By HUGH M. SMITH.

The herring ( Clupea Inarengus) has justly been called the “king of fishes.”
Although its importance is now relatively less than it was several centuries ago, it is
to-day a leading fish in the United States, Canada, Newfoundland, England, Scotland,
Ireland, Holland, France, Norway, Sweden, and Russia. A species very similar to
that of the Atlantic Ocean is found in the North Pacific Ocean, and is caught in
large cpiantities in Japan and Alaska. In point of number of individual fish taken
for market, no species exceeds the herring. The annual value of the herring fish-
eries is about 125,250,000, representing 1,500,000,000 pounds of fish.

In 1900 the writer visited the principal herring-fishing centers of England, Scot-
land, and Holland. The following notes, based on the observations then made and
the information there collected, are presented chiefly because of the large consump-
tion of European herrings in the United States and because of the desirability of
applying the foreign fishing and preserving methods to the herring industry on our
east and west coasts. No attempt is made herein to furnish a complete account of
the herring fisheries of the countries mentioned.

The writer was very courteously assisted in his inquiries by the following persons,
to whom special acknowledgments are due: Mr. Charles E. Fryer, London, one of
the three Government inspectors having jurisdiction over the fisheries of England
and Wales; Mr. W. C. Robertson, Edinburgh, secretary of the fishery board for
Scotland; Mr. J. R. Nutman, of Great Yarmouth, England; Mr. James Ingram, Gov-
ernment fishery officer, Aberdeen, Scotland; Mr. E. A. Man, United States consular
agent at Schiedam, Holland, and Messrs. C. Van der Burg & Son, Ylaardingen,
Holland.

Although the capture of herring is already one of the leading fisheries of the
United States, the writer believes that the industry may be increased and the trade
made more profitable by the adoption of foreign methods with a view (1) to supply
from local fisheries the very large quantities of pickled herring now imported from
Europe and Canada, and (2) to open a large trade with southern Europe and other
regions.

The following letter from Mr. F. F. Dirnick, secretary of the Boston Fish
Bureau, dated April 7, 1900, is pertinent to this subject:

The herring imported from Norway, Holland, and Scotland are of a different quality from
the herring found on our coast. They are fatter, and great care is taken of the fish when caught
and in packing them. The herring caught on our coast of the same size are not so fat. Our
fishermen generally find a good demand for their herring at from $1.50 to $3.50 per barrel fresh for

1

F. C. B. 1902— X

2

BULLETIN OF THE UNITED STATES FISH COMMISSION.

bait for the cod and haddock fisherman. The packers generally receive from $3 to $4.50 per barrel
for United States shore herring, and there is generally not enough to supply the demand. The
foreign herrings are consumed principally by foreigners, and sell at from $8 to $14 per barrel.

The quantities of pickled herring imported into the United States in 1900 from
the countries stated were as follows:

Countries.

Smoked.

Pickled.

Pounds.

Value.

Pounds.

Value.

United Kingdom

299,323

69,123

§12,043
3, 661

8, 960, 272
12,191,397

§375, 586
674, 665

Holland

THE ENGLISH AND SCOTCH HERRING FISHERIES.

GENERAL NOTES.

Grimsby and Yarmouth, the important English fishing centers on the shore of
the North Sea, are extensively engaged in the herring fishery. At Grimsby the
herring fishery is overshadowed by the beam-trawl fishery for bottom fish, but at
Yarmouth the herring fishery predominates. The great herring markets of Scotland
are Aberdeen, Fraserborough, and Peterhead, also on the North Sea. While these
notes are based primarily on visits to Yarmouth and Aberdeen, supplementary
information was gathered in London, Edinburgh, and other places. Although the
same general methods are adopted in the herring industry in different parts of
each country, it must be remembered that the accompanying notes are especially
applicable to the particular places mentioned.

The principal fishing season at Yarmouth is from some time in October to the
last of December. A few fish are caught earlier in special nets; these are small,
well-flavored fish known as “longshore herring,” and are for local consumption.
There is also a spring fishery, involving most of the month of March and lasting five
or six weeks. The fish then caught are small and poor, and are used for bait in the
line fisheries.

The Yarmouth herring fishing is carried on with sailing vessels called luggers,
and also by steamers. The use of the latter is increasing, 50 to 70 being operated in
1900. The crew of each kind of vessel consists of 11 men. Each vessel carries 200
gill nets, 30 yards long, 20 yards (or 260 meshes) deep, and costing £2 apiece. The
number of meshes to a yard is from 28 to 31, the average size of mesh, bar measure,
being equal to a shilling. The nets are cotton, machine-made, and, with proper
care, may last seven years. They are first tanned with a solution of hot “kutch”
or catechu from Burma, which is said to be better for this purpose than tan bark.
After drying they are thoroughly soaked in linseed oil, and again dried by spreading-
on the ground. Oiling is done only once, but soaking in the astringent solution is
repeated from time to time.

The entire complement of nets is shot at one time, the nets being tied together.
A duplicate set of nets is held in reserve. The water is thick for 15 or 20 miles off
Yarmouth, and fishing may be done at any hour of the day or night, but the best
times are about sunrise and sunset. The herring do not remain long in the nets
unless storms prevent hauling. Some vessels run fish in fresh, others dry-salt their

HERRING FISHERIES OF ENGLAND, SCOTLAND, AND HOLLAND.

3

catch at sea, and store it in the hold in compartments. A vessel may leave port,
set nets, make a catch, and be back the same day, or it may be out two weeks.

There has never been a failure of the Yarmouth herring fishery; although fish
are less abundant some years, they have never been so scarce as to make fishing
unprofitable. Some of the Yarmouth fishermen think that the herring frequenting
that part of the English coast constitute a distinct body which spawn and remain off
that coast, and do not come down from the North. Herring are taken much earlier
in Scotland and northern England than here.

When a vessel arrives in port the fish are lifted out cf the hold in baskets and
spread on deck, where they are counted into baskets by hand, 100 fish to a basket.
These baskets are then passed over the rail to the dock and emptied into large,
peculiarly shaped baskets holding 500 fish, arranged on the dock in lines or tiers of
20 baskets each. The fish are heaped in 10 piles over the edges of adjoining baskets
to facilitate counting. A line of the large baskets constitutes a last, which is the
unit of measure in the herring trade. A last represents about 1-J tons of herring,
or, theoretically, 10,000 fish; but, as a matter of fact, 13,200 fish of any size, as 132
fish are called 100 in counting.

Herring are sold at public auction by lasts. The buyer puts his card or tag on
the first basket of the tier, and his drayman comes shortly afterwards and takes the
fish to the pickling-liouse or smoke-house. Sometimes, at the height of the fishery,
1,000 lasts (or 3,000,000 pounds) are landed and sold daily in Yarmouth, and the
wharves present scenes of great activity and excitement.

The Scotch herring fishery is rather uncertain. In 1900 it was poor on the
Scotch coasts, except about the Shetland Islands, where there was a phenomenally
large run. A number of years ago, after expensive curing establishments had been
built in those islands, the fish disappeared, fishing had to be abandoned, and the
packers lost all they had invested.

Different races of herring are recognized as frequenting different parts of the
Scotch seaboard. Thus, according to Mr. W. C. Robertson, of the fishery board, the
best herring are those taken near Barra and Loch Fyne, on the west coast. These
are fine, fat fish, which have brought as much as £6 per barrel.

The different kinds of cured herring to which reference may be made are
ordinary pickled fish, kippered herring, bloaters, and red herring.

BLOATERS.

A favorite form of preserved herring for local consumption is the bloater. In
the United States this term has come to mean a large, lightly smoked herring, but
in Great Britain a herring of any size may be a bloater, which may be defined
as a round herring, lightly salted and lightly smoked, and intended for immediate
consumption.

The extensive trade in Yarmouth bloaters which formerly existed with London
and other cities away from the coast has to a great extent died out, owing to the
fact that the smoking is now done at the place of consumption. The fish bear
the rail shipment better before smoking than after, so that the bloater trade now
consists largely in shipping lightly salted fish to cities where there are smoke-houses.
Bloaters remain in good condition for two or three days, but are regarded as being
best when smoked and eaten the day after being caught.

4

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The essential steps in the preparation of bloater herring are as follows: Imme-

diately after being canglit the fish are dry-salted from 12 to 24 hours if fat, or only 6
hours if lean. They are then smoked for 4 to 16 hours and are ready for consumption.

Yarmouth bloaters bring a good price; sometimes as much as 17s. 6cZ. is
received for 100 fish.

RED HERRING.

A special grade of salted and smoked herring is known to the English and
Scotch trade under this name. The fisli are destined chiefly for the Italian, Grecian,
and general Mediterranean trade, but some are sold in London and other parts of
Great Britain. Some of the herring dealers handle only red herring; but, as a rule,
the preparation of red herrings is incidental to t he packing of other grades.

The fish which are destined to be made into red herrings are often those which
have been kept at sea for several days to a fortnight, and hence have become too
hard, from prolonged salting, to be made into bloaters, kippers, or regular pickled
herring. If they have been salted too long on the vessels they are spitted on sticks
and softened by steeping them in fresh water. The special peculiarities of red
herring are that they are round, are rather heavily salted, and are smoked for a long
time to give them a good rich color.

When intended for export to very warm countries red herring are salted for 30
to 48 hours in strong brine and are then smoked for a fortnight or three weeks.
For temperate or cold countries the fish are kept for a shorter time in pickle and are
smoked 10 or 11 days. Hard-wood sawdust and hard-wood sticks are considered
necessary in producing the smudge and heat required to give to these fish their
peculiar flavor.

Red herring for the Mediterranean trade are packed with their heads against
the barrel and their tails at the center, in dry-ware barrels holding 500 to 600 fish,
half barrels holding 300 to 350 fish, and kegs or third barrels holding 180 to 200 fish.
The average gross prices for these packages in recent years have been 10s., 5s. to 6s.,
and 31s. to 4s., respectively. The expenses on a barrel for freight and commissions
are about 3s. , the fish being sent by rail to Liverpool and thence by water to the
Mediterranean. They are sometimes packed in tin cases when destined for especially
warm countries, and for the London market they are packed in flat boxes holding 50
to 60 fish.

KIPPERED HERRING.

Among the various kinds of prepared herring none ranks higher than kippered
herring. The essential characteristic of kippered herring (and of all kippered
fish) is that, before being salted and smoked, they are split and eviscerated. Fish
intended to be made into kippers should be very fresh when received from the
vessels. At Yarmouth large fish are preferred for this method, while at Aberdeen
small, fat fish are preferred.

As soon as received they are split down the back from tail to head, eviscerated,
and then salted in strong brine of Liverpool salt for 15 to 60 minutes, according to
fatness. They are then spread on square sticks by means of hooks, and smoked
over allot fire of hard-wood shavings for 6 to 8 hours (Aberdeen) or 10 to 16 hours
(Yarmouth), requiring constant attention. The color imparted to the skin is either
golden or light, to suit the markets. After cooling they are packed in boxes

HERRING FISHERIES OF ENGLAND, SCOTLAND, AND HOLLAND.

5

holding 4 to 6 dozen, and are sold at good prices throughout the United Kingdom.
They have longer life than bloaters, will easily keep for three to five days, and in
cool weather a fortnight, but they should be eaten as soon as possible.

WHITE-CURED HERRING.

Under this name are officially recognized the herring brine-salted and packed in
barrels and half barrels. Such fish are more extensively prepared than all others
combined, and give to the English and Scotch herring trades the importance they
have attained.

Various grades of herring are recognized by the salt-herring trade. These
grades are sharply distinguished and are usually indicated on the outside of the
barrel by a brand. Branding is more generally practiced in Scotland than elsewhere.

The grades of salt herring in England are “ mattie,” “mat full,” “full,” and
‘ ‘ spent ” or ‘ ‘ sliotten. ” ‘ ‘ Matties ” are the smallest herring, 8 J to 9 J inches long, with
undeveloped reproductive organs; “mat fulls” are fish to 10^ inches long, with
the ovaries and spermaries left in; “fulls” are fish 10J inches or more in length,
with roe or milt; and “spent” fish are at least 1 0 r inches long, with eggs or milt
discharged.

The grades as recognized in Scotland are “mattie,” “mat full,” “full,” “large
full,” and “spent,” and several other minor grades. The word “mattie” originally
meant a maiden herring, “mattie” being the terminology of the east coast and
“matje” of the west coast. “Matje” still retains the original meaning, the herring
so designated being caught in May and June; all such fish when salted are now
sent to Russia. “Mattie,” however, represents a small herring, full of either roe
or milt, or even spent. The official requirements of the herring of the various
grades are as follows: “Mattie,” not less than 9 inches long; “mat full,” not less
than 9:[- inches long, with roe or milt well developed and clearly seen at throat;
“full,” not less than 10Jf inches long, with roe or milt; “large full,” not less than
11J inches long, with roe or milt; “spent,” not less than 10^ inches long, without
roe -or milt.

The lengths of salted herring specified under the different grades apply to the
fish after shrinking, and are measured from the end of the snout to the end of the
compressed tail fin. Special measuring sticks or gages are employed by the fishery
officers.

The continental markets require fish that are gilled and gutted but not split.
Herring are gutted through the gill cavity, the heart, liver, and reproductive organs
being left in situ , but the greater part of the stomach and intestines being removed.

Gutting is done as soon as the fish are landed, by a crew of three women, two
of whom do the gutting while the third first “rouses” the fish (i. e., stirs them by
hand in “rousing” tubs of water to remove dirt, blood, etc.) and then packs them
in barrels with the proper amount of salt. In gutting, a small knife is inserted
through the isthmus and, with the forefinger or thumb, draws out the viscera. The
roes and spermaries are always left in, as they are considered food delicacies and,
in addition, give the fish a fat or plump appearance. Sometimes the roes are so
large that in packing they rupture the abdominal wall. Although excellent fish,
they can not, in this condition, be sold as the best grade.

6

BULLETIN OF THE UNITED STATES FISH COMMISSION.

For rousing, Liverpool salt is used; but for packing, coarser Spanish salt is
employed, about 100 pounds of salt being required for each barrel of fish.

In packing herring, it is customary to pack 7 barrels with a ton of fish (2,100
pounds), there being 300 pounds in a barrel. Each barrel contains from 850 to 1,100
fish, according to size. In packing, each herring is carefully arranged in a definite
position, with the abdomen upward and with the head against the side of the barrel,
the fish in a given layer or tier being parallel. The fish in the next layer are
arranged in the same way, but their long axis is at right angles to that of the fish
in the adjoining layer. The barrels are filled with alternate layers of fish and salt,
and then headed. In packing, the fish are compressed vertically and their bellies
are flattened, giving them the appearance of being larger and rounder. Laterally
compressed fish are not in demand.

During the process of curing, the fish shrink considerably and the barrels have
to be refilled. In Scotland the law requires that the barrels rest on their side and
be refilled after 11 days. In England, where there is no law, about 8 days are
allowed to elapse. A bunghole is bored 13 inches above the bottom, the barrel is
placed on end, the head is removed, and the pickle is allowed to run off; then the
hole is closed, 2 to 5 tiers of fish of the same catch are placed on top, and the
barrel is closed, placed on its side, and the original pickle is returned through
the bunghole. No new pickle is introduced, and under no circumstances are the
fish washed in water. After branding, the barrel is ready for market. A well-
cured and well-packed barrel, after the lapse of 10 full days, should contain no more
undissolved salt than would fill a cylindrical tub 9 by 9 inches.

The prices of salt herring vary greatly, depending on the supply. The average
price of the best grades is usually about 30s., but it may drop to 20s. or rise to 40s.
In 1899 the prices in the German, cities of Stettin, Konigsburg, and Danzig, and also
in Russia, were the best ever known, “matties” bringing 24 to 34 marks per barrel,
“mat fulls” 32 to 36 marks, and “fulls” 36 or more marks. From these gross
prices, expenses amounting to about 4-? marks per barrel Avere deducted. In 1896-97,
Avhen there Avas a large catch in Scotland, the average prices of salt herring in
Germany were 13 or 14 marks for “matties,” 16 or 17 marks for “mat fulls,” and 22
or 23 marks for “fulls.”

The authorities and fishermen of Scotland fully appreciate the importance of
plainly designating on the barrels the quality of salted herring, and the fishery board
has formulated a very complete system of regulations governing branding. In view
of the benefits which have accrued to the Scotch herring fishery from the operation
of the branding regulations, and because of the importance with which the present
writer regards branding as applied to the United States herring trade, the folloAving
detailed references to the subject are made.

The official branding of barrels of salted herring is not compulsory, and only
about half the packers resort to branding, but it is generally regarded as facilitating
the sale of fish. A good judge of herring Avould be able, from personal inspection,
to buy just as good fish without the brand as Avith it; but in distant markets the
brand carries a guaranty. The fee charged by the government for affixing the
official brand, certifying to the quality of the fish, is 4 d. (10 cents) per barrel.
During the years 1898 and 1899 the fees from this source aggregated £11,500, or
about $57,500.

HERRING FISHERIES OF ENGLAND, SCOTLAND, AND HOLLAND.

7

The following are the regulations now in force governing the official branding
of “white-cured” herring in Scotland. They are presented in extenso because of
their thoroughness and the model they afford:

Fishery Board for Scotland.

Regulations for examining barrels and half barrels intended to be filled, and branding and stencil-
ing barrels and half barrels filled with cured white herrings, for the guidance of fishery officers

and the fish-curing trade.

I. Capacity and mode of construction of barrels and half barrels filled or intended to be

filled with cured white herrings.

1. Capacity: (1) Every barrel shall be capable of containing 26£ gallons imperial measure,
being equal to 32 gallons English wine measure. (2) Every half barrel shall be capable of con-
taining 131 gallons imperial measure, being equal to 16 gallons English wine measure.

2. Tightness: Every barrel and half barrel shall be perfectly tight.

3. Staves and ends:

(a) Thickness: The staves and ends of every barrel and half barrel shall, when completed,
be not less than one-half part of an inch, and not more than three-fourths parts of an inch, in
thickness throughout.

(b) Breadth: (i) The staves of every barrel and half barrel shall not exceed 6 inches in breadth
at the bulge, (n) The head end of every barrel and half barrel must contain not less than three
pieces and the bottom end not less than two pieces.

(c) Quality, etc., of staves: The staves of every barrel and half barrel shall be well seasoned
and well fired, so as to bring them to a proper round. The staves shall not be cracked, broken, or
patched, and there shall not be a double croze. The chime shall not be less than 1 inch in length.

(d) Fitting of ends in crozes: The ends of every barrel and half barrel shall fit properly in
the crozes, and shall not be turned inside out, nor bent outwards nor inwards so as to affect the
sufficiency of the barrel or half barrel.

4. Hooping: (1) Every barrel or half barrel shall be hooped in one of the three following ways,
viz: (a) Entirely with wooden Poops; ( b ) entirely with iron hoops; or (c) partly with wooden
hoops and partly with iron hoops.

(a) Entirely with wooden hoops:

Every barrel or half barrel hooped entirely with wooden hoops shall be hooped in either
of the two following ways, viz: (i) Every barrel and half barrel shall be full-bound at the bottom
end and have at least three good hoops on the upper quarter, and every barrel shall have four good
hoops and every half barrel three good hoops on the head end; the distance between the nearest
hoops on opposite sides of the bulge of every barrel shall not exceed 11 inches after the hoops have
been properly driven; the distance for half barrels-shall be in like proportion. Or, (ii) every barrel
and half barrel shall be quarter-hooped, the barrels with four good hoops on each end and three
good hoops on each quarter, and the half barrels with three good hoops on each end and three good
hoops on each quarter.

- ( b ) Entirely with iron hoops:

(1) Every barrel hooped entirely with iron hoops shall be hooped in either of the two
following ways, viz: (i) Every barrel shall be hooped with at least four hoops, one of those to be
on each end of the barrel and not to be less than 2 inches wide, of wire gage No. 16, and the other
two to be on the quarters of the barrel and not less than If inches wide, of wire gage No. 17, the
four hoops to be placed at proper relative distances on the barrel. Or, (ii) every barrel shall be
hooped with six hoops, one of these to be on each end of the barrel and not to be less than If inches
wide, of wire gage No. 16, one to be on each of the quarters and not to be less than 1 inch wide, of
wire gage No. 18, and one to be on each side of the bulge and not to be less than If inches wide,
of wire gage No. 17, the six hoops to be placed at proper relative distances on the barrel.

(2) Every half barrel hooped entirely with iron hoops shall be hooped with at least four
hoops, one of these to be on each end of the half barrel and not to be less than If inches wide, of

BULLETIN OF THE UNITED STATES FISH COMMISSION.

wire gage No. 17, and the other two to be on the quarters of the half barrel and not less than 1J
inches wide, of wire gage No. 18, the four hoops to be placed at proper relative distances on the
half barrel.

(c) Partly with wooden hoops and partly with iron hoops:

(1) Every barrel hooped partly with wooden hoops and partly with iron hoops shall have

either (a) the hoop of the head end alone, or (b) the hoops of both ends, made of iron at least 2
inches wide, of wire gage No. 16. (a) If the hoop of the head end alone be of iron, the remaining

portion of the barrel shall be bound with wooden hoops in either of the two following ways, viz,
the bottom end frill bound, with at least three good hoops on the upper quarter, or quarter hooped,
with three good hoops on each quarter and four good hoops on the bottom end. (6) If the hoops
of both ends be of iron, each of the two quarters shall be bound with at least three good wooden
hoops.

(2) Every half barrel hooped partly with wooden hoops and partly with iron hoops shall

have either (a) the hoop of the head end alone, or (b) the hoops of both ends, made of iron at least
1| inches wide, of wire gage No. 17. (a) If the hoop of the head end alone be of iron, the remain-

ing portion of the half barrel shall be bound with wooden hoops in either of the two following
ways, viz, the bottom end full bound with at least three good hoops on the upper quarter, or
quarter hooped, with three good hoops on each quarter and three good hoops on the bottom end.

( b ) If the hoops of both ends be of iron, each of the two quarters shall be bound with at least three
good wooden hoops.

II. Marks which curers are required to put , or are prohibited f rom putting , on barrels
and half barrels filled , or meant to be filled, with cured white herrings.

(a) On the outside of the bottom of every barrel and half barrel, at the time when they are
given by the curer to the packer to be packed with herrings, there shall be legibly written or ,
marked with red keel or black lead a description of the herrings to be packed, the date of their
cure, and the number of the packer; and neither chalk nor any other substance shall be allowed
to pass as a substitute for red keel or black lead, and no barrel or half barrel unmarked as here
prescribed shall be examined for branding.

(b) When any barrel or half barrel has been emptied of the herrings it contained, the old
marks on the bottom shall be obliterated, and the barrel or half barrel, at the time it is given to a
packer to be again packed with herrings, legibly marked anew, in red keel or black lead, with the
description of herrings it is intended to pack therein, the date of the cure, and the number of the
packer.

(c) The curer’s name and the name of the port or place of cure shall be branded on the side of
all barrels or half barrels presented for the crown brand, and in addition the name of the district
may be added thus: to Sandhaven may be added Fraserburgh, and to Boddam Peterhead, and
above these impressions there shall be legibly scrieved a description of the herrings contained in
the barrels or half barrels, and the date of their cure — the month of cure to be expressed by the
first letter thereof, except in the cases of January. April, May. and June, which shall be designated
by JA, AP.MA, and JE, respectively; the following being given as examples of scrieving: 12th July
1895, La. Full, L 12 J/95; Full, F 12 J/95; Mat. Full, MF 12 J/95; Spent, S 12 J/95; Mattie, E 12 J/95.
On crown-branded barrels of herrings the year need not be branded, as that is given in the scriev-
ing and also in the crown brand, which should be placed in close proximity to the curer’s name
and the name of the port or place of cure.

( d ) No descriptive mark or marks shall be placed on the ends of barrels or half barrels of
crown-branded herrings under penalty of removal of the crown brand without return of fees.

III. Heading up of barrels and half barrels after filling them up with cured white herrings.

After filling up, according as the barrel or half barrel has been opened at the head end or the
bottom end, it shall be flagged round the head or bottom, made perfectly tight to contain the pickle,
and pickled at the bungliole. The bungliole shall be bored within II inches of the foremost hoop
of the left end; and both chime and quarter hoops of each end of every barrel or half barrel shall
be properly nailed.

9

HERRING FISHERIES OF ENGLAND, SCOTLAND, AND HOLLAND.

IV. Quality , method of cure, packing, etc., of white herrings necessary to secure the official brand.

Quality: The herrings shall be of good quality.

Chitting, curing, and packing: They shall be gutted with a knife, and cured, and packed in
barrels or half barrels within twenty-four hours after being caught.

They shall be well cured and regularly salted, and all fish broken or torn in the belly shall be
excluded.

They shall be carefully laid in barrels or half barrels, each tier being completed with head
herrings, and the herrings’ in each successive tier being arranged transversely to those in the next
tier underneath, and drawn closely together, care being taken that the heads of the herrings are
kept close to the sides of the barrel or half barrel until it is completely filled.

None of them shall be laid in bulk after being cured in barrels or half barrels.

They shall, if intended for the La. Full, Full, or Mat. Full brands, be pined in salt for not less
than ten free days; and, if intended for the Spent or Mattie brands, they shall be similarly pined
for not less than eight free days, these periods to be exclusive of the day of cure and the day of
filling up for branding; and this requirement shall apply to the herrings used in filling up as well
as to those in the original packing.

Filling up:

(a) The surplus pickle shall be run well off through the bunghole, and the seastick herrings
then left in the barrel or half barrel shall be pressed down by the cooper steadily and uniformly,
by daunt or otherwise (use of daunt preferable) , thus testing the firmness of the original packing,
and whether the surplus pickle has been sufficiently poured off or not. Pickle shall alone be used
for the purpose of washing herrings offered for the crown brand.

(b) The space left in the head end of the barrel or half barrel shall then be tightly packed with
herrings carefully laid, regularly and lightly salted, the barrel or half barrel being firmly filled
with herrings round the sides, as well as in the center. The herrings shall be pressed firmly to the
sides of the barrel or half barrel with both hands, each tier being completed with head herrings,
and the herrings in each successive tier being arranged transversely to those in the next tier under-
neath, and the weight of the hands being pressed on each tier when finished, care being taken that
the heads of the herrings in every tier are kept close to the sides of the barrel or half barrel until
it is completely filled.

(c) No herrings which have lost their original pickle shall be used in filling up.

V. Conditions on which cured white herrings which have lost their original pickle may

secure the official brand.

No herrings which have lost the original pickle shall be accorded the crown brand unless they
have been repacked, washed in pickle, and presented separately for inspection, when if found
worthy in every other respect they shall, in addition to the crown brand, receive the “ Repack ” iron
across the St. Andrew’s cross on the shoulders of the crown, so that it can not be removed without
effacing the crown brand. If barrels or half barrels of repacked herrings, instead of being offei’ed
separately, are found mixed up with any parcel of bung-packed herrings presented for the brand,
the whole parcel shall be rejected.

VI. Reassortment of rejected herrings for the official brand.

When herrings once presented for branding have been rejected by an officer for bad quality,
bad cure, bad gutting, or for being mixed with overday’s fish (see penalty for presenting overday’s
fish, etc. , on the back of request note, and at the end of these regulations) , they can not be reassorted
and presented again for branding.

Herrings rejected for bad selection, or for too many undersized herrings for the standard of
the iron applied for, may be reselected and presented anew, but they must be pickled with original
pickle, when they may be crown-branded, if found otherwise satisfactory, with the “ Repack ” iron
added. The daunt must be used with all repacked herrings.

Early-caught herrings: Herrings caught on the north and east coast of Scotland and on the
coasts of Orkney and Shetland before 12th July shall not be crown-branded with the ” Mattie”
iron, while those caught on the coasts mentioned from 12th to 19th July, both days inclusive, shall
have the long gut taken out before being eligible for the “ Mattie ” brand.

10

BULLETIN OF THE UNITED STATES FISH COMMISSION

Winter-caught herrings: Winter herrings may, from the 1st November of the one year to
1st April of the succeeding year, be crown-branded, with the word “ Winter ” branded right across
the St. Andrew’s cross on the shoulders of the crown, so that it can not be removed without effacing
the crown brand.

VII. Examination of barrels and half barrels in respect of their capacity and mode of constr uction.

(a) Barrels and half barrels intended to be filled with cured-white herrings: Officers shall
examine at least four in every hundred barrels or half barrels intended to be filled with herrings,
the capacity of one being tested (if necessary) by liquid measure, and the capacity of the remain-
ing three by diagonal rod 23 inches long for barrels and 18| inches long for half barrels, measured
from the croze of the bottom end to the croze of the head end; the examination to be made at a
time or times suitable for the officers themselves.

(5) Barrels and half barrels filled with cured white herrings: Officers shall examine all barrels
and half barrels filled with herrings, and (if necessary) shall empty the herrings out of at least
one barrel or half barrel in every hundred and test its capacity by liquid measure and test the
capacity of at least three others by callipers.

VIII. Examination for branding and stenciling barrels and half barrels in respect of the
quality, method of cure, packing, etc. , of the white herrings they contain.

(a) The barrels or half barrels presented for branding shall be laid out so that the bottom ends
come at once under the eye of the branding officers.

(b) The curer, or his authorized manager at the place of cure, having delivered to the officer the
proper account of cure of the herrings presented for branding, and a request note containing the
number of barrels and half barrels to be presented, the officer shall see, first, that the number of
barrels and half barrels is correctly stated in the request note; second, that the request note is
signed by the curer or his authorized manager (as the case may be) ; and third, that the branding
conditions attached thereto are likewise signed by the curer or his authorized manager.

N. B. It shall be understood that no manager can be recognized as an authorized manager
except under authority obtained from the board upon application previously made by the curer.

(c) Brand fees (at the rate of 4 d. per barrel and 2d. per half barrel) corresponding to the
correct number of barrels and half barrels in the request note shall be deposited with the officer
before branding, subject to the condition that if the parcel be not branded the amount of brand
fees so deposited shall be returned to the curer; or, if only a portion of the parcel be rejected, the
brand fees corresponding thereto shall be returned.

(d) The curer 's declaration shall then be taken and signed by him.

(e) The minimum number of barrels to be examined per hundred shall be seven. Two barrels
per hundred or smaller parcel shall be examined down through the original, the remaining five
down to the lower quarter hoop of either end.

Officers are not restricted to this scale, but if need be shall open as many more barrels or half
barrels as they may deem requisite to satisfy them that the herrings are fit for branding, for
which they will be held responsible to the board; but they shall understand that in no case what-
ever shall fewer barrels or half barrels than what is prescribed in the above scale be opened for
examination previous to branding.

(/) The barrels or half barrels selected for examination shall, as a general rule, be opened at
the bottom and head end alternately — that is to say. No. 1 shall be opened at the head end. No. 2
shall be opened at the bottom end, and so on until the whole examination is concluded. The
herrings in all barrels or half barrels opened shall be searched down to the lower quarter hoop
of either end, two barrels per hundred or under as in note (e), and as much farther as may be
deemed necessary.

But where an officer, from any cause, sees reason to examine a larger proportion of barrels or
half barrels at the one end than at the other, he shall be at liberty to substitute the examination
of such larger proportion for the above alternate examination, only observing that not less than
the full proportions per hundred or smaller parcel which are laid down for the minimum scale of
examination shall be examined in all. and that as many more shall be examined as he may see fit.

(g) In examining a parcel the work of different packers shall be selected, as well as herrings
of different dates of cure.

HERRING FISHERIES OF ENGLAND, SCOTLAND, AND HOLLAND.

11

( h ) All objectionable herrings shall be removed, from the barrels examined before affixing
the crown brand.

( i ) When, on the first examination of herrings for branding, they are found of bad quality,
badly cured, or badly gutted, refusal of the brand shall be final and absolute. When, however,
this refusal has been entirely owing to the barrels or half barrels being too slackly packed with
herrings, or the filling up badly selected, the case shall be treated exceptionally and shall be reme-
died by filling up or reselection only of the filling up, and the herrings may thereafter be branded,
if in accordance with the following conditions:

(i) They shall be presented for renewed inspection only to the officer who previously rejected
them, who shall satisfy himself by full examination that they are, apart from slack
packing or bad selection of the filling up, worthy of the brand.

(ii) The filling up shall have been properly completed; but. failing this, the herrings shall be
finally rejected and no further examination permitted.

(ill) The officer shall retain the fees until the herrings are branded or finally rejected; in the
latter case returning the fees.

(iv) He shall state upon the request note the particulars of the first refusal; and if the her-
rings be afterwards branded, the date of branding.

When, however, a parcel is small, and upon the first inspection the deficiency in filling up is
seen to be so very trifling that it can be supplied at once in the presence of the officer without
difficulty or detention, the above conditions shall not apply, but the filling up may be done upon
the spot and the branding proceeded with immediately afterwards, the officer being careful to
satisfy himself previous to the branding that the herrings are in all other respects entitled to
be branded.

( k ) The officers shall see that the barrels opened are filled up and headed with proper care.

( l ) The officers shall put a double crown on the bilge of the barrel examined and toward the
end examined.

(m) Oversalting shall be determined by the measure known as the cog; and the quantity of
salt left in any barrel emptied of fish must not exceed this measure.

IX. Branding and stenciling barrels and half barrels in respect of the quality, method
of cure, packing, etc. , of the white herrings they contain.

Every barrel or half barrel containing white herrings presented to one of the officers for exam-
ination shall, if the capacity and mode of construction of the barrel or half barrel, and the quality,
cure, selection, packing, etc., of the herrings are, in his opinion, such as to satisfy the requirements
of these regulations, (1) have branded in his presence, by means of a hot iron, on the bilge, in close
proximity to the curer’s name and the name of the port or place of cure, a crown surrounding the
word “ Scotland,” a description of the herrings, viz: La. Full, Full, Mat. Full, Spent, or Mattie
(as the case may be) , the initial letters of the examining officer’s name and the year; and (2) have
stenciled in his presence, on the head end, a crown surrounding the same word, description, and
letters as those branded on the bilge, with the words “ Fishery Board, Crown Brand,” stenciled
below.

X. Requirements of the different brands.

In addition to what are contained in the foregoing regulations, the requirements in respect of
the different brands shall be as follows:

Crown “La. Full” Brand.

Barrels or half barrels of herrings for this brand shall contain large full herrings of not less
than 1 1 ;} inches in extreme length, as measured by the fishery officer’s gage, made for the purpose.
Rejections under this brand shall be:

(1) On original packing for spent, torn, broken herrings, or herrings of bad or indifferent
quality if more than fifteen; or, on filling up, if more than six.

(2) On original packing if the undersized amount to more than fifteen; or, on filling up, to
more than six. And the parcel shall also be rejected if it should appear that the larger herrings
suitable for this brand have been previously taken out.

12

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Crown 1 ‘ Full ' ’ Brand.

Barrels or lialf barrels of herrings for this brand shall contain full herrings of not less than
10J inches in extreme length, as measured by the fishery officer’s gage, made for the purpose.
Rejections under this brand shall be:

(1) On original packing for spent, torn, or broken herrings, or herrings of bad or indifferent
quality, if more than eighteen; or, on filling up, if more than nine.

(2) On original packing if the undersized amount to more than eighteen; or, on filling up, to
more than nine. And the parcel shall also be rejected if it should appear that the larger herrings,
suitable for this brand, have been previously taken out.

Crown “Mat. Full ’’ Brand.

Barrels or half barrels of herrings for this brand shall contain full herrings well developed —
the roe or milt being clearly seen at neck or throat without pressure — of not less than 9f inches in
extreme length, as measured by the fishery officer’s gage, made for the purpose. Rejections
under this brand shall be:

(1) On original packing for spent, torn, or broken herrings, or herrings of bad or indifferent
quality, if more than twenty-one; or, on filling up, if more than nine.

(2) On original packing if the undersized amount to more than twenty-one; or, on filling up, to
more than nine. And the parcel shall also be rejected if it should appear that the larger herrings
suitable for this brand have been previously taken out.

Crown “Spent’’ Brand.

Barrels or half barrels of herrings for this brand shall contain spent herrings of not less than
10:} inches in extreme length, as measured by the fishery officer's gage, made for the purpose.
Rejections under this brand shall be:

(1) On original packing for torn or broken herrings, or herrings of bad or indifferent quality,
if more than eighteen; or, on filling up, if more than nine.

(2) On original packing if undersized amount to more than eighteen; or. on filling up, to more
than nine. And the parcel shall also be rejected if the larger herrings have been previously
taken out.

Crown “ Mattie ” Brand.

Barrels or half barrels of herrings for this brand shall contain herrings not eligible for any of
the foregoing brands, and of not less than 9 inches in extreme length, as measured by the fishery
officer’s gage, made for the purpose, but shall not contain headless herrings. Rejections under
this brand shall be:

(1) On original packing for torn or broken herrings, or herrings of bad or indifferent quality,
if more than thirty; or, on filling up, if more than twelve.

(2) On original packing if undersized amount to more than thirty; or, on filling up, to more
than twelve.

“ Repack ’’ Brand.

For exportation out of Europe:

1. The herrings for this brand shall have been pined in salt for not less than ten days, exclusive
of the day of catch and the day of beginning to repack for branding.

2. They shall be emptied out of each barrel or half barrel in which they were originally cured,
and they shall be washed clean.

3. They shall have the crown gut, if adhering to them, removed.

4. They shall be repacked into the barrels or half barrels from which they were emptied and
into as many additional barrels or half barrels as may be necessary.

5. They shall be salted sufficiently, and be pickled with strong pickle made of clean salt.

6. Every barrel or half-barrel shall be full bound at the head end as well as at the bottom end,
and shall have at each end an iron hoop of 1 inch in breadth.

HERRING FISHERIES OF ENGLAND, SCOTLAND, AND HOLLAND.

13

“ Lozenge ” Brand.

1. This brand shall he applied to herrings previously branded which have been repacked in the
manner required for the “ repack ” brand, and the lozenge shall be stamped immediately under and
close to the crown brand already upon the barrels or half barrels.

2. Upon the additional barrels or half barrels derived from the repacking the “ repack ’’ brand
shall be affixed, subjoining thereto the lozenge brand as above.

Note. — By the strict letter of the act the curer or proprietor of the herrings ought to give
twenty-four hours’ notice, in writing, of his intention to repack for this brand, but, of
course, where the officer can accomplish his examination of the herrings sooner he
should endeavor to do so and accommodate the curer as far as he can.

Under any crown brand, if the officer is satisfied with the cure, quality, etc., of the herrings,
but considers them generally too flatly packed, in addition to the crown brand he shall cause the
lozenge brand to be affixed to cover the St. Andrew's cross on the top of the crown.

SCOTLAND,

1 —
<c

2

I

-H

m

(WB

95 J

SCOTLAND,

) LA-1

FULL/

UM

95i

FISHERY BOARD
CROWN BRAND

SCOTLAND

) SPI

ENT/

U.R

95 J

AkkUA 44

SCI

dtland;

^MAT-FULL/

(j.B

95 J

J.L 95

SCOTLAND! (SCOTLAND
FULL

\ LA-1

rULL

U.M

95 J

REPACK/! WINTER

Representations of Brands used in Scotland in marking Barrels of Cured Herring.

About, one-half actual size.

L4

BULLETIN OF THE UNITED STATES FISH COMMISSION.

THE DUTCH HERRING FISHERY.

Vlaardingen, situated on the Maas, a few miles below Rotterdam, is the center
of the Dutch herring trade. There the herring boats fit out, there they land their
catch, and there are the houses in which the fish are prepared for shipment.

The herring fishery is conducted by steam and sail vessels, which use tanned
cotton gill nets 360 meshes deep and 720 meshes long, the size of mesh being 2-inch
stretch. From 80 to 150 nets are carried by each vessel, this outfit usually costing
from 5,000 to 7,000 guilders. The nets are set about 6 feet below the surface, being
held in position by leads and by corks (5 inches long, 3 inches wide, and 2 inches
thick) at intervals of a foot, the cork line being attached to the top of the net by
numerous gangings 8 inches long. At times, when the herring are close inshore,
some surface fishing is done.

The dressing and salting of the fish immediately after the nets are hauled on
the vessels are considered of great importance by the Dutch herring fishermen, and
no doubt contribute largely to the quality of the cured fish.

The Dutch method of cleaning herring is similar to the Scotch. Provided with
a short knife, attached to the fourth and fifth fingers of the right hand by a string
tied to the handle, the fishermen take the herring in the left hand, with the belly
up and the head forward, and thrust the knife crosswise directly through the gill
cavities, entering the left side and emerging from the right. The edge of the knife
being turned upward or outward, the knife is pulled directly through the tissues,
cutting and tearing away the gills, branchiostegals, heart, oesophagus, stomach, and
often a part of the intestines; the pectoral fins, with the skin and muscle at their
base, also come away with the same movement. There is apparently little effort
made to remove anything except the gills and pectorals, the other organs coming
away incidentally. The men become very expert in cutting, and some of them can
handle 1 ,200 fish per hour (20 per minute).

The removal of the gills and heart results in opening the large blood vessels, and
free bleeding ensues; this leaves the flesh pale or white, in contrast with the dark
reddish color of the Scotch herring, in which the blood is allowed to clot. It some-
times happens in the Dutch fishery that when there is a large catch the blood has
clotted in the last fish handled. The chief and only genuine benefits of cutting as
now practiced are the bleeding and the opening of the abdominal cavity to the brine.

Some herring examined by me as brought in by the fishing vessels at Vlaardingen
contained pyloric cieca and part of the intestines, as well as the liver and reproductive
organs. The intestine, with or without the cseca, often hangs outside the wound
made with the dressing knife; it is called the “zeele” (soul), and is frequently eaten
by the packers of salt fish, being regarded a choice morsel.

The packing of herring is done w'hile the vessels are still at sea. The fish are
first rolled in salt and then carefully packed in straight rows, with backs down.
The fish in a given layer are at right angles to those of adjoining layers. One barrel
of St. Ubes salt is required to pack four barrels of herring at sea, the salt being-
disposed between the layers of fish. The barrels are headed up and stored in the
hold until the fishing trip is ended or all the barrels filled. On reaching port the
catch is unloaded and sold at auction.

HERRING FISHERIES OF ENGLAND, SCOTLAND, AND HOLL AND.

15

The buyer almost always repacks the fish in order to sort them by size and
grades of quality, no sorting being attempted on the vessels. Sometimes purchasers
or agents prefer the sea-packed, unsorted fish, but as a rule the dealers or jobbers
wish to know how many fish are in a barrel and what their quality is. Some
shrinkage ensues; this is usually made up with fish of the same lot before the sale,
the refilling being done either on shore or on the vessel.

After coming into the hands of the packer the herring are emptied into large vats
or tanks, from which they are repacked according to the prevailing practice. The
original brine (called “blood brine” or “blood pickle”) is considered much better
than any newly-made brine, and is always saved and poured back on the fish after
repacking. The fish are placed in the barrels in the same manner as at sea, and
fresh salt is added in the proportion of 1 barrel to 8 barrels of fish. The shrinkage
in repacking is about 8 per cent — that is, 100 barrels of sea-packed herring will
make about 92 barrels of fish ready for market.

The Dutch herring barrels contain about 125 kilograms of fish, and most of the
catch is marketed in such barrels. Smaller receptacles — i, i, jr, and barrels —

are also used, but are not nearly so much in demand as they were a few years
ago. The barrel staves are oak and are imported from New York, Baltimore, and
Newport News in the form of rough pieces, which are cut into proper shape at
Ylaardingen, where several thousand coopers are regularly employed. The hoops
are made from willow trees grown on the dykes. White, clean barrels are required
for the American trade; dark, dirty barrels are accepted by the continental countries.
The Scotch herring barrel is regarded as a very good, strong barrel, and is imported
by the Dutch packers. The preferred arrangement of the hoops is to have four
between the bung and each end as well as four at each end, so that when the barrel
is rolled its weight rests on the hoops. Idle bung is large and central. Some barrels
have a single iron hoop at the top.

In Holland there is no official regulation of packing or branding, but the packers
have a standard which is generally observed, as it is to their interest to have the
fish properly packed and labeled. The different grades of herring recognized are
similar to the Scotch, and are based on the spawning condition of the fish. The ripe
or full fish are branded “ vol” ( = full); the matties (maatjis) are branded “m”; the
spent herring (ijlen) are branded “u”or “ijle.” Of each of these there are several
qualities designated No. 1, No. 2, etc., and there are several other grades. The
barrels are usually marked with a stencil.

The Dutch herring trade is not restricted to the fish caught and packed by the
Dutch fishermen. Considerable quantities of salt Scotch herring are received at
Vlaardingen, which, after repacking, are sold as Dutch herring. Furthermore, the
Dutch sell some of their own herrings in Scotch barrels in the continental countries,
where the Scotch pack is well known.

16

BULLETIN OF THE UNITED STATES FISH COMMISSION.

SUGGESTIONS IN REGARD TO THE AMERICAN HERRING FISHERY.

The experience of the European herring-packers has resulted in a prepared
product, which meets with ready sale throughout the world at better prices than are
received for other cured herring. If American lierring-curers wish to supply the
home markets and to establish a profitable trade with other countries, they must
take cognizance of the demands of those markets and make their fish conform thereto.
That there is an opportunity for a large increase in both the domestic and foreign
trade in American herring there can be no doubt; and the following suggestions are
made to this end:

1. While the demand for fresh herring for bait, for smoking, and for canning
takes a large part of the catch on certain parts of the New England coast, there are
localities where the salting of herring could be made very profitable. Even in the
canning and smoking districts it may prove more remunerative to the weir fishermen
to salt their large-sized herring. It seems probable that the excellent herring of the
Pacific coast can be salted to great advantage and ought to find a read}' market.

2. Care must be exercised in all steps of the curing and packing processes.
Oidy plump fish in the best condition should be salted, and only sound fish should be
packed. Herring of different grades should not be packed in the same barrel.

3. Different standard qualities of salt herring should be recognized and con-
scientiously adhered to. The organization of the United States Government would
probably not warrant the Federal authorities in exercising jurisdiction in the matter
of inspecting and branding fish. While this jurisdiction could doubtless be acquired
(as has been done in the case of meats intended for export), it can not be regarded
as essential. Each State is competent to superintend the inspection and branding
of its own fish, to adopt special regulations and brand marks, and take such other
measures as will tend to promote the salt-herring trade. This systematic branding
under State authority is regarded as one of the most essential factors in the develop-
ment of the salt-herring trade.

4. The establishment of a large trade with southern Europe, the Philippines,
Australia, and elsewhere in salted river herrings or alewives is entirely feasible.
These fish, which are in excellent condition when they ascend the Eastern rivers in
untold millions each season, should, if properly prepared, sell almost as well as the
sea herring. An especially good opportunity for promoting the alewife fishery
appears to exist in the Middle and South Atlantic States, where the catch is only
imperfectly utilized and where labor is abundant and cheap. The river herrings
might be prepared as white-cured herring and also as red herrings.

Bull. U. S. F. C. 1 902. (To face page 16.)

Plate 2.

HERRING VESSELS AND HERRING-PACKING ESTABLISHMENTS AT VLAARDINGEN, HOLLAND,

JAPANESE OYSTER-CULTURE.

By BASHFORD DEAN,

Adjunct Professor in Zoology , Columbia University.

European oyster-culture, especially as practiced in France and Holland, is gen-
erally regarded as the most refined development of this ancient art; and from our
knowledge of this — if we admit that those cultural methods are the most perfect which
produce the greatest number of oysters in a given area — we can reasonably conclude
that some of the devices at least of the European culturist will ultimately come to
be adopted in our own system. In view, accordingly, of the prospective value of
foreign methods in the development of American oyster-culture, the United States
Commission of Fish and Fisheries has already published (in its bulletins® for 1890
and 1891) reports upon the practical workings of the best forms of European oyster
parks. From the character of the methods there in use, we can, I believe, conclude
positively that similar establishments could be operated successfully at suitable
points — e. g., in Chesapeake Bay or Long Island waters, as soon, that is, as the
demand for oysters will warrant the use of what will prove at first a more expensive
system.

While these European methods are applicable on our Atlantic coast, it still
remains to be determined whether they include the best that could be employed
along the Pacific, should artificial oyster-culture be here attempted. For in these
waters different conditions have produced oysters which differ widely from those of
the North Atlantic. The Pacific culturist may therefore feel a more lively interest
in the oysters of Japan, for not merely are they closer akin to his own, in structure
and in habitat, and therefore more readily acclimatable, but they are larger, better
shaped, and certainly of greater value, commercially speaking, than the local product,
Ostreci calif ornica. Moreover, the Japanese oysters have long been cultivated, and
with great success. Indeed, by some experts the Japanese methods have been
commended as the simplest and most practicable of all forms of ‘ ‘ ai’tificial ” oyster
culture, and thus of possible interest in somewhat broader lines.

Unfortunately there is no literature accessible dealing in detail with the culture
or living conditions of this western Pacific oyster, and it is witli the aim of filling
this gap that the present report was prepared. Its material was collected by the
writer during a stay in Japan in 1900-1901. He there acted under special instructions
from Commissioner George M. Bowers, and in aid of his inquiries Avas designated as
a biologist of the United States Commission of Fish and Fisheries.

« Report on the Present Methods of Oyster-culture in Prance. Bulletin XJ. S. Commission of Pish and Fisheries,
1890, pages 363-388, plates LXVIII-LXXVIII.

Report on the European Methods of Oyster-culture. Op. cit. for 1891, pages 357-406, plates LXXV-LXXXVIII.
(Italy, Spain, Portugal, Germany, Holland, Belgium, and England.)

17

P. C. B. 1902—2

18

BULLETIN OF THE UNITED STATES FISH COMMISSION.

For the rest, Japanese oyster-culture proved to he worthy of careful study, not
only for its merits, but because of the suggestions it affords for cultural experiments.
One may frankly doubt whether it can at once be employed profitably — for example,
at many points on the Oregon coast — in view of the expense for labor which it
entails, but I believe that there is a reasonable chance that it could be made profitable
if employed in a favorable locality. In any event, so far as the Pacific coast is
concerned, the Japanese methods are the most practicable, and experiments with
them could be made readily and at little expense, and would soon demonstrate
whether artificial oyster-culture can here be employed commercially.

The Japanese industry is largely seated along the north shore of the Inland Sea
near Hiroshima, in the gulf-like Sea of Aki, famous for its oysters. From what
period, indeed, this oyster-culture has been carried on is not known accurately, but
from its present condit ion it is evidently the product of centuries. As early as 1708
records show that concessions at Osaka were granted to an oyste r company or to
oyster companies of Aki for storage of their output pending the final marketing.®

Regarding the origin of the oyster-culture in this region I may here quote a
paragraph from a tract on fishery matters published by Hiroshima-Ken.

In ancient times certain shellfish, Tapes , were gathered in great numbers on the flats of Aki;
and while awaiting their shipment to market the fisher people came to keep them in shallow-water
inclosures, the fences of which they formed of bamboo stalks. The discovery was then made that
the brushy fences became incrusted with young oysters, and thus it soon became evident that
under certain conditions and at certain places it would be more profitable to plant bamboo and to
cultivate oysters than to continue the Tapes industry. This was the first instance, it is said, that
bamboo collectors, or “ shibi,” were employed in oyster-culture.

The first detailed report upon the oyster industry of Aki was prepared for the
Japanese government by Prof. K aki chi Mitsukuri, the head of the department of
zoology of the Imperial University of Tokyo. It was published in 1894 (Tokyo) by
the department of agriculture and commerce, a royal octavo of about 50 pages,
containing many figures and several plates.* * 6 Unfortunately for the foreign reader
it has not yet been translated into a European language. My own knowledge of it
is due to my friend (who has also kindly drawn for me the text figures here repro-
duced), Mr. Naohide Yatsu, Rigakushi, a pupil of Dr. Mitsukuri. I have availed
myself freely of its substance, and if there is anything of value in the present paper
it owes its merit to my Japanese colleague. He has, moreover, given me generous
assistance in connection with my visit to Aki, advising me as to ways and means
and providing me with personal and official letters to the local authorities.

In Hiroshima I was courteously received by his honor Governor Senshi Egi, of
the prefecture of Aki, and to him and to his staff I am indebted for favors extended
me in many ways. To Mr. Shinobu Suda, official engineer of the oyster properties,
and to Mr. Masugi Shiraishi, a government expert in oyster matters, I am under
especial obligation for information regarding details in cultural methods; and
finally to Mr. Y. S. Murai, for many personal courtesies.

a A probable trace of this early privilege is seen at tlie present day, for the people of Aki are granted the test

places in the river mouths of Osaka.

6 All the present text figures, except fig. 27, are reproduced with unimportant changes from the Japanese i eport.
Plates f and 5 are from photographs taken by an artist in Hiroshima, and are copied by Dr. Mitsukuri; Plates 3 and
(j are original.

JAPANESE OYSTER-CULTURE.

19

THE JAPANESE OYSTER -ITS KINDS AND NATIVE CONDITIONS.

Three kinds of Japanese oysters are to be considered in this connection. First,
a small one, probably a dwarfed salt-water variety of Ostrea cucullata Born., is
abundant along the southern and eastern coasts of Japan — by far the most abun-
dant of its kind. This is a bay oyster, occurring in shallow water of specific
gravity of about 1.020 to 1.026, where it forms an almost barnacle-like incrustation
upon the tidal rocks. It is collected in great numbers for local consumption; fisher
people open them on the spot, not detaching the shell from the rock, and market
them by bulk. In actual size this oyster is rarely larger than one’s finger nail; but
its flavor makes amends for its size. It is plumbeous in color. The shell measures
rarely more than 2 inches in length; it is deeply crenulate, Gryphsea-like, at the
margin. In size and flavor it suggests very closely the California oyster.

The second form, Ostrea cucullata (cf. pi. 6, fig. D), is the important one from the
culturists’ standpoint. Although not large, it averages the size of a “Blue Point,” or
of an English “native.” The oyster itself is cream-colored, its shell delicately nacre-
ous, well shaped, thin, deep, and with a series of imbricating, horn- like outgrowths,
which suggest the shell of the European oyster, O. edulis. This species occurs
abundantly throughout the Inland Sea, in the small bays along the northeast coast
of the main island and at certain points in the Hokkaido (Yezo). It thrives best
in the bays well tempered by fresh water, of specific gravity of 1.017 to 1.023. Its
young are more abundant in the shallow and fresher water. The best that are mar-
keted grow at a depth of a fathom or two below low-water mark; it is practically
absent in water deeper than 8 fathoms. It is this species which will be considered
through the remainder of the present report in connection with cultural methods.

The third form, Ostrea gigas Thunb., is of large size, specimens weighing with
shell 4 or 5 pounds being not infrequent. It rarely occurs in water less than 2
fathoms deep and is most abundant in about 10 fathoms. The specimens which I
examined were taken by divers in water of about 35 feet. It is a typical sea oyster,
occurring in water of specific gravity of about 1.026. As far as I have been able to
ascertain, its value is purely local, no region producing sufficient numbers to warrant
a definite fishery. A large bank of oysters occurs in the Hokkaido, off the north-
east coast, not far from the town of Akkeshi. The oysters here are said to be of
extraordinary size, but during my visit to the Hokkaido I was not able to ascertain
whether they represent this third species or whether they are large examples of O.
cucullata. The latter species certainly occurs in the neighborhood.

The oyster-producing region of Japan is par excellence the Inland Sea, and it is
here that the culturists have carried on their industry with greatest success. This
body of water can indeed be looked upon as one of the most important natural pre-
serves of fish and shellfish in the world. It can be compared to a deep marine lake,
but it is sufficiently open to the sea to insure favorable conditions of density and of
renewal of its waters, while its occupants are free from the dangers of an open gulf.
From the oyster-culturists’ standpoint the Inland Sea is remarkable in that its con-
nection with the ocean is established both at its ends and near its middle point.
Thus at the extreme east it opens to the ocean through the Straits of Naruto, as
well as at the mouth of the Izuminada. At the west, 240 miles away, it opens again,
this time to the Japan Sea, through the Straits of Shimouoseki, and to the south

20

BULLETIN OF THE UNITED STATES FISH COMMISSION.

again to the Pacific through Bungo Channel. In this middle region the large island
Shikoku approaches closely the mainland, and the Inland Sea is broken up by a
maze of islands extending from Shodoshima on the east to Iwai on the west, a
stretch of 130 miles; and it is here that the most favorable conditions exist for the

Fig. 1. — Map of the region of Japanese oyster-culture on the north shore of the Inland Sea near Hiroshima
(Sea of Aki). Oyster parks are indicated in hlack areas. Particular reference in the present report to
Kaida Bay, Kusatsu, and Nihojima. The distance from Kaida to the island Itsukushima is 12 miles.

growth of shellfish. Everywhere are bays and harbors, and the density of the
shallow water is favorably tempered by the incoming streams of Aki, Bingo, Bitchu
and Bizen on the north and Iyo and Sanuki on the south. Add to these advantages
that there is a favorable tide fall of from 10 to 15 feet and an abundance of sandy

HIROSb

KAIDA

NIHO-

JIMA

HATSUKA1

I) ITSUKU
SHIMA

ETAJIMA

'tj K.URE

ONDO,

1IJIMA

JAPANESE OYSTER-CULTURE.

21

and gravelly bottom, thus enabling the culturist to operate his submerged farms
conveniently. Throughout this entire region oyster-culture is carried on more or
less generally, but the most important seats of the industry are at Okayama in the
east and near Hiroshima (prefecture of Aki) in the west. In the former locality a
small nearly- inclosed bay, which suggests that of Arcachon in France, proves very
productive and supplies no little part of the market of Kobe and Osaka. Here
also are canning factories. Near Hiroshima, however, the industry is conducted on
a somewhat larger scale, although on the same general lines.

It is in the latter region, as already noted, that Professor Mitsukuri secured the
material for his report upon Japanese oyster-culture.

Fig. 2. — Fisherwoman opening oysters. The drawing shows the block and opener, oyster basket and trays.

OYSTER-CULTURAL METHODS IN THE NEIGHBORHOOD OF HIROSHIMA.

In addition to its natural advantages the region of Aki is especially favorable
for oyster-culture, since close by is Hiroshima, a city of nearly 100,000, to furnish a
ready market for its product and to provide the necessary capital and labor for the
growth of the industry. The culture is carried on in a gulf-like area, known as the
Sea of Aki, a dozen miles in width, whose mouth is protected from southern storms
by the islands of Itsukushima (Miyajima), Nomijima, and Etajima (fig. 1). The
most favorable points for the cultural work are along the northern coast on either
side of the Otagawa, which flows through the city and tempers the salinity of the
neighboring water. On the east of Hiroshima are the establishments of Niliojima,
and somewhat further up the bay, Ivaida; to the west, the grounds of and near
Kusatsu. The entire extent of these most favorable points is about 10 miles. This

22

BULLETIN OF THE UNITED STATES FISH COMMISSION.

cultural area, it may be noted, corresponds closely to that of Tarente in the south
of Italy, of Arcachon in France, and of the best part of Long Island Sound.

It is an interesting fact that the culturists in Aki have at certain points developed
independently branches of the industry which are strikingly similar to those employed,
for example, in France. We thus find that a clear distinction is made between the
regions in which young oysters — “spat” — can be obtained and those having the best

Fig. 3. — Hand pick used for
making sockets in gravelly
bottom for insertion of
skibi.

Ftg. 4. — Bamboo collectors, sliibi, arranged so as to form a boundary hedge
of an oyster farm.

conditions for different stages of growth; also regions in which the final touches are
given in preparing the oyster for market. It is convenient, therefore, to describe
the Japanese cultural methods from the standpoint of locality. Thus at Nihojima,
where the water is freshened by the entrance of the Otagawa, the “production” of
young oysters is an especial feature of the industry. At Kaida Bay there is a region
favorable for a combination of production and growth (elevage), and at Kusatsu, and
further along in the direction of Miyajima, are the best conditions for elevage.

JAPANESE OYSTER-CULTURE.

23

In tlie description of the methods employed in these three localities, it will
to consider them in the order from the simplest and the most complex, (1)
(2) Ivusatsu, (3) Nihojima,
for at Raid a Bay all stages
of the industry are repre-
sented in the same oyster
park; at Ivusatsu the
methods become more
complicated,' and finally,
in the region of Nihojima,
specialization in the cul-
tural devices has reached
a point surpassed in but
few European localities.

be best
Ivaida,

CULTURAL METHODS OP
RAID A BAY.

Pig. 5. — Arrangement of branched collectors as a close-set fence to
form one of the lines of an oyster park.

In this well-protected
region (cf. the accompa-
nying map of the sea cf
Aki, fig. 1) there is a large
area of shallow water, and
at low tide great flats are
exposed. Here it has been
ascertained that the con-
ditions of water density
are favorable rather for
the growth of young oys-
ters than for the produc-
tion of “ spat,” but at cer-
tain points production is
carried on with marked
success. The greatest dis-
advantage of the region is
the lack of space in which
the oysters can be kept
covered by water during
all times of the month.

Probably it is for this rea-
son that the growth of the
oyster comes to be checked,
since it is well known that
they rarely increase in size
after the end of the second
year. At this age, then, they are marketed, their small size distinguishing them
from the oysters of Aki cultivated in other localities. The bay of Raida is, however,
so fertile in its class of production that it takes a high place among Japanese oyster-

Ftg. 6.— Arrangement of
branched collectors in
close-set hedge, com-
mon in most types of
Japanese oyster farms.
Vertical projection.

Fig. 7.— Shibi of different rows, new and old, in
boundary hedge, showing how they are im-
planted to give mutual support.

24

BULLETIN OF THE UNITED STATES FISH COMMISSION.

grounds and its concessions are keenly sought. At low tide it bristles with closely
set oyster farms and from a distance reminds one, save in color, of a region of
European vineyards. Each farm is a simple inclosure formed by “ shibi ” or bamboo
stalks, with or without interlacing branches. (Of. tigs. 4 to 9.) Bamboo in this, as
in many other arts and trades of the Japanese, possesses many advantages. It is
durable even in salt water (good material lasting three years or thereabouts) ; it is

The black lines in figs. 8 and 9 represent newly arranged bamboo collectors, the dotted lines
the collectors of the second year. Direction of current is indicated by arrows.

light and strong, gives an interlacing series of branches and leaves which in texture
serve admirably for the attachment of spat, and which give, moreover, a great extent
of attachable surface. In addition the bamboo stalks can be readily put in place and
removed ; they are easy to obtain in any locality, and their cheapness is not one of
their smallest virtues.

In the present farms shibi are planted in position every spring at a time which the

Tlie photograph was taken during a favorable spring tide.

Bull. U. S. F. C. 1902. (To face page 24.)

Plate 3,

JAPANESE OYSTER-CULTURE.

25

culturists have determined by experience to be most favorable for the set of spat,
usually about the middle of April. Their arrangement and mode of replacement
are as follows: Shibi are brought to the oyster-grounds in skiffs and deposited on the
flats as they become exposed by the receding tide. The culturist will have had the
boundaries of his concession staked out, and lie has but to construct his fences of
shibi as quickly as possible to take advantage of the hours of low tide. As a time-
saving device, he has already had the ends of the shibi sharpened so that they can
be thrust deeply (one-quarter or one-fifth their entire length) into the soft bottom.
Should the bottom prove hard, however, holes are first made for the shibi by means

Fig. 10. — Diagram of an oyster farm in which are combined the longitudinal and transverse modes of
arranging the bamboo collectors. Kusatsu.

of an iron-shod pick, shown in fig. 3. This the workman sometimes presses down
with his foot, hence the lateral support near the head of the implement.

According to the usual type of oyster farms in Kaida, the main boundaries are
planted nearly parallel to the tide marks. Similar rows of shibi are next thrown
out in the direction of the middle line of the park. (Figs. 8, 9. ) They do not, however,
meet, but leave an open median space passing through like an aisle. Thus on either
side of the main aisle of the oyster park there are rows of transverse alleyways, each
about 6 feet wide, which terminate often blindly at the main fence of the park. Details
in arrangement are given in figs. 4, 5, 6, and 7, fig. 6 showing a horizontal projection
of the shibi of fig. 5. The fences and partitions of shibi stand about waist high.

26

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Such an inclosure in Kaida Bay is practically the whole stock in trade of the
culturist, for in it spat is taken and the oysters grow attached to the bamboo until
the time when they are sent to market. From the permanent character of the park,
therefore, the culturist has been led to ingenious arrangements by which the shape
of the park can be retained and the oysters grown, young and old, side by side. To
accomplish this each fence is usually formed of a double row of shibi, as indicated in

figs. 8, 9, of which one row is of the second
year and the other a new one. The latter
is often added as the older row is removed.
The fences in this locality are made of
the more delicate species of bamboo,
“hachikn.” The stalks are trimmed in
lengths of about 5 feet and are inserted a
foot deep in the bottom. The distance
between the double rows of shibi is inten-
tionally narrow, so that their opposing
branches can interlace. (Fig. 7.)

The foregoing description refers to

Fig. 11. — Diagram of a small oyster farm of labyrinthine
pattern. Ondo.

the commoner type of park in Kaida.
Occasionally some are seen of a more
complicated pattern, as in fig. 9, and,
rarely, some in which the double rows of shibi are placed parallel instead of trans-
verse to the main axis of the park, as in fig. 10. An excellent example of this type
is pictured in plate 3, from a photograph which was taken at the end of a season, the
shibi showing well-grown oysters (from park at Tanna). Rarer still is a form in
which a labyrinth-like arrangement of the hedges of shibi prevails (fig. 11), or even
a concentric pattern (fig. 12). In all these forms, however, the arrangement is such
that many eddies will be formed about and within the rows of the shibi, since these
eddies have been found con-
ducive to the attachment of
the young oyster.

In t he foregoing types of
parks the visitor notes that
the height and strength of
the shibi, their simple or
branching character, to-
gether with the closeness in
their arrangement, vary
somewhat widely in different
localities. Occasionally an
arrangement which alternates old and new shibi in the same row is found to be
adopted advantageously.

At the close of the cultural process, that is, at the end of the second year, it
remains only to remove the marketable oysters from the shibi. This is done during
the favorable tides, the culturist using a pick-like instrument (fig. 13) with one hand
and seizing the shibi with the other. For the protection of this hand a curious
heavy but open mitten is used, figured in fig. 17. After the oysters are detached

Fig. 12.— Diagram of an oyster farm in which the collectors are arranged
in circular and concentric order. Middle clear space used as a living
ground. Itsukushima.

JAPANESE OYSTER-CULTURE.

27

they are taken from the ground by means of the rake shown in fig. 15, placed in
baskets, fig. 18, and carried thence in the usual oyster boat often to the mouth of
some adjacent river, where they are thrown out and raked over. By the latter pro-
cess, “ drinking ” the oysters in fresher water, they increase in size and become
cleaner, a process, by the way, quite similar to that employed in France, in England,
and often in America.

Ftg. 13. — Oyster hook used for dislodging well-grown oysters from the sliibi.

At Kanawa, an important cultural ground, a similar method to that of Ivaida
Bay is employed. The cultural area is not large but it is very productive, and here
they have found it profitable to plant shibi in close rows at right angles to the coast-
line, as shown in fig. 19.

OYSTER-CULTURAL METHODS OF KUSATSU.

The oyster-grounds of Kusatsu are the best of those situated west of Hiroshima,
but all of these, and of this entire region, are essentially the same, as far as cultural

Ftg. 14.— Oyster rake, gofuze-guwa, used for “cultivating” the oysters, i. e., stirring them about roughly as they
lie on the living grounds, so as to break off the delicate shell margin.

Fig. 15.— Oyster rake, nihon-zume, used to gather oysters fallen from the shibi, or to select oysters to be detached.
Ftg. 10.— Oyster rake, yatsugo, used to collect marketable oysters from the living ground.

methods are concerned. They extend along the western coast 7 or more miles from

Hiroshima, at points indicated on the map (fig. 1). In these localities oysters are
cultivated at greater depths than in other waters of Aki, for it has here been found
that under the deeper conditions the shellfish continue to increase in size after the

28

BULLETIN OF THE UNITED STATES FISH COMMISSION.

second year, unlike those at Kaida. Therefore, at Ivusatsu, to the end of growing
a larger oyster, the culturist divides his parks into three classes: those of shallow
water (largely for spat collecting), early rearing, and deep water (for late rearing).

The cultural concessions accordingly have come
to he arranged with their long axes at right
angles to the shore, thus providing a range of
water passing from shallow water to deep, each
lot measuring, in round numbers, a thousand
feet by fifty. In shallower waters, tempered by
fresh streams from the direction of Hiroshima,
are the best conditions for spat-collecting.
Here the specific gravity is very nearly the same
as in Kaida Bay — in summer about 1.018; in
winter 1.020. The density of the water rises
gradually and attains about 1.025 in the deepest
zone. Accordingly, the shallowest region in
each park usually becomes laid out in a zone of
collectors, or shibi-ba, and resembles somewhat
Kaida Bay. In the next and deeper ground
are the rearranged and oyster-bearing shibi (of
the shallower zone), toya-ba, and in the deepest
water are the typical oyster beds, or miire-ba.
Of course such an arrangement is sometimes
modified, since practice demonstrates that the
local conditions of an oyster park, e. g., water
currents, are apt to warrant widely different
treatment.

(a) In the shallow zone an arrangement of shibi in lines parallel to the shore
is the common one. Between the rows are intervals of about 4 feet, the park in
this case resembling one of the common type of Kusatsu (fig. 8). Often, however,
the shibi are shorn of their branches and
planted like canes (3 to 5 feet in length)
in close-set rows. Such, for example, are
the shibi photographed in plate 6, A, B,
and C. The first of these, A, has had the
oysters attached to it for about a month;
the second, B, for about 6 months, and the
third, C, for about 18 months. It is at
the last period that the masses of oysters
come to separate somewhat readily from
the bamboo. In these parks the arrange-
ment and treatment of the shibi varies
greatly in accordance with local condi-
tions. In rapid currents, which distinguish
the region of Kusatsu, short and branchless shibi are commonest, these, too, made
of the strong and short-jointed species of bamboo, “madake.” Their arrangement
is frequently in clumps, or toya, as indicated in figs. 20 to 23. Of these the toya

Pig. 17. — Mitten of heavy sack cloth, open at
thumb and finger tip, used to hold the oyster-
bearing shibi while separating the oysters.

Pig. 18.-

-Basket, taruyama kayo, for collecting and
storing marketable oysters.

OYSTER PARK NEAR NIHOJIMA. TYPICAL LIVING GROUND, OR IKE-BA, WITH HEDGE OF SHIBI.
Maximum tide fall of about 15 feet.

Bull. U. S. F. C. 1902. (To face page 28.)

Plate 4.

ii'jF

JAPANESE OYSTER-CULTURE.

29

of figs. 20 and 21 are specialized for shallow water, and that of fig. 22 to a rapid
current. In such clumps branching shibi frequently occur, and in this event the
tips of the stalks are more apt to diverge than in Kaida (contrast figs. 23 and 24),

Pig. 19.— Diagram of oyster farm in which shibi are planted in rows parallel and at right angles to shore line.

another adaptation to more rapid current. In general the bases of the component
shibi are implanted about a foot. Thus made the toya remain in position for about
two years.

Fig. 20. — Arrangement of collectors in clumps, dotted parallel to the shore, leaving straight shoreward alleyways
between. Clumps 3 or i feet apart, arranged for sluggish current, the direction of the latter denoted by an arrow.

( b ) In the next deeper zone, typical toya-ba, the grouping of shibi becomes
more massive, and when at low tide this region is exposed (10 to 15 feet is the tide
fall) one sees them in long rows which suggest diminutive haycocks. An excellent
idea of such toya-ba is had in plate 5. Closer inspection shows that each toya is

30

BULLETIN OF THE UNITED STATES FISH COMMISSION.

in nearly every case a complex of shibi of many ages. Transplanted shibi, with
oysters of one or two years’ growth, usually form the nucleus of the cluster, and
around them are planted concentric rows, one, two, or three, of branched shibi, old
and new, to the end that all ultimately bind or mat themselves into a living, springy,
cone-shaped mass, well suited to resist currents or storms. (Tig- 25 and PI. 5. ) Such

• • •

• •

• • f

• •

• • •
% •

• •
• •

• •

• t •
• •

» • •
• •

• • •
• •

• • •
• •

• • ft

• «

• •

» • •

• •

• • •
• •

• •

• • •
• •

• •

* • •

♦ •

• • •
• •

Fig. 21. — Arrangement similar to fig. 20, but in oblique order, adapted to a somewhat more rapid current.

toya are made, or remade, toward the end of each spawning season, i. e., during
the end of August or early part of September. They are then pulled to pieces (in
this work the rake shown in fig. 14 is used), and from each bamboo there are shaken
and broken off the oysters which are least securely attached. During this process
the shibi found to be still useful are put aside to form the nucleus of new toya;

• • • •

• • • • • •
• • • •

• • • •

• • • • • «

• ♦
t • •

• t

« %
• •

• •

• • •
• •

« •
« • •
• •

• •

• # •
• #

• ♦

• t •
• •

• •

• • •

• •

• • •
t •

• ♦ ♦ •

• • • • • •
• • • •

♦ ♦ • •

• • • • • •
• • • •

• • • •

• • • • • •
• • • •

Fig. 22. — Arrangement similar to fig. 20, but adapted to rapid current, rows of collectors separated

by intervals of 8 or 10 feet.

the gleanings, twigs, detached oysters, and all, are now raked up and carried to a
third locality, the living grounds (ike-ba). Here in swifter current the debris washes
away and the oysters remain, the shells usually becoming clean in the meanwhile.

(c) The living ground, or ike-ba, of which a good example is pictured in plate 4,
is generally located in a zone of deeper water. It has a clean, gravelly bottom,

JAPANESE OYSTER-CULTURE.

31

which in most cases becomes exposed only at the lowest tides. Here the oysters are
brought together which have become detached from shibi in various parts of the
farm, and thus they are retained in classified beds until they have attained their
second year’s growth. During this time the culturist has but to keep them well
spread out and to see that the beds are kept clean. Always at lowest tides, and
sometimes as often as fortnightly, the laborers give the oysters a vigorous raking
(using for this purpose the short-toothed rake shown in fig. 14), which scatters them
about the bed, removes foreign bodies, and, best of all, gives the shells a better
shape and a firmer rim, for in this treatment the delicate, cuticular outgrowths of
the shell are removed and a more symmetrical growth results. Of especial impor-
tance is the process of raking in cases where the oysters are sent directly from the
living ground to market; for it has been found (here as elsewhere) that those oysters

Fir;. 23.— Bamboo collectors arranged after the fashion common in Kusatsu. The shibi stand about
3 feet above the bottom and their tips diverge; the clumps are set i or 5 feet apart.

whose shell rims are strong and accurately fitted together fare better during trans-
portation and in the market. Those with delicate and brittle margins soon suffer
injury and lose their fluid through leakage and evaporation.

In some farms, on the other hand, it is maintained that oysters of different,
sizes should be mixed on the same living ground. For it is claimed that the
young oysters grow better side by side with the older ones, and even that if the
more perfectly grown oysters of different grades can be mixed together during the
process of raking, the better will be the general output.

( d ) The final stage in the culture of Kusatsu takes place in maturing grounds,
or miire-ba. Here the larger oysters of the second year’s growth are laid down,
transplanted from the one or more ike-ba of each establishment. Usually they are
in the deepest water cultivated, i. e., in water a few feet in depth at lowest tides
up to water of 3 or 4 fathoms. I was told that in one miire-ba oysters were culti-

32

BULLETIN OE THE UNITED STATES FISH COMMISSION.

vated successfully iu water of 50 feet. Each farm iu Kusatsu lias its separate
miire-ba, aud these adjoin one another, forming a continuous zone in deeper water,
each miire-ba designated by a number. In the shallowest portions the oysters are
usually protected against displacement and invasion of mud by means of low fences
arranged with wing-like expansions, as shown in ground plan in fig. 26, and in detail
in fig. 27. This device has been developed largely in view of the storms of the typhoon
season. The effect of the maturing ground is to give the oysters greater size and

Fig. YA. — Collectors arranged in a way common in Kaida Bay. The shibi stand about 4 feet
from the bottom and their tips converge.

weight, and to give the meat a whiter color. The finished product resembles closely
a well-grown oyster of Long Island Sound. Marketing takes place after the oysters
have remained from six months to a year in the deeper ground, making their total
age about three years.

Each oyster farm has its separate houses, situated usually on the adjacent
shore, and the details in handling and packing the oysters appear to be closely
similar to those of Continental Europe — baskets, blocks, rakes, arrangement, and

OYSTER PARK NEAR NIHOJIMA. GENERAL VIEW, SHOWING NEWLY ARRANGED TOYA.

Bull. U. S. F. C. 1902. (To face page 32.)

Plate 5.

JAPANESE OYSTER-CULTURE.

33

storage. I may note in passing the curious “knife,” a combination of knife, mallet,
ancl lever, with which oysters are opened with surprising rapidity. The modus
operandi is shown in fig. 2, page 21.

OYSTER-CULTURE AT NIHOJIMA.

In this locality (cf. fig. 1), finally, oyster-culture has been developed on more
special lines than anywhere else in the East. As at Kusatsu, the industry embraces
three distinct branches — (a) spat collecting; ( b ) rearing the young, and (c) maturing
and fattening the oysters for market. But, unlike at Kusatsu, these three branches
of the industry are carried on, not on the same shore reach, but at points widely

Pig. 25. — Ground plan of a mound toya of collectors. Sliibi with well-grown oysters are
indicated by the black spots within the two circles of branching shibi.

separate. In other words certain special tracts are taken advantage of in collecting
spat; others are specially arranged for early rearing; others, in turn, for maturing.
In these regards oyster-culture at Niliojima resembles closely that of the coast
of Brittany or of Holland. The details of the management of the farms are in
essentials, however, like those previously described.

(a) Spat collecting. — The spat is collected in very shallow water, less than a
fathom deep at the usual tides, tempered considerably by incoming streams. The
specific gravity is said to rarely exceed 1.017. In such a region shibi are put in
position, usually in very close order, at the beginning of each spawning season, say
from the middle of April to the middle of May. After a period of about three
months the entire mass of shibi will be uprooted and transplanted, sometimes a
mile or more, to a locality better fitted for rearing the young oysters. This trans-
portation, I was told, is the most difficult part of the work of the culturist of
Niliojima, for the minute oysters are, as everywhere, peculiarly liable to injury;

F. C. B. 1902—3

34

BULLETIN OF THE UNITED STATES FISH COMMISSION

careless handling will destroy great numbers of the delicate shells; hasty packing of
the shibi on the scows, whereby the branches are allowed to rub together, is another
palpable source of loss; drying, direct sunlight, changes of temperature, are all
deadly, and, above all, severe thunderstorms — the latter, according to my informant,
causing death by fear. I suspect, however, in the last regard, that a fresh, cold
shower bath is more apt to be a moving cause, although I was assured that the scows
are covered with the ever-present Japanese oiled paper to guard against such a
contingency.

(b) Rearing the young oysters. — The shibi, covered with young spat, are now
arranged in toya-ba, like those of Kusatsu, but closer in arrangement and usually
of many varieties. Here they remain from one to two and a half years. In the

Fjq. 20. — Diagram of living ground of two to three year old oysters. The spots represent oysters and the dark
lines are rows of shibi placed so as to provide shelters against currents or storms. The arrow denotes direc-
tion of strongest current.

case of the older and rearranged toya a long, mound -like type is commoner than
the circular ones described above. As a rule the toya are covered with water save
at lowest tides.

(c) Later rearing ; Maturing . — The living inclosures in deeper water, ike-ba,
correspond to those of Kusatsu. They contain the oysters which have been separ-
ated or are readily separated from the brush of the toya. A similar process prevails
of raking the oysters roughly, and I was shown some shells of the older oysters
from this region which were of very regular shape. At. favorable tides, further-
more, the oyster beds are cleaned and the boring whelks — Purpura clavigera Kaster

JAPANESE OYSTER-CU LTURE

35

and Rapana bezoar L. — are removed. Starfish are not troublesome. Further trans-
planting takes place, usually at the end of the second season, and in still deeper
water the oyster attains finally its marketable size.

REGULATION OF OYSTER-CULTURE BY THE JAPANESE GOVERNMENT.

All cultivable grounds, whether for oysters, other shellfish, or seaweed, are — in
Aki, at least — the property of the prefecture and can be neither sold nor sub-
rented. All cultivable tracts are surveyed, the lots tending to decrease in size as
the estimated value of the property increases. The mode of laying out concessions
can perhaps be best understood by reference to plate 7, copied from a Government

chart. The farms are rented auction fashion to the highest bidder, and the tenants
have the privilege of renewing their leases indefinitely at the original rentals, a
privilege, however, which can not be used speculatively. When at the termination
of a lease the property passes again into the hands of the prefecture it is at once
advertised and rerented. In addition to the yearly rental the property is subject
to a small local tax upon the total area of each farm, and to a charge of 1 per cent
of the rental to cover the expenses of administering the oyster-cultural bureau of
the prefecture. This series of taxes impresses the stranger as formidable, until he
learns that it does not represent an accumulation of taxes, but rather an itemized
statement as to the apportionment of public, funds made thus in accordance with
local custom. Rental of concessions from the prefecture, it should further be said,
brings with it certain privileges in marketing the oysters in Osaka.

Fig. 37. — Detail of the living ground shown in ground plan in preceding figure.
Bamboo rods 3 to 4 feet high.

36

BULLETIN OF THE UNITED STATES FISH COMMISSION.

THE QUESTION OF INTRODUCING AND CULTIVATING THE JAPANESE OYSTER

IN THE UNITED STATES.

To wliat degree the Japanese oyster would flourish on our coast can be deter-
mined, obviously, only by experiment; and if experiments are to be made, the
following suggestions seem to me of some practical value :

On the Pacific coast, on grounds which have been found especially favorable for
the reproduction and growth of our Western species, O. calif ornica, consignments
of Japanese oysters may be planted — in the north with oysters from the Hokkaido
beds, preferably from the region of Mororan or from Akkeslii Bay, and in the south
from the region of Hiroshima. In this way similar conditions of temperat ure will
be obtained. To fulfill a second favorable condition an effort should be made to
secure oysters from water of nearly the same specific gravity as in the chosen
American localities. There would be, I fancy, little difficulty in the matter either
of securing oysters from equivalent localities or of transporting them. Through the
Imperial Bureau of Fisheries, under the able headship of Dr. Iv. Kishinouye, one
could promise, ft priori , prompt and efficient aid in getting in touch with the Japanese
oyster-culturists whose establishments are known to be favorably located. And
with fast freight service from Yokohama the oysters could be transported with a
minimum of loss, as similar exportation (e. (/., of American oysters to England)
demonstrates. In a case of this kind, however, extra precautions would not be out
of place. Large oysters should first be selected, and, preferably, treated with the
raking process on ike-ba for a few weeks. By this means the shell margins will be
thickened, and thus the oysters will lose a minimum amount of fluid during
transportation. Care should also be taken to pack the oysters each with the more
concave valve downward and to mark the cases so that during shipment they shall
be kept in the right position. Other details — if not indeed the above— can safely
be left to the skilful Japanese work people.

A further suggestion is that the shipments be made during the months of Feb-
ruary and March, to the end that the coldest season on our coast would be avoided,
and thus the oysters would have the chance of becoming in good condition and
somewhat acclimated by the following winter. Moreover, at this season of the year
it has been found that the oysters have laid up the maximum supply of nutriment
against the breeding time, a supply which could be used as reserve nutriment dur-
ing transportation.

No experiment, however, could be regarded as a fair one, I believe, which did
not deal with an adequate number of the imported oysters. If but a few oysters — a
score or two — are laid down in each experimental locality they may be lost through
accidents which would not befall a larger number. For in oyster banks there is
certainly strength in numbers, an oecological feature in oyster-culture which
governmental regulations in France, Holland, and Germany keenly recognize. By
this numerical factor it appears a true paradox that one thousand planted oysters
have more than ten times the chances of survival than a hundred. As oysters are
cheap, it seems to me better economy therefore to make the more convincing
experiment.

To what degree is it practicable to introduce the methods of Japanese oyster-
culture into the United States? This, I take it, is a question which can be answered

Bull. U. S. F. C. 1902. (To face page 36.)

Plate 6.

BAMBOO OYSTER COLLECTORS, OR SHIBI, AFTER HAVING BEEN IN USE ABOUT ONE MONTH, A- SIX MONTHS, B-

EIGHTEEN MONTHS, C.

Detached oysters are shown at V. Figures about one-third natural size.

JAPANESE OYSTER-CULTURE.

37

as well by individual culturists as by the U. S. Fish Commission. The culturist who
takes keen interest in his work should be looked to, I think, to select some especially
favorable spot on his grounds and lay out an experimental farm, which need not be
large and would not be costly. Brush to take the place of the bamboo shibi can be
obtained readily, and one has merely to follow the broad lines of procedure laid down
by the Japanese.

Culture with the Japanese oyster on the Pacific coast is clearly out of question
until it can be demonstrated conclusively that the imported oysters are breeding
prolifically. The greater importance of these methods is as clearly in the line of
improving our native Western oyster, for if they are grown by means of shibi and
subsequently laid down in <1 eeper and deeper water they will, in all probability,
increase in size and improve in flavor. Similar experiments with the Virginian
oyster would also be of considerable interest.

In the general problem of artificial oyster-culture one is not surprised to find that
Japanese methods have paralleled closely those developed in Continental Europe.
The principles are clearly the same, the means alone have varied. In Japan, as in
Europe, those localities alone possess oyster- culture which are peculiarly favorable
for it. Not every locality where oysters naturally occur offers a place for developing
oyster farms, and he will surely be disappointed, whether in the United States, Italy,
France, or Japan, who believes that his collectors are powerful magnets capable
of attracting like iron filings, the entire yield of young oysters of the neighborhood.
Tiles, fascines, .jingle shells, or shibi are excellent mediums for the attachment of
oysters, but are valueless save in the event of an abundance of fry.

He , therefore, who wishes to experiment in this form of oyster-culture will fare
badly, I am sure, in spending time and substance in attempts to lay out a collecting
ground in any locality in which spat does not occur every year and in great abun-
dance. In fine, the experience of foreign culturists has gone so far as to determine
that even in a good collecting ground there are tracts which, owing largely to cur-
rents, are of great value, and that neighboring ones, sometimes but a few yards
away, are to be avoided. But such delicate adjustments are the true product of a
long-used system, whether European or Japanese.

In the matter of elevage the Japanese do not rank as high as the French par-
queurs, for they have not developed a system of tidal closures, or claires. These, I
believe, would prove an important adjunct of the establishments of the shallower
region of Aki, especially if carried out after the fashion of Brittany.

Bull. U. S. F. C. 1902. (Page 38.)

Plate 7.

AAP OF THE OYSTER AND SEAWEED CONCESSIONS IN ONE OF THE ESTUARIES OF NIHOJIMA, TO SHOW HOW COMPLETELY THE CULTURAL AREA IS DEVELOPED.

I, Shibi bearing purple seaweed; II, Shibi bearing oysters; III, grounds for rearing oysters; IV, cultural area for other mollusks, Tapes, Area, etc.

THE HABITS AND CULTURE OF THE BLACK BASS.'

By DWIGHT EYDELL.

This paper sets forth the experiences of nine seasons, beginning with 1894,
during which I have had charge of the black-bass work of the Michigan Fish Com-
mission. The work was begun at Cascade, Mich., and after four seasons was
transferred to Mill Creek, where it is now carried on. The methods of pond culture
finally adopted are based on a knowledge of the breeding habits of the fish under
natural conditions. The account has reference to the small-mouthed bass, unless
the large-mouthed is specified.

In studying the habits of the bass, it is necessary to distinguish the males from
the females; ordinarily this is not possible except by dissection, but at spawning
time the female is distinguishable, even at a distance of 10 or 20 feet, on account of
her distension witli eggs, and this makes it possible to determine the part taken by
eaclx sex in nest-building and the rearing of young.

The nests of the black bass are built by the male fish working alone. The
small-mouthed bass prefers a bottom of mixed sand and gravel, in which the stone
ranges from about the size of a pea to that of one’s fist. As the spawning season
approaches the males are seen moving about in water of 2 or 3 feet depth seeking a
suitable resting-place. Eacli male tests the bottom in several places by rooting into
it with his snout and fanning away the overlying mud or sand with his tail. If he
does not find gravel after going down 3 or 4 inches, he seeks another place. Having
found a suitable place, he cleans the sand and mud from the gravel by sweeping it
with his tail. He then turns over the stones with his snout and continues sweeping
until the gravel over a circular spot some 2 feet in diameter is clean. The sand is
swept toward the edge of the nest and there forms a few inches high, leaving the
center of the nest concave like a saucer. The nest is usually located near a log or
large rock so as to be shielded from one side. If the bank is sheer and the water
deep enough, the nest may be built directly against the bank. If possible, it is
placed so that the fish can reach deep water quickly at any time.

During nest-building no females are in sight, but when the nest is done — and
this takes from 4 to 48 hours — the male goes out into deep water and soon returns
with a female. Then for a time— it may be for several hours — the male exerts
himself to get the female into the nest and to bring her into that state of excitement
in which she will lay her eggs. If she lies quiet, he turns on his side and passes
beneath her in such a way as to stroke her belly in passing. If she delays too long,
he urges her ahead by biting her on the head or near the vent. If she attempts to
escape, he heads her off and turns her back toward the nest. If, after all, she will
not stay in the nest, he drives her roughly away and brings another female.

a Read at the thirty -first annual meeting of the American Fisheries Society.

39

40

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Some 15 to 30 minutes before the female is ready to enter the nest and spawn,
her excitement is made evident by a change of color. Ordinarily she appears to be
of a uniform dark olive or brown above, changing to a light green below. The only
markings readily seen are four stripes on each cheek; but in reality the sides of
the fish are mottled with still darker spots on the dark-olive background. The spots
are arranged so as to form irregular, vertical bands like those on the perch, but
these are not usually visible. Now, as the excitement of the female increases the
background becomes paler and finally changes to a light-green or yellowish hue, so
that the spots and bands stand out in strong relief. The whole surface of the fish
becomes thus strongly mottled. This is a visible sign that the female will soon
spawn. The male undergoes a similar but less pronounced change of color.

Soon after this the female enters the nest and the male continues to circle about
her, glide beneath her, and to bite her gently on the head and sides. At times he
seizes her vent in his mouth and shakes it. When this has continued for a time
spawning takes place. The two fish turn so as to lie partly on tlieir sides with their
vents together and undergo a convulsive fluttering movement lasting 3 to 5 seconds.
During this time the eggs and milt are extruded. The circling movements are then
resumed to be interrupted after a few seconds by spawning. This alternate circling
and spawning continue for about 10 minutes. The male then drives the female
away, biting her and showing great ferocity. She does not return.

The male, and the male only, now continues to guard the nest, fanning sedi-
ment from the eggs and repelling enemies. At 66° F. the eggs hatch in 5 days and
the young fish swarm up from the bottom in 12 to 13 days from the time of hatching.

Henshall in his “More About the Black Bass,” published in 1898, quotes, with
approval, Arnold’s observations to the effect that the nests are built and then guarded
by the female. The Manual of Fish Culture, published in 1897 by the United
States Fish Commission, speaks of the nests as being built by the mated fish, some-
times working together, sometimes separately. These seem to be the latest published
observations, and are not at all in accord with my observations in Michigan.

Shortly after the young small-moutlied bass rise from the nest they scatter out
over a space 4 or 5 rods across — not in a definite school with all the fish moving
together, but as a loose swarm, moving independently or in small groups. This
makes it impossible to seine the young fry, as upon the approach of the seine,
instead of keeping together, they at once scatter and escape the seine. The fry
may be at the surface or on the bottom in weeds or clear water and are attended by
the male until they are 14 inches long. The swarm then gradually disperses and
the young fry, which were previously black, take on the color of the old fish.

The breeding habits of the large-mouthed black bass are similar to those of the
small-mouthed, but differ in some respects, which are of importance in pond culture.

1. The nests of the large-mouthed are not made on gravel, but by preference on
the roots of water plants. These are cleaned of mud over a circular area, and on
them the eggs are laid. As the eggs of the large-mouthed bass are smaller and
more adhesive than those of the small-mouthed they are apt, when laid on gravel, to
become lodged between the stones and to stick together in masses, and are then likely
to be smothered. When laid on fibrous roots of water plants this does not occur.

2. The young large-mouthed bass remain together in a compact school very
much smaller than that of the small-mouthed, and the fry usually move all in the
same direction. This makes it easy to seine the large-mouthed fry when wanted.

Bull. U. S. F. C. 1902. (To face page 40.)

Plate 8.

BLACK BASS POND, SHOWING FRY RETAINERS IN USE— BED FRAMES REMOVED.

FRY RETAINER, TO BE PLACED AROUND BED JUST
BEFORE THE FRY RISE.

BASS BED, TO BE PLACED ON POND IN EARLY
SPRING AND FILLED WITH GRAVEL.

HABITS AND CULTURE OF THE BLACK BASS.

41

CULTURE OF BLACK BASS.

Ponds and, stockfish. — After some experimenting all our ponds, both for stock fish
and fry, are built on the model of a natural pond. There is a central deeper portion
or kettle about 6 feet deep, and around the shore a shallow area where the water is
about 2 feet deep. The bottom is the natural sand, and water plants are allowed to
grow up in the ponds. All ponds are supplied with brook water, and silt from this
furnishes a rich soil for the aquatic plants. The water of these ponds contains
Daphnia, Bosmina, Corix, and other small aquatic forms in great numbers. These
furnish food for the bass fry. The ponds run in size from 120 feet by 190 feet to
100 feet by 100 feet.

At first we were unable to feed the stock fish on liver, but after a time we found
that by cutting the liver into strips about the size and shape of a large angleworm
and by throwing the strips into the water with the motion that one uses in skipping
stones they wriggle like a worm in sinking and are then readily taken. The liver
must be fresh. If bass are fed on liver alone they do not come out of winter
quarters in good condition. Of eleven nests made by bass thus fed only three pro-
duced fry. Although eggs were laid in all, they seemed to lack vitality, owing to
the poor condition of the parent fish, and in eight of the nests the eggs died.

In order to bring the fish through the winter in good condition it is necessary to
begin feeding minnows in September and to continue this until the fish go into
winter quarters. The bass eat minnows until they go into winter quarters, after
which they take no food until spring. The minnows are left in the ponds over
winter, so that the bass, when they come out of winter quarters, find a plentiful
supply, which lasts them until the spawning season. At this time the minnows are
seined from the pond, as their presence interferes with the spawning. Before this,
however, some of the minnows have spawned, and their fry later serve the young-
bass as food. Bass fed in this way come out of winter quarters in fine condition
and their eggs are found to be hardy.

Artificial fertilization. — During the first two or three seasons of our work
numerous attempts were made at artificial fertilization, but only twice with success.
On one occasion the female was seined from the nest after she had begun to spawn.
She could then be readily stripped. The male was cut open and the eggs were
fertilized with the crushed testes. About 75 per cent of the eggs hatched on a wire
tray in running water, the eggs being fanned clean every day with a feather.

In the second case the fish were seined while spawning, and it was found that
in the case of one female pressure on the abdomen caused a reddish papilla to
protrude from the vent. This had the appearance of a membrane closing the vent.
It was pinched off, and the female Avas then stripped readily and the eggs fertilized
and hatched.

Pond culture. — Having abandoned artificial fertilization, our attention Avas
turned to pond culture, and this we have carried on for about six years. Our earlier
ponds not furnishing natural spawning-grounds, Ave constructed alongside each of
the large ponds six smaller ponds for use as spawning-ponds, each about 16 by 24
feet, 16 inches deep, Avith gravel bottom, and connected to the central pond by a
4-foot channel.

The fish entered these and spawned. In one case we had eight nests in a single
pond of this sort. Where so many nests were made, usually but one or two of them

42

BULLETIN OF THE UNITED STATES FISH COMMISSION.

came to any good, the others being destroyed by the fighting of the males. Ordi-
narily but one or two nests were built in eacli spawning pond. The male first to
enter and begin the construction of a nest generally regarded the whole pond as
his property and held it against those that tried to enter after him. On one occasion
the male thus holding the pond was attacked by 10 or 20 other males at one time
and after a long struggle was killed and his nest destroyed.

The attempt to use small spawning-ponds was then abandoned and all the ponds
were made of good size and with a central kettle and shallow shore area, as already
described. The problem now Avas to prevent the fighting of the male fish and the
consequent destruction of nest and eggs. I finally hit upon remedies for Avhat
seemed to be the two chief causes of this fighting. I had noticed that in the natural
water the nests of the small-mouthed bass Avere frequently built against a stone or
log, so as to be shielded on one side. When so built the nests might be quite close
together, as near as 4 feet, and the fish did not fight, because they did not see one
another when on the nest. On the other hand, if a bass nest Avas built where it Avas
not shielded the bass on that nest would prevent any other bass from building
within 25 or 30 feet of him. It occurred to me to try to construct artificial nests and
shield them so that the fish on the nests could not see one another, placing the nests
so near together as to fully utilize the pond area.

In the spring, before the spawning season opened, the ponds were drawn doAvn
so as to expose the shallow terrace along the shore. The terracie was then cleaned to
a depth of about 2 inches of sediment and vegetation Avhich had accumulated since
the previous summer. Rectangular nest frames of inch board AA'e re made 2 feet
square and without bottoms. On two adjacent sides these frames were 4 inches
high, Avhile on the other sides they Avere 16 inches high. They Avere set AAdiere there
would be about 2 feet of water when the pond Avas filled, and so placed that the
corner formed by the junction of the two loAver sides pointed to the center of the
pond, Avhile the opposite corner, formed by the higher sides, pointed toward shore.
The frames were set directly on the bottom, not in excavations, and each Avas filled
with gravel containing sand suitable for nest-building. A board Avas laid diagonally
across the tAvo higher sides and a heavy stone laid oh this to keep the frame in place.
The two higher sides form a shield on two sides of the nest, while the board across
the top affords shade. The frames Avere set in two i-oavs about the pond, parallel
to the shore line.

The i-oavs Avere about 6 feet apart and the nests in each row about 25 feet apart,
alternating with those in the other row. There was thus about one nest to each 100
square feet of suitable bottom, or in each area of 10 by 10 feet. When the bass
Avere on the nests no one was able to see any other and the fighting from this cause
Avas practically eliminated. The number of rows of nests may be increased to three
or four, or more where the area of shallow water is wide enough.

The bass selected these nests in preference to any other spawning-ground.
They cleaned up the gravel and behaved in the nests in every particular as they
would on natural spawning-grounds. The first time Ave tried these shielded nests
not a single bass made a nest outside of them, though there Avas plenty of good
gravel bottom available for this purpose.

As to the second cause of fighting: In 1900, when these nests Avere first tried,
from 475 stock fish Ave obtained 315,000 fry and 750 fingerlings. In the season of
1891 the output Avas very much less and there Avas considerable fighting among the

HABITS AND CULTURE OF THE BLACK BASS.

43

fish. This remained unexplained till the ponds were drawn down after the spawning
season, when it appeared that although the fish had been sorted, the number of
male fish was considerably in excess of the number of females, and these excess
males, banding together, went about breaking up the nests of their more fortunate
brothers. It is now the practice when setting the nests to seine out the stock fish and
sort them, putting about 40 males to 60 females, thus removing the second source
of fighting.

During the present season from 493 adult fish we had produced 430,000 fry up
to May 26, and we believe that we can do as well every year.

Up to the present year there have been two sources of loss incident to the water
supply. The supply is a spring-fed brook, which runs over an open country before
it reaches us. The water in this brook becomes quite warm on a hot, sunny day and
cools off at night. The temperature thus falls at night sometimes as much as 13° F.
and becomes as low as 46° F. This is disastrous, since when the temperature gets
below 50° F. the adult fish desert the nests and the eggs or young fry are killed by
the sediment. By watching the temperature of the water and, when it approaches
50° F., shutting off the supply until the water warms up, this difficulty is obviated.
Since the ponds are well stocked with water plants the fish do not suffer from lack
of oxygen when the water is shut off. Indeed, if the water did not leak out of the
ponds, I doubt if it would be necessary to introduce any running water into them
during the breeding season.

The second difficulty with the water supply is from sediment brought down by
the brook after heavy rains. This sometimes accumulates over the nests so thick
as to smother the eggs and drive away the parent fish. By shutting off the water
supply whenever the water is much roiled this trouble is avoided.

The water supply, however, must be kept fairly constant. If the level lowers
more than about 6 inches the fish leave their nests and the eggs die. For the purpose
of maintaining a constant water level it would probably be best to have the ponds
made with clay bottoms. The difficulties arising from roily water of variable tem-
perature are, however, local, and would probably not be usually encountered.

Handling the fry after they rise from the nest. — The small-mouthed fry have
the habit of scattering into a large swarm when they leave the nest and it is con-
sequently difficult to seine them when wanted. It is therefore desirable, just before
the fry rise from the bottom, to set over each nest a cylindrical screen of cheese-
cloth supported on a frame of band iron, first removing the wooden nest frame.
The screen keeps the fry together. They thrive and grow within it and may be left
there until one desires to ship them. The old fish stays outside and watches the
screen. When this supply is gone other Crustacea may be taken from the pond
with a tow net and placed inside the screen. The fry are removed from these
screens directly to the shipping cans, as wanted.

Raising the fingerlings.- — The water in one of the ponds is lowered, the old fish
seined out of the kettle and transferred to another pond; the pond is then refilled,
and the fry, now about one-half to three-fourths of an inch long, are put in. The
water in the pond is thick with Daphnia and other Crustacea, and these do not get
out when the water is drawn off. The fry feed on them and the supply is usually
sufficient; but if it gives out, a fresh supply may be gathered from another pond and
placed in the nursery pond. As the young bass grow they eat not only the Daphnia
but young Corixa, and doubtless other aquatic animals.

44

BULLETIN OF THE UNITED STATES FISH COMMISSION.

In 1901, fry one-half to three-fourths inch long were introduced into the nursery
pond on July 12; on August 5 they were seined out and shipped, and were then 2 or
3 inches long. They had had none but the natural food. In three months these
fish, under the same conditions, are 4 to 6 inches long.

I have spoken so far of the small- mouthed bass, and it remains to say something
of the large-mouthed, with which my experience is more limited. It is less necessary
to resort to pond culture with them since, owing to the habit of the fry of keeping in
a close swarm, t hey may be readily seined from their natural waters shortly after
they have left the nests. In cultivating them in ponds I use the shielded nests
already described, but make the bottom of some fiber, preferably Spanish moss
bedded in cement, as has been suggested by Mr. Stranahan. This imitates the
natural nest bottom and gives better results in our locality than the gravel nest. I
do not place screens about the nests, since the young fry are so small that it is
difficult to hold them ivitli a screen, and since they may be readily taken with a
seine when wanted. I allow the large-mouthed fry to leave the nests with the parent
fish and seine them when wanted.

Finally, I will sum up what seem to me to be important points in pond culture
of small-mouthed black bass, the ponds being constructed, as is usual, on the model
of a natural pond with a central kettle and shallow shore region, well grown up
with water plants, and supplied with lake or brook water:

1. Fish should be so fed (with minnows) as to be in good condition in the spring.

2. They should be sorted into the ponds in the spring in about the proportion of 4 males to 6

females.

3. Shielded nests should be used, arranged as already described — about 1 to each 100 square feet of

shallow water.

4. The gravel in the nests should be carefully selected; it should contain sand and plenty of small

stones.

5. Water on the nesting-grounds should be kept constantly at a level between 18 inches and 2 feet.

G. The water temperature should be kept constantly between GO and 75° F. (in our locality).

7. Roily water should be, as far as possible, kept out of the ponds during the spawning season.

8. Fish should not be disturbed until the eggs are hatched.

9. The nests of the small-mouthed bass should be screened just before the fry rise from the bottom.

10. The water should contain an abundance of natural food for the fry.

The processes described are perhaps susceptible of improvements, viz:

1. Special nursery ponds might be provided for rearing finger lings.

2. It is perhaps desirable to have the nest frames shielded on three sides instead of two sides, and

made with a bottom; then when the fry rise from the nest, close the fourth side of the
nest frame by sliding a screen into it. In this way it would not be necessary to remove
the nest frame and put a screen over it, but the frame could be left in place and the open
side closed with a screen.

3. If the ponds were made with clay bottoms, the water supply could be entirely shut off during

the breeding season, if necessary.

Contributions from the Biological Laboratory of the U. S. Fish Commission
Woods Hole, Massachusetts.

HEARING AND ALLIED SENSES IN FISHES.

By G. H. PARKER,

Assistant Professor of Zoology , Harvard University.

It is a well-known fact that many fishes are extremely sensitive to disturbances
in the water such as are caused by splashing with an oar, stamping in a boat, or
striking the side of an aquarium. When, for instance, the opaque wall of a fish
tank containing young king-fish, sea-robins, or killi-fish is struck a vigorous blow
with the fist, the fishes usually respond by giving a short, quick leap, and, if such
blows are frequently repeated, surface fishes are often driven to the bottom and kept
there. Notwithstanding the sensitiveness indicated by such reactions, most of these
fishes appear to be unaffected by loud talking or other like noises originating in the
air. Fishermen are familiar with these peculiarities and often take them into
account in the practice of their art.

Such facts as these are also usually accepted as evidence that fishes can hear
(as an example, compare the statements made by W. C. Harris in Dean Sage’s
“Salmon and Trout,” 1902, p. 311), but a simple experiment will show, I believe,
that this assumption is not necessarily correct. If one end of a wooden rod is
vigorously tapped while the other is beneath the level of the water a disturbance
is produced that vdll call forth an obvious response from most fishes of moderate
sensitiveness. Such a disturbance will likewise affect a human being, for if one
holds the head beneath the water the vibrations from the rod can be easily heard,
and if the hand be placed in the water near the rod they can be distinctly felt.

Since, as Muller (1848, p. 1229) long ago pointed out, we can feel as well as hear
these vibrations, it follows that such evidence as that already given can not be
accepted as conclusive proof that fishes hear, for it is conceivable that their responses
maybe entirely through their sense of touch, i. e., dependent on their skins. More-
over, fishes possess a special system of tegmentary sense organs, the lateral -line
organs, which are completely absent from us, and it may be that these are in some
way the recipient organs for the disturbances already described. When, therefore,
a fish responds to water vibrations of the kind mentioned, we are not justified in
concluding that it hears, for it may respond through the skin or the lateral-line
organs and not through the ears.

45

46

BULLETIN OE THE UNITED STATES EISH COMMISSION.

It may be reasonably asked at this point, What constitutes hearing? Every-
one will agree, I believe, that the sensation we get through the skin from a vibrating
rod in water should not be called hearing, and what is true for us should hold for
the lower vertebrates. Hearing in these animals may therefore be defined as that
sensory activity resulting from a stimulation of the ear by material vibrations. This
is in essential accord with the definition given by Kreidl (1895, p. 461) to the
effect that hearing is that sensation which is mediated by the nerve that is homolo-
gous with the auditory nerve of man. When, therefore, a fish responds to sound
vibrations the question at once arises whether the stimulus is received by the skin,
the lateral-line organs, or the ear. And until this question can be answered, at least
so far as the ear is concerned, the query whether fishes hear or not must remain
open. In dealing with this general subject I shall take up, first of all, the question
whether fishes respond to sound vibrations through the ears.

THE EARS.

Introductory. — The internal ears of fishes were described as early as 1610 by
Casserius, and were studied in some detail in the following century by Geoffroy,
Scarpa, Comparetti, and Hunter. The attitude taken by many of these early
workers on the question of the ability of fishes to hear or not is well illustrated by
a quotation from Hunter (1782, p. 388), who at the conclusion of his paper on the
organs of hearing in fishes made the following statement:

As it is evident that, fish possess the organ of hearing, it becomes unnecessary to make or
relate any experiment made with live fish which only tends to prove this fact; but I will mention
one experiment to shew that sound affects them much and is one of their guards, as it is in other
animals. In the year 1762, when I was in Portugal, I observed in a nobleman's garden, near
Lisbon, a small fish-pond full of different kinds offish. Its bottom was level with the ground and
was made by forming a bank all round. There was a shrubbery close to it. Whilst I was lying
on the bank, observing the fish swimming about, I desired a gentleman, who was with me, to take
a loaded gun and go behind the shrubs and fire it. The reason for going behind the shrubs was
that there might not be the least reflection of light. The instant the report was made the fish
appeared to be all of one mind, for they vanished instantaneously into the mud at the bottom,
raising, as it were, a cloud of mud. In about five minutes after they began to appear, till the whole
came forth again.

This passage shows very clearly that in the opinion of Hunter the internal ears
of fishes, like those of the higher vertebrates, are organs of hearing. Without further
experimental evidence this view was accepted by Muller (1848, p. 1238) in his well-
known chapters on the physiology of the senses, and by many other eminent authori-
ties, such as Owen (1866, pp. 342 and 346), Gunther (1880, p. 116), and Romanes
(1892, p. 250). To these investigators the presence of the internal ears seemed, as
it did to Hunter, sufficient ground for concluding that these animals could hear.

Within recent years, however, this opinion has been called in question, or even
denied. Some of the grounds for this change of view may be stated as follows:
Bateson (1890, p. 251), in some investigations on the sense organs and perception of
fishes, observed that the report from the blasting of rocks caused congers to draw
back a few inches, flat-fishes (like the sole, plaice, and turbot) to bury themselves,
and pouting to scatter momentarily in all directions; other fishes seemed to take
no notice of the report. When the side of a tank containing pollock or soles was
struck witli a heavy stick, the fishes behaved as they did toward the report of the

HEARING AND ALLIED SENSES IN FISHES.

47

blasting. Pollock did not respond, however, to the sound made by rubbing a wet
linger on the glass window of an aquarium or to the noise made by striking a piece
of glass under water with a stone, provided the means of producing the noise was
not seen by the fishes. Bateson concluded that, while it may be regarded as clear
that fishes perceive the sound of sudden shocks and concussions when these are
severe, they do not seem to hear the sounds of bodies moving in the water but not
seen by them.

Without knowledge of Bateson’s observations, Kreidl (1895) carried out a series
of experiments with the view of testing the powers of hearing in the gold-fish, Car-
assius auratus. This species was chosen because of the ease with which it could be
kept in the laboratory and, further, because it is one of those fishes that have long-
been reputed to come at the sound of a bell. After an extended series of experi-
ments, Kreidl (1895, p. 458) concluded that normal gold-fish never respond to sounds
produced either in the air or in the water, though they do react to the shock of a
sudden blow given to the cover of the aquarium. Individuals rendered abnormally
sensitive by strychnine gave no response to the sound of a tuning-fork or a vibrating-
rod even when these were in contact with the water, though the fishes responded at
once to such slight shocks as tapping the aquarium, etc., or even clapping the
hands vigorously in the air.

To test whether these responses were dependent upon the auditory nerves,
Kreidl removed these nerves and the attached ear-sacs from a number of individ-
uals, and, after poisoning them with strychnine, subjected them to stimulation by
sound. In all cases they were found to respond precisely as the poisoned animals
with ears did. Kreidl, therefore, concluded that gold-fishes do not hear by the
so-called ear, but that they react to sound-waves by means of an especially developed
cutaneous sense, or, to put it in other words, the gold-fish feels sound but does not
hear it (Kreidl, 1896, p. 581).

After having reached this conclusion, Kreidl was led to take up a specific case
of the response of fishes to the sound of a bell, and an opportunity for doing this
was found at the Benedictine monastery in lvrems, Austria. Here the trout of a
particular basin were said to come for food on the ringing of a bell. Kreidl (1896,
p. 583), however, found that they would assemble at sight of a person and without
the ringing of the bell. If they were not then fed, they soon dispersed and no
amount of bell-ringing would induce them to return. If, however, a pebble or a
small piece of bread Avas thrown into the water they immediately swam vigorously
toward the spot where the disturbance had occurred. Moreover, if a person
approached the basin without being seen and rang the bell vigorously no response was
observed. From these facts Kreidl (1896, p. 584) concluded that the assembling of
the fishes was brought about through sight and the cutaneous sense, and not through
hearing, and that the conclusion reached with the gold-fish might be extended to
other kinds of fishes.

Kreidl’s conclusions were supported by the observations of Lee (1898), who
studied the reactions of several species of fishes to such sounds as the human voice,
the clapping of hands, and the striking of stones together in air and under water.
In all of his experiments Lee (1898, p. 137) obtained no evidence whatever of the
existence of a sense of hearing, as the term is usually employed, although he found
that the fishes were exceedingly sensitive to gross shocks, such as the jarring of their

48

BULLETIN OF THE UNITED STATES FISH COMMISSION.

tank or concussions upon its walls. Lee, moreover, called attention to the fact that
the papilla acustica basilaris, which is the special organ of hearing in the internal
ears of the higher vertebrates, did not occur in the fishes. From the observations
and experiments of Bateson and of Kreidl, and from his own work, Lee (1898, p. 138)
believed that the conclusion was justified beyond doubt that fishes do not possess the
power of hearing, in the sense in which the term is ordinarily used, and that the
sole function of the ear in fishes is equilibration.

The generalization to be drawn from the work just summarized, viz, that fishes
do not hear, though they may respond to sound-waves by the skin, has seemed to
me not wholly in accord with certain well-known facts in the natural history of these
animals. Among these facts may be mentioned the undoubted ability on the part
of some fishes to make sounds. If a fish has this power it might naturally be sup-
posed to hear the sounds it makes. Lee (1898, p. 137) has called attention to the
small number of sound-producing fishes as evidence against the view that fishes in
general hear. But the fact that there are such fishes has always appealed to me
in quite the reverse way and should, in my opinion, serve to indicate the species
most worthy of attention in any investigation of the sense of hearing. It must be
admitted, however, that fishes may possibly produce sounds that they themselves can
not hear, but that other animals may hear and take warning from. Thus when small
swell-fish, Chilomycterus scJiiepji, are thrown into a tank containing hungry scup,
Stenotomus chrysops , they are immediately, set upon by the latter. In defense the
swell-fishes inflate themselves with sea water till their tegmentary spines stand out
rigidly, and at the same time they make a peculiar sound by gritting the two front
teeth of the lower jaw against the inner surface of those of the upper jaw. It is not
known that this sound is heard by the swell-fish, though it may be. All that one can
say with certainty is that the sound seems to be directed against the foe, for it is
made, so far as I know, only when the swell-fish is molested. Granting, however,
that the swell-fisli does not hear its own sound, one would still be rash to conclude
that this was an argument against the hearing of fishes, for the vast majority of
animals toward which the sound is directed are fishes themselves, and these pre-
sumably hear the sounds.

Another good instance of the production of sound by a fish is found in the
squeteague or weak-fish, Cynoscion regalis. The grunting noise made by this fish is,
however, produced only by the males, and this specialization is very difficult to under-
stand unless one assumes an ability on the part of one or other sex to hear. Since
the sounds made by both the swell-fish and the squeteague are in no sense shocks or
concussions but resemble more closely, in rate of vibration and in intensity, such
sounds as might be obtained from the ordinary action of an instrument like a tuning-
fork of low pitch, it seems to me that they afford evidence in favor of the sense of
hearing rather than the reverse.

A second reason for questioning the generalization advocated by Kreidl, and
by Lee, is the character of the observations upon which it is based. Both authors
state that no positive evidence in favor of hearing could be obtained. But it must
be borne in mind that in many animals known to possess a sense of hearing the
auditory reflexes are perhaps the least conspicuous of any connected Avith the more
important sense organs, and that consequently the most careful scrutiny of the
movements of fishes must be made before one can with certainty declare that hear-
ing is absent. A perusal of the papers already summarized led me to the conclusion

Bull. U. S. F. C. 1902. (To face page 48.)

Plate 9.

1.

2.

3.

4.

5.

Side view of aquarium, showing sounding apparatus at right-hand end and suspended glass cage in which the fish were confined.
End view ol aquarium, showing sounding apparatus.

Dorsal view of brain of Fundulus heteroclitus, dissected to show positions of the roots of the fifth and seventh nerves (V), the roots of
the ninth and tenth nerves (X) , and the internal ear as indicated by its otolith. Enlarged about twice.

Dorsal yew of head of J'lmdulus heteroclitus , to show the region where the following nerves were cut: The fifth and seventh (V), the
_ eighth (VIII), and the lateral-line nerve (X). Enlarged about twice

Side view of Fundulus heteroclitus , showing the region where the lateral-line nerve was cut (X). Slightly enlarged.

HEARING AND ALLIED SENSES IN FISHES.

49

that something more might he attained in this direction, and I therefore resolved to
give particular attention to the reactions of a few fishes with the view of ascertaining
whether or not they showed any evidence of hearing.

At the outset I thought it best to experiment on some common sound-producing
species, and for this purpose I did some preliminary work on the swell-fish ( Ckilo -
mycterus schoepft), the squeteague ( Cynoscion regalis), and the sea robin ( Prionoius
carolinus) . To all of these, practical objections were found, and I was at. last obliged
to abandon them for fishes that produce no sounds. Among these, three species
were found to be especial ly sensitive to slight vibrations — the king-fish ( Menticirrhus
saxatilis), and the two common species of killi-fish ( Fundulus majalis and F. hetero-
clitus). Because of the great abundance of F. heteroclitus, the ease wil l: which it
could be operated upon, and its great hardiness, I chose it for study, and the observa-
tions recorded on the following pages, unless otherwise stated, refer to this species.

The ears in Fundulus heteroclitus. — When a tank containing a number of Fun-
dulus heteroclitus is struck with the open hand so that the fish can not see the
movement of striking, they respond to the vibrations by springing suddenly an inch
or so through the water. The question to be considered is whether these vibrations
stimulate the fishes through the skin, the lateral-line organs, the ears, or some
combination of these. If it could be shown that the ears were not stimulated by
the vibrations, it seems to me that we would have evidence pointing to the conclusion
that the fishes did not hear. If on the other hand it could be demonstrated that the
vibrations did stimulate the ears, the evidence would lie conclusive that the animals
possessed the sense of hearing. To test these points considerable experimentation
was necessary.

Much of the work that has been carried out heretofore has been done with
sound generated in air but intended to affect fishes in water. That this method is
extremely inefficient I found by trying the following experiment. If a dinner bell
is rung in the air by a person standing breast-deep in water, it will, of course, be
heard easily by a second person standing in a similar way a yard or two off. If,
however, the second person puts his head under the water during the ringing of the
bell the sound seems to cease almost entirely and is not again heard clearly by the
diver till he emerges. In like manner a bell rung or hit with a stone under water
is heard, at best, very faintly by a person standing in the water unless his head is
under the surface. In other words, the plane separating air and water is, under
ordinary circumstances, an almost impenetrable one for most sounds, whether they
are generated on one side or the other of it, and many of the negative results
obtained by previous investigators on the sense of hearing in fishes may have been
due not so much to the absence of hearing in the animals experimented upon as to
their inaccessibility to the sound, or at least to sound of an intensity sufficient to
stimulate. This difficulty has been recognized by Kreidl, and in devising apparatus
I have profited by his experience and used sound-producing appliances that were
in direct contact with the water containing the fishes.

The chief piece of apparatus that I used consisted of an ordinary marine
aquarium (pi. 9, fig. 1) with a slate base, two heavy glass sides, and originally two slate
ends, one of which, however, I replaced by a piece of deal board free from knots, to
serve as a sounding-board. The inside dimensions of the aquarium were as follows:
depth, 40 cm. (16 in.); breadth, 37 cm. (15 in.); and length, 87 cm. (35 in.). To the
middle of one edge of the sounding-board a stout beam of wood was attached

F. C. B. 1902—4

50

BULLETIN OF THE UNITED STATES FISH COMMISSION.

so that it stood out horizontally about. 1 meter (TO in.) in the plane of that#end
(fig. 2). From the free end of the beam a bass-viol string was stretched to the
opposite side of the sounding-board. This string could be tightened by a bolt and
nut at the free end of the beam, and it was made to pass over a bridge placed near
the middle of the sounding-board. The length of the string from the attached end
on the sounding-board to the bridge was 25 cm. (10 in.), and from the bridge to
the attachment near the free end of the beam 1.15 meters (45 in.). Thus the end
of the aquarium might be regarded as something like a large one-stringed bass viol
resting sidewise, with the sounding-board for a body and the beam for a neck.

When the string was tightened and plucked or bowed a good tone was obtained,
which was transmitted directly through the sounding-board to the water within the
aquarium. On keying the string up to a good clear tone, I found by writing off its
vibrations on a revolving cylinder that it produced on an average 40 per second,
and I retained this pitch by frequent adjustment for the experiments that I subse-
quently performed. I was led to adopt this low tone since most of the noises that. I
have heard fishes make were in the nature of low-pitched grunts.

Each time the string was plucked the note began with maximum intensity and
then gradually died away. It was, consequently, impossible to get any very signifi-
cant record of the intensity, but I endeavored to use the apparatus in a uniform way
by drawing the string out a constant distance from its position of rest each time
I plucked it. The distance usually employed was about 1.5 cm. (0.6 in.). The
amount of weight required, when hung at the middle of the longer segment of the
string, to depress it 1.5 cm. was found to be about 2.15 kilograms (4.75 pounds),
so that each time the string was liberated on being plucked in the usual way, it
moved forward with an initial force equal to the pull of 2.15 kilograms, a rough
measure of the maximum intensity of the sound produced.

The fishes to be experimented upon were not allowed to swim unrestricted' in
the aquarium, but they were placed in a small cage (fig. 1) suspended from a cord
attached at its ends to the Avails of the room. Thus the support for the fish cage
Avas entirely independent of the walls of the aquarium and any vibrat ion that reached
the fishes must have done so almost entirely through the water. The cage could be
moved in a horizontal direction back and forth on the cord, and thus the fish could
be placed at any desired distance from the sounding-board up to 75 cm. (30 in.).
The inside measurements of the cage Avere as follows: Height, 10 cm. (4 in. ) ; length,
20 cm. (8 in.); and breadth, about 10 cm. (4 in.). The bottom of the cage Avas wood,
padded on the inside Avith cotton wool covered with cloth to provide a deadened
surface on which the fishes might rest. The top and three sides were glass; the
fourth side was made of coarse netting to retain the fish but to interfere as little as
possible Avith the entrance of sound, and this side Avas always directed toward the
sounding-board. As the fishes averaged about 7 cm. (2f in.) in length, the cage
gave them ample room for moving about.

My plan was to introduce fishes in various conditions into the cage, and, after
they had become accustomed to their surroundings, to subject them to stimulation
by sound and observe their reactions. I found it desirable to experiment Avith three
classes of fishes; first, normal ones for a basis of comparison; secondly, fishes from
which the ears had been removed; finally, fishes in which the general integument
had been rendered insensitive, but in which the ears were intact. The methods of

HEARING AND ALLIED SENSES IN FISHES.

51

obtaining fishes in these conditions and the responses that they showed will be
described for each class of fishes.

Normal fishes. — When a normal fish is first liberated in the cage it swims vigor-
ously about for a few moments, after which it may, sooner or later, come to rest on
or near the bottom. The animals are extremely quick-sighted, and, if after they
have come to rest the observer makes any sudden movements near the aquarium,
they are very likely to begin active swimming anew. It is, therefore, extremely
necessary to work in such a way that all movements, and particularly quick ones,
are made out of sight of the fish. When the fish is resting on the bottom of the
cage, two sets of motions will usually be observed : first, the respiratory movements
of the operculum; and secondly, the alternate vibratory movements of the pectoral
fins. The opercular movements, as might be expected, always continue, but the
movements of the pectoral fins, which seem to be connected also with the respiratory
function, often cease entirely.

When a fish has become quiet, except for the respiratory movements, the vibra-
tions from the string may call forth any of four kinds of responses. The first of
these is the vibratory movement of the pectoral fins, either a few slight beats, if the
fins were previously at rest, or an increased rate or extent of swing if they were
previously in motion. The vibration of the string at the intensity ordinarily
employed almost invariably called forth this reaction; thus, in ten observations
taken from each of ten fishes at a distance of about 25 cm. (10 in. ) from the sounding-
board there were 96 pectoral-fin responses and 4 failures. Since this response is so
readily observed, it has afforded one of the most satisfactory criteria of stimulation.

The second form of response is a change in the rate of the respiratory movements.
In a quiescent fish measuring 8 cm. (3.2 in.) in length the respiratory rate was 114
per minute. On stimulating by sound this rate rose suddenly to 138 per minute for
some ten or a dozen movements and then fell rapidly to about the former number.
This is probably a very usual form of response, perhaps quite as much so as the
movement of the pectoral fins, but the shortness of its duration and its inconspicu-
ousness make it less satisfactory as an indication of stimulation than that afforded
by the pectoral fins. If the sound from the string is of considerable intensity, the
third form of response may appear, a slight motion of the caudal fin, beginning
usually at the dorsal edge and proceeding as a wave ventrally. Finally, with strong
stimulation, the fish may make a short but quick spring forward.

All these reactions have been obtained from fishes even at 75 cm. (30 inches)
from the sounding-board, although the springing movements are more frequently
observed when the animals are not so far from the source of sound. One very inter-
esting fact about these reactions is that they can not be repeated rapidly for even a
short period. A fish that responds to the first stimulus by a spring, may react to the
second or to the third only by moving the pectoral fins, and to the fourth in no
observable way. It is only when a considerable period of rest intervenes that the
reactions may be repeatedly obtained; and I have found that the minimum period
of rest is not far from one minute, though, even then, reactions may sometimes fail
to appear.

Earless fishes. — The removal of the ears from a fish is a serious operation, but it
is one which, after a little practice, may be accomplished with success and from
which the fishes generally recover. These animals are easily etherized by putting

52

BULLETIN OF THE UNITED STATES FISH COMMISSION.

them in sea water containing enough ether to give it a strong ocjor. On being trans-
ferred to pure sea water they quickly recover, and an individual that I etherized six
times in the course of one afternoon finally recovered without showing any ill effects.
The first method I used in operating on the etherized fishes was to open the cranium
in the region of the ears and, after cutting the auditory nerves, to remove those
nerves and the attached ear-sacs. These parts were easily identified from the fact
that the auditory nerve emerges from the medulla almost exactly ventral to the cleft
between that organ and the optic lobe, and the ear-sac, which is only partly sur-
rounded by cartilage, lies in the cranial cavity only slightly peripheral to the point
where the nerve leaves the medulla (pi. 9, fig. 3). After the operation the fishes
were returned to pure sea water and, notwithstanding the exposure of the brain, a
considerable number recovered and survived. One of these I kept for more than six
weeks, and, though its swimming was characteristically irregular, it was alert and
active and, except for a brief intervening period, it fed normally.

From the operation just described about one fish in ten recovered. This pro-
portion was greatly increased by a second form of operation in which the auditory
nerve was cut without opening the cranium (fig. 4). After a little practice I found
that this could be done with great certainty and about eight out of ten fishes usually
recovered. All fishes that had been operated on were kept at least twenty-four
hours before they were subjected to experimentation.

Fishes in which the auditory nerves have been cut have very characteristic
reactions. When resting or when swimming slowly they behave for the most part
as normal fishes do, and, in fact, are often undistinguishable from individuals upon
which no operation has been performed. When, however, they are stimulated to
rapid locomotion, they swim either in irregular spirals, the same individual revolving
sometimes to the right and sometimes to the left, or they turn over and over in
irregular circles without accomplishing much real progression. This loss of
orientation on attempting rapid locomotion has for some time been recognized as
indicative of one of the chief functions of the ears in fishes — i. e., equilibration.
It is probable that in resting or in swimming slowly the fish depends upon the eye
for orientation, but in quick movements the ears come more into play, and hence
after their loss quick movements are accompanied with lack of orientation. The
forced movements thus observed may be taken with perfect certainty, so far as my
experience goes, as evidence of the successful outcome of an attempt to cut the
auditory nerves, for in the few cases where these movements failed to appear,
subsequent dissection showed that the nerves had not been cut, and in all instances
where the movements were observed and the animals afterwards dissected, the
nerves were found severed.

A second feature of interest that generally characterized fishes with severed
auditory nerves was the color that they finally assumed. Under ordinary circum-
stances the color of this species is a light greenisli-gray. When etherized the fishes
become very dark, with a mottling of blue-green on the sides and belly. After
recovery from cutting the auditory nerves, the dark coloration disappears and the
fish assumes a tint even paler than that of a normal individual. This tint is retained
throughout life. Etherizing probably influences the chromatophores of the skin
directly, but cutting the auditory nerves introduces changes that are probably
dependent upon the nervous control of the chromatophores.

HEARING AND ALLIED SENSES IN FISHES.

53

When the earless fishes were tested in the sounding apparatus, they yielded
very interesting results. Unlike the gold-fishes experimented on by Ivreidl (1895),
they differed markedly from normal fish. In an extended series of observations on
over 20 fishes I never once observed with certainty the springing reflex as a result
of sounding the bass-viol string. The fishes were usually very active, and I was
never able to ascertain with certainty whether they showed a change in the respira-
tory rate on stimulation. The pectoral-fin movements, however, were observed with
much certainty. On 10 earless fishes I succeeded in getting 10 observations each
to sound stimuli at about 25 cm. (10 inches) from the sounding-board. The total
result was that in 82 observations there were no reactions and in the remaining 18
the reactions were at best slight ones. As the fishes often moved the pectoral fins
without apparent cause, some of the 18 reactions may have been accidental coinci-
dences, but others were so precise and typical that I am convinced they were due to
stimulation. Earless fishes, therefore, differ from normal ones in that their pectoral-
fin responses to vibrations from the bass-viol string are enormously reduced, though
not entirely obliterated.

Fishes with insensitive skins. — For reasons already given it is imperative, before
drawing conclusions from the condition of earless fishes, to examine the evidence
afforded by those whose general surface has been rendered insensitive. In this way
it is possible to ascertain what part the integument plays in the reception of sound
vibrations. I had hoped that the integument of Fundulus lieteroclitus could be
rendered insensitive by immersing the fish for a short time in a solution of cocaine,
but all attempts in this direction proved failures, since the drug acted much more
vigorously as a poison than as an anaesthetic, and I was finally obliged to abandon
this method altogether and resort to nerve-cutting.

The following operation performed on etherized fishes insures an almost complete
insensibility of the surface. The fifth and seventh cranial nerves can be cut just
posterior to the eyeball (pi. 9, figs. 3 and I), the lateral-line branch of the tenth nerve
can next be cut at the posterior edge of the pectoral girdle (fig. 5), and finally the
spinal cord can be severed at the fourth or fifth vertebra. Severe as this operation
is, almost all fishes recover from it and respire and feed normally, though they
seldom live beyond two weeks after the operation.

Fishes that have recovered from this operation show certain well-marked charac-
teristics. The integument, particularly l.hat of the dorsal surface, is unusually
dark, as a result of the expanded condition of the chromatophores. The fish’s
mouth is gaping and motionless in consequence of the motor portion of the fifth
nerve having been cut. This condition, however, does not interfere with respiration
or with the sucking in of pieces of food, an act which the fish performs with avidity.
Since in cutting the fifth and seventh nerves, the three small nerves to the muscles
of the eyeballs, the third, fourth, and sixth, must also be cut, the eyes are motion-
less and usually protrude somewhat. Finally, as a result of cutting the spinal cord,
the whole trunk of the animal is, as a rule, passive and is drawn after the head, the
swimming being performed by the pectoral fins. Since the greater part of the cord
is intact, a more or less vigorous stimulus applied to the trunk is followed by move-
ments in the dorsal, anal, and caudal fins, or even by a locomotor response of the
whole trunk, but such movements are made only after special stimulation, and the
trunk is ordinarily carried passively, like a paralyzed appendage. As a result of

54

BULLETIN OF THE UNITED STATES FISH COMMISSION.

having so little of the normal locomotor apparatus intact, the fishes often swim ven-
tral side up, for the action of the pectoral fins is not always sufficient to overcome
the physical effects of the specific gravity of the fish’s body.

Fishes that have recovered from the operation just described have intact the
ears, the central nervous organs from the anterior end of the brain caudad to the
fourth or fifth vertebra, and the sensory and motor apparatus for the region of
the gills and the pectoral fins. Excepting in these two rather restricted regions, the
whole integument is insensitive, at least so far as its capacity to originate impulses
to movements in the gills or pectoral fins is concerned. Such fish, therefore, are in
a condition to receive stimuli through the ears and to respond by respiratory or
pectoral-fin movements.

The reactions that these fishes showed to the sound apparatus were surprisingly
clear and decisive. From the nature of the operation one would not expect them to
be able to give the sudden spring that the normal fishes often showed, and, as a
matter of fact, such responses were never observed. Were the skin of the trunk
sensitive, it is conceivable that the caudal-fin reaction might occur, for the cord,
though severed from the rest of the central nervous organs, was in itself intact.
Caudal-fin reactions were, however, also never observed. The respiratory reactions
and the pectoral-fin responses occurred with great regularity. When the bass-viol
string was made to vibrate, the respiratory rate increased for a very brief period.
In a fish 7 cm. (2f inches) long the rate previous to stimulation was 120 per minute;
immediately after stimulation it was 156. The reactions of the pectoral fin were also
well marked. In ten observations on each of ten animals at a distance of about 25
cm. (10 inches) from the sounding-board the pectoral- fin responses occurred 94 times
in the total hundred. This is in close agreement with the normal fishes and in strong
contrast with the earless ones. So far, then, as reactions to the vibrating chord are
concerned, these fishes show the essential characteristics of normal individuals.

Discussion of the results of the experiments. — It is clear from the experiments
described in the preceding sections that fishes whose ears were rendered functionless,
but whose skins were normally sensitive, reacted only slightly to the stimulus from
the sound-producing apparatus, whereas those with insensitive skins but functional
ears responded to this stimulus, as far as their conditions would permit, almost exactly
as normal fishes did. It might be assumed that the failure to respond on the part of
earless fishes was due not to the loss of the ear, but to the shock of the operation
they had undergone. This, however, does not seem to be the case, for, after the
fishes had recovered from the immediate effects of the operation, they were active,
fed well, and sometimes lived many weeks. Moreover, if the operation were as
severe as is implied in the above assumption, one might expect some indications of
this in fishes in which only one auditory nerve had been cut. As a matter of fact,
immediately after this operation fishes with only one ear intact did swim irregularly,
but in from six to eight hours this tendency disappeared entirely, and the fish in its
quickness, precision, and normality of response became, so far as my obsei’vations
went, absolutely indistinguishable from a normal individual. Further, fishes with
the fifth, seventh, and lateral-line nerves and spinal cord cut have without doubt
suffered a more severe shock than those that have had only the eighth nerve cut,
and yet the pectoral- fin reactions of the former were essentially normal. It therefore
seems, to me that the great reduction in the number of pectoral-fin reactions of

HEAKING AND ALLIED SENSES IN FISHES.

55

earless fishes is due to the loss of the ear as a sense organ and not to secondary
complications accompanying the operation.

Although some of the observations recorded on the preceding pages make it
certain that in these fishes the ears are stimulated by disturbances such as those
set up in the water by the sounding apparatus, it may still fairly be asked whether
these disturbances are in the nature of sounds. When the bass-viol string attached
to the aquarium was plucked, a series of sound waves of diminishing intensities
was delivered to the water. To ascertain something of the nature of this sound I
immersed my head in the water of the aquarium and had an assistant pluck the
string in the usual way. The sound that I thus heard was, so far as I could judge,
of nearly the same pitch as that which the string gave to the air and of only slightly
greater intensity. This sound certainly reached the fishes.

The sounding apparatus, however, did more than give rise to this sound. When
the string was plucked two things besides the production of sound certainly happened :
First, the whole aquarium, including its supporting table, trembled slightly, and,
probably as a consequence of this, ripples started from the ends and sides of the
aquarium and proceeded toward the center. These ripples, though chiefly surface
effects, indicated a wave motion that penetrated the water to some extent, and that
was doubtless the cause of the very slight swaying movement of the fish cage occa-
sionally noticed after the string had been vigorously plucked. Moreover, a distinct
tremor could be felt in the water when the hand was held 5 to 8 cm. (2 to 3 inches)
from the sounding-board and the string was plucked. The question naturally arose
whether the fishes did not respond to the movement of the aquarium as a whole or
to the wave movement indicated by the ripples rather than to the true sound waves.

To answer this question, at least so far as the ripple movement was concerned,
I was led to study the reaction time of the fishes. Unfortunately circumstances
prevented me from reducing this to a very accurate process ; but, by listening to the
beat of a chronometer and at. the same time watching the fish, I am confident that
the fin reactions occurred in less than 0.2 second after the string had been plucked.
The sound waves and ripples mentioned above traveled from the sounding-board
toward the fish at very different rates. The sound waves must have passed over
the 25 cm. of water between the sounding-board and the fish almost instantly. The
surface ripple traveled much less rapidly and its rate could be easily measured.
This proved to be a meter (39f inches) in t.8 seconds; hence, to traverse 25 cm. (10
inches) the ripple required about 1.2 seconds. Since the fishes responded in less
than 0.2 second, they must have reacted to something other than the disturbance
indicated by the ripples.

Having eliminated the ripples as the initial stimulus for the fishes, it remained
to be shown whether this stimulus was the movement of the whole aquarium or the
sound waves proper. I succeeded in doing this by substituting an electric tuning-
fork for the bass-viol string. The tuning-fork was placed so that its base was within
about a millimeter (dT inch) of the sounding-board. The iron frame holding the fork
rested on supports made of rubber bottle-stoppers. These flexible supports allowed
the fork to be moved enough to bring its base into contact with the sounding-board
without moving the supports over the surface on which they rested. As this could
be done without any initial jar, it was possible to communicate to the water in the
aquarium a sound of uniform intensity and pitch without moving the aquarium as

56

BULLETIN OF THE UNITED STATES FISH COMMISSION.

a whole and also without producing any ripple. The fork, moreover, produced atone
much purer than that obtained from the string. It had a pitch of 128 vibrations per
second.

Earless fishes, when subjected to sound waves from the tuning-fork, showed
nothing that I could identify as a reaction. Normal fishes and fishes with normal
ears but insensitive skins very usually reacted by pectoral -fin movements. The
occasional failure to respond was attributed by me to the faintness of the vibrations,
for the most intense sound obtained from the fork was much less than that produced
by the bass-viol string. That the fishes, however, always did react, even to this
relatively faint tone, was pointed out to me by my friend Dr. F. S. Lee, who while
watching one of the experiments thought lie detected an increase in the respiratory
rate even when no pectoral-fin reaction occurred. Subsequent study showed this to
be entirely correct, for, irrespective of pectoral-fin responses, at each sounding from
the tuning-fork an increase of the respiratory rate did take place for a very short
period. There is, then, no question but that these fishes respond to sound waves,
and, since this response is through the ear, I conclude that Fundulus heteroclitus
may be said to hear. Since I never succeeded in getting reactions of any kind to
the tuning-fork from earless fishes with skins and lateral lines intact, I have no reason
for believing that these parts are stimulated by true sound waves, and I attribute
the responses that earless fishes occasionally showed to the vibrating bass-viol string
not to the action of its sound waves on the skin or the lateral-line organs, but, as
will be shown later, to the influence of the accompanying movement of the whole
aquarium and its contained water on these parts.

Although the experiments already described remove every reasonable doubt
from my mind as to the ability of these fishes to hear, the objection may still be
raised that the conditions under which they were carried out were so artificial that
they may be said to have almost no bearing on the ordinary habits of Fundulus, and
it must be admitted that the relatively small volume of water in the aquarium and
the character of its walls as reflecting surfaces for sound, may possibly have
introduced factors to which the fishes, in their natural surroundings, were not
accustomed. To ascertain how much weight should be given to this objection the
following experiment was tried. The sounding apparatus, consisting of the sounding-
board and the bass-viol string, was taken from the aquarium and set up in the open
water of the outer pool at the Fish Commission wharf. The fish cage was hung at a
distance of 50 centimeters (20 inches) from the sounding-board and toward the center
of the pool, which is about 100 feet wide. The sound, therefore, was as unrestricted
as that which naturally reaches these fishes. On experimenting with normal fishes,
fishes without ears, and those with insensitive skins, results were obtained essentially
like those observed in the aquarium, and I therefore concluded that the restriction
of the water in the aquarium played no essential part in the results obtained from
that apparatus. There is, thus, good reason to believe that Fundulus heteroclitus
not only hears, but that for it hearing is a normal process.

Having determined that hearing was one of the normal functions of the ears in
Fundulus, I had hoped to be able to ascertain by experiment the particular part of
the ear, if such there be, that was concerned with this sense. The internal ear in
Fundulus heteroclitus is like that in most teleosts. It consists of the usual three
semicircular canals and a large sacculus, at whose posterior end a well-developed
lagena is present. The sacculus is a thin-walled chamber, vertically flattened and

HEARING- AND ALLIED SENSES IN FISHES.

57

containing a thin, flat otolith of considerable size. Sometimes this otolith is repi’e-
sent.ed by two pieces — -a small one at the anterior end of the sacculus and a much
larger one occupying the more central part of this chamber. The lagena, which is
well separated from the sacculus, also contains an otolith. On the median face of
the sacculus is an extensive macula acustica sacculi, formed by the termination of
the major part of the eighth nerve. There is also a well-developed papilla acustica
lageme, as well as the usual three crista; acustica; ampullarum. I am unable to
state whether other sensory patches, such as the macula acustica neglecta, occur
here or not.

Having made a preliminary study of the anatomy of the internal ear, I had hoped
to be able to cut in different individuals different branches of the eighth nerve, and,
by further experimentation on fishes thus prepared, to determine the functions of
the several sense organs of the internal ear. After numerous unsuccessful attempts
I was at last obliged to abandon this plan because of the small size of the branches
and their somewhat intricate relations, and I am, therefore, in possession of no
observations that show which part or parts of the internal ear are concerned in
hearing.

THE LATERAL-LINE ORGANS AND THE SKIN.

Introductory . — The lateral-line canals were regarded by most of the earlier
investigators as glands for the production of the mucus so characteristic of the skins
of fishes. About the middle of the last century Leydig (1850, p. 171) discovered the
numerous sense organs contained in these canals, and declared that the whole system
represented a sensory apparatus peculiar to fishes. Subsequently Leydig (1868,
p. 2) expressed the opinion that these organs implied the possession of a sixth sense,
one in addition to the five usually attributed to vertebrates, though he admitted that
this sense was probably closely related to touch. Two years later the lateral-line
organs were investigated by Schulze (1870), who demonstrated that true lateral-line
organs ivere found only in the water-inhabiting vertebrates. From a study of their
structure Schulze (1870, p. 86) was led to the belief that they were stimulated by
the mass movement of the water, as when a current passes over the surface of a
fish or when the fish swims through the water. He further believed that they were
stimulated by sound waves whose length was greater than that of waves to which
the ear was adapted. In this respect they were organs somewhat intermediate in
character between those of touch and of hearing. These opinions were opposed by
Merkel (1880, p. 5-1), who pointed out the inaccessibility of the organs to moving
water in many cases, and who regarded them merely as organs of touch. The
opposite extreme was taken by P. and F. Sarasin (1887-1890, p. 54), who designated
them accessory ears, a view suggested some years previously by Emery (1880, p. 48).

The opinions thus far given were based for the most part on an interpretation
of the anatomy of the lateral-line organs, and not upon any positive experimental
evidence as to the function of these parts. Fuchs (1895, p. 467) seems to have been
the first to attempt work in this direction. His experiments were made chiefly on
the torpedo, a fish in which, in addition to the lateral line proper, two other sets or
organs, the vesicles of Savi and the ampulla; of Lorenzini, may be regarded as parts
of the lateral-line system. In an active torpedo Fuchs cut the nerves connected with
these two special sets of organs without, however, being able to detect any significant
change in the subsequent movements of the fish. He then exposed the nerve inner-
vating the vesicles of Savi, and having placed it in connection with the appropriate

58

BULLETIN OF THE UNITED STATES FISH COMMISSION.

electrical apparatus, lie found that on pressing lightly upon the vesicles a negative
variation in the current from the nerve could he detected. As this negative varia-
tion is evidence of the momentarily active condition of the nerve, it follows that
pressure differences may he assumed to he a means of stimulating the vesicles of
Savi. No such results were obtained from the nerves distributed to the ampullae
of Lorenzini, hut the nerves from the unmodified lateral-line organs in Raja clavata
and R. asterias showed negative variations when their terminal organs were sub-
jected to pressure. Dilute acids and changes of temperature were not stimuli for
any of the terminal organs tested, and Fuchs (1895, p. 474) concluded that pressure
was the normal stimulus in the skate for the lateral-line organs, and in the torpedo
for the vesicles of Savi, hut not for the ampulbe of Lorenzini.

Apparently without knowledge of the work done by Fuchs, Richard ( 1 89G, p.
131) performed some experiments on the gold-fish. These consisted in the removal
of the scales from the lateral line and the destruction of the sense organs under
these scales by cauterizing with heat, silver nitrate, or potassic hydrate. After this
operation some of the fishes were unable to keep below the surface of the water, and
though they soon died, Richard (1896, p. 133) believed that lie had evidence enough
to show that the lateral-line organs were connected with the production of gas in
the hydrostatic apparatus.

Richard’s conclusions were called in question by Bonnier (1896, p. 917), who
pointed out the severity of the operations employed by the former and intimated
that the results were more probably dependent upon the excessive amount of tissue
removed than upon the destruction of the lateral line. Bonnier (1896, p. 918)
further recorded experiments of his own in which the lateral-line organs were
destroyed by electro-cautery. Fishes thus operated upon showed two character-
istics — they could easily be approached by the hand and even seized, and they
failed to orient themselves in reference to disturbances caused by bodies thrown
into the water. Bonnier concluded from his experiments that the lateral line, in
addition to other functions, had to do with the orientation of fishes in reference to
centers in the water from which shock-like vibrations might proceed.

Lee (1898, p. 139), whose experimental methods were much the same as those
used by Bonnier, obtained some significant results, particularly with the toad-fish,
Batrachus tau. When the pectoral and pelvic fins of this fish were removed, so that
the animal might be said to be without its usual mechanical support, and the lateral-
line organs wei‘e destroyed by thermo-cautery, the animal would lie quietly for some
time, either on its side or back, and acted as though it had lost its “sense of equili-
bration.” That its condition was not due to excessive injury was seen from the fact
that a Unless fish in which an equal amount of skin had been cauterized, but in
which the lateral-line organs were intact, showed no lack of equilibration, and in its
general behavior closely resembled a normal fish. Moreover, stimulation of the
central end of the lateral-line nerve resulted in perfectly coordinated fin movements,
and Lee (1898, p. 144) therefore concluded that the organs of the lateral line are
equilibrating organs. How these are stimulated Lee does not attempt to decide,
though he suggests (1898, p. 143) that pressure changes in the surrounding medium
may be the means of stimulation.

From this brief historical resume it must be evident that thei’e is still very little
unity of opinion as to the functions of the lateral-line organs.

HEARING AND ALLIED SENSES IN FISHES.

59

The lateral line in Fundulus heteroclitus. — The lateral-line system in F. heteroclitus
presents a condition typical for teleosts. Its sense organs are contained in canals
that open by pores on the surface of the skin. A lateral-line canal as indicated by
its pores (pi. 9, fig. 5) extends along the side of the trunk from near the tail forward
to the head. Here the arrangement of the pores (figs. 4, 5) gives evidence of a man-
dibular, a suborbital, a supraorbital, and an occipital branch. By cutting the fifth
and the seventh nerves behind the eye (fig. 4), and the lateral-line nerve near the
pectoral girdle (fig. 5), the innervation of this whole system, except a small tract
above the gills, can be rendered inoperative; the sense organs in the small tract can
be easily excised. Fishes that have undergone this operation recover almost inva-
riably and in a very short time; the integument of their heads is insensitive owing
to the necessity of cutting the fifth as well as the seventh nerves; but that of
their trunks, which is of course innervated from spinal sources, retains its normal
sensitiveness, except so far as the lateral-line organs are concerned. In seeking
for evidence as to the function of the lateral-line organs, I compared carefully the
reactions of normal fishes with those in which the nerves of the lateral-line organs
had been cut.

When a normal fish is liberated in an aquarium, it swims at once to the bottom.
Here it may move about excitedly for some minutes, after which it usually begins to
make upward excursions. At first it will swim only part way to the top, returning
each time quickly to the bottom. Eventually it may make several quick excursions
to the upper surface of the Avater, and ultimately may remain there playing about
close to the top. If now any disturbance is made the fish will again swim at once
to the bottom, and only after some time Avill it return to the top, in the same cautious
way as before. Almost any disturbance seems to drive the fish to the bottom — a
flash of light on the water, a quick but noiseless movement of the observer, or an
unseen blow on the aquarium, conditions all of which suggest that the movements
of the fish are of a protective nature.

To one form of disturbance the fishes were particularly sensitive, and this was
the slight movement of the whole aquarium that occurred whenever the bass-viol
string was plucked. This movement could lie produced without the accompanying-
sound by giving a slight vibratory motion to the beam attached to the sounding-
board on the aquarium (fig. 2). It Avas remarkable how accurately the fishes
responded to this stimulus. If the fish Avas playing at the top of the Avater, the
slightest movement of the aquarium as a Avhole would cause it to descend immedi-
ately to the bottom; if it Avas on its upward course, it could be checked and made to
descend at any point; and if it was near the bottom, it could be kept there as long
as the movement continued. In all of the several hundred trials of this kind that
I made, I never found a normal fish that would remain high in the water or swim
upward while such movements were being imparted to the aquarium. Whenever
the fish was above the bottom, the response was an instantaneous downward course.

With fishes in which the nerves to the lateral-line organs had been cut, the
reactions Avere totally different. Such fishes, when left to themselves in an aqua-
rium, were scarcely distinguishable from normal ones. As with the toad-fishes
observed by Lee (1898, p. 140), the loss of the lateral-line organs seemed to interfere
in no essential respects with the movements of the animals; they were active and
quick, returned at once to a normal position when displaced, and oriented with

60

BULLETIN OF THE UNITED STATES FISH COMMISSION.

accuracy, so far as I could see, in that they at once swam away from such centers
of disturbance as come from dropping a stone in the water. In this last particular
they were very unlike the fishes reported on by Bonnier (1896, p. 918). In one
important point they differed absolutely from the normal fishes; they would swim
upward and remain near the top during even a considerable agitation of the whole
aquarium, though they would dart downward at any sudden movement on the part
of the observer. Hence these fishes must have lost their capacity to be stimulated
by the mass movement of the water, and since this defect was observed only after
the lateral-line organs had been rendered inoperative, I concluded that the normal
stimulus for these organs was a very slight mass movement of the surrounding
water. Since such movements always accompanied the sound produced by the bass-
viol string, it follows that the disturbances set up by this string must have acted as
a stimulus for the lateral-line organs as well as for the ears, and it is therefore not
surprising that earless fishes sometimes reacted when the string was plucked.

If the lateral-line organs are stimulated by a slight mass movement of the water,
it occurred to me that I ought to be able to separate, in a mixed school of fishes, those
with lateral-line organs intact from those in which the nerves to the organs had
been cut, by simply imparting a slight mass movement to the water. Under such
conditions the normal fishes ought to swim to the bottom, leaving the defective ones
above; but on trying the experiment I found that the fishes were so accustomed
to form a school that when the normal ones started for the bottom the others did
the same, and I was entirely unable by this means to separate the normal from
the defective individuals. But I finally succeeded in doing this by modifying the
experiment, in that I used only two individuals, one normal and one defective, and
agitated the aquarium only when they were widely separated. The result was very
decisive in that the normal one invariably took the initiative in descending, and in
fact was often not followed by its defective companion.

Having found the conditions under which the lateral-line organs were stimu-
lated, it is natural to inquire as to the exact nature of the stimulus. Ordinarily the
fishes were induced to react by making the whole aquarium swing at about ten
vibrations per second ; but a like reaction was obtained from the normal fishes when
a single swing, or what was as near as possible a single swing, was given to the
aquarium. The stimulus therefore is not necessarily of a vibratory kind, but con-
sists in a slight movement of the body of water as a whole. It might be supposed
that since the fish was suspended in the water, the motion of the aquarium as a
whole could have no influence on it. But it must be remembered that the fish was
somewhat heavier than the water, and that each time the aquarium was moved the
fish, from its inertia, must have lagged a little behind or, once set in motion, moved
a little ahead, and it is this slight difference in the rate of movement of the fish and
of the adjacent water that, in my opinion, induces stimulation. I am not prepared
to say how this affects the sense organs in the lateral-line canals; but it is not
impossible, as Schulze (1870, p. 85) suggested, that slight currents are thereby set
up that move and thus stimulate the bristle cells of the lateral-line organs.

The extreme sensitiveness of animals to slight motions of this kind has already
been pointed out by Whitman (1899, pp. 287 and 302) in the leech and salamander,
and I suspect that the sensitiveness of the blind fish, as observed by Eigenmann
and quoted by Whitman (1899, p. 303), may also be in the nature of a lateral-line
response.

HEARING AND ALLIED SENSES IN FISHES.

61

Having reached the conclusion that the downward swimming of the fishes could
be brought about by stimulating the lateral-line organs through slight mass-move-
ments of the water, I next attempted to ascertain the relative importance of different
parts of the lateral-line system in this reaction. I prepared one set of the fishes in
which the lateral-line nerves were cut close to the pectoral girdles, thus rendering
ineffective the lateral-line organs of the trunk while those of the head were left
intact. These individuals responded in all respects, so far as I could see, as normal
fishes did, and I therefore concluded that the lateral line proper was not an essential
part of this system of sense organs.

In the second set of fishes I cut the fifth and seventh nerves of both sides, thus
preventing the lateral-line organs of the head from acting. These animals always
descended when the aquarium was shaken, but with noticeably less precision than in
the cases of normal individuals. It therefore seemed probable to me that the por-
tion of the lateral-line system on the head was more effective than that on the trunk,
but as this experiment involved cutting the general cutaneous nerves of the head
as well as the lateral-line nerves, the experiment is not wholly conclusive.

Finally, in a third set of fishes, I cut the lateral-line nerves and the fifth and
seventh nerves of the right sides only, leaving the left sides intact. These fishes,
though a little sluggish, reacted in an essentially normal way. From these three sets
of experiments I conclude that the lateral-line organs maybe considerably reduced
without seriously impairing the action of the system as a whole, though the portion
of the system on the head is less easily dispensed with than that on the trunk.

The shin in Fundulus heteroclitus. — While I was experimenting on fishes in
which the lateral-line organs had been rendered inoperative I was at times puzzled
by getting reactions that seemed contradictory to the general conclusion that such
fishes were not stimulated by a slight movement of the whole mass of water. Occa-
sionally on making the whole aquarium move slightly a fish without lateral-line
organs would swim rapidly to the bottom. On watching for instances of this kind
I soon found that they occurred only when the fishes were close to the top of the
water, and in fact were within the range of wave action. When the whole aquarium
was moved, even only slightly, the upper surface of the water was thrown into small
waves. These waves, as could be seen by the motion of small suspended particles,
extended only a few centimeters below the surface of the water, but they established
a region into which the fishes without lateral-line organs would not ascend, and
from which, if overtaken by the waves there, they immediately escaped by swimming
downward. As fishes without ears as well as without lateral-line organs were stimu-
lated by these surface waves, I concluded that in this instance the motion of the
water must affect the general cutaneous nerves (touch).

If the motion of surface waves is a stimulus for the general cutaneous nerves,
it would seem probable that currents in the water would also affect these nerves and
that the ability of a fish to head up a stream might depend rather on the stimula-
tion of its skin than, as Schulze has implied, on the stimulation of its lateral-line
organs. Fundulus is in a marked degree rheotactic, i. e., it swims vigorously against
a current, and I therefore resolved to test this fish to ascertain whether its rheotaxis
depended on its lateral-line organs or not. Six specimens, in which the nerves to
the lateral-line organs had been cut, were placed one after another at the open end
of a large glass tube through which a moderately strong current of sea water was
flowing. All swam energetically up the tube, and, so far as this reaction was

62

BULLETIN OF THE UNITED STATES FISH COMMISSION.

concerned, they were in no observable respect to be distinguished from normal fishes.
Their rheotaxis certainly did not depend upon their lateral-line organs, but was
undoubtedly the result of cutaneous stimulation. Unfortunately I was unable so
to operate on other individuals that I could obtain active specimens whose cutaneous
nerves were severed but whose lateral-line systems were intact, and hence the only
conclusion I can draw is that the general cutaneous nerves are stimulated by wave
and current action and that this is sufficient to account for rheotaxis, but I can not
state whether or not the lateral-line organs are also stimulated by these means.

Conclusions concerning the lateral-line organs and the skin. — The observations
on Fundulus recorded in the preceding pages give no support to the view of P. and
F. Sarasin that the lateral-line organs are to be regarded as accessory ears, for
individuals in which the eighth nerves had been cut and in which the lateral-line
organs were intact did not respond to the sound-waves from a tuning-fork to which
fishes with ears reacted with certainty. I have also seen no reason to suppose that
the lateral-line organs are especially connected with the production of gas in the
air-bladder, as suggested by Richard, or that they are particularly concerned with
equilibration, as advocated by Lee. Since they are stimulated by slight disturbances
in the water that do not affect the general cutaneous sense organs, I can not agree
with Merkel in classing them as tactile organs. Their appropriate stimulus is a
slight mass-movement of the water, which may or may not be vibratory, and which
induces the fish to swim into deeper regions. This form of stimulus is of precisely
the kind that was attributed to these organs by Schulze (1870), but I have not been
able to confirm Schulze’s further opinion that current and surface wave movements
stimulate these parts. Such stimuli certainly do affect the general cutaneous sense
organs, but whether or not they influence the lateral-line organs I am unable to say.

GENERAL REACTIONS OF OTHER FISHES.

Although my studies were made almost exclusively on Fundulus heteroclitus, I
tested, as opportunities offered, other species of common fishes. These were placed
without being operated upon in the aquarium with the bass-viol string as a means
of producing sound. Because of the mixed character of the stimulus produced by
this apparatus and also because of the fact that the fishes were not operated upon
in any way, the results are significant in only one or two instances.

Young mackerel, while swimming in the aquarium, always moved downward
when the string was vibrated. The same was found true of adult mackerel, but
whether this reaction was an ear or a lateral-line response was not determined.

Menhaden, after they became somewhat accustomed to the aquarium, gave a
sudden leap each time the string vibrated, but showed no tendency to descend. In
tins instance, too, no clew was obtained as to the organs stimulated.

Three specimens of smooth dog-fish, each about 18 inches long, were tested.
When these fish were resting quietly on the stone bottom of the aquarium, the
vibration of the string would cause them to move their pectoral and pelvic fins,
or even begin swimming, but when they rested on some 3 inches of cotton wool
covered with a cloth to afford a deadened surface on the bottom of the aquarium,
no reaction of any kind was ever obtained. Apparently the ears, lateral-line
organs, and skins of these fishes are not open to any of the stimuli produced by the
vibrations of the bass-viol string and transmitted through the water, and they thus
differ markedly from the other fishes examined.

HEARING AND ALLIED SENSES IN FISHES.

63

These few notes serve to show that different species respond very differently
to the same forms of stimuli and emphasize the importance of refraining from
generalizations on the functions of the lateral-line organs and the ears in fishes
before a considerable number of species have been fully examined.

SUMMARY.

1. Normal Fundulus heteroclitus reacts to the sound waves from a tuning-fork
of 128 vibrations per second by movements of the pectoral fins and by an increase
in the respiratory rate. It probably also responds to sound waves by caudal-fin
movements and by general locomotor movements.

2. Individuals in which the eighth (auditory) nerves have been cut do not respond
to sound waves from the tuning-fork.

3. The absence of responses to sound waves in individuals with severed eighth
nerves is not due to the shock of the operation or to other secondary causes, but to
the loss of the ear as a sense organ.

4. Fundulus heteroclitus therefore possesses the sense of hearing.

5. The ears in this species are also organs of prime importance in equilibration.

6. Normal Fundulus heteroclitus swims downward from the top of the water
and remains near the bottom when the aquarium in which it is contained is given a
slight noiseless motion.

7. Individuals in which the nerves to the lateral-line organs have been cut will
swim upward or remain at the top while the aquarium is being gently and noiselessly
moved.

8. The lateral-line organs in this species are probably stimulated by a slight
mass movement of the water against them. They are not stimulated by sound waves
such as stimulate the ears.

9. Individuals in which the nerves to the lateral-line organs have been cut swim
downward and thus escape from regions of surface wave action. They also orient
perfectly in swimming against a current. Since surface waves and current action
stimulate fishes in which the nerves to the lateral-line organs and to the ears have
been cut, these motions must stimulate the general cutaneous nerves (touch).

10. The vibrations from a bass-viol string when transmitted to water stimulate
the ears and the lateral- line organs of Fundulus. They also stimulate mackerel and
menhaden, but not the smooth dog-fish, which responds only when in contact with
solid portions of an aquarium subjected to vibrations.

The work recorded on the preceding pages was done at the biological laboratory
of the United States Fish Commission at Woods Hole, Mass., and I take this oppor-
tunity of expressing my indebtedness to the Director, Dr. Hugh M. Smith, and
to his assistants for much help rendered me. I am also under obligations to Prof.
W. C. Sabine, of Harvard University, for advice and assistance in connection with
the sound-producing apparatus, and to Prof. F. S. Lee, of Columbia University, for
friendly criticism and many suggestions.

64

BULLETIN OF THE UNITED STATES FISH COMMISSION.

LIST OF REFERENCES.

Bateson, W.

1890. The sense-organs and perceptions of fishes: with remarks on the supply of bait. Journal
of the Marine Biological Association of the United Kingdom. New series, vol. 1, pp.
225-256, pi. XX.

Bonnier, P.

1896. Sur le sens lateral. Comptes rendus des seances et memoires de la societe de biologie.
Serie 10, Tome 3, pp. 917-919.

Emery, C.

1880. Le Specie del Genere Fierasfer nel Golfo di Napoli, e Regioni limitrofe. Fauna und Flora
des Golfes von Neapel. 2 Monographic, Leipzig, 76 pp., Tav. i-ix.

Fuchs, S.

1895. Ueber die Function der unter der Haut liegenden Canalsysteme bei den Selachiern.
Archiv fur die gesammte Physiologie. Bd. 59, pp. 45U478. Taf. vi.

Gunther, A. C. L. G.

1880. An introduction to the study of fishes. Edinburgh, xvi + 720 pp.

Hunter, J.

1782. Account of the organ of hearing in fish. Philosophical transactions of the Royal Society
of London. Vol. 72, pp. 379-383.

Kreidl, H.

1895. Ueber die Perception der Schallwellen bei den Fisclien. Archiv fur die gesammte

Physiologie. Bd. 61, pp. 450-464.

1896. Ein weiterer Versucli fiber das angebliche Horen eines Glockenzeichens durch die

Fische. Archiv ffir die gesammte Physiologie. Bd. 63, pp. 581-586.

Lee, F. S.

1898. The functions of the ear and the lateral line in fishes. American Journal of Physiology.

Vol. 1, pp. 128-144.

Leydig, F.

1850. Ueber die Schleimkanale der Knochenfisclie. Archiv ffir Anatomie, Physiologie und wis-
senschaftliche Medicin. Jalirgang 1850, pp. 170-181. Taf. iv, figs. 1-3.

1868. Ueber Organe eines sechsten Sinnes. Dresden. 108 pp., Taf. i-v.

Merkel, F.

1880. Ueber die Endigungen dersensiblen Nerven in der Haut der Wirbeltliiere. Rostock. 214
pp. Taf. i-xv.

Muller, J.

1848. The physiology of the senses, voice, and muscular motions, with the mental faculties.
Translated by W. Baly. London, xvn-j - pp. 849 to 1419+32-]- 22 pp.

Owen, R.

1866. On the anatomy of vertebrates. Vol. i. London, xlii+650 pp.

Richard, J.

1896. Sur les functions de la ligne laterale du Cyprin (lore. Comptes rendus des seances et
memoires de la societe de biologie. Serie 10. Tome 3, pp. 131-133.

Romanes, G. J.

1892. Animal intelligence. International Scientific Series, vol. 44. New York, xiv+520 pp.
Sage, D., C. H. Townsend, H. M. Smith, and W. C. Harris.

1902. Salmon and Trout. New York, x+417 pp.

Sarasin, P., und F. Sarasin.

1887-1890. Zur Entwicklungsgeschichte und Anatomie der ceylonesischen Blindwfihle Ichtliyo-
pliis glutinosus, L. Ergebnisse naturfwissenschaftlicher Forschungen auf Ceylon.
Bd. 2. 263 pp. Taf. I-IH.

Schulze, F. E.

1870. Ueber die Sinnesorgane der Seitenlinie bei Fischen und Amphibien. Archiv ffir mikro-
skopische Anatomie. Bd. 6, pp. 62-88, Taf. iv-vi.

Whitman, C. O.

1899. A.nimal Behavior. Biological Lectures from the Marine Biological Laboratory, Woods

Holl, Mass., 1898, pp. 285-338.

NATURAL HISTORY OF THE (JllINNAT SALMON.

A REPORT OF INVESTIGATIONS IN THE SACRAMENTO

RIVER, 1896-1901.

By CLOUDSLEY RUTTER,

Naturalist , United States Fish Commission Steamer Albatross.

F. C. B. 1902—5

65

TABLE OF CONTENTS.

Origin and method of the investigation

General results of the investigations

The Sacramento River as a salmon stream

The general life history of the Pacific salmons

Salmon eggs and milt

Experiments on the vitality of spermatozoa

Vitality in water

Activity in normal salt solution

Vitality in air

“Watery” milt

Amount of milt required in artificial fertiliza-
tion

Experiments with ova

How to test fertilization

Short exposure to water detrimental

Effect of exposure to air

Fish slime not deleterious

Fertilization in hody fluid

Immersion in normal salt solution

To make artificial spawning complete

An aid to spawn-taking

Dry process of fertilization

Killing the female before spawning

Quality of bloody eggs.

Foamy eggs

Granular eggs

Eggs dead when spawned

Eggs from dead fish

Spotted eggs

Yellow eggs

The critical period for eggs

Fungus in the hatchery

The alevin

Under natural conditions

Observations on alevins artificially reared

General account

Planting alevins

Enemies

Kesults of observations

The fry

Notes on younger fry

Planting fry from the hatchery

Observations of a particular fry

Enemies

Migration of fry

In Olema Creek

Battle Creek station

Balls Ferry station

Walnut Grove station

Observations at Benicia

General investigation during 1898

Movements in estuary and bay

Effect of sea water on alevins and fry

66

Page.

The fry — Continued.

Change of color during migration 100

Summary of observations on migration 100

Summer residents in the rivers 100

General account 100

Upstream movement 101

Migration during summer 102

Summary on number and movements 105

Growth in fresh water 105

Gastric parasites 109

Diseased parrs 109

Mature male parrs 109

Temperature notes 110

Conclusions from study of summer resi-
dents 110

Food of young salmon 110

General study of food in fresh water 110

Food in brackish water 117

Classification of the insect food of young sal-
mon 118

The period of growth 120

The adult salmon 121

Migration 121

Do salmon return to their native streams 121

The two runs of salmon 121

Details of migration 122

Downstream movement 124

Relation between weather and migration 125

Changes in salmon after entering fresh water 125

The alimentary canal 125

The skin 126

Loss in weight 127

Comparison of the sexes 130

Relative changes in fresh water 130

Two forms of males 130

Hermaphrodites 132

Relative number of males and females 132

Relative weight 133

Natural propagation 134

Spawning habits 134

Percentage of fertilization 135

Mortality among ova 136

Natural versus artificial propagation 137

Injuries and diseases 138

General effects of spawning 138

Spent salmon 139

Diseases of the intestine 139

Fungus 139

Gill parasites 140

Diseased ova 140

Length of life of fall salmon after reaching

spawning-grounds 140

Death 141

Page.

67

68

70

71

72

73

73

73

74

74

74

74

74

74

75

75

75

75

76

76

77

77

77

77

77

77

78

78

78

78

80

81

81

81

81

83

83

85

86

86

86

87

87

88

88

88

90

92

94

95

97

97

;• '.r .

Bull. U. S. F. C. 1902. (To face page 67.)

Plate 10.

LOWER McCLOUD RIVER FALLS, WHICH PREVENT THE
ASCENT OF SALMON.

SACRAMENTO RIVER AT DUNSMUIR.

PIT RIVER FALLS.

NATURAL HISTORY OF THE QUINNAT SALMON.

A REPORT ON INVESTIGATIONS IN THE SACRAMENTO RIVER 1896-1901.

By CLOUDSLEY RUTTER,

Naturalist , United States Fish Commission steamer Albatross.

ORIGIN AND METHOD OF THE INVESTIGATIONS.

The superlative value of the quinuat salmon as a food-fish has long been recog-
nized, and the great market demand for it was, until recently, rapidly leading to
its extermination. Within the last few years, however, artificial propagation has
reached such efficiency that the species is again increasing, especially in the Sacra-
mento River, California. Two divisions of the work of artificial propagation — spawn-
ing and hatching — have received much attention and are especially well understood.
The first includes securing the spawning fishes and taking and fertilizing the ova;
the second includes the care of the eggs from the time they are fertilized until they
are hatched.

At Battle Creek hatchery in 1897 nearly 50,000,000 eggs were taken, which
hatched over 40,000,000 alevins. With such an enormous output of young salmon
the question of the best method of planting them became of vital importance. It
had usually been the custom to plant the alevins within a few days after they were
hatched, and with such large numbers as were produced in 1897 this was an absolute
necessity, owing to lack of space in which to care for them. But there was evidently
a great risk in planting so many helpless young fishes when the knowledge of their
habits and enemies in the streams was so limited. It was for the purpose of supplying
such knowledge that these investigations were undertaken. Begun originally with
the one purpose in view, the investigation has grown till many other questions have
been involved. In order to determine the best method of planting it was necessary
to study the habits of the young salmon under natural conditions. This led to
a study of the spawning habits and finally the general life-history. In studying
the natural propagation it seemed advisable to compare it with artificial propagation,
which led to various experiments bearing on the latter subject. Most of the work
was carried on in the Sacramento basin. Every mile of the Sacramento River from
source to mouth has been visited, except 5 or 6 miles below the head of Box Canyon
near Sisson. Nearly 1,000 miles have been traveled in skiffs, and twice the entire
distance from Redding to Sacramento has been thus covered, seining stations being-
established every 17 miles on an average. In exploring Pit basin and the sources
of Feather River about 500 miles were ti'aveled by wagon and about 30 seining
stations were made. (See chart of explorations, plate 18.)

67

68

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Daily observations of the migrations of the young salmon in the main river
were made at two stations, about 325 miles apart, from January to May, 1899. The
hatchery experiments were carried on during two seasons, 1897 and 1898, at Battle
Creek hatchery. Twenty-five thousand eggs were hatched at the Hopkins Seaside
Laboratory at Pacific Grove during the winter of 1898-99, and the young were used
in experiments testing the effect of sea water on alevins and fry. Experiments in
planting alevins and fry were made at Olema, Marin County, and at Sisson, Siskiyou
County, in 1897 and 1898. The investigations in 1900 and 1901 consisted of observa-
tions on adult salmon at Pacific Grove and at various points on the Sacramento,
especially at Black Diamond, Rio Vista, and Mill Creek and Battle Creek fisheries.

Observations of the habits of the young were made first by watching them in the
water. This, of course, could be done only in the small streams, but was very suc-
cessful in the work the first year at Olema. Specimens were secured by means of
the ordinary Baird seine; the one most used in this investigation was 50 feet long
and 7 feet deep; smaller seines, 20 and 15 feet long, were used in small streams; in
some of the work in San Pablo Bay a net 150 feet long was used. Traps constructed
to suit particular purposes were employed where occasion required.

The work was carried on under the direction of the United States Fish Com-
mission, with the cooperation of the California Fish Commission, through Mr. N. B.
Scofield, during the first two years. Mr. F. M. Chamberlain, of the United States
Fish Commission steamer Albatross , assisted in the work from May, 1898, to April,
1899. Mr. A. B. Alexander, also of the steamer Albatross , began the work at Olema
in 1897. Much of the success of the investigation is due to the interest and counsel of
Mr. J. P. Babcock, of the California Fish Commission; and Prof. Charles H. Gilbert,
of Stanford University, has aided much in planning the work and in affording
facilities for studying the collections.

The author is under obligations to the agents of the Southern Pacific Company
at the shipping-points along the Sacramento River, to the various fish-dealers in
Sacramento, and to the salmon-packing associations at Benicia and Black Diamond
for statements of the catch of salmon at various places along the river and bays
and for other courtesies; also to the officials of the United States Weather Bureau at
San Francisco and Red Bluff and to the agent of the Southern Pacific Company
at Sacramento Bridge for river statistics. The directors of the Hopkins Seaside
Laboratory at Pacific Grove gave the free use of laboratory facilities for carrying
on the experiments at that place.

GENERAL RESULTS OF THE INVESTIGATIONS.

A great many points of more or less interest have been considered in this series
of investigations, of which the following have the most practical value and deserve
special mention :

1. The original object of the investigation has been carried out in determining
that young salmon should be released from the hatcheries soon after the yolk has
been entirely absorbed and that they should not be released in the headwaters late
in the spring.

2. A method has been found for removing and fertilizing the eggs left in the
fish after artificial spawning, thus increasing the take of eggs from a given number
of fishes by about one-fiftli.

3. The site for a new hatchery, Mill Creek station, has been discovered.

NATURAL HISTORY OF THE QUINN AT SALMON.

69

Other points of interest determined are :

a. The spermatozoa of the milt are active for only 3 to 5 minutes after the
milt is mixed with water.

b. Ova become incapable of fertilization after 5 minutes’ immersion in water,
and good results can not be obtained after 1 minute.

c. Ova can be exposed to air for half an hour provided they are kept moist by
the ovarian fluid.

d. Ova may be fertilized while immersed in the ovarian fluid, or in the slime
from the skin, or in unclotted blood.

e. Ova are not affected by immersion in normal salt solution for half an hour,
and are capable of fertilization thereafter.

/. Between the ages of 6 and 16 days, when the water temperature is about 50°,
the embryo is especially sensitive and liable to injury. During this period the eggs
should be handled with the greatest care.

g. Fungus is not a great pest at Battle Creek hatchery, probably owing to the
considerable amount of silt carried in the water and deposited on the mat of filaments.

li. Ale vins have many enemies in the streams ; fry but few.

i. The fry begin feeding and commence their downstream migration as soon as
the yolk is absorbed and they are able to swim.

j. The fry drift downstream tail first, traveling mostly at night and averaging
about 10 miles a day. They are 4 or 5 months old when they reach the ocean.

k. A few of the later winter fry, about 10,000 to the mile in the Upper Sacra-
mento, remain in the headwaters all summer, which is deleterious on account of
slow growth.

l. The food of young salmon at all places and seasons is insects, larval or adult.

to. Salmon spend from 2 to 4 years in the ocean.

n. They usually return to the river through which they reached the ocean,
because during their ocean life they do not get far away from its mouth.

o. The later fall salmon ascend the Sacramento River at the rate of 4 or 5 miles
a day, being about 65 days reaching Tehama from Rio Vista. The spring salmon,
without doubt, travel faster.

p. Salmon do not eat after leaving the ocean, and the stomach shrivels up to
about a tenth of its normal size.

<j. Salmon lose from 15 to 20 per cent of their weight in migrating, and from 10
to 15 per cent more in spawning.

r. The sexes can not be distinguished in salt water, but they differ greatly in
fresh water. The males develop the long hooked jaw, the large canine teeth, the
deep slab-sided body, and the color usually becomes more or less reddish. The
females do not change in appearance except as is due to the loss of flesh, the
development of the ova, and to the change in color from silvery to olive.

s. The males vary more in size than the females and are of two forms, adult and
grilse; the grilse resemble the females, but are much smaller.

t. The percentage of fertilization in natural propagation is high, probably about
85 per cent.

u. The injuries received in fresh water are mostly due to exertions in spawning
the last few ova.

v. It is well known that all Pacific salmons die immediately after spawning
once, and this investigation simply bears out the fact.

70

BULLETIN OF THE UNITED STATES FISH COMMISSION.

THE SACRAMENTO RIVER AS A SALMON STREAM.

The Sacramento is a large river, navigable for boats as far as Red Bluff, which
is 225 miles by rail from San Francisco. It is quite crooked, and the distance by
water from Red Bluff to the Golden Gate is about 375 miles. The river rises in
several small lakes in the mountains about 20 miles west of Sisson, in Siskiyou
County, California, and for nearly half its length flows through a narrow canyon.
The upper portion is a typical mountain stream, with innumerable pools and rapids
and gravel beds, forming ideal spawning-places for the salmon, though it lias not
been visited by many of them during the past few years. Near the lower end of
the canyon it receives Pit River from the east.

Pit River is a much larger stream than the Sacramento above their union. Its
lower portion lies in a canyon and except in size is similar to the Upper Sacramento.
About 75 miles above its mouth are the Pit River Falls, which, until a fishway was
blasted out, were impassable for salmon. The upper portion of Pit River lies on a
plateau and during the summer is a very unimportant stream.

The salmon that pass Pit River Falls spawn in Fall River, which enters Pit
River a few miles above the falls. When seen in August, 1898, Fall River flowed
several times as much water as Upper Pit River, though it is only 12 or 15 miles in
a direct line from its mouth to its source. It is about 100 feet wide and 3 to 4 feet
deep, flowing through a level plain and taking its rise in several large springs.

Hat Creek, draining Mount Lassen on the north, empties into Pit River a few
miles below the falls. It is a considerable stream, but its ascent is difficult for
salmon on account of very steep rapids.

McCloud River, draining Mount Shasta on the south, empties into Pit River
near its mouth. It is two or three times the size of the Sacramento River above the
mouth of Pit River and is an important salmon stream. Baird hatchery is located
near its mouth.

The most important salmon stream of the basin, excepting the main river as
noted below, is Battle Creek, which drains Mount Lassen on the west and empties
into the Sacramento between Redding and Red Bluff:'. Battle Creek hatchery is
located at the mouth of this stream.

A few miles below the mouth of Pit River, and just above Redding, the Sacra-
mento emerges from the canyon through which it runs from its source and widens
into a broad, shallow stream, though the current continues swift. Below Redding,
for perhaps 100 or 150 miles as the river winds, it continues broad and shallow, with
many short riffles and usually a gravel bank along one side. In ordinary years
when the river is in its normal low-water condition the principal spawning-beds of
the fall salmon are in this portion of the main river, notably in the vicinity of Red
Bluff and Tehama. In November, 1900, the river was examined carefully between
the mouth of Battle Creek and Tehama. Few salmon were seen until within a few
miles of Red Bluff, but from that point on every riffle was covered with spawning-
beds and dead salmon were everywhere abundant in their vicinity. Seventy-five
dead fishes were counted at one time in the lower 100 yards of Mill Creek and in the
river within 50 yards of its mouth.

A few miles above Red Bluff the river cuts through a range of hills, and for 2
or 3 miles consists of a series of rapids, the longest of which is known as Iron Canyon.
After passing Iron Canyon the river again assumes the character found at and

NATURAL HISTORY OR THE QUINNAT SALMON.

71

below Redding. Farther downstream the channel becomes deeper, gravel ban Its
disappear, sand banks become less frequent, and rapids are wanting. Such is the
character from Colusa to Sacramento. Below Sacramento it runs through a level
country and for most of the distance is affected by tides. There are many sloughs,
some connecting it with the San Joaquin. The Sacramento and San Joaquin rivers
join as they empty into Suisun Bay.

The water of the upper part of Sacramento River and the upper tributaries is
quite clear, and continues so until the mouth of Feather River is reached, from
which point to the mouth it is very muddy. It is in the muddy water between the
mouth of Feather River and Vallejo that the salmon for the markets are taken.

The only species of salmon regularly frequenting the Sacramento River is the
quinnat. The dog salmon is found occasionally, four specimens having been seen
during this investigation. Mr. Chamberlain reports finding single specimens of the
blueback and humpback at Baird in 1899. The humpback has also been reported
by others. The only record of the silver salmon from the Sacramento River is that
given by Jordan & .Touy (Proc. U. S. National Museum 1881).

THE GENERAL LIFE-HISTORY OF THE PACIFIC SALMONS.

The salmon under consideration in this report is the Sacramento or quinnat
salmon ( Oncorhynclius tschawytscha) , also known as Columbia River salmon, king-
salmon, and chinook salmon, and is the largest and most important of the five species
of Pacific salmons. The others, in the order of their importance, are (1) blueback
(O. nerJca), also called red salmon, redfish, and sockeye; (2) silver salmon (O.
kisutch ); (3) humpback salmon (O. gorbuscha), and (4) dog salmon (O. keta).

The Pacific salmons, as above named, are different from the Atlantic salmon,
which is related to the steelhead or salmon trout of the Pacific coast. One of the
important characters that separate the Pacific salmons from the steelhead and its
relatives, the Atlantic salmon and the rainbow and cut-throat trouts, is the larger
number of rays in the anal fin — the unpaired fin on the under side of the tail. The
steelhead and its allies have fewer than 13 rays (usually 11) in this fin, while the
Pacific salmons have more than 13, the number for the quinnat being 16.“

An equally fundamental, though physiological, difference lies in the fact that
the Atlantic salmon and steelhead trout spawn several times while the Pacific
salmon, of whatever species, dies as soon as it spawns once. This is a very striking-
difference and its importance can hardly be overestimated. A further difference
lies in the habits of the young. The young of the Pacific salmon seek the ocean as
soon as they are able to swim ; their migration is accelerated by high water and
retarded by low water, and they do not return to fresh water till mature. On
the contrary, young trout do not seek the ocean for several months after they are
able to swim, low water is an incentive to migration, and they run back and forth
between fresh and salt water seeking food.

The quinnat salmon is found throughout the Pacific coast from Monterey Bay
northward, but is less abundant north of Puget Sound. At spawning time it fre-
quents the larger streams, especially those with estuaries. The blueback salmon
is the most abundant of the Pacific salmons, and is most numerous in Alaska. Its
favorite spawning streams are those tributary to lakes. The silver salmon prefers

« For other anatomical characters, see Jordan & Evermailn, Fishes of North and Middle America, Bulletin 47,
United States National Museum.

72

BULLETIN OF THE UNITED STATES FISH COMMISSION.

the larger streams, though not necessarily those emptying through estuaries. The
other two species are of little importance. They spawn in any kind of stream,
frequently in mere brooks that empty directly into the ocean.

The Pacific salmon, of whatever species, passes most of its life in the ocean, and
upon arriving at maturity ascends the rivers to spawn. Sometimes enormous
numbers pass upstream together; stories of their being “ thick enough to walk
across on” are often told, but I have never seen them quite so numerous as that.

The upstream migration occurs sometime during the warmer half of the year,
the earlier fish going farther upstream. Spawning occurs late in the summer or
during the fall, and as soon as it is completed the salmon dies.

Most of the eggs deposited are eaten by other fishes, or are killed by being cov-
ered with sand and gravel. Those not destroyed hatch in from seven to ten weeks,
according to the temperature of the water. In the cold waters of Alaska they are
four or five months in hatching. It requires about six weeks more for the yolk-sac
to be absorbed, when the fry are able to swim and are ready for their seaward
migration. Most of the alevins, however, are devoured by other fishes before they
are able to swim. It is to prevent this great mortality among eggs and alevins that
artificial propagation lias been employed.

The young salmon start downstream as soon as they are able to swim, and
reach brackish water when three to five months old, according to the distance they
have to go. Those from the vicinity of Battle Creek hatchery reach Benicia in
about seven weeks. It is not known when they reach the ocean, but probably soon
after. The variation in the time of spawning and hatching makes the period of
migration very long. The fry from the summer run begin passing Battle Creek in
September, and from that time until April following there is a continuous stream of
young salmon, about I t inches long, passing that point.

Although a large majority go downstream as soon as they can swim, many,
especially those hatched in the spring, remain in the pools in the headwaters all
summer and fall. There were estimated to be from 700 to 1,000 in each of several
pools in the vicinity of Sims during the summer of 1898, and there were probably as
many as 10,000 to the mile in that portion of the river. These remained in the
headwaters until the first of December, when rains caused a rise in the river.

The food of young salmon in fresh water, at all times, places, and ages, consists
of insects, either larval or adult.

SALMON EGGS AND MILT.

The following notes embody observations and experiments made at Battle Creek
hatchery in 1897 and 1898. Although dealing largely with artificial propagation,
they are not intended to give even a general account of the methods of fish-culture
as applied to the Pacific salmon. For such an account reference is made to the
Manual of Fish-culture issued by the United States Fish Commission.®

For facilities placed at my disposal I am under obligations to Mr. G. H. Lambson,
superintendent of the station. Special acknowledgments are due to Mr. William
Shebley, superintendent of Sisson hatchery, and to Mr. Robert Radeliff, of Baird,
who had charge of the spawn-taking operations in 1897 and 1898, respectively, and
heartily cooperated in the experimental work.

" Report United States Fish Commission 1897, pages 1-340.

Bull. U. S F. C. 1902. (To face page 72.) PLATE 1 1 .

SACRAMENTO RIVER NEAR SIMS, “POOL B,” REFERRED TO ON PAGE 102.

SACRAMENTO RIVER IN THE VICINITY OF PRINCETON

NATURAL HISTORY OF THE QUINN AT SALMON.

73

EXPERIMENTS ON THE VITALITY OF SPERMATOZOA.

Vitality in water. Experiment No. 1. — Milt was mixed with water until the
mixture became of a milky appearance, which was the condition desired for fertiliz-
ing the ova at the spawning platform. Small quantities of the mixture were taken
from time to time and examined under a compound microscope. After several
repetitions of the experiment it was found that the spermatozoa remained active in
the water from 3 to 5 minutes, the length of time varying slightly in each experi-
ment. Each time the spermatozoa were placed on the slide they were very active for
about 30 seconds, after which time nearly all became attached by the tail to the
slide or cover. The head continued to move for a few seconds longer, but all
motion ceased after 65 seconds from the time the milt was placed on the slide.

Experiment No. 2. — Water was mixed with milt as in the above experiment and
a small quantity used to fertilize eggs every half minute for S.} minutes, with the
following results, the eggs being freshly spawned in each case:

Time milt had been in the water.

Percent-
age of
fertiliza-
tion.

Time milt liad been in the water.

Percent-
age of
fertiliza-
tion.

93

4.5 minutes

r

88

0

38

5.5 minutes

0

4

(> minutes . . .

0

2 minutes

8

6.5 minutes _

0

2.5 minutes

3

7 minutes .....

2

3 minutes .

8

7.5 minutes

0

3.5 minutes

1

8 minutes .

0

4 minutes

0

8.5 minutes

0

The experiment was tried three times, but the results were practically the same.
At one time a number of eggs were immersed in water taken from the top of the
can in which the spawn was taken from the spawning platform to the hatchery.
This water was white from the superfluous milt which had been spawned from
3 to 10 minutes. None of the eggs so treated were fertilized.

From these experiments it will be seen that the milt and eggs should be thor-
oughly mixed while in the spawning pan and within 30 seconds from the time
the milt is mixed with water.

Activity in normal salt solution /' — Milt was mixed with normal salt solution
until the liquid was distinctly whitish, and a portion of it was at intervals poured
over freshly spawned eggs. After a short time the eggs were washed with fresh
water. The per cent fertilized in each case is given in the following table:

Time solution had been spermatized.

Percentage of fer-
tilization.

First

attempt.

Second

attempt.

1 minute

74

98

2 minutes .

83

4 minutes _ ...

55

5 minutes ...

79

6 minutes

70

8 minutes

34

15

“Normal salt solution, 0.75 per cent common salt in water.

74

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Tlie results of this experiment are too varying to he of any practical value. It
seems probable that both spermatozoa and ova remain passive in the salt solution
and that fertilization takes place only after the addition of fresh water in washing.
Where a considerable quantity of fresh water was added and the ova mixed well with
it before it was poured off the percentage of fertilization was high. When the water
was poured off immediately and without mixing the ova well the fertilization was
incomplete. Normal salt solution apparently pi*eserves the vitality of the sperma-
tozoa longer than fresh water.

Vitality in air. — It was found that milt kept in an open, large-month bottle for
24 hours fertilized 74 per cent of the eggs it was mixed with. Milt that had been
so exposed 48 hours did not fertilize any eggs, nor did that kept in a tightly corked
vial for 24 hours.

“Watery” milt. — Milt when taken from the fish varies greatly in consistency.
That from some fishes is very thin and is known as watery milt. Experiment
proves that it fertilizes eggs as well as any and that no larger quantity is needed.

Amount of milt required in artificial fertilization. — In taking spawn it was
the custom at Battle Creek to express the eggs from one female into a pan con-
taining about a pint of water and add enough milt to make the water distinctly
whitish. The amount of milt necessary for this varies, depending on the amount
of abdominal fluid mixed with it, but is never less than 3 or 4 fluid ounces. This
method gives good results and should be followed when there is an abundance of
males, which is always true at Battle Creek after the first few days of the season.
A smaller amount of milt, however, will suffice. Ninety-six per cent of the eggs from
one female were fertilized by a tablespoonful of milt; 85 per cent were fertilized by
a teaspoonful; 35 and 57 per cent were fertilized by spawning fishes in the creek and
letting the milt float over the eggs, which had been caught on a screen.

Of course, all that is necessary is to bring a very minute quantity of milt in
contact with each egg. A single drop of milt if thoroughly disseminated through
the water would be sufficient to fertilize all the eggs from one female. In the experi-
ments above noted the milt could not have been thoroughly mixed until after it had
become inactive. It is not advisable to use less than a fifth of an ounce of milt to
fertilize 1,000 eggs. More water is necessary where a small amount of milt is used
in order to facilitate thorough mixing.

EXPERIMENTS WITH OVA.

How to test fertilization. — The quickest and surest way to determine whether ova
have been fertilized is to put them into a dilute (5 to 10 per cent) acetic acid. This
can be made from commercial acetic by adding from two to five parts of water. A
few minutes after the ova have been placed in the acid the embryos turn white,
while the yolk remains clear. The embryo can be distinguished in this manner
within 15 hours after the ovum lias been fertilized. In making this test for the first
time it is best to make a comparative test with unfertilized ova that have been kept
in water during the same period.

Short exposure to water detrimental. — A quantity of eggs were spawned into a
pan of water and some were removed and spermatized every half minute for several
minutes, and for various periods up to several hours. The milt was, of course,
taken fresh each time.

NATURAL HISTORY OF THE QUINN AT SALMON.

75

The results were as follows:

Time eggs had been in water.

Percent-
age of
fertiliza-
tion.

Time eggs had been in water.

Percent-
age of
fertiliza-
tion.

0.25 minute __

98

4

0.50 minute

96

5

95

4

1.5 minutes

68

5.5 minutes

1

2 minutes

57

6 minntes __

0

33

6.5 minutes

0

17

i

17

7+

0

The susceptibility of ova to fertilization decreases rapidly after they are placed
in fresh water, and the milt can not be added too quickly. Fifteen seconds is as
long as they should be in the water before the milt is added, and it is preferable to
add the milt at the same time that the ova are spawned.

On one occasion the eggs remaining in the body cavity after artificial spawning
were removed by cutting the fish open, which mixed them with much blood. They
were washed as quickly at possible and then spermatized. Only 11 per cent of the
eggs were fertilized. In another experiment 85 per cent were fertilized in the blood
without washing it off.

Effect of exposure to air. — A number of eggs were spermatized after having
been exposed to the air, temperature 76°, for various periods with the following-
results, the ordinary method giving fertilization of 99 per cent:

Time ova were exposed to air.

Percentage
of fertil-
ization.

12 minutes

100

99

76

50 minutes

Apparently there is no injury to eggs by an exposure to the air for half an hour.
It must be noted that although the eggs were in an open pan they were practically
immersed in the liquid from the body cavity.

Fish slime not deleterious. — It is sometimes said that the slime on the fishes is
fatal to the spermatozoa, and at some stations much care is taken to wipe the fish
dry before spawning. To test the truth of the supposition, a pan of eggs were cov-
ered with slime scraped from several fishes, and then spermatized without the use
of water. Out of 174 eggs examined only two were unfertilized, which is as good as
is obtained by t he ordinary process.

Fertilization in body fluid. — At another time, 392 eggs were spermatized “dry,”
and the milt entirely washed off with normal salt solution before water was added.
All but six, or 98.5 per cent, were fertilized. This proves that water is not neces-
sary to excite the activity of the spermatozoa, and that fertilization may be effected
in the abdominal fluid alone.

Immersion in normal salt solution. — A quantity of eggs were immersed in
normal salt solution and at the end of certain periods were taken out and fertilized
in the usual manner. The results as shown in the table, while as good as could be
desired, are scarcely better than are obtained by the ordinary method, which gives
a fertilization of 99 per cent. The value of the experiment lies in the fact that it
gives us a method of washing bloody eggs without preventing their fertilization, as

BULLETIN OF THE UNITED STATES FISH COMMISSION.

76

will be noted below. By making a chemical analysis of the fluid of the body cavity,
a liquid could probably be prepared that would be entirely passive and in which the
ova could be kept for days. This, however, is unnecessary, as a saltness of three-
fourths of 1 per cent gives a liquid sufficiently passive for washing out bloody eggs.

Time eggs were in salt solution.

Percentage
of fertil-
ization.

2 minutes

99

4 minutes _ _ _

100

(5 minutes

100

8 minutes

99

15 minutes _

25 minutes

97

86

To make artificial spawning complete. — Even the best spawn-takers can not get
all the eggs from the fish. Often the fish is not entirely ripe; but whatever the
condition may be many of the eggs are entangled in the folds of the ovary and
viscera and are not spawned. Under natural conditions the ovary shrivels up and
does not obstruct the outward passage of the eggs. The number of eggs remaining
in the fish after the artificial spawning varies from 200 to 1,500, depending upon the
size and condition of the fish. I have found an average of 900 eggs remaining in
55 fishes after artificial spawning. The spawning was done by experienced men
and could not well be improved upon.

The average number of eggs taken from a fish in ordinary spawning is 5,000.
This was the average during the season of 1897. By removing the remaining 900
eggs the yield can be increased about 18 per cent. They can all be removed only
by slitting the abdomen from the pectoral fins backward, but this allows a large
quantity of blood to mix with them. It is possible to fertilize 85 per cent of these
eggs in the blood, but in this case the unfertilized 15 per cent have to be picked
out of the hatching baskets, which would.be a considerable expense if the plan
were followed. But the blood can be removed from the eggs without any detriment
to fertilization by washing them in normal salt solution (one ounce of common salt
to one gallon of water). They can then be fertilized in the ordinary manner.

This method has been used at Battle Creek hatchery since 1900 with satisfactory
results, the loss with the “remnant” eggs being but little greater than with the
ordinary take. By care in handling the loss need not be any greater. By such
means the take of eggs can be increased from 10 to 20 per cent without increasing
the cost appreciably.®

An aid to spawn-taking .—It was found that fishes were much more easily and
rapidly spawned after cutting the body walls across the opening of the oviduct.
Unless the cut was made considerably in advance of the vent no perceptible amount
of blood issued. A greater percentage of the ova were spawned than if the gasli
had not been made, and no eggs were broken in spawning, which is an important
point. The shells or ‘ ‘ shucks ” from eggs broken in spawning are a great nuisance in
the hatching basket, being difficult to pick out and forming a basis for the growth

a From his study of the physiology of the Sacramento salmon in 1902, Prof. C. W. Greene, of the University
of Missouri, has determined the amount of salt in the ovarian fluid to he 0.94 per cent, which, therefore, is the
density of the solution that is normal for salmon ova, and should be used in washing the blood from eggs, rather
than 0.75 per cent as used in the experiments here noted. A solution of 0.94 per cent can be made by adding If ounces
of pure dry salt to 1 gallon of water.

NATURAL HISTORY OF THE QUINNAT SALMON.

77

of fungus if they are not removed. It was found by counting the number of broken
eggs in several lots spawned in the ordinary manner that they averaged nearly 1 per
cent of the entire take.

In the ordinary method of expressing the eggs they leave the oviduct under
considerable pressure and strike the spawning pan with as much force as if they
had fallen several feet. This manner of spawning, as already seen, breaks nearly 1
per cent of the eggs, and there may be many among those not broken that are
injured; this may account for the heavy loss of the first day in the hatching-house.

Dry process of fertilization. — The method of fertilization used at Battle Creek
in 1897 and 1898 was to spawn the eggs into a pan containing a little less than a pint
of water, spermatizing them at the same time. They were then allowed to stand
about 2^ minutes, when they were poured into a large bucket and gradually washed
by adding fresh water. Basket No. 6 of the table given in the notes below on the
critical period experiment was treated in this manner. The eggs of basket No. 5
were fertilized without any water, but otherwise were treated the same as No. 0.
There was a difference of only 0.3 per cent in fertilization. There was a difference
of 0.2 per cent between baskets No. 6 and No. 7, and they were from the same fishes
and treated in the same way, so far as fertilization was concerned.

The method used at Battle Creek seems the better, as the eggs can be mixed with
the milt more easily. A half minute, or just long enough to mix the eggs thoroughly,
is an abundance of time for them to remain in the spawning-pan.

Killing the female before spawning. — It has been claimed by some fish-culturists
that killing the female before spawning causes deformed fry. Basket No. 1 (see
table on p. 79) contained eggs from fishes killed by a blow on the head. There were
not even so many deformities in it as in others. This method of procedure is not
recommended, however, as green fishes would sometimes be killed, and their eggs
therefore lost.

Quality of bloody eggs. — Occasionally a female has been injured before spawn-
ing, and the eggs when pressed from the body were mixed with blood. Eggs from
three such fishes were kept separate; 7.7 per cent of the eggs died within five days;
of the remainder, 2 per cent (3 out of 154) were unfertilized. The fertilization was
about as good as the average, and a small amount of blood seems not to be detri-
mental to fertilization. Several females were opened after spawning, and the
eggs remaining were removed. The eggs were mixed with a great deal of blood,
and only 85 per cent could be fertilized, so that a large amount of blood is detri-
mental to fertilization, probably because clots of blood prevent thorough mixing
with the milt rather than from any injurious effect upon the ova or spermatozoa.

Foamy eggs. — Often the ovarian liquid becomes foamy as the eggs are spawned.
It was not known whether such eggs were fertilizable. In the foamy eggs experi-
mented with 99 per cent were fertilized.

Granular eggs. — The eggs from a certain small salmon, owing to the arrange-
ment or superabundance of oil globules, had a peculiar granular appearance. Fer-
tilization by ordinary process was 99 per cent; apparently healthy when 26 days old.

Eggs dead when spawned. — Occasionally eggs at the time of spawning have a
dull, yellowish appearance, and are evidently not healthy. They are always thrown
away. Eggs of this kind from one fish were kept. Seventeen days after spawning
30 per cent had died, and of the remainder 23 per cent were unfertilized. They
were not kept for further observations. Such eggs were not found in 1898.

78

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Eggs from dead fish. — On two occasions a ripe female was removed from the
water, and after it had been dead 2 hours a few eggs were spawned and fertilized in
the usual manner. A few eggs were spawned from time to time until the fish had
been dead 34 hours. The following table gives the results :

N umber of hours fish had been dead.

Percentage of fer-
tilization.

Percent-
age of eggs
that died
within 10
days.

First

fish.

Second

fish.

99

98

92

94

93

89

27

10

62

3]

31

0

100

0

100

The first four lots of eggs of 2, 4, 6, and 8 hours were kept 26 days and were
apparently entirely healthy at the end of that time.

At another time eggs were taken from two fishes that had died in the water.
One had been dead 1 hour, the other over 6 hours. Of the first, over 97 per cent
hatched and were healthy fry; of the other, 85 per cent.

From the above it is evidently safe to take the eggs from fishes that have been
dead less than 5 hours, and fairly good results can be obtained up to 8 hours.

Spotted eggs. — Sometimes a considerable number of eggs, a few weeks after
fertilization, have a small, irregular white spot about the size of the head of a pin
in the yolk near the surface. This does not mean that the egg is about to die.
Fifteen such eggs were put into a separate basket, and all hatched as perfectly
healthy fry excepting one, which died in breaking through the shell. The spot did
not appear on the yolk-sac.

Yellow eggs. — When eggs are nearly ready to hatch a yellowish fluid sometimes
collects around the embryo. This does not affect them very seriously, as most of
them hatch into healthy alevins.

The critical period for eggs. — It is a well-known fact that at a certain stage in
development, from about the sixth to the sixteenth day, eggs are much more liable
to injury than at other stages. When first taken they can be handled with com-
parative roughness with impunity. At Battle Creek in 1897 the spawning platform
was about half a mile from the hatchery. The eggs were hauled this distance over
a road that lacked much of being smooth, yet the loss traceable to such handling
was slight. Of nearly 700,000 eyed eggs sent from Battle Creek to Olema at one
time, less than 300 were killed in shipping. They were hauled about 10 miles in a
heavy wagon, were on the train some 15 hours, and out of the water 48 hours. At
the time of shipment the eggs were 43 days old. But at an earlier date, when 6 to
16 days old, such treatment would have killed every egg.

For purposes of comparison 60,000 eggs from several fishes, fertilized in the
ordinary manner, were mixed in a can at the spawning platform, and at the hatchery
were equally divided between two baskets. The eggs of one of the baskets were
picked over daily regardless of results in order to remove the dead or addled eggs.
The eggs in the other were picked over in the same manner on the first, third,
twenty- second, twenty-fourth, and forty-first days, and occasionally after that date.
The former of these experiments was called No. 7, the latter No. 6. As a further test

NATURAL HISTORY OF THE QUINNAT SALMON. 79

another basket, No. 1, of 30,000 eggs was picked daily. The loss of the baskets picked
daily was from three to seven times greater than that of baskets not so treated.

The following table indicates the comparative loss in baskets No. 6 and No. 7,
and shows that eggs are very sensitive between the sixth and sixteenth days, and
that they should be disturbed as little as possible during that period. Basket No.
6 lost but 785 eggs from the fourth to the twenty-second day, while basket No. 7,
which was picked daily, lost over 8,000. After the twenty-second day the loss in
this basket, No. 7, was only 613, while it was 1,369 in No. 6, the one not picked
daily. The loss on the forty-third day was the result of being shipped from Battle
Creek to Olema, in which case the loss in No. 6 was nearly eight times that of No. 7.
It is evident, therefore, that daily picking takes out nearly all the weak eggs, but it
is also strikingly evident that it takes out a very great many that are not weak.

Table showing loss of eggs in baskets No. 6 and No. 7, taken November 15, 1897.

Age, days.

Tem-
pera-
ture of
water.

Loss.

Age, days.

Tem-
pera-
ture of
water.

Loss.

No. 6.

No. 7.

No. 6.

No. 7.

° F.

° F.

i _.

49

303

30(5

35

47

16

2 ...

49

54

43

665

84

3

52

56

in

18

1

52

60

47

5

5 -

53

131

48...

2

6

51

203

49

64

3

7

50

«267

50

5

8

51

«225

51

3

9

50

a 420

52..

4

in.

49

'0,240

53

9

n

47

1>1,000

54

11

12..

48

d/97

2

13

48

1,200

57

2

14

49

1,582

58

62

15

52

469

60

14

m

51

331

61...

23

17

48

132

62...

20

18

46

47

63

25

lit

46

1(H)

64 ..

115

20

48

54

65

17

21

51

28

68.

21

13

22

50

785

31

69

15

23

50

18

72. .

32

24

24

48

430

24

74..

11

15

25

50

31

75. ..

5

14

51

28

77

5

9

27

51

24

78 ..

1

28

49

12

79

3

29

49

41

80

2

30

47

28

83 ..

31

46

23

84 ..

2

32

45

23

33

46

16

2,497

9,404

34

44

14

a Many killed while picking (went over basket but once). ('Not so many killed while picking.

c Died almost as fast as one person could pick them out.

The table below gives a summary of the loss in four baskets; two, No. 1 and No.
7, picked daily, and two, No. 5 and No. 6, picked so as to avoid the critical period :

Picked daily.

No. 5.

No. 6.

No. 1.

No. 7.

Loss of eggs

16,889

7

35

16,931

56

0.3

9, 404
14
4

9,422

32

0.6

3,292

21

38

3,351

11

0.7

2,497
« 112
56
2,665
9
0.4

Loss of alevins .

Deformed fry

Total loss

Percentage of loss .

Percentage unfertilized .

" The large loss of alevins of No. 6 was caused by accidental smothering.

80

BULLETIN OF THE UNITED STATES FISH COMMISSION.

There were 30,000 eggs in each basket at first. The eggs of basket No. 1 were
from fish that were killed before spawning. Those of No. 5 were fertilized accord-
ing to the dry process. Those of No. 6 and No. 7 were fertilized by the ordinary
process used at Battle Creek. Nos. 1 and 7 were picked daily. Nos. 5 and 6 were
not disturbed during the critical period. The number unfertilized among the dead
eggs was not determined before the twenty-eighth day; therefore the last item of
the table is only relative. To get the true percentage of unfertilized eggs it would
probably be about right to double that given. (It can be determined whether
addled eggs are fertilized by putting them in about 8 per cent acetic acid. The yolk
of the egg becomes clear, and the embryo, if there be one, turns white. A strong-
solution of common salt will clear addled eggs, but it also disintegrates very young-
embryos. )

Even after the critical period has passed, the most careful handling kills some
fertilized eggs. Several tests show that from 10 to 20 per cent of the loss after the
critical period is in fertilized eggs that have been killed in handling. They should
therefore be disturbed as little as possible.

Fungus in the hatchery. — Fungus is a considerable pest in a hatchery, but the loss
of eggs at Battle Creek traceable to this cause is very small. Numerous experiments
were made in order to determine if the fungus would attack and destroy living eggs.
Only on one occasion have I found a live egg attacked by fungus. This one had
a few filaments of fungus growing on one side, and the egg had begun to die where
the fungus was attached. Whether the egg had started to die before the fungus
attacked it or whether it was attacked first I can not say, but all other observations
and experiments indicate that the fungus attacks only the dead eggs.

Fungus grows rapidly on dead eggs, and the filaments extend in all directions
and entwine the adjacent eggs in a thick mat. This interrupts the circulation of
the water and often smothers the eggs so matted. When eggs are smothered the
embryo turns white before the yolk becomes addled, so that death from that cause
can be distinguished.

At Battle Creek, in 1897 and 1898, the baskets of eggs were gone over on the
second and third days after spawning and all of the dead eggs picked out. They
were not disturbed again until after the critical period, or about the twentieth
day. This method was followed even where baskets (size, 23 by 15 £ by 6 inches)
contained 40,000 eggs each. The number of eggs that died after the third day was
small, and at the “breaking out,” that is, the first picking after the critical period
at about the twentieth day, the few dead eggs were found scattered here and there
through the baskets. Each dead egg was covered with fungus, the filaments of
which had entangled the live eggs lying in its immediate neighborhood, holding them
together in a bunch. It was seldom that more than fifteen eggs were held together
in such a bunch, and the dead eggs never exceeded three or four.

Such treatment is not recommended for other stations, as the difference in the
character of the water supply makes it necessary to carry on separate investigations
for each station in order to determine methods of treatment.

The reason Battle Creek is so free from fungus is that the water contains a
considerable quantity of silt or dirt, and if the eggs are left undisturbed a couple
of days they become covered with a fine sediment. This collects on the fungus,
which acts as a kind of filter, making of it a black muddy mass and impeding its

NATURAL HISTORY OF THE QUINNAT SALMON.

81

growth. Clay or other dirt free from organic matter is often mixed in the water to
destroy a growth of fungus on fry.

While carrying on the experiments at Pacific Grove in January, 1899, when
they were 38 days old the fry Avere attacked by a very serious growth of slime, some-
times called gill-fungus by fish-culturists. This slime was composed mostly of a
microscopic unicellular animal with a silicious shell, belonging to the order Flagel-
lates. Some other microscopic animals and unicellular plants, such as diatoms, were
present. The slime collected on the gills of the fish and killed about two-thirds of
them. They Avere treated with a 25 per cent mixture of sea Avater, which Avas very
effective. Those which had been removed to a mixture of sea water before the
appearance of the slime were not affected at all.

It must not be supposed from the statements given above that hatching troughs
and baskets must never be touched during the critical period, nor that fungus is
the only disease to which salmon ova and the alevins are liable. As has been
stated above, I am not giving a general method of fish-culture, but an account of
certain investigations. If the deposits of sediment on the eggs and troughs show
traces of decaying organic matter, especially if there is a growth of slime on the
Sides of the trough, everything must be cleaned immediately. There is no doubt
that microscopic plants and animals, such as bacteria and those mentioned above
as having injured the fry at Pacific Grove, are very injurious to the eggs, alevins,
and fi*y and must be scrupulously guarded against.

THE ALEVIN.

UNDER NATURAL CONDITIONS.

The eggs that are not destroyed in one way or another when deposited by the
spaAvning fishes lie among the rocks, where they lodge and hatch in from 6 to 9
Aveeks, the time depending on the temperature of the water. The alevins also
remain among the rocks at the bottom for a few weeks, and their movements, slight
though they are, expose them greatly to such fishes as the sculpin and trout. Dur-
ing this time the yolk supplies them with what nourishment they need, but about
four weeks after hatching the quantity of yolk has become so small that it is not
absorbed rapidly enough by the blood to meet the needs of growth. At this time
also the alevin is able to swim a little, and it frequently leaves the bottom to snap
at some floating object. Its movements are necessarily awkward on account of the
unabsorbed yolk, and it therefore attracts predaceous fishes. This is the most
critical period in the life of the salmon after hatching. The yolk-sac disappears
entirely at the age of 5 or (i Aveeks, when the young are known as fry. This is the
age at which they begin feeding.

OBSERVATIONS ON ALEVINS ARTIFICIALLY REARED.

General account. — In December, 1896, 855,000 eyed eggs were shipped from
Battle Creek hatchery to Bear Valley hatchery in Marin County. Here they AArere
hatched early in February, 1897, under the care of Mr. Frank Shebley, of the Cali-
fornia Fish Commission. After the yolk-sac Avas absorbed, which was about 35
dgys later, they were fed for a feAv days on curds of milk, and then, in the second
week of March, Avere turned into Paper-mill Creek and its tributaries, Nicasio,

F. C. B. 1902—6

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Olema, and Hatchery creeks. The fry were strong and healthy, and were turned
into the streams in the best of condition. The young salmon were watched day
after day, and systematic observations were made of their movements, habits,
enemies, and rate of growth. The work was first begun by the United States Fish
Commission and carried on until the middle of May. After a break here of three
weeks the California State Commission carried it to completion.

In the winter of 1897-98 eggs were again shipped from Battle Creek to Bear
Valley. This time t he number was increased to 2,000,000, necessitating the plant-
ing of the alevins as soon as they began to hatch. All were planted before the
yolk-sac was absorbed. This, when taken in connection with the previous year’s
work, gave an opportunity to study the comparative effectiveness of planting alevins
and fry. The alevins were transported from the hatchery to the planting-grounds
in 20-gallon cans. It was possible to carry 40,000 alevins in such a can for two hours
at a temperature of 40° without loss, though 20,000 or 30,000 to the can was the usual
number carried. They were carried by wagon or rail as the ease required. A
wagon was found to be preferable, the jolting being an advantage, as the splashing
in the cans kept the water well aerated. If the road was smooth or if carried by
rail, the water had to be aerated frequently, and it was necessary to put fewer in the
can. Although alevins appear to be very delicate, they stand transportation much
better than fry, and a much larger number can be safely carried in each can.

Paper-mill Creek and its tributaries, where the young salmon were planted, were
never visited by the quinnat salmon. This was one reason that they were selected
for the experiment, as any young of that species that we might find would be known
to have been. planted there. The streams are rich in aquatic insect life, affording an
abundance of food for the salmon fry. Trout and sculpins ( Cottus ) are the only
predaceous fishes. The streams do not flow directly into the ocean but through
several miles of brackish tidewater into Tomales Bay, and the transition from fresh
to salt water is very gradual, removing the danger of the fry being rushed too quickly
from river to ocean water. It was thought that if the fry could thrive in these
streams and pass successfully into salt water it would be of advantage to utilize
coast hatcheries and plant in the smaller streams, where the young salmon would
not be subjected to their supposed enemies during the long journey from the Upper
Sacramento to the sea. The thing feared in the experiment was that the streams
would prove too short and that the young salmon would arrive at salt water before they
were ready to conform to the conditions of life they would have to encounter there.

Observations indicate that fry can be as safely planted in Paper-mill Creek and
its tributaries as in the Sacramento River, and they reach the ocean six weeks earlier.
If it is true, as the experiments made at the Clackamas hatchery in Oregon indicate,
that most of the salmon return to fresh water to spawn after being in the ocean two
years, a difference of one or two months in the time of reaching the ocean is worth
considering. If the full growth is attained in 24 to 36 months, the average gain in
weight is from 12 to 16 ounces a month. As the gain is necessarily slight at first, it
must be much more than a pound a month later. Any extension of time for living-
in salt water is an increase of the rapid-growing period, as the early period of slight
increase in weight must be passed through in any case. This argument holds good
only on the supposition that the individual would leave the ocean in a particular
month. But the great variation in the time in which the Sacramento salmon leaves

NATURAL HISTORY OF THE QUINN AT SALMON.

83

tne ocean makes it <juite certain that the time is determined by other influences than
the season. If so, it is doubtful whether those planted in Paper-mill Creek would
have any advantage over those planted in the Sacramento River.

Planting alevins. — Alevins on being liberated in swift water swim frantically
anti scatter in all directions as they are swept downsteam. Most of them seek the
bottom and crowd into crevices between the pebbles or get into quiet places under
or behind large bowlders. Others find their way into still water along the edge of
the stream, where they remain exposed to view. In moderately swift water some
find a lodging-place on the bottom or near the shore before they have been carried
a hundred feet downstream, and it has to be very swift water that will carry them
a hundred yards. For several hours after being planted in swift water many of
them keep moving about. Often the place where they first lodge is too much
exposed to the current, and they are repeatedly swept downstream, lodging here and
there for a few moments, until they finally reach a quiet place where they can stay.
After a few hours this moving about ceases and they remain quiet, retaining
their places for at least several days. In one instance 60,000 alevins were liberated
on a very swift riffle in Sacramento River 200 yards above a quiet pool. The riffle
was shallow, at no place over a foot deep, but so swift as to make it almost impos-
sible for a person to stand. The alevins all found shelter before they were carried
a third of the distance to the pool. On visiting (lie riffle a day later none could be
found much over a hundred yards below the place of planting and none was found
in the pool below, which was seined thoroughly. All had found sheltered places
and had ceased to move down with the current. When alevins are planted only a
few yards above a pool, even in moderately quiet water, large numbers will drift
into it, where they remain if they are not eaten by trout or other fishes.

When alevins are liberated in a pool or pond they at first scatter out near the
surface, but soon settle to the bottom, where they keep up a constant wriggling of
the tail and pectoral fins. Within a day or so they collect in bunches, appearing
as brilliant salmon-colored blotches over the bottom of the pond. The constant
motion of the individuals stirs up the silt until it is washed away from them and
each bunch rests on the solid bottom. Alevins in a hatchery, by this constant
motion, keep the hatching-trough free from sediment.

During a freshet at the California State hatchery at Eel River a thick sediment
of sand was washed into the hatching-troughs and came so fast that the alevins,
just hatched, were unable to keep it from settling. It covered the bottom of (lie
troughs to a depth of 2 inches, becoming hard and compact. The alevins, instead
of being covered, were found above the cement-like deposit, and none of them had
been lost. This interesting incident demonstrates their ability to keep from being
covered by sediment during a freshet, whether they be in pools, ponds, or troughs.

Alevins in the pond at the Bear Valley hatchery began swimming about in
schools before the yolk-sac was entirely absorbed. The presence of predaceous
fishes might have caused them to do otherwise.

Enemies. — When alevins are planted on riffles they are inclined to congregate
in eddies and sheltered places behind bowlders. In these places several thousand
of them may be found huddled together in a bunch plainly exposed to view. They
are not very shy at this age a-nd do not appear to try so much to get out of sight as
to get out of swift water. The brilliant salmon color of the yolk makes them very con-

84

BULLETIN OF THE UNITED STATES FISH COMMISSION.

spieuous, especially when t hey collect in bunches. One such bunch in Olema Creek,
in which there were four or live thousand, could be plainly seen from a point 50 yards
distant. In such instances they are very much exposed to the ravages of ducks
and geese — both tame and wild — herons, cranes, and other wading birds, and even
of hogs, to say nothing of the fish in the stream. Alevins are very tempting mor-
sels, and there is scarcely an animal that will not eat them when given a chance.
However, most of the alevins can get into crevices between and under the pebbles
and bowlders, where they are much safer from the attacks of the fish that may be
in the stream than they would be if they were in the pools or quiet places.

Observations were also made on planting alevins in Sullaway Creek near Sisson.
This is a very favorable stream in which to release young salmon, so far as preda-
ceous fishes are concerned. The only fishes of the stream are rainbow trout, scul-
pins, and quinnat salmon parrs remaining from the season before. The plants were
made on the riffles, but as these were all rather short, many of the alevins drifted
into the pools. They were liberated in the morning, and in the afternoon the pools
were seined with a small-meshed net. The fish caught were examined to find to
what extent they had eaten the alevins. The result is shown by the following table,
which giv.es the length of each fish examined and the contents of its stomach:

Species ex-
amined.

Size, in
inches.

Number

alevins

eaten.

Other material in
stomach.

Species ex-
amined.

Size, in
inches.

Number

alevins

eaten.

Other material in
stomach.

Trout

6

17

1 small pebble.

Sculpins

3.5

3

5.5

9

6 insect larvae.

3. 5

4

5

7

1 caddis larva.

3.3

2

4.5

6

3.3

2

3 insect larvae. >

2.5

0

Insect larvae.

3.3

3

2.5

0

Do.

3

2

Salmon parrs.

3.8

1

3

2

1 insect larva.

3.5

1

2.8

2

3.5

1

2.8

2

3.5

2

2.8

2

3.5

2

2.7

1

3.5

2

2.4

0

3.5

2

1 insect larva; 1

2.4

1

water bug.

2.4

1

3.4

1

2

0

3.4

1

2

0

3.4

2

Species and size

3

i

2 winged insects.

unknown

3

In the case of the largest trout, 6 inches long, 7 of the 17 alevins were in its throat
and mouth. It had evidently gorged itself to the limit.

In all cases where salmon parrs had eaten two alevins, the tail of the second
remained sticking out of the mouth, their stomachs being large enough to accommo-
date only one. The sculpins also had gorged themselves in the same manner. All
of the fish caught were examined and only three had not eaten alevins, being too
small. Three alevins had been disgorged by some of the fish. Evidently alevins
are a favorite food for trout, sculpins, and salmon parrs; and when they remain
exposed to such enemies from 2 to 1 weeks, it is a wonder that any escape.

Alevins planted in the Marin County streams in 1898 met even a worse fate. Here
the trout are more numerous and larger. The sculpins are also larger and more
abundant. There were no salmon parrs to feed on alevins, but there were myriads of
sticklebacks, which, though unable to swallow an alevin, killed many by nibbling at
the yolk. The only other fish in these streams was the roach ( Kutilus symmetricus),
which as far as could be learned did not feed on the alevins. Four is a moderate

NATURAL HISTORY OF THE QUINNAT SALMON.

85

estimate of the average number of alevins that a trout will eat in a day, at which
rate each trout would destroy about 150 before the absorption of the yolk-sac; and
1,000 trout would destroy 150,000 alevins. The lesson is obvious.

Just here it may be well to state that in 1897, although only 150,000 fry — not
alevins — -were planted in Olema Creek, large numbers of them were yet to be found
in June following, and quite a number in August. In 1898, 850,000 alevins were
planted in Olema Creek, and in June following there was a smaller number left in
the stream than was found in August the year before. There are two ways to
account for this. One is that the alevins were washed out to sea before they began
swimming; but it is more probable that they were eaten by trout and sculpins.

In the spring of 1898, 7,000,000 salmon were planted at the hatchery on Battle
Creek about two weeks before the yolk-sac was absorbed. Although trout are not
numerous there, the stream swarms with sculpins ( Coitus gulosu-s), salmon fry
remaining from the season before, Sacramento pike ( Ptyclioclieilus grandis), black
pike ( Ortliodon microlepidotus ), hitch (. Lavinia exilicauda), split-tail ( Pogonichthys
macrolepidotus), and suckers ( Catostomus occidentalis). All of these, though they
do not feed exclusively on animal matter, take salmon eggs and alevins when they
can get them. The Sacramento pike is very destructive to young fish. The split-tail
is the most numerous species, and lives on salmon eggs during the spawning season.

Each day while the hatchery was in operation the bad or addled eggs picked
from the hatching baskets were thrown into the stream. Usually they were thrown
into a small brook near its entrance to the creek. In a very short time after
emptying a can of eggs the split-tails always began to appear, running in from the
creek. In a few minutes the water would be alive with them, almost a solid mass
tumbling one over the other, splashing the water and crowding each other in their
frantic efforts to get the eggs, until some were forced into the mud at the edge,
while others were lifted upward till their backs or bellies were out of water, or one
might get into a vertical position with its head or tail out of water. Frequently
one would gorge itself till throat and mouth were so full that the passage of the
water over the gills was shut off and it suffocated. It usually required about 5
minutes to consume 5 gallons of eggs.

Alevins are almost as helpless as eggs and fully as palatable, and there can be
little doubt of their fate when planted in such an environment.

Results of observations. — The egg and alevin stages are the periods in the life
of the salmon when the care of the fish-culturist is most needed. The art of taking
and caring for the spawn has been so perfected that the loss in hatching need not
be over 10 per cent, and is often less. The loss of alevins, if thej" are retained in
the hatching- troughs or nursery ponds, need not be over 2 per cent. If the young-
are planted during the alevin stage, the loss is very great. If large numbers of
alevins are released in unsuitable places, where the bottom is comparatively free
from stones, and where such predaceous fishes as the split-tail and trout abound, the
loss may even be greater than if the parent salmon had been allowed to take their
natural course in spawning.

Young salmon should never be planted until the yolk-sac lias entirely disap-
peared and their swimming power has fully developed, even though they have to be
fed a few days. There is no advantage in holding them after this time.

86

BULLETIN OF THE UNITED STATES FISH COMMISSION.

THE FRY.

NOTES ON YOUNGER FRY.

Planting fry from the hatchery. — Fry are transported from the hatchery to the
streams in the same manner as already described for alevins, but it is not practicable
to carry over 10,000 in a can, even for a short distance. They require more care
than alevins, it being necessary to aerate the water constantly.

When fry are liberated in running water, they immediately head upstream and
try to stem the current. Owing to their being more or less faint from confinement
in t he can the current nearly always carries them downstream a short distance, but
they soon find their way into the more quiet water along the edge of the stream, in
the eddies or quiet pools, or among the stones at the bottom. Some even move a
few yards above the place of planting before they come to rest. On gaining quieter
water they rest themselves, moving only enough to keep from drifting downstream.
When in such position they begin feeding on any particles of food that float within
their vision, often snapping viciously at insects half as large as themselves.

In a small stream there is no marked tendency of the fry to form schools, each
appearing to act independently; but in a larger stream, and especially in the large
pools, they often swim about in schools. It appears, too, from our observations
in the Sacramento, that they run in schools after gaining the main river in their
migration to the sea.

After planting, the fry soon begin to drift downstream from one resting-place
to another. This movement in small streams is not in schools. If many are planted
at one place the movement downstream is quite rapid, and within 24 hours they will
be scattered evenly along the stream for over a mile below the place of planting.
The movement, though marked in the daytime, is more general at night. In one
instance a sci-een was placed across a small stream a quarter of a mile below where
50,000 fry were released. Although but few reached the screen that day, the
following morning apparently every one had reached it. Other observations have
shown the same thing. Muddy water hastens the movement downstream, as does
also high water, which is usually muddy.

In Hatchery Creek, in Marin County, 150,000 fry 10 weeks old were released.
They gradually scattered downstream, floating tail first. In four or five days they
were about evenly distributed along the creek for 14 miles below the hatchery. At
the end of this time a net with a 10-incli circular mouth was placed' in the current
in the daytime with mouth upstream. In one hour 30 or 40 fry were caught. This
illustrates well the decided movement downstream after planting.

When released in a large pool or pond the fry collect in schools immediately
and travel toward the inlet.

In 1898, 150,000 alevins were placed in a pond at the Bear Valley hatchery.
These remained in the pond without being fed until four weeks after the absorption
of the yolk-sac. As it had but 600 square feet of surface and was only 2 or 3 feet
deep, there were obviously too many in the pond to do well without being fed. As
would be expected, they grew but little, though few, if any, died. At the end of
four weeks all were very nearly of one size — 1.4 inches long. Those of the same
age in the creek a mile below the pond varied from 1.5 to 1.9 inches; specimens
from Olema Creek only two weeks older were from 2 to 2.4 inches long.

NATURAL HISTORY OF THE QUINN AT SALMON.

87

At any time during the four weeks that the fry were so crowded in the pond they
could have gone out, as the overflow trough was unobstructed. Very few if any of
them did so, however. Indeed, it was difficult to get them to go out at all, very few
escaping till nearly all the water was drawn off. As soon as they came near enough
to the overflow to feel the course of the current they would dart back into the pool
again. It has often been noticed that fry have an aversion to going over a water-
fall or swift rapid, 'flic observations at Sims during the summer of 1898 indicate
the same thing. On account of this, fry should not be planted above falls or swift
rapids, especially in small streams, as it is desirable that they should move down-
stream as soon as possible.

Observations of a particular fry . — Fry were observed daily from September 18 to
October 3, 1900, in a pool between a rock and the shore in Battle Creek. The pool
was about 18 inches across, 4 feet long, and 2 or 3 feet deep. There was but one
fry until the 25th, when another appeared. It is probable that only two individuals
were seen during the observations, though we can not lie sure that such was the case.

When first seen the fry swam near the surface, but after a few days it remained
a few inches below. It stayed most of the time in the rather strong current, and
was continually snapping at minute floating objects. When swimming near the
surface it made from two to ten strikes a minute. Observations could not be made
so easily after it began swimming deeper. It was seen to make at least 150 strikes,
but each time whatever was caught was immediately ejected. Apparently it had to
make a great many efforts before finding anything edible.

One of the fry was seen to leave the pool and resume its migration. It had
been in the lower portion of the pool all day, and as evening approached allowed
itself to be carried down into the shallow and swift water of the outlet, always
keeping its head upstream. Several times it was carried halfway through the out-
let, but darted back into the pool. Once it got entirely through the outlet and into
the deep water below the rock and then darted back, but finally it was carried out
into the main current, tail first, and was lost sight of.

Enemies. — As already stated, 855,000 young salmon were planted in the streams
of Marin County, Cal., in 1S97, after having been kept in the hatchery until the
yolk-sac was absorbed and I hey had begun to feed. In order to determine to what
extent they were preyed upon by the other fishes of the stream, large numbers of
trout and a few sculpins ivei’e caught and examined, being the only fishes in the
stream that could be suspected of eating salmon fry. Beginning at the time the
plants were made and continuing for three weeks, 30 or 40 trout, ranging from G to
10 inches in length, were daily caught and examined. In not one instance had a
salmon been eaten. The only fish eaten by them was the small minnow (Rutilus
symmetricus) , and no more than 10 of these were found in about 700 trout examined.
Of the sculpins ( Coitus gulosus), only 25 of size large enough to eat a salmon fry were
caught. None of these had eaten fish of any kind.

In 1898, after the young salmon planted that year had absorbed the yolk-sac, a
number of trout were examined. None were found to have eaten salmon fry.

On one occasion a small pool 8 feet across and about 18 inches deep was seined.
Over 100 young salmon were caught, averaging 2.1 inches in length. Along with
them about a dozen trout from 6 to 8 inches long were taken. It would seem that
if ever trout ate young salmon it would be here. These trout were examined, and

88

BULLETIN OF THE UNITED STATES FISH COMMISSION. '

it was found they had eaten only caddis larvae and periwinkles. On the Upper
Sacramento River I have examined many trout taken while the stream was full of
the small salmon fry, hut have never found that they had eaten young salmon.
The same is the case with the sculpin, and these are the only fishes to be feared in
the Upper Sacramento. Farther down stream many of the smaller Sacramento pike
have been examined, but none of them were guilty of eating young salmon.

In the spring of 1899, while observing the migration of the salmon fry on the
lower Sacramento River by means of a fyke-net trap, we occasionally caught cat-fish
along with the young salmon. In every case it was found that the cat-fish had
eaten salmon fry. Their capacity for young salmon was greater even than that of
the trout for alevins. Several cat-fish 9 inches in length were found with over 60
salmon fry in their stomachs, and one of this same size had eaten 86 of the fry
which averaged a little over TV inches long. To determine whether the cat-fish
captured the fry only while in the bag of the net we caught nearly 50 with hook and
line. The stomach of none of them contained a young salmon. Thus it is evident
that the cat-fish likes salmon fry and would catch them regularly if it could. It is
too sluggish a fish, however, to catch salmon fry under ordinary circumstances.

The only other fish at all likely to prey upon the young salmon in fresh water
is the striped bass ( Roccus lineatus), which is found in the lower river and in large
numbers in the brackish water of Suisun Bay. It is also found in San Pablo and
in San Francisco bays. I have no information on the subject, except that the
striped bass preys to a large extent on the carp in the sloughs of the lower rivers
and in the salt or brackish water feeds almost exclusively on small crabs. It is
significant, however, that both striped bass and salmon are increasing in numbers
in California waters, the former enormously, and it can not, therefore, be very
detrimental to salmon. Young pike, suckers, and split-tails are abundant in the
waters inhabited by the bass, and all are sluggish in comparison with the salmon.
It would seem that young salmon would be the last fish upon which 1 hey would prey.
A young salmon is very active and strong and much more shy than even a trout of
same size; after it has begun to swim about and feed it is perfectly able to take care
of itself, and the number killed by enemies in the Sacramento is very small.

MIGRATION OF FRY.

In 01 'em a Creelc. — The first year at Olema 150,000 fry, and the second year
.850,000 alevins, were released in Olema Creek. The stream was seined about a
month after the fry were planted in 1897, and in 1898 about a month after the time
when the fry should have begun swimming. Very few young salmon were taken in
either year, and the results show that over 95 per cent had left the stream within
the month.

Battle Creek station. — The observations in Battle Creek were made while we
were engaged with the hatchery experiments during October and November, 1898.
In obtaining data concerning the young salmon we used a 50-foot seine such as was
employed in nearly all of the investigations; but the most important device for this
work was a trap which caught the young salmon as they were going downstream.
The trap was made by sewing a piece of fine-meshed webbing across the mouth of
the bag of a 30-foot seine and fixing a funnel to extend back into the bag from the
middle of the webbing. It was set in a strong current just below the upper rack at

NATURAL HISTORY OF THE QUINNAT SALMON. 89

the Battle Creek fishery, with the wings extending obliquely upstream, their ends
being about 10 feet apart. The fish were deflected by the wings to the middle por-
tion of the net, and found their way through the funnel into the bag. No effective
means could be devised to prevent the funnel from becoming choked with leaves or
other trash, which often happened within an hour or two after the net was set.
There were many adult salmon below the rack, and they often tore the net with their
teeth and frequently got fast in the funnel. Part of the time the net was set during
the day, more often during the night. Sometimes it was set for only an hour or two
during the night. The following is a record of the catch, showing the date, the time
of day, and the number and size of the fry taken:

Record of salmon fry taken in trap at Battle Creek fishery , Oct. 7 to Nov. 30, 1898.

X ote. — The numbers in the column headed 1.5 ± and 5. ± indicate the number taken that were about
1.5 inches long, or about 5 inches long, as the case may be, but were released without measuring.

From the preceding record it will be seen that all of the fry (not including the
parrs) were practically of the same size, 1.5 inches long. Of the 322 fry examined,
only two were over 1.6 inches long, one being 1.7, the other 1.8. The 1.5-inch
specimens had just absorbed the yolk-sac. Indeed, there was often a small amount
of yolk remaining in the body, although the sac had disappeared. The size of these
specimens shows that they begin their downstream migration as soon as they begin
swimming, or at the age of six weeks; their continuing the same size during the
two months shows that practically all start downstream at the same age. If part
of them had held back for two or three weeks, this would have been indicated by
a greater variation in size.

90

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The record also shows that ordinarily the young salmon travel at night. The
trap was so set that they could not have avoided it had they traveled during the
day. That they can he caught during the day is proved by their being taken in the
open tow net set in Hatchery Creek, as noted above, under “Planting fry from
the hatchery . ” F ry were seen quite often lying in a pool near the shore during the day,
and were seen to rise to small insects that lighted on the water. They probably feed
more during the day, which makes their migration slower, or stops it altogether.

On November 30 there was a rise in the creek and the water was muddy. The
catcli from 8 to 0 a. m. was larger than the average night catch at other times, showing
that high and muddy water induces salmon fry to travel during the day. This fact is
borne out also by the work at Walnut Grove, an account of which is given below.

The great variation of the catch when the net was set for an hour or two during
the night indicates that they travel in schools. »

A trap similar to that used in 1898 was set in Battle Creek in 1900 from Septem-
ber 13 to October 4. Salmon fry 1.5 inches long were taken, two or three at a time,
from September 18 on. The downstream migration, therefore, begins at least as early
as the middle of September.

Adult salmon can be found in some part of the river throughout the year, and
the spawning season is therefore very long. It is probable that there are salmon
spawning at some place in the river or its tributaries in every month of the year.
They are spawning in considerable numbers from July till January, inclusive. With
such an extensive spawning period, it is obviously difficult to separate the young
according to size, and say that t hose of a certain size belong to the spring or fall run
of a certain year. A variation in rate of growth, noted elsewhere, adds to the difficulty.
However, in the following table of measurements of specimens taken with the
seine at Battle Creek fishery during October and November, 1898, three sizes may
be distinguished, which doubtless represent three runs of adults. Those from
1.4 to 2.2 inches in length were from the summer run of 1898; the 3.7 to 4.7 inch
specimens from the fall of 1897 (and they doubtless were among the last to hatch);
and the 6.2-incli specimens an earlier run, probably the summer run of 1897.

Measurements of young salmon taken ■ with the seine. Battle Creek, October 18 to December 1, 1898.

Size.

No.

Size.

No.

1

13

1

1

1

3.9 inches

1.5 inches

1 inches

4.1 inches .

1

4.2 inches

1.8 inches

4.3 inches

4.4 inches

2

1

2

1

2

None.

2

4.5 inches .

2.1 inches

.2

1

None.

1

4.6 inches . .

2.3 .to 3. 6 inches

4.8 to 6.1 inches

Balls Ferry station. — An observation station equipped with a trap similar to that
used at Battle Creek fishery was established on the river at Balls Ferry, about 3
miles above the mouth of Battle Creek. Observations were made by Mr. Chamber-
lain, beginning January 6 and closing April 25, 1899. The following table gives the
data obtained at this station.

NATURAL HTSTORY OF THE QUINN AT SALMON

91

Record of salmon fry taken in trap at Balls Ferry, January 6 to April 25, 1899.

Date.

Catch.

Size, inches.

Aver-

age

size.

A. M.

M.

P. M.

1.4

1.5

1.6

1.7

1.8

1.9

2.0

2.1

3.1

Jan. 6

8

9

10

13

11

21

22

23

24

25
20

27

28

29

30

31

Feb. 1
2

3

4

5

6
7
9

10

11

12

13

14

15
10

17

18

19

20

23

24

25

26

27

28

Mar. 1
2

3

4

5

6

7

8
9

10

11

12

13

14

15
21
22
31

Apr. 1
2

3

4

5

6

7

8
9

10

LI

12

13

14

15

16

17

18

19

20
21
22

23

24

25

24
37
14
33
39

25
9

4

18

2

1.49

2

11

0

1

— -

;i

10

7

24

5

1

1

1.51

35

44

22

13

1

6

12

13

15

18

5

5

15

5

6
0
1

15

29

15

2

6

4
6

11

5
3

3

4
2
0

(.)

25

14

1

1

1

1.53

19

250

158

104

39

56

42

87

183

173

192

322

326

241

239

131

70

100

117

213

157

217

237

143

163

100

137

100

59
77

128

227

137

45

166

70

53

95

35

60
27
86

109

121

130

65

114

45

76

254

30

89

30

2

1.51

1

2

7

....

....

0

2

1

■ 1
0
0
10

4

2

9

}“5

41

4

2

1.48

0

82

70

61

01

45

41

12

28

5

14

14

29

30

31
28

15
10
13

36

24

14

5

15

20

55

14

25
5

20

26
24

29

78

23

3

1

1

1.51

7

0

0

34

101

88

13

1.52

1

5

8

2

3

0

0

1

1

0-

0

1

1

1

0

0

5

['

2

6

7

1

1

1.58

4

1

1

1

3

3

5

0

1
0
0
0
0

7

5

0

2

1

1

0

2

2

2

1

0

0

(J

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

92

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The preceding table needs but little explanation. It indicates that the greater
part of the young salmon from the fall run passed Balls Ferry between the middle
of January and the middle of March. Practically all had passed by March 20.
Measurements taken January 6, 13, 21, and 30, February 19, March 7, 15, and 31, and
April 1, 2, 3, and 4 show that the average size of those taken on the dates specified,
during a period of 34 months, varied but one-tenth of an inch. The average of all
measurements is 1.53 inches. No satisfactory estimate of the number passing could
be made, except that there were probably many millions. This record also proves
that salmon fry begin migrating as soon as they are able to swim, and that practically
all start downstream at that age, otherwise the later ones would have been larger.

It was also ascertained that a large migration was not coincident with remark-
ably high water. It is probable that when the fry once enter the main river their
migration is not impeded by low water; but it seems probable, from observations
noted in another place (see “Summer residents”) that many of the late fry that
hatch in the headwaters are detained there during the summer by low water.

Walnut Grove station. — Our knowledge of migrations through the lower part
of the river was gained from the general investigation of 1898, and especially from
observations made by means of a trap established at Walnut Grove from January
to May, 1899. This trap was constructed especially for the work, but was hardly
more efficient than traps made from seines and used at Battle Creek and Balls Ferry.
It consisted of a bag with a short funnel hung to a 4-foot hoop, with wings 20
feet long. At Walnut Grove the Sacramento makes a sharp bend, changing its
direction from southeast to southwest. At this bend Georgeanna Slough breaks off
and continues the southeasterly direction of the river above. It thus gets a large
amount of water, probably half as much as the river below, and is in the direct
path of the migrating fry. The trap was set about 150 yards from the head of the
slough, which at that place is about 75 feet wide and 15 to 20 feet deep. The banks
are abrupt and covered with bushes. One end of the trap was fastened by a long-
rope to a tree on the bank, the other to a buoy anchored about the middle of the
stream. It was sometimes set in other positions in the slough or in the river, but
without results of particular value. During a sudden rise in the river it could not
be set on account of the great amount of trash in the water.

The following gives the record of the catch. In the column headed “A. M.” is
given the number of fry found in the trap at 8 a. m., and in the “P. M.” column
the number caught between noon and 5 p. m.

Record of salmon fry taken in the trap at Walnut Grove. January 7 to May S. 1899.

NATURAL HISTORY OF THE QUINN AT SALMON

93

Record of salmon fry taken in the trap at Walnut Grove, January 7 to May S, 1899 — Continued.

Date.

Catcli.

Size, inches.

Aver-

age

size.

A.M.

M.

P.M.

1.4

1.5

1.6

1.7

1.8

1.9

2

2.1

2.2

2.3

2.4

2.5

2.6

2.7

2.8

2.9

3

3.1

3.2

3.3

3.4

3.5

3.9

4.1

Feb. 7
10
11
12

13

14

15

16

17

18

19

20
21
22

23

24

25

Mar. 1
2

3

4

5

6

7

8
9

10

11

12

13

14

15

16
17

19

20
21
22

27

28

30

31

Apr. 1
2

3

4

5

6

7

8
9

10

11

12

13

14

15

16

17

18

19

20
21
22

23

24

25

26

27

28
30

May 1

3

4

5

6

7

8

3

1

0

2

3

0

L

4

2

3

1.79

1

5

5

5

i

3

0

6

2

1

3

2

2

1

1

1

0

4

0

4
1

5

4
0

5
0
1

i

1

5

2

4

3

3

1

3

1

1.81
1. 81

2

2

5

3

i

1

i

1

1

1

1

1

1

1

1

1

1

1.82

2.00

(

1

2

1

1

2.11

8

2

14

19

29

5

13

30

128

4

45

T

1

1

5

1

3

6

14

12

6

3

5

17

27

16

1

1

1

T

3

2

1

1.85
1. 75
1.81

1.73

1.74
1.79
1.79

1.75

5

17

89

44

3

9

24

13

3"

8

5

2

3

1

3

1

1

1

1

1

81

246

36

39

175

183

101

12

17

80

19

o

2

7

10

7

10

4

8

2

1

300

200

138

168

51

43

43

92

67

33

25

l

3

1

225

...

2

1

3

6

12

10

6

4

1.70

1.68

1

64

36

21

4

7

316

153

64

2

1

12

376

263

48

1

3

3

8

2

3

1

1

1.77

1.70

16

278

1

1

11

10

8

12

10

14

25

18

13
17
48

24

14
19

15

7

8
22

7

8
7
5
2

25
10

5

3
2
5
7
5

4

4

5
4
3
1
3

1

1

4

1

2

2

2

1

2

1

1

1.70

1.85

1

2

1

1

1

2

3

2

1

1

3

3

2

3

1

1

3

2

1

1

10

1

5

1

4

2
2
2

5
4
2
2

i

1

1

1.94
1.87
1.85
1.97
1.99
2.03
2. 07
2.03
2.07
2.19
2.11
2. 27

2.33

2.34
2.48
2.26
2. 50
2. 48

1

3
1

4

5
5
5
7
7
4
4
3

1

1

3

4
3
2
8

3
2

4
1
1

5

1

1

l

2"

1

2

3
2
2
2

4
1
3
1
1
1
1

2

1

y

1

1

1

1

1

2
1

1

1

4

3

2

2

2

Y

1

1

2
1

2

1

3

2

2

T

1

1

1

1

1

2

1

2

1

2

1

2

1

1

1

3

1

1

1

2

1

3

2

'1

1

2

1

T

2

1

1

1

1

2

1

1

2

2

1

2.50
2.69
2. 63
3.03
2.68
3.00
3.08
3.00
2. 74
3. 00
2.60
2.60
2.78
3. 17
3.00
2. 77

1

1

2

y

1

6

i

1

y

3

1

1

3

1

3

1

1

1

1

1

1

1

1

1

1

2
1

2

"i"

1

1

T

1

1

2

1

1

2

y

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

i 1 1

1 1

.1

From the data above given it wilL be seen that —

1. From the middle of .January to the middle of May there were salmon fry in
various numbers passing Walnut Grove.

94

BULLETIN OF THE UNITED STATES FISH COMMISSION.

2. 'I’he height of migration was from March 4 to about the 24th, about 20 days.

3. On March 8 and 20 there were two large runs of fry.

4. Practically all had passed by the 22d of April.

5. The average size of those taken during January was 1.6 inches, during
February 1.8 inches, during March 1.7 inches. From the 30th of March till May 7
the size gradually increased from 1.7 inches to 3 inches.

6. No fry were taken during the first 9 days in January.

7. Young salmon traveled as much during the day as during the night.

Comparing the information for Walnut Grove with that for Balls Ferry, as

shown in the accompanying diagrams, plate 12, it appears that —

(a) The Balls Ferry run of February 2 reached Walnut Grove March 8 and was
34 days making the distance. The fry increased in size 0.3 inch.

(b) The Balls Ferry run of February 14 reached Walnut Grove March 20, and
was 34 days making the distance. The fry increased in size 0.3 inch..

(c) The runs that passed Balls Ferry February 25, and later, were caught by
high water the latter part of March, which probably carried them down faster. The
runs were not noticed at Walnut Grove; the net could not lie worked during the
beginning of high water, March 23 to 29, during which time they may have passed.

(d) The fry taken at Walnut Grove after April 1 had grown more than 0.3 inch
since starting downstream, and were therefore the stragglers from the regular
migration. Those taken during May had probably been three months on the way.

It is evident, therefore, that the fry of the regular migration require about 34
days to pass from Balls Ferry to Walnut Grove.

The distance between the two stations is about 350 miles. An object floating
as fast as the current would make the distance in about 9 days. It requires 8 days
for a rise in the river to travel from Red Bluff to Sacramento. If the fry traveled
only at night, and simply kept with the current, they would make the distance in
18 days. There is no doubt that in migrating the fry drift downstream tail first,
keeping the head upstream for ease in breathing as well as for convenience in catch-
ing food floating in the water. In this way they would drift much more slowly than
the current. At Battle Creek hatchery fry have been observed traveling with the
current, and always with the head upstream unless frightened.

The later and larger specimens found had simply been longer on the way. The
larger they became the more slowly they drifted, as they swam against the current
more strongly. Those taken at Walnut Grove in January were but 1,6 inches long,
being brought down by the high water in January, the short time they had traveled
being indicated by their smaller size.

The failure to catch any fry during the first 9 days in January indicates that the
fry from the summer run had all passed and that those from the fall run had not
yet reached Walnut Grove. Without doubt there were a few passing at that time,
for there were some passing Battle Creek as late as December 6, but they were so few
that none, were taken in the trap. It is possible that there are a few passing down
the river all summer, though we have been unable to find any after June.

Observations at Benicia. — February 21 and 24 and March 3, 1899, five specimens
1.8 to 1.9 inches long were taken in Carquinez Straits at Benicia. The average size
at Walnut Grove after February 10 was 1.8 inches and the size of the Benicia speci-
mens indicates a short passage between the two places, probably not over a week.
This would make the time from Battle Creek hatchery to brackish water 6 weeks.

Bull. U. S. F. C. 1902. (To face page

Plate 12.

44

<b

*

Sa/mon_

Jan

Apr//.

21

20

19

18

17

16

15

14

13

12

II

10

9

8

7

6

5

4

3

2

1

330

320

310

300

290

280

270

260

250

6

7

8

9

10

ii

12

13

14

15

16

17

18

5

6

7

8

9

10

II

12

13

14

15

16

17

18

19

20

21

22

23

24

25

-

240
230
220
210
200
190
180
170
160
150
140
130
120
1 10
100
90
80
70
60
50
40

-

-

-

l

1

1

»

i 1

K

!

1

1

*

'» 1

»

|#

'

/

--

-•

A

r\

—

-

...

-

~

30

20

/

V/

\

X

J

10

Bull. U. S. F. C. 190:

Pi atf 1 0

DIAGRAM SHOWING NUMBER AND SIZE OF YOUNG SALMON TAKEN AT WALNUT GROVE, 1899.

s indicate number of fry; dotted lines indicate height of river; vertical spaces indicate 10 fry each, or one foot for height of water, or tenths of inch for length of fry; the vertical lines of diagram

Heavy lines

J? represent the exact average length of the fry taken.

NATURAL HISTORY OF THE QUINNAT SALMON.

95

General investigation (luring 1898. — The following notes give the data obtained
by the general investigation in 1898.

Table showing greatest number of f ry taken at one haul of the seine.

Station.

Apr. 28
to

May 4.

May IS
to 30.

•T uly 8 to
27.

Station.

Apr. 28
to

May 4.

May 18
to 30.

July 8 to
27.

50

66

8

Princeton

15

6

0

Colusa

0

60

2

4

0

Red Bluff

150

1

0

4

0

Six miles below Red
Bluff

6

(1

Twenty miles below

0

2

0

100

31

0

ii

0

0

1

0

Sacramento

10

6

<1

0

2

(1

8

0

Rio Vista. - __

0

28

(1

Collinsville __

0

0

10

II

2

0

10

0

4

(1

Table showing size of young salmon taken in the Sacramento River in May , 1898.

Dates and localities.

Size in inches.

Aver-

1.4

1.5

1.6

1.7

1.8

1.9

2.0

2.1

2.2

2.3

2.4

2.5

2.6

2. 7

2.8

2. 9

3.0

3.1

3.2

3.3

3.4

3.5

3.6

3.7

3.8

3.9

4.5

age

size.

Sisson, May 15

2

2

5

2

2

1

2

4

2

1

2

3

3

i

2

1

1

2.4

Sims, May 17

2

3

2

1

2

5

i

3

2

3

2

3

1

2

Hazel Creek, May 17

2

3

i

4

2

2.0

Redding, May 4 and
IS

1

3

1

3

3

4

2

o

3

1

3

2

2

1

2

i

i

i

2.4

River at Battle
Creek, April 30. ..
Red Bluff, April 28,
May 20..

1

1

4

3

4

4

4

i

2

i

2

1

1

2.1

2

1

1

1

4

4

4

1

i

1

1

i

2

i

2

2.1

Tehama, May 5 and

21

i

3

4

1

i

i

2

i

i

9;

i

1

1

2.1

River at Thomas
Creek, May 22

1

4

]

•>

1

2 5

River at Deer Creek,
May 22

i

i

i

i

1

2

i

2.2

Chico Bridge, May
23

1

i

1

1

i

1

1

2.7

Jacinto, May 24

1

i

i

1

i

1

2 4

Butte City, May 25 .

i

1

1

2

2.1

Princeton, May 25..

3

1

1

2.6

Colusa, May 26 . ...

1

i

i

2.4

Grimes, May 27

i

1

2 3

Wilson farm. May

27

i

i

i

2.3

2.3

Knights, May 29 . .

i

3

2

Sacramento, Apr. 23
Rio Vista, May 11
and 30

2

1

1

i

3

2

1.9

2.3

2

2

Benicia, May 13and
23.

1

1

1

1

1

1

i

2. 6
2.7

San Pablo Bay, May
18, June 17

2

2

i

i

i

1

From the tables given it will be seen that —

1. Young salmon were abundant in the river the first of May, at least between
Redding and Tehama. As a few were taken at Sacramento April 23, it is probable
that they were distributed throughout the river.

2. The last of May they were nowhere so abundant as they had been at Red
Bluff and Tehama three weeks previously.

3. A few were found throughout the river May 18 to 30.

4. The fry found in the river in May had an average size of about 2.2 inches.

5. In July there were no fry found except a few at Redding and Battle Creek.

96

BULLETIN OE THE UNITED STATES FISH COMMISSION.

It is known from the work at Balls Ferry in 1899 that practically all the fry
leave the river before March 20. It is evident, therefore, that the fry found between
Redding and Tehama in May, 1898, were the stragglers left from the regular migra-
tion. Their size, 2.2 inches, precludes their belonging to the regular run. They
had collected in the pools where we did our seining, as they do in the headwaters
during the summer (see notes on “ Summer residents”), which made them appear to
be more abundant than they probably were. The rise in the river, which occurred
from May 15 to 25, and the accompanying muddy water caused them to pass down-
stream. This is indicated by finding fewer in the pools in the latter part of May
(when migrating they would not be collected in the pools) and by finding none at
all in July.

The conditions in 1898 were exceptional on account of the early occurrence of low
water — from the middle of March till the middle of May — though doubtless there are
always a few stragglers from the regular migrations. These decrease in number
and increase in size (slightly) till the rains of the following winter, when all leave
the river.

In the mountain streams the young salmon prefer the pools, where they are
often abundant. Nearly 500 were taken at one haul of the seine in a pool at the
head of Box Canyon, near Sisson, in August, 1897, and it was not at all uncommon
to catch over a hundred at a time in many of the pools of the headwaters. The
rapids have been fished a number of times, but young salmon were scarcely ever
caught unless the water was at least 2 feet deep.

Below Redding more salmon were found in the water with moderate current,
gravelly bottom, and a depth of over 2-1 feet, but none was found in absolutely still
water, and none over a soft mud bottom. Not much seining was done over rocky
bottom, on account of the strong current and the injury to the seine by its picking
up cobblestones. A few salmon were caught by putting enough floats on the seine
to keep it at the surface and then hauling in water 15 feet deep.

The following table indicates the various characters of stream in which young
salmon were found in the main river, with the number taken in one haul of the seine
under the various conditions.

Table showing number of young salmon taken in various stream conditions.

Localities.

Month.

Current.

Bottom.

3l-feet depth.

Slight.

Medium.

Strong.

Mud.

Clay.

Sand.

Gravel.

Rocks.

Under.

>

o

116

116

116

60

60

60

60

60

60

Red Bluff

do

36

178

178

314

do . . .

31

100

31

100

100

31

..do ...

13

13

13

do . ..

34

54

32

46

18

60

... do . . .

33

3

10

13

1

13 .

ii

i

23

16

7

8

15

40

8

48

17

17

17

do . .

8

8

8

4

4

4

3

1

2

5

10

10

10

4

4

4

Total

11

334

580

46

233

375

161

209

606

NATURAL HISTORY OF THE QUINNAT SALMON.

97

Movements in estuary and bay. — Much seining was done both years at Olema
in trying to learn something of the movements of young salmon in brackish water.
None was found in 1898. A few were caught near the mouth of Paper-mill Creek in
1897, and one was taken 2 or 3 miles from the mouth of the creek, across the head
of the bay. The net was stretched across the mouth of the creek for 15 minutes
during the flood-tide, and two salmon fry were taken, indicating that they run back
and forth with the tide.

The fishermen at Marshall, on Tomales Bay, about 20 miles from the mouth of
Paper-mill Creek, reported having taken young salmon inconsiderable numbers the
last of April, 1897, about 50 days after they were liberated in the streams near
Olema. At that time the salmon were about 100 days old and were large enough to
be taken in the seines used by the fishermen. They caught as many as 15 or 20 at
a haul for about a week, and caught them occasionally till the middle of June. I
think the report reliable, as the salmon was a new fish for the bay, and would
attract much attention. This indicates that the fry may reach the ocean at the age
of three months. The water at Marshall is pure sea water.

Specimens have been taken in brackish water in Suisun and San Pablo bays,
but not enough to determine their movements. A few about 10 weeks old have been
taken at Benicia in water that was about 20 per cent sea water.

Effect of sea ivater on alevins and fry. — To determine the effect of sea water on
alevins and fry, 25,000 eyed eggs were taken from Battle Creek hatchery to the
Hopkins Seaside Laboratory at Pacific Grove. The eggs were received at the labora-
tory December 10, 1898, and most of them hatched on the 17th, which date was taken
as the basis for determining their age in the various experiments. Those not being
experimented with were cared for as alevins and fry ordinarily are.

The first experiments were made by putting a few alevins directly from fresh
water inf o battery jars filled with various mixtures of fresh and sea water. In the
later experiments glass tanks 2 and 3 feet long were used, and the water was kept
running. The experiments were begun when the alevins were 6 days old. It was
found that at this age they could live indefinitely in water that was 25 per cent sea
water. Those about 40 days old could live in 50 per cent sea water, and at 50 days
75 per cent. Those 60 days old could live in 95 per cent, though there was consid-
erable loss. Ninety-live per cent was as nearly pure sea water as could be obtained,
the laboratory pump being broken and the tank partly filled with fresh water. The
loss was much less when the density alternated between a high and low percentage,
which indicates the value of the change of density in the estuaries with the rise and
fall of the tides.

When the younger alevins were placed in 50 per cent sea water or stronger,
the yolk was solidified, becoming much like soft rubber. The blood was driven
from the body, making it appear bleached, and the adipose membrane at each edge
of the tail adjacent to the caudal fin turned white. The circulation was retarded
and the fish became sluggish. The only noticeable effects on the older alevins were
sluggish movements and an inability to keep a horizontal position. Sometimes
death was immediately preceded by violent and spasmodic swimming in any and all
directions. The same actions were noticed in minnows placed in a strong mixture
of sea water.

F. C. B. 1902—7

98

BULLETIN OF THE UNITED STATES FISH COMMISSION,

Tlie following table gives a record of the experiments. The percentage of sea
water is reckoned by taking pure sea water as a standard of 100, a mixture of equal
parts of fresh and sea water being 50 per cent. The percentage of dead alevins at any
one time is hot based on the original number with which the experiment began, but on
the number left in the vessel the previous day, e. g., if we start with 40 alevins and
10 die the first day, that is 25 per cent; if 15 die the second day, that is 50 per cent.

Table showing effect, of salt water on alevins.

Percentage died.

Day of
experi-
ment.

Exp. 1.

Exp. 2.

Exp. 3.

Exp. 4.

Exp. 5.

Exp. 6.

Exp. 7.

Age.

Presli

water.

25 per
cent sea
water.

50 per
cent sea
water.

75 per
cent sea
water.

Sea water.

Few
changed
from No.
2 to 50 per
cent.

Few
changed
from No.
6 to 75 per
cent.

i

6

0

0

0

0

0

0

0

20

100

0

3

8

0

2

0

100

0

0

4

9

0

0

100

0

100

10

0

0

0

6

11

12

(1

0

100

0

0

8

13

0

0

9

14

0

0

in

15

o.

ii

16

0

0

3

12

17

0

13

18

19

(1

0

14

0

0

15

20

0

0

10

21

2

17

17

??■

0

0

IS

23

(1

19

24

0

0

20

25

0

0

21

20

0

0

23

27

0

8

23

28

0

0

24

25

29

2

0

30

0

0

20

31

0

0

27

28

33

33

0

1)

2

0

39

34

0

0

30

0

0

31

36

Discontinued.

Day of
experi-
ment.

Age.

Percentage died.

Density raised gradually to —

Placed
directly
into sea
water.

Experiment 12.

25 per cent.

50 per cent.

75 per cent.

Age.

Percents
age sea
water.

Percent-
age died.

Exp. 8.

Exp. 9.

Exp. 10.

Exp. 11.

i

13

0

0

0

0

19

50

0

2

13

2

0

85

100

20

50

0

3

14

0

7

50

21

50

50

4

15

5

8

100

22

50

7

10

0

7

23

50

0

6

17

29

89

24

50

33

50 per cent.

7

18

ii

10

25

50

43

8

19

9

100

26

50

0

9

20

0

27

50

50

10

21

10

28

75

0

11

22

15

29

75

50

12

23

45

30

75

100

13

24

8

14

25

63

15

26

50

15 per cent.
25

16

27

17

28

33

18

29

100

NATURAL HISTORY OF THE QUINN AT SALMON.

99

Table showing effect of salt water on alevins and fry.

Experiment 13.

Experiment 15.

Experiment 16.

experi-

Percent-

Percent-

Percent-

Percent-

Percent-

Percent-

ment.

Age.

age sea

age

Age.

age sea

age

Age.

age sea

age

water.

dead.

water.

dead.

water.

dead.

1

26

10

0

33

0

0

33

0

0

2

27

15

0

34

10

2

34

5

1

3

28

25

1

35

35

0

35

15

5

4

29

35

1

36

40

3

36

20

5

5

30

50

7

37

50

5

37

25

5

6

31

65

84

38

55

7

38

30

1

7

32

75

100

39

60

7

39

35

2

8

40

65

34

40

40

t

9

41

65

58

41

45

8

10

42

70

83

42

50

2

n

43

80

66

43

52

6

12

44

90

100

44

70

10

13

45

70

20

14

46

80

41

15

47

65

23

16

48

60

2

17

49

60

8

18

50

60

7

19

51

65

0

20

52

85

0

21

53

70

0

22

54

65

0

33

55

93

2

34

56

50

0

25

57

93

0

26

58

93

0

27

59

95

8

28

60

95

0

29

61

Discontinued.

Experiment 1 was made to cheek the others, showing that the loss was not the
result of being confined in battery jars.

Experiment 2 shows that 25 per cent sea water lias but little deleterious effect
upon alevins over 5 days old.

Experiments 3 to 7 show that alevins of 6 to 10 days of age can not live in sea
water of 50 per cent or over, either when put directly from fresh water into the
mixture or when the density is gradually raised.

Experiments 8 to 11 show that alevins 12 days old can live longer in 50 and 75
per cent sea water than those only 6 days old. Some of the 12-day alevins lived
7 days in 50 per cent sea water, while the 6-clay ones lived but 3 days. The older
alevins lived one day longer in 75 per cent, but died the first day in pure sea water,
as did the 6-day individuals.

Experiment 12 shows that when 19 days old they live longer yet in 50 per cent.
Two lived 2 days in 75 per cent after having been in 50 per cent for 9 days. The
one that lived till the twelfth day was a week older.

Experiment 13 shows that a gradual rise from a density of 10 per cent sea water
when 26 days old to 75 per cent when 32 days old was fatal. In a similar way a
gradual rise from fresh water when 33 days old to a density of 90 per cent when 44
days old was fatal, as is shown in experiment 15.

Experiment 16 indicates that alevius 50 to 60 days old can bear a high density
of sea water, but that they can withstand it better if the density, instead of increas-
ing regularly, alternates between high and low. It does not show the exact age at
which the fry can live in sea water, and it is doubtful whether this can be determined
accurately in aquaria. Whether it can or not, the time was not at my disposal to
carry the observations further.

As a whole, the experiments show three important points. First, the fry can
not live in sea water until several Aveeks after the yolk-sac is absorbed; second, when

100

BULLETIN OF THE UNITED STATES FISH COMMISSION.

able to live in sea water they can not go directly from fresh water to sea water, but
must pass gradually; third, they are greatly aided by an alternation of densities such
as is obtained by passing through an estuary. For these reasons it would not be well
to plant fry in a stream that does not reach the ocean through an estuary.

Change of color daring migration. — The color of young salmon depends much
on the character of the water in which they live. Those in small, cold streams are
much more dusky and have the parr marks strongly developed. They become
lighter in color upon entering the main river. Those in brackish and salt water are
bright silvery on the sides, with the back sea green. Parrs 4 to 6 inches in length,
found in Battle Creek and similar places, have the sides bright silvery, the back
olive brown, with the upper end of the parr marks making regular shadings along
the back. Specimens 2.6 inches in length from Rodeo, San Pablo Bay, May 18, have
distinct parr marks; 2.7 inches, from Benicia, May 13, have lost them. Sometimes
the caudal fin is reddish; sometimes there are yellowish stripes on the ventral and
anal fins, especially with fry about 1.5 inches long.

Summary of observations on migration. — The fry begin their downstream migra-
tion as soon as they are able to swim. In the Clearwater they travel more at night;
in muddy water, as much or more during the day. Much of the time they float
downstream tail first, and in the larger streams they travel more or less in schools.
In the larger streams their downstream movement is not dependent upon the height
of the water, but upon age. From October to April, inclusive, over 99 per cent that
pass the vicinity of Battle Creek are of the same size, 1.5 inches long. They pass
down the river at the rate of about 10 miles a day, and are about 6 weeks reaching
brackish water, being 3 months old at that time. They are probably 4 or 5 months
old when they reach the ocean. The ebb and flow of the tide in the estuary, causing
an alternation in the density of the water, is apparently beneficial.

SUMMER RESIDENTS IN THE RIVERS. “

General account. — In the upper portion of the Sacramento River there yet
remained, after the winter and spring migration in 1898, a large number of young
salmon. In the vicinity of Sims we found from 700 to 1,000 in the various pools.
We found them common in the McCloud at Baird in September, and in Fall River
in August. These summer residents, as they may be called, are confined to the
headwaters — the clear streams with rocky bottoms. They do not stay much of the
time in the very swift current or riffles, but remain in the more quiet pools. Here
they feed on aquatic insects and take the angler’s fly the same as trout.

Most of the data concerning the summer residents was obtained from investi-
gations near Sims, in Hazel Creek, and the river below its mouth. Hazel Creek is
a small mountain stream, with many pools and gravelly riffles, and is a favorite
spawning stream both for salmon and trout. The two lower pools, which are about
a quarter of a mile from the mouth, were seined several times during the summer
and fall, and it was from this work that we learned much that we know of the habits
of the fry remaining in the streams during the summer. In the table below one of
these pools is called the upper and the other the lower.

The Sacramento River in the vicinity of Sims is about 40 or 50 feet wide, and
during the summer has an average depth of about 3 feet. It is very swift except

ii These notes on the “ summer residents ” are given largely as a matter of record. While the conclusions drawn
in some cases are scarcely warranted, yet the available data point toward them, and too many of the estimates closely
approximate each other to be the result of mere chance.

"NATURAL HISTORY OF THE QTTINNAT SALMON".

101

in tlie pools, which were the only places that could be seined. Seven of these pools
were seined frequently, and for convenience in keeping notes we numbered them
A, B, C, D, E, F, and G, beginning with the upper.

In July and August all specimens taken in Hazel Creek and in the river near
by were marked by cutting oft the adipose fin with a pair of small curved scissors.
This enabled us to know when we were taking specimens that had been taken before.

Upstream movement. — The following is a record of the seining in Hazel Creek.
The data for each seine haul consists of the date, the pool where it was made, the
number of fry previously marked in lliis pool, number of days since the last were
marked, total number taken, and number of marked lislies taken.

Record of seine hauls in Hazel Creek.

Date.

Pool.

Previ-

ously

marked.

Days
since last
marking.

Total
catch .

Catch of
marked
fish.

Date.

Pool.

Previ-

ously

marked.

Days
since last
marking.

Total

catch.

Catch of
marked
fish.

May 17.

14

Sept. 18..

Lower .

104

32

52

39

20

Upper..

107

32

40

30

July {)

do _ . -

55

Oct. 18. . .

Lower .

104

62

125

87

Aug. 17.

....do...

48

39

49

12

Upper. .

107

62

94

48

....do...

48

39

45

5

Nov. 18. _

Lower .

104

93

93

22

do

48

39

30

2

Upper. .

107

93

113

54

Upper. .

0

78

i

Dec. 18..

Lower

104

123

4

1

1

Upper.

107

123

II

0

do...

37

0

It will be seen from the table that in seining the lower pool in August, 1!) young
salmon were obtained that had been marked in July. As there had been but 48
marked in this pool in July, it indicates that at least 40 per cent of the fishes that
were in this pool July 9 remained until August 17.

In the upper pool we found two July-marked fishes in August, where none had
been marked in July. As the lower pool is about 100 yards from the upper, this
indicates that at least 4 per cent of the fishes in the lower pool had ascended the
stream that distance.

There were 104 specimens marked in the lower pool during July and August;
39, or 38 per cent of these, were found there in September; 36, or 35 per cent of the
107 marked in the upper pool in August, were found in September. Only one haul
of the seine was made in each place in September.

In August four-sevenths of the marked fishes found in the lower pool were taken
in the first haul of the seine. (When more than one haul was made the marked fishes
taken were held till the seining was over, in order that they might not be counted
twice.) Assuming that the same proportion was taken in the one haul in September,
we would reason that there were 68 marked salmon in the lower pool that month; 68
would be 65 per cent of the number marked — that is, 65 per cent of the fishes in Hazel
Creek on August 17 remained until September 18. This approximates the estimates
made for the pools in the river below the mouth of Hazel Creek. (See notes below.)

In a similar way, 36 per cent of the 104 marked in the lower pool were found in
one haul in October, and 45 per cent of the 107 marked in the upper pool. Thus
there was a loss of 2 per cent over the previous month in the lower pool, and a gain
of 10 per cent in the upper, indicating an upward movement.

But the upward movement is indicated better by the simple statement of num-
bers, as given in the table. In September there were 39 marked fishes taken in the
lower pool to 36 in the upper; in October the ratio was 37 to 48, and in November it
was 22 to 54. It is difficult to see how this can mean anything else than that the

102

BULLETIN OF THE UNITED STATES FISH COMMISSION.

young salmon in Hazel Creek continued to work their way upstream during Sep-
tember, October, and November.

This table shows an increase in the number of salmon in October. The average
of two hauls in May was 17; one in July gave 55; six in August gave an average of
19; two in September averaged 46; two in October 110; two in November 103. The
simple fact of there being more fishes found on later dates would not indicate an
upward movement; they might have come downstream. The larger percentage of
marked fishes in the upper pool, however, would indicate an upstream movement.

Several places above the upper pool were fished in October — one place within 15
yards of it — but no marked salmon were found. This would indicate a lack of upward
movement; but, all the data being considered, it is evident that there was at least
a slight upward movement in October and November.

In December the young salmon had all disappeared from the upper pool, and only
1 were found in the lower. One of these was a specimen marked in July or August.
That is, 1 out of 200 remained after the December rise. It is evident from this that
practically all the young salmon left the creek between November 18 and December
18. There was a heavy rain in the vicinity on November 28 and 29 (precipitation
over 2 inches), and there can be little doubt that the salmon all left at that time.

Migration during summer. — Pool B (see plate 11) of the river is separated from
Pool A (at the mouth of Hazel Creek) by a rapid about 150 yards long with a fall of
about 1 feet. It is over 6 feet deep, with large angular rocks along one edge afford-
ing excellent hiding-places for young salmon. It required three men to seine this
pool well, one to throw out the seine from a large rock at the upper end, and two to
pull it in. When there were but two of us, one would pay out the seine from a
riffle above the pool; the other would wade out as far as possible in the lower end
of the pool and pull the seine down with a rope. When the seine was stretched
through the pool it was pulled ashore. Obviously such work was not very satisfactory.

The following table gives a record of the catch at each haul of the seine in Pool
B in July and August. Four hauls were made in this pool in May, about 50 young-
salmon being taken in one haul. The specimens caught in May represented two
sizes, such as were found in Hazel Creek on the same date, but there were very few of
the smaller size. These smallest were probably from a few late-spawning individuals.

The table gives for each haul of the seine in July and August, (1) the date, (2)
number of young salmon caught, (3) number previously marked and released in the
pool in July, (4) time since the Juty-marked fishes were released, (5) number of July-
marked fishes caught, (6) number previously marked in August, (7) time since the
August-marked fishes were released, (8) number of August-marked fishes taken.

These data may be used to estimate the number of young salmon in the pool by
making the following proportion for any particular haul of the seine: The number
of marked fishes taken is to the total number of marked fishes known to be in the
pool (having just been released), as the total number taken is to the toted number
in the pool. The results, of course, are variable, and it is only by a number of trials
that Ave can get near the probable truth. The value of the estimates is not enhanced
by there being no marked fishes taken at certain hauls. In such cases, however,
there were but few of either kind. The estimate made from each seine haul' is given
in the table.

The freshly marked specimens could be distinguished from those marked a
month previously by the latter having the scar healed.

NATURAL HISTORY OF THE QUINNAT SALMON.

103

Table of data obtained from seining Pool B.

Date.

Num-
ber of
young
salmon
caught.

N umber
previ-
ously
marked
in July.

Time since
J uly mark-
ing.

N umber
of July-
marked
fishes
taken.

Number
previ-
ously
marked in
August.

Time since
last August-
marked
fishes were
released.

N umbel1 of
August-
marked
fishes
taken.

Estimated
number
of young
salmon
in pool.

July 9

177

23

167

0

25

190

6

792

22

209

5

29

209

do"

9

15

1 to 4 hours.

1

435

Aug. 15

IT

209

3T days

4

34

1 day

1

5T8

9

209

. do .

0

46

(»

29

209

....do

6

55

do

1

1,595

9

209

do

1

do

1)

Aug. 10.

30

209

38 days

9

83

1 day

6

415

56

209

....do

13

98

4 hours. ..

9

621

11

209

do

2

132

30 minutes -

3

616

Aug. 18

3

209

40 days . ...

0

141

2 days

1

423

685

The chief value of this table is in the estimates given in the last column. The
estimate of the number of young salmon in the pool ranges from 415 to 1,595. Half
of the estimates come within 107 of the average, which is 685. This average is prob-
ably not far from the actual number.

Pool C is quite similar to Pool II, and was seined in much the same way. When
two worked it, one had to hang the net on the rocks on one side, then swim across
to the other side, when both pulled the seine off the rocks and hauled it inshore at
the lower end of the pool.

The following table gives a record for Pool C similar to that given for Pool B.
In the only haul made in July, two of the four fishes taken had been marked.
They had evidently come down from the pool above. Likewise in August four
August-marked fishes were taken in the first haul, though none had yet been marked
in this pool. In making the estimates of the number in the pool these four are
considered, being added to the “number previously marked in August.”

Table of seine hauls in Pool C.

Date.

Number
of young
salmon
taken.

Number
previ-
ously
marked
in July.

Time

since

July

marking.

N umber
of July-
marked
fishes
caught.

N umber
previ-
ously
marked in
August.

Time since
last mark-
ing.

Number of
August-
marked
fishes
caught.

Estimated
number in
pool in
August.

4

0

2

92

;>

3

0

4

26

2

do'. . .

2

83

2 hours

2

1,131

39

2

....do...

2

105

31 ) minutes .

2

1,908

2_

2

do . . -

0

140

....do

1

288

August 16

32

2

37 days.

0

141

1 day

5

928

III

2

do . . .

0

168

30 minutes .

5

344

August 18

32

2

39 days.

1

173

2 days

10

566

861

The data obtained from Pool C gives a larger estimate for the total number
in the pool than that for Pool B (C, 861 ; B, 685). In the seven hauls of the seine in
this pool in August there were 8 July-marked fishes secured. In the ten hauls in
Pool B 49 were secured. From these two statements we determine that 18 per
cent of the July-marked fishes of the two pools were in the lower pool in August.
(8=7x10=11, 49+11 = 60, 1 l-r-60=. 18.) Only 1 per cent was released there, leaving-
17 per cent to migrate. Some of these may have drifted over while faint from being
confined in the net, but we think not many. We never saw any do so, though we
often watched them for that purpose. It is safe to say that most of them went

104

BULLETIN OF THE UNITED STATES FISH COMMISSION.

over voluntarily. We would expect this, as the connection between the two pools
is quite deep, though swift. It is even remarkable that no more than 17 per cent
passed to the lower pool. It is conceded that the estimates in the tables above are
liable to considerable errors. There are always some unknown quantities in the
equations, yet the results appear trustworthy.

Pool I) is below the lower railroad bridge, and was quite unimportant. It was
seined but once, a large rock at the lower end of the pool making seining impracti-
cable. Six young salmon were taken. It was one of these six that was taken in
the first haul in Pool E. The two pools are continuous.

Pool E is a portion of the river about 75 yards long, immediately below and not
separated from D, ending above a long riffle. It was the only pool that could be
entirely covered by the seine. The bottom is mostly covered with cobblestones, and
there are large rocks along one shore. These afforded hiding-places for the young
salmon while the seine was being drawn. It was seined many times in August.

The following table gives the record of the catch at each haul of the seine made
in the pool, giving (1) the date, (2) the number caught, (3) the number previously
marked and released in the pool, (4) (lie number of marked fishes taken, (5) the
estimate of (lie number of young salmon in the pool, and (6) the variation of this
estimate from the average estimate. On the 16th and 17th of August (lie seining
was carried on continuously, the time required for making a haul and counting and
marking the parrs being from 20 to 30 minutes. The marked specimen taken in the
first haul was one of ( lie six from Pool D.

Record of seine hauls in Pool E.

Date.

N umber
of young
salmon
caught.

N umber
pre-
viously
marked.

Number

marked

fish

caught.

Estimated
number
of young
salmon
in pool.

Variation
of esti-
mate from
average
estimate.

66

1

1

14S

66

13

751

-271

.1411

197

30

992

- 30

August 1 7, a. m

82

312

38

682

-340

47

357

19

883

—139

1!)

385

4

1 , 804

+782

64

4011

25

1,024

+ 2

1411

439

62

.1,055

+ 33

71

525

41)

932

- 90

35

556

15

1,297

+275

August 1.7, p. m

111

576

9

1,216

+194

3

586

3

586

-436

Total for 24 hours

850

586

259

■ ■ 1.022

August 18

0

(1

September 18

178

97

« 1,075

October 18

29

"Average.

The purpose of the work in this pool was to determine the number of young
salmon that might be found in a pool. The estimates are made in the same way as
in the case of Pools B and C. The third haul, as noted in the table, may be taken
as an example: 30, the number of marked fishes taken, is to 197, the number marked
previous to this haul, as 146, the total number taken, is to 992 , the total number 1 in
the pool. The estimates vary from 586 to 1,829, but several of them are not far from
a thousand. The average of the estimates is 1,022, which is probably near the truth.

September 18 we seined the pool again, catching 178 young salmon, 97 of which,
a little over half, had been marked. By looking at (lie table above it can be seen
that a little over half of the estimated number in the pool were marked in August —
586 out of 1,022. If an estimate is made of the number in the pool in September, by

NATURAL HISTORY OF THE QUINNAT SALMON.

105

assuming that all the marked fishes remained in the pool, the result will be the fol-
lowing proportion: 97, the number of marked fishes taken in September, is to 586,
the number marked during August, as 178, the total catch in September, is to 1,075,
the total number in the pool, which is remarkably close to 1,022.

If there had been much of a migration between August 18 and September 18 we
would have taken a larger proportion of unmarked fishes. If there were no migration
we would expect to get marked fishes in proportion to the total catch as 586 to 1,022,
which was the proportion of marked fishes in the pool in August. The record for
September, 97 marked out of a total of 178, is very nearly that ratio, indicating
that there was little or no migration between August 18 and September 18.

August 17 we fished two pools, F and G, about a quarter and a half mile,
respectively, below where any fish had been marked, catching 13 and 115 salmon.
We also caught 57 in Pool G in September. Neither at this nor at any other 'time
have we found marked salmon below the pool in which they were released, except
in the case of Pool C, above referred to. The one marked salmon taken in the first
haul in Pool E was released in the upper end of that pool.

It is especially worthy of note that none of the 591 fishes taken in Pool E had
been marked in the pools above the previous month. If any had left the upper
pools, they had not stopped in the vicinity.

Summary on number and movements. — The estimate of the number of young
salmon in August in Pool B is 685, in C 861, and in E 1,022. Pool E is much longer
than the others, and might very well have more fishes than either. From these
'‘estimates it is probable that there were about 10,000 young salmon to the mile in
the Upper Sacramento during the summer of 1898, or between a half and three-
quarters of a million in all the headwaters of that stream.

There is little migration of the young salmon between May and December.
Where pools are separated by shallow riffles, no evidences of migration could be
found. If connected by deep water, it was found that about 17 per cent of those in
the upper passed to the lower. In Hazel Creek there was an upward migration of
4 per cent during July, and a larger — about 12 per cent — during September. There
was no diminution in number either in Hazel Creek or the river up to November —
even an apparent increase. There was a slight migration, however, during the
whole period, indicated by the disappearance of the larger marked specimens.

This residence in the headwaters during the summer is probably due to low
water. It has been noticed many times, both in the streams and in the hatcheries,
that- young salmon dread going over a fall. When the river is very low, as it is
every summer, the rapids become almost like waterfalls, thus preventing down-
stream migration. A slight rise obliterates the fall and at the same time makes it
difficult to find food; hence the decided migration in December. The abundance of
food appears to be of some importance when we notice that there was a scarcity
of food in September and also a slight increase in migration that month, although
the water was the lowest of the season.

Young salmon were reported abundant in the pools near Sims the 1st of May,
1899, and they doubtless remained during the summer, as was found during 1898.

Growth in fresh water. — It has been shown above (“Migration, general investi-
gations of 1898”) that the size of young salmon found in May was the same for all
parts of the river. This was true also in July for the portions of the river in which
they were found.

106

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The following shows the measurements of young salmon taken in the Sacramento
in J uly, 1898. Bold-faced type indicates where the average sizes fall. For Dunsmuir,
Redding, and Battle Creek the numbers given indicate the total number of fish taken.
None were taken below the latter point. All were taken between July 9 and 13.

Table showing size of young salmon taken during July, 1S9S.

Number of specimens.

N umber of specimens.

Size.

Duus-

rnuir.

Sims.

Hazel

Creek.

Red-

ding.

Battle

Creek.

Size.

Duns-

muir.

Sims.

Hazel

Creek

Red-

ding.

Battle

Creek.

1

3 inches

1

1

1

3.J inches

1

O

1

1

3.2 inches

1

1

3.3 inches . . .

1

1

3.5 inches _ __

1

1

2

1

3

1

1

1

1

3.7 inches . . .

i

1

1

i

o

3.9 inches _

1

i

1

0

4.3 inches .

1

Note that the average sizes in the above table are from 2.8 inches in Hazel Creek
to 3.1 inches at Redding. In comparing this table with that for the month of May
(see above), it will be seen that the average size increased from 2.4 inches in May to
3 inches in July, an increase of 0.6 inch in two months. This is also the amount
of increase if only the smallest specimens in Hazel Creek are considered. They
increased from 1.5 inches in May to 2.1 inches in July. This is an increase of 0.3
inch per month for fishes averaging under 3 inches in length. Each table shows
that there was a greater variation at the upper stations.

The growth of 0.3 inch per month is also shown bjT the following table of meas-
urements of specimens taken at Sisson in May and August. The average sizes are
indicated by heavy-faced type. In computing the average for May, the four largest
fish are not counted, as they evidently belonged to the summer run of adults instead
of the fall. Measurements of specimens taken in the river at the mouth of the creek
in August ai'e given for comparison with those from the creek. The average size
in May was 2.2 inches; in August, a little over three months later, it was 3.3 inches,
the increase in size of those remaining in the creek being one-third inch per month.
The growth was probably a little greater than that, the larger specimens migrating.

Table showing increa.se in size of young salmon at Sisson.

Number of specimens.

Number of specimens.

Size.

Sulla way Creek.

River,
Aug. 19.

Size.

Sulla way Creek.

River,
Aug. 19.

May 15.

Aug. 19
and 25.

May 15.

Aug. 19
and 25

1.6 inches

2

2.9 inches „ _

i

2

1

1.7 inches

2

3 inches

1.8 inches . _

5

2

3.1 inches

2

4

3.2 inches

1

4

2

3 3 inches . _ .

1

5

2.1 inches

1

3.4 inches

2

2

2.2 inches

o

3 5 inches

1

1

2.3 inches

4

3.6 inches

2

3

2 4 inches

2

I

2

2.5 inches

3.8 inches

3

1

2 6 inches

3

1

2

1

A

1

3

3

1

The above shows the amount of variation in the young salmon of approximately
the same age. All were released from the Sisson hatchery. The oldest were hatched

NATURAL HISTORY OF THE QUINNAT SALMON.

107

December 23, 1897, and the youngest January 23, 1898. The largest were 2.9 inches
long, the smallest 1.6, a difference of 1. 3 inches, which can not be accounted for by the
one month difference in age. There was still a difference of 1 inch in August, when
they were 3.8 and 2.8 inches long.

The following is a table of measurements of specimens taken in Hazel Creek
during various months. They are thought to be representative, though the specimens
were selected. We picked out extremes and what we thought to be average sizes.

Table of measurements of specimens from Hazel Creek.

Size.

Number measured.

Size.

Number measured.

May.

July.

Sept.

Dec.

May.

July.

Sept.

Dec.

2

3.0 inches

1

1

3.1 inches

1

3

3.3 inches

1

1

3.7 inches

1

2

1

3.8 inches

i

1

3.9 inches

1

1

4.2 inches

1

1

4

1

i 4.3inclies

2

2.8 inches

2

1

The above measurements indicate two ages in May, but the youngest were doubt-
less from a few, probably a single pair of fishes, that spawned much later than usual.
The oldest were from the regular fall run of adults. The difference between these
two sizes in May was 0.9 inch. The two sizes are not discernible after May, those
shown in the table being due to the selections of specimens, which is not the case
for May, however. The growth is indicated by the increase in size of the smaller
specimens. The smallest specimens were: In May 1.5 inches, July 2.1 inches, Sep-
tember 2.8 inches, December 3.1 inches, the intervening period in each case being
2, 2, and 3 months, and the increase being 0.6, 0.7, and 0.3 inch, respectively, or
0.30, 0.35, and 0.10 inch per month. The total growth in 7 months, as shown by the
smallest specimens, was only 1.6 inches, and for the last 3 months 0.1 per month.

Pool A is at the mouth of Hazel Creek It is a semicircular pool of quiet water
at one side of, but not at all separated from, the main channel. It is over 6 feet
deep, and the seine had to be hauled by means of ropes. As the seine Avas stretched
across the mouth of the pool and hauled in at the upper end, with the ends close to
the banks, there was but little chance for the fish to escape. The pool was fished
monthly, beginning with August.

The following table gives the number and size of the young salmon taken in-
Pool A during the season, and also indicates how many were marked fishes and had
therefore remained since August. As the measurements were made on live fishes
they could not be made accurately enough to be given in tenths. There was one
2-inch fish taken in October, but it was not counted, as it evidently belonged to a
different run. We marked and returned to the pool all of the fishes taken in
August. None were marked in any other month. At another haul in September,
not recorded in the table, 82 specimens were taken, 32 being marked, which was a
larger proportion by 7 per cent than in the haul recorded in the table for that
month. No marked fishes were taken in December. The record of one haul in pool
E for September is given for comparison with one made in pool A on the same date.
It sIioavs that specimens from pool E Avere smaller than those from pool A, Avhich
was the deeper pool.

108

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Table giving size of young salmon taken in Pool A, Sacramento River near Sims,
with a comparative record from Pool E.

The main value of this table is in showing the size of the young salmon in
different months. The average size in August ivas 3.91 inches; that of the August-
marked specimens taken in September, 3.86 inches, a very slight decrease. It will
be noted later that there was a scarcity of food in September, which would account for
a slight migration. In October the average size of the marked specimens increased
to 4.20 inches, but remained the same in November. In the one month, September 18
to October 18, there was an increase of 0.34 inch, but 1 lie total increase for the three
months, August 18 to November IS, was but 0.29 inch, or 0.10 inch per month. This
small increase for the total period indicates either that the growth was very slow or
that the migration during that time almost compensated for the growth. During
September, probably owing to a scarcity of food, the migration was a little greater,
and as a result the fishes were smaller September 18 than they were a month before.
There was a. decrease of 0.6 inch in December, when nearly all had left the pool.

From measurements made on specimens taken at Olema, we have the following-
table for age and size of young salmon remaining in fresh water:

Age.

Size, inches.

Increase
per month,
average.

Smallest.

Largest.

Average.

Three months .

1.7

2.6

2 2

0.4

Five months

2.8

3.3

:i. n

A

. 36

Seventeen months.

7.0

.35

The growth in fresh water is, therefore, very slow, and in artificial propagation
every effort should be made to prevent their remaining in the river over summer.
The growth in salt water is much more rapid. The salmon should reach the ocean

NATURAL HISTORY OF THE QUINNAT SALMON.

109

when about 3 inches long, and grow to be 36 inches in twenty-four months, which
would be an increase of about 1.4 inches per month.

Gastric parasites. — Of 209 fresh- water specimens examined in the investigation
of food of young salmon, 31 had parasites in the stomach. The parasites were of two
or three kinds, one elongated, the others short and grain-like. They have not been
studied, except to note the date and size of the fish. It is evident that residence in
fresh water is conducive to the growth of parasites in the stomachs of young salmon.

Month.

Number

examined.

Number

with

parasites.

Percent-
age with
parasites.

Size.

N umber
examined

N umber
with

parasites.

Percent-
age with
parasites.

January

9

i

11

1.1 to 2 inches _ - .

01

3

5

February

ki

0

0

2.1 to 3 inches . . .

57

3

5

March

10

o

0

3.1 to 4 inches . . .

53

10

19

April

15

0

0

4. 1 to 5 inches _ . .

30

12

40

May

50

1

8

5.1 to 0.3 inches..

8

3

38

August

20

3

15

Total

209

31

15

September

18

3

17

< Ictober

80

8

23

November

15

3

20

December

11

8

73

Total

209

31

15

Diseased parrs. — Only two diseased young salmon from the streams have been
met with. One was found dead, covered with fungus, near Sims in 1898; the other,
5 inches long, was taken in the trap at Battle Creek, September 28, 1900. The upper
lobe of the caudal fin was wanting, and the remainder, with the caudal peduncle, was
covered with fungus.

Mature male parrs. — In October, 1897, several mature males, between 4 and 5
inches long, were taken at Sisson. In January, 1898, two males, 5.5 inches long and
known to be only a year old, were taken above the Bear Valley dam near Olema; one
was mature. In August, 1898, a 4-inch mature male was taken at Sisson. Four of
the 6 young salmon taken at Fall River Mills in August, 1898, were males, all with
the genital organs mature. Mature male parrs were frequently taken at Battle Creek
fishery in October and November, 1898. The sex of a number of parrs, 4 to 6 inches
long, from the general collection was determined; 15 were mature males, 2 immature
males, and 12 were females. These mature male parrs can usually be distinguished by
their more dusky color and by the slightly distended abdomen. Examined under the
microscope, the milt is apparently the same as that from adults. A few eggs from a
female of ordinary size were fertilized by milt from a 4.7-inch male. The fertilization
was complete, all of the eggs hatched, and the alevins were of normal appearance.

No explanation of this early maturing of males can be made, and nothing is
known of their future history. They feed the same as other young salmon and appar-
ently are not attracted by mature females as the adult males are. It may be that
they return from the ocean as the stunted form known as grilse. It is probable
that several months’ residence in fresh water causes the generative organs to mature
both in the adults and in the young males.

Temperature notes. — The following table shows the number of young salmon
taken at one haul of the seine in water of various temperatures. It indicates but
little, except that young salmon may reside during the summer in water having a
temperature of 64 degrees. The 25 taken in Thomas Creek, with a temperature of
68 degrees, were landlocked in a shallow pool.

no

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Table of seine-haul and water-temperature records.

Place.

Date.

53°

54°

55°

56°

57°

58°

59°

60°

61°

62°

63°

64°

65°

66°

67°

68°

July 8

May 17

July 0

May 17

July 9

Aus. IS

6

14

Do

55

River at Sims

50

Do

177

Do

148

Do

Sept. 18

Oct. IS

178

Do

81

May 4

May 18

July 11

Aug. 13

Apr. 30

July 13

Apr. 28

July 14 to 31

May 22

May 22

May 23

May 27

May 11

May 13

June 17

50

Do.l

66

Do._.

8

Do

0

60

60

Do...

2

Redblulf

150

Below Redblulf

0

0

0

Mouth of Thomas

4

25

8

■ 28

6

Rio Vista _

2

2

1

Conclusion drawn from study of summer residents.. — It seems evident from
these observations that the later fry that hatch in the headwaters, or are planted
there after the spring freshets have passed, are liable to remain till the rains of the
following winter. This means a slow growth for at least 6 months, or about a fifth
of their growing period. It means the precocious maturing of the males, which may
be responsible for the great number of dwarfs known as grilse; and it means that
15 per cent will become infested with gastric parasites. For these reasons it is
imperative that the fry from our hatcheries should not be released above Redding-
after the spring freshets, though they may be released in the headwaters earlier
without any detriment, and they certainly should not be held after this time merely
for the purpose of feeding. Superintendent Shebley of Sisson Hatchery states, as
this paper is going to press, that there are not nearly so many young salmon
remaining in the Sacramento River near Sisson during the summer since he has
quit holding the fry in the hatchery during the spring for feeding as there were
when he did so hold them. There is no advantage in holding fry in the hatcheries
for feeding.

FOOD OF YOUNG SALMON.

General study of food in fresh water.- — The young salmon feed principally upon
floating or drifting insects, either immature or adults. When feeding they often
take a station below a stick or rock and catch their food as it floats down on either
side. They eagerly catch small insects and larvae if thrown into the water. Fry 1.5
inches long have been observed to rise to a small fly that alighted on the water. They
hardly ever eat encased caddis larvae, although that is the main food of the trout.

The following is a tabular statement of the stomach contents of 225 young
salmon, being based on an examination of about five specimens from each locality
each month in the year in which any were taken. The record for each fish examined
consists of the station, date, size of fish, and number of specimens of each kind of
food or other material found in the stomach. Four forms of insects are recognized
in the table, viz, larval; pupae, including nymphs; flying insects; and “terrestrial
insects,” including adult wingless insects and spiders.

NATURAL HISTORY OF THE QUINNAT SALMON

111

Table of stomach contents of young salmon, Sacramento Basin.

+ indicates presence of certain objects, the number of which was not determined. The totals in last two
columns indicate number of fishes in whose stomachs parasites or indigestible material was found.

112

BULLETIN OF THE UNITED STATES FISH COMMISSION,

Table of stomach contents of young salmon , Sacramento Basin — Continued.

Station.

Specimen number.

Size of salmon,
inches.

8

a

5

fu

j Plying insects.

Terrestrial insects.

Crustacea.

Gastropods.

a

o

£

| Unidentifiable.

No. of speci-
mens with —

o

o

o

Parasites.

Wood, seeds,
etc.

Sims, Nov. IS, 1898. _

42

4.3

6

19

o

2

Sims, Dec. 18, 1898 _

43

44

45

46

47

48

49

1.5

1.6
3.5
3.5
3.9
4.3
4.7

7

2

4

37
23
IS

38
26

+

21

16

16

13

20

i

+

+

+

+

3

23.0

3.3

93

13

148

21

4

1

5

Average . ...

Hazel Creek, July 9, 1898

Total .

1 60
1 61
1

1 63

2.0

2.6

3.7

4.3

2

3

3~

8

3

3

—

—

—

—

"

—

—

=

—

1
i.

2

12.6

3.2

5

1

17

4

1 64
65
06
67

1 68

2.8

3.0

4.1
4.4
4.9

6

4

5

2

7

3

2

6

1

Total

1

2

4

2

1

1

+

+

2

19.2

3.8

13

3

13

3

20

4

3

Average

Hazel Creek, Oct. JK, 1898

1 69

70

71

72
1 73

2.6
2.8
3. 6
3.9
4.3

6

5

9

1

4

1

17

1

2

~2

2

1

2

T

3

—

Total

17.2

3.4

21

4

25

5

7

1

4

1

Average ..

1 74

2.9

3.9
4.1
4.7

3

7

iT

9

3

3

2

1

—

—

—

—

75

nazei i^ieeiv, jnov. id, l&yo

Total

1 ??

15. 6
3.9

10

3

18

4

~ r

Average

\

HT
1 80
1 81

3.1

3.2
3.9

4.2

1

.....

7~

1

1

—

—

+

+

+

+

Total

i

+

14.4

3.6

2

1

9

2

2

1

1

1

3

1

Pall River at Dana, Aug. 28, 1898 . . .
Pall River at mouth, Aug. 29, 1898..
Total.. . .

=

—

50

3.8

45

5

+

1

52
] 53
1 54

5.2

5.2

5.3
5.6

48

32

10

18

18

26

13

+

+

1

i

+

—

21.3

5.3

1

108

27

62

16

1

2

Average ..

McCloud River at Baird, Sept. 16,

1 55

56

57

58
1 59

3.3

3.7

3.8

4.4

4.8

—

—

—

—

—

3

1

+

+

+

4

Total

20. 0
4.0

lT
1.8
2.2
2 5

3

1

1

|

Average

Redding, May 4, 1898 .

Total

1 ~82~

83

84
1 85

86
[ 87

=

=

4

1

5

9

12

1

2

+

3.2

3.7

4

15

i

....

15.1

2.6

5

1

45

8

4

1

1

NATURAL HISTORY OF THE QUINNAT SALMON.

113

Table of stomach contents of young salmon, Sacramento Basin — Continued.

0

O

n

O

0

No. of speci-
mens with —

Station.

Specimen nun

Size of sain
inches.

SB

>

3

p.

ffi

P.

£

Flying insects

.9

3

tn

0

0

Eh

| Crustacea.

[ Gastropods.

tn

a

3

£

"ci

£

0

’3

£

j No food.

j Parasites.

Wood, seeds,
etc.

1 88

2.6

1

6

3

1

Redding, July 11, 1898

1 89

3. 0
3 3

2

+

s

4

\

3.5

i

12.4

8

6

4

1

1

Average

3.1

2

2

i

( 92

1.4

2

2

6

4

—

—

93

1.5

2

■Rsi/IIq Tforvy .Tan 21 189ft

! 94

1.5

2

1 95

1.6

i

1 96

1.6

2

1

i

7.6

14

1

2.5

3

3

—

( 97

1.5

T

1

1

1 98

1.5

| 99

1.5

1 100

1.7

2

[ 101

1.7

4-

Total

7.9

2

3

i

1.6

1

1 102

1.5

i

4

103

1.5

4-

Ralls Forry Mar 15 189ft

■ 104

1.5

2

1 105

1.6

1

2

{ 106

1.6

i

Total

2

ft

i

1.5

2

[ 147

2.9

1

if

12

24

9

1

1 148

3.1

10

1

1

1

Battle Creek at Longs, Sept. 14, 1898.

149

3.5

10

1

150

4.0

15

1

4-

l 151

4.0

ii

12

1

17.5

47

63

6

1

i

1

1

i

3.5

9

13

i 107

2.3

4

i

1 108
1 109

2.4

2

3

{ 110

3.5

i

48

1

11.7

3

56

3

2.9

1

14

1

1 HI

1.5

5~

6

112

2.1

1

13

Rat, tip, flrfip.lr Apr 80 1898

113

2.5

2

8

1

3

1

114

3.0

12

12

1 115

3.5

+

12.6

15

3

31

17

1

i

2.5

6

3

130

1.5

1

i

4-

138

1.6

1

3

139

1.7

2

2

140

4.2

2

3

+

Battle Creek, Nov. 4 to 13, 1898

141

4.4

4-

142

4.6

2

4-

143

4.8

+

144

5.4

2

3

i

145

6. 2

2

4-

146

6. 3

Total

40.7

5

13

6

2

3

3

4.1

J.

1

1

F. C. B. 1902—8

114

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Table of stomach contents of young salmon, Sacramento Basin — Continued.

Specimen number.

O

1 Flying insects.

4^

o

03

03

CD

"ci

eg

<D

[d

No. of speci-
mens with —

Station.

r“| 03

03

■H§

CD

SI

GQ

SB

>

i-l

s

P4

5

Ph

•2

03

CD

s

EH

o2
a >
o
c3
0)
p

5

03

rd

o

P4

o

03

cS

03

a

8

£

»d

o

o

4-4

o

£

03

03

*03

c3

P4

c3

PH

03

■d

a)

03 .
03 0

d Id

o

o

£

116

1.4

1

+

117

1.8

+

(a)

118

1.4

(«)

119

1.5

1

120

1.5

-1-

121

1.5

1

122

4

123

1.5

1

( b )

124

1.6

i

1

Battle Creek, Oct. 9 to 25, 1898

125

4.1

i

1

i

-f

126

4.1

2

22

2

+

-f

127

4. 3

.10

128

4.3

8

2

-1-

129

4.3

3

4-

130

4.6

3

1

131

4.6

3

4

132

4.7

5

8

15

13

133

4.7

2

+

134

4.8

1

4-

+

135

5.2

+

53. 0

9

61

34

2

13

5

1

6

1

3.1

j

'

3

1

I 152

1.7

153

1.8

26

{ 154

2. 1

i

8

1

+

155

2.5

4

4

+

{ 156

2.7

29

1

Total

10.8

67

2

1

i

Average ..

2.2

1

13

f 157

1.5

18

2

158

2.2

1

+

Tehama, May 5,1898

159

2.4

9

3

160

3.1

4

1

l 161

3.5

1

3

i

Total

12.7

28

12

i

2

1

Average

2.5

6

2

( 162

1.8

+

+

163

1.9

Chico, May 23, 1898

{ 164

2.6

3

4

3. 1

7

g

l 166

3.6

9

1

+

Total

13. 0

19

11

1

1

2

Average

2.6

4

2

f 167

2.1

3

2

\ 168

2.3

3

Total...

4.4

3

5

Average

2.2

2

2

I 169

1.8

16

3

1

Butte City, May 25, 1898

i 170

2.1

4

20

[ 171

2.5

8

2

Total

6.4

28

25

1

Average _

2.1

9

8

7 172~

2.1

30

i

Colusa, May 27, 1898

| 173

2.4

-f

1 174

2 7

8

+

Total

7.0

38

i

2

Average

2.4

13

Yolk not yet all absorbed.

b Salmon egg, waste from spawning platform.

NATURAL HISTORY OF THE QUINNAT SALMON

115

Table of stomach contents of young salmon , Sacramento Basin — Continued.

116

BULLETIN OF THE UNITED STATES FISH COMMISSION.

This study shows that young salmon in fresh water feed exclusively on insects,
and that immature aquatic insects form by far the larger portion of their food. The
general summary of the table shows that approximately half of the food of the speci-
mens studied consisted of pupae (or nymphs, which were not distinguished from
pupae), one-third of larvm, and one-sixth of adult winged insects.

There was an increase of flying insects in the food of specimens taken in Sulla-
way Creek in August, and an increase in amount of food in specimens from Sims during
July and August. It was during September, when apparently there was a scarcity
of food, that the larger young salmon disappeared from Sims. There was an increase
in flying insects in food of specimens from Hazel Creek in September; a scarcity of
food and a noticeable lack of larvae in specimens from Battle Creek in October and
November, and a smaller amount of food in specimens from the lower stations. (See
summary for May, page 117.)

Two specimens from above Bear Valley Dam, near Olema, taken January 18,
1898, had stomachs gorged with larvae and pupae, one having about 50 of the former
and 25 of the latter, but no indications of adult insects.

Three specimens, Nos. 117, 118, and 137, of food table, were taken in October and
November before the yolk was yet absorbed. One had nothing in its stomach;
another had some food, but it was unrecognizable; the third had eaten one larva and
two adult insects, besides some other food that was unrecognizable. This indicates
that they begin feeding even before the yolk is all absorbed.

The food data, if arranged according to size of fish, would give the following-
average amounts per fish. This table shows that pupae and nymphs are the favorite
food for all sizes. Those from 1.4 to 2 inches in length feed very little upon adult
insects; the largest size feed very little upon larvae.

Size.

N umbei’
examined.

Average number in

sti imach.

Larvae.

Pup*.

Adult

insects.

1.4 to 2 inches

59

4

3

(“)

3.1 3 inches

53

4

ti

3

3.1 4 inches _

48

r>

8

2

4.1 5 inches

32

3

4

1

5.1 ti.3 inches .......

8

1

14

8

<’ Indicates an average of less than one.

The following table brings together a statement, of the average amount of food
found in the stomachs of the young salmon from various stations for the month of
May, the only month in which we secured young salmon from many of the lower
stations. The table indicates that the important food of the young salmon through-
out, the basin in May was larva; and pupa;, of which there was an average of 4.4 of the
former and 6.5 ot the latter per fish. The fish examined averaged 2.5 inches. It
also shows there was a slightly smaller amount of food in specimens from the lower
portion of ihe river. They were not starving, however, and there is no evidence
that the passage down the river is detrimental on account of the lack of food.
a indicates an average of less than one. The numbers in the columns headed
“No food'' and “ Paiasites’’ indicate the number of fishes examined that had
empty stomachs or parasites, as the case may be. The totals are taken from the
complete table ot food, but only for the month of May, and are not the sums of the
averages given in this table

NATURAL HISTORY OF THE QUINNAT SALMON.

117

Table showing the average amount of food in stomachs of young salmon from the

various stations in May.

Station.

No. of
speci-
mens.

Size.

Larva?.

Pupae .

Winged

insects.

Terres-

trial

insects.

Crusta-

cea.

Clastro-

pods.

Uniden-

tifiable.

No

food.

Para-

sites.

3.3

9

(a)

(«)

(«)

3

4

2.9

5

10

1

(a)

6

2.0

1

3

1

(«)

2.5

3

0

3

(a)

(a)

Red Blu If

2.2

i

13

(a)

(a)

(«)

1

1

2.5

6

2

(a)

(a)

2. 0

4

2

(«)

(<>)

2

2.2

2

2

3

2.1

9

8

(«)

3

2.4

9

(a)

(a)

2

2.3

1

4

1

2.4

2

(a)

2

1.9

1

2

(a)

(a)

Rio Vista

2.3

i

4

(«)

i

Total

60

262

381)

60

6

2

1

0

5

Average

2.5

4.4

6.5

.1

(«)

(«)

.1

,i

. i

a Indicates an average of less. than one.

Food in brackish water. — Relatively few specimens of young salmon have been
obtained from brackish water, and the following table gives a list of the food found
in nearly all that were caught:

Station.

No.

Size.

Amphi-

pods.

Cope-

pods.

Fish.

Adult

insects.

Seeds.

Para-

sites.

206

1.8

25

207

1.9

+

208

1.9

1

30

209

1.9

100

10

210

1.9

1

211

2.2

150

15

212

2.7

2

100

Total

14.3

3

350

81

1

Average -

2.0

+

50

12

1 213

~ 2A

2~

NT

214

2.3

1

13

Benicia, May 13-20, 1898

215

2.5

31

216

2.7

1

200

1

1

| 217

3.3

1

Total

11.9

200

50

1

2.4

1

40

10

-f

218

1.6

-f

219

2. 1

2

220

2.4

221

2.4

c

222

2.5

8

+

223

2.8

3

+

224

3.0

2

8

225

3.4

1

5

Total

20.2

7

8

1

53

2

Average

2.5

1

1

Summary for 20 brackisli-water specimens.

Total

15

558

1

104

3

Average _

2.3

1

29

9.2

The chief food of the few brackish and salt water specimens studied were adult
insects. Only 5 of the 20 specimens had fed to much extent on copepods, and only
1 had eaten a fish; the species of the fish could not be determined, though it was
evidently a smelt ( Osmerus ). No aquatic insects were found, such as were found
in specimens from the lower river, which indicates that the fish had been in brackish
water at least long enough to digest all fresh-water food.

118

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Classification of the insect food of young salmon. — Tlie following is a classified
list of the insects found in the stomachs of young salmon from the Sacramento
River, collected during this investigation, as reported by Miss Bertha Chapman.
It is evident that in many cases the fish confined themselves largely to food collected
at the surface of the stream, as is the case with fish taken at Rio Yista in May, or
those taken at Fall River Mills in August; others sought the immature forms living
under water, as can be seen from the majority of cases in the list. But in no case
can this distinction in feeding habits be definitely made. They seem to have fed
indifferently on water and surface forms. These surface forms are almost invari-
ably insects living about streams, and which might therefore easily have fallen into
the water from overhanging plants. Much of the stomach contents had been so far
digested that it was not possible to identify t lie insects. Other insects have been
partially determined by single wings or particles of the body; but it seemed not so
important to carry the classification to species as to determine the types of insects
forming the food of young fish. The results of the study are given in the following
table :

Classification of insect food for young salmon.

Place.

Date.

ft
c n
ft

*o

5

ft

a

j Libellula.

Ephemerida.

Plecoptera.

oi

5

-*->

o

ft

5

$

d

o

fi

d

;s

o

O

9*

d

ft

d

£

| Notonectidse.

82

d

’55

ft

d

O

82

m

d

d

S

d

'v

d

'd

<D

ffi

CD

cu

82

d

2

ft

d

S

-e>

ft

O

ft

o

0>

u

H

Lepidoptera.

oj

s

-e»

5

8

(d

S

82

[d

<D

o

o

d

ft

ft

<D

s

82

d

"o

5

Chironomidte.

Ceratopogon.

82

d

d

9

/ 5

12

do

39

X

1

t

1

XI

February

do....

+ 2

3

+

+29

t

2

t

1

2

+

March .

1

+23

t

1

/I

t

t

2

do....

+30

+

*

Benicia (brackish )

do _ ..

4

12

::::

+

April

‘ do

2

i

2

p3

+1S

t

- 2

±

.. .do

+28 L-

*

/ 5

p2

s

1

1

pi

i

XI

13

do....

4

(1

xi

XI

1

3

t

2

14

do

(1

Xp

+

tp

1

Red Bluff

... do

5

p2

/ 2

do .

2

o

2

13

12

do

L '

tp

tl

1

... .do

2

1

Xp

do

3

+1

... do . . .

3

1

XI

..do _

2

pi

Knights

...do....

1

do...

5

x

2

+

Benicia (brackish)
Pinole (brackish ) .
Point Richmond

""

3

3

t

2

t

+

...do ...

2

t

9

1

2

+ 6

XI

+

+

...

1

X

X

3

5

...

+

+7

Hazel Creek

do

+ 6

t

XI

X

2

4

1

1

1

p2

+z

.. do. .

1

+

...

3

1

5

ti

1 1

14

.. .do ...

1

1

+

pi

+

4

1

i

1

+

1

X

12

5

+

X

xi

x

XI

Sul la way Creek . .

September
do ...

+12

2

2

2

+9

2

3

xp

pi

i

3

1

+

X

do ...

+12

n

i

n

2

6

1

«

n

tl

X

x

xl

. .do

5

20

6

XI

Xp

xi

do

+

+

1

2

i

i

t

t

i

1

1

2

p2

n

13

10

2

2

5

r

. . .do

+14

+

t

2

x

XI

NATURAL HISTORY OF THE QUINNAT SALMON.

119

Classification of insect food for young salmon — Continued.

Place.

Date.

Number of fish.

&

[d

+3

P+

<D

t-3

8

'S

&

a

w

8B

3,

O

r3

PH

SB

;d

o

m

d

SB

d

M

A

| Anthomyiin®.

$

_d

’o

w

d

Coleoptera.

SB

[d

d

A

P+

&

55

d

fH

<D

P+

O

d

<D

a

>>

w

SB

^d

"d

o

pj

D

d

rd

o

SB

;d

d

o

o

d

2

SB

[d

*o

g

o

2

d

^d

A

o

d

'<1

2

< h

d

o

‘pi

o

u

m

o

"d

<D

CO

Ph

Crustaceans.

Parasites.

j Woody particles.

j Unclassified.

A

E

CO

b£)

be

W

Walnut Grove

January .
February
.do

39
+ 2
t29

3
+23
t30

4

5

tin

+28

5

4

6

5

l

t

3

1

+

2

+

+

... do

2

+

March

do

...do....

+

Walnut, Grove

Benicia (brackish)

i

2

i

1

+

+

i

+

+

1

1

2

April

+

2

1

./..do

t

4

... do

+

May

t

i

+

12

n

+

..'do ..

1

i

....

3

... do

. . do _ .

2

4

13

do

2

5

5

Y

+ 0
+ 0

4
1

5
1

4

5

+12
+ 9
5

+12

5

5

20

10

i

do

P 2

i

i

t

1

Benicia (brackish )
Pinole (brackish) .

t

2

1

2

i

i

2

+

1

do

...do ...

1

2

....

....

....

j —

"T

3

1

...

. . . do . . .

do

2

1

1

X

t

t

X

August- . .
do

2

1

X

4

Fall River Mills.. -

do....

2

"5

1

"T

5

5

3

1 1

Sullaway Creek . . .
Sims

September
.do

3

1

1

1

2

3

1

t

t

P 3

3

+

1

X

2

3

3

+

October . .
... do

t

Hazel Creek

2

....

2

2

Battle Creek

1

12

1

i

i

Battle Creek . .

N ovember
Decembei

2

Sims

t Means more than 10. x Means present, but number not known. p Means pupa.

$ Means more than 5. I Means larva. a Means adult.

The following list of common names will he of assistance to those who are not
familiar with the scientific terms used in the above table:

Odonata:

Libellulidse Dragon-flies.

Ephemerida . May-flies.

Plecoptera Stone-flies.

Orthoptera:

Acrididse Grasshoppers.

Hemiptera ... . _ . Bugs, lice, aphids, etc.

Corisidse .Water-boatmen.
Notoneetidse . . Back-swimmers.

Capsid* Leaf-bugs.

Jassidae . . Leaf -hoppers.

Membracidae . . Tree-hoppers.

Psyllidae. ... .Jumping plant-lice.

Aphididae Plant-lice or aphids.

Trichoptera Caddis flies or caddice

worms.

Lepidoptera ... . . Moths, skippers, butterflies.

Coleoptera ... Beetles.

Staph ylinidae . . . Rove beetles.

Diptera:

Tipulidae . . Crane-flies.

Blepliaroceridte Net-winged midges.

Culicidae _ _ _ Mosquitoes.

Chironomidae Midges.

Ceratopogon ... Punkies.

Simuliidas . ........ Black flies.

Leptidae . .Snipe flies.

Empididae . . Dance flies.

Phoridae Humpbacked flies.

Muscidae Muscids.

Dexiinas ... Nimble-flies.
Anthomyiinae . Anthomyiids.

Muscinae Common flies.

Hymenoptera Ants, bees, wasps, etc,

Ichneumonidae Ichneumon flies.

Braconidae Braconids.

Formicidae Ants.

The following is a list of insects having the early stages passed in the water,
selected from the set worked over:

Ephemerida.

Plecoptera.

Corisidae (entire life in water).
Notonectidae (entire life in water).

Trichoptera.

Tipulidae.

Blepharoceridae.

Culicidae.

Chironomidae.

Simuliidae.

Leptidae.

Coleopterous larv*.

120

BULLETIN OF THE UNITED STATES FISH COMMISSION.

THE PERIOD OF GROWTH.

Of the ocean life of the salmon very little is known, although it comprises two-
thirds of their existence. They have been taken in Tomales Bay when four months
old, about 20 miles from the mouth of the stream in which they had been planted,
and a few of about the same age have been taken in San Pablo Bay. These were
on their way to the ocean and were already in nearly pure sea water. Specimens 8
to 1 5 inches long are sometimes taken by anglers in San Francisco Bay. In the
estuary and along the beach at Karluk, Kadiak Island, Alaska, the writer has seen
schools of several hundred 8-incli red salmon that had come inshore with the adults.
These were feeding and were not dwarfs on their way to spawning-grounds.

Salmon are sometimes taken in the ocean near San Francisco in paranzella
nets and are also captured in large numbers with the troll in Monterey Bay, where
they appear in February and are found until the middle of August.

Something of the ocean life of the salmon might be learned by making a study
of the food found in the stomachs of individuals taken when they first enter Mon-
terey Bay. At such times a portion of the food taken on their offshore feeding
grounds should yet be in identifiable condition. The presence of deep-water fishes
or crustaceans would indicate a deep-water life for the salmon.

During its life in the sea the salmon is, of course, not entirely free from
enemies, and something might be learned of them by studying the scars left by the
injuries they inflict where they fail to kill. Injuries and deformities received before
entering fresh water were of frequent occurrence among spawning fishes at Battle
Creek in 1897, but no particular attention was given them. The males often had
the snout twisted or split, or even cut off. Very often there were one or more scars
extending obliquely backward and downward on the side above the anal fin. Some-
times two or three were parallel, as if they were scratches made by teeth of some
other fish while the salmon was smaller. These scars were more often present on
the females. The dwarf females were always injured in some way. Very few
injured fishes were observed in 1898. Two had lost the ventral fins, one had lost
the lower two- thirds of the opercle, two had deformed backbones. Only one fish, a
female, had the oblique and parallel scars mentioned; they were slightly curved and
in two series of seven each. This subject is worthy of further consideration.

A very characteristic scar, and one that always attracts attention, is that left by
the sucking disk of the lamprey. The lamprey has no lower jaw, its mouth being
circular and of the size of the end of its head. The gullet ends in the middle of the
mouth and is bounded on the upper and lower sides by hooked teeth. There are
other smaller hooked teeth above and below these on the disk, and on each side of the
disk there are about four slioi't cross rows of teeth, and the whole circumference of
the disk is beset with small teeth. When the lamprey attaches itself to another fish
the outer row of small teeth leaves a scar somewhat resembling the milling on a coin,
which lias led imaginative persons to see in the whole scar a resemblance to a brand
made by a heated coin. The illustration on plate 13 is made from the photograph
of a lamprey scar on the opercle of a blueback salmon found at Karluk, Alaska.

In 1896 5,000 marked fry were released in a tributary of the Columbia River;
about 400 of these were taken in 1898, and a few more each in 1899 and 1900. This
indicates that most salmon remain in the ocean two years, though a few remain three
or four years, as will be seen from the following chapter.

Plate 13.

ADULT AND GRILSE FORMS OF MALE SALMON, WITH GENITAL ORGANS MATURE.

The upper is the adult, the lower the grilse. The measure shown is ‘24 inches long.

DIED FROM GILL PARASITES, LAST OF
SUMMER RUN, SEPTEMBER, 1900.

LAMPREY SCAR ON OPERCLE OF SALMON.

NATURAL HISTORY OF THE QUINNAT SALMON.

121

THE ADULT SALMON.

MIGRATION.

Do salmon return to tlieir native streams? — There is a widespread belief that
when a salmon returns to fresh water to breed it seeks the stream in which it was
hatched, though there is very little evidence that such is true. Various fishermen
claim that they can distinguish the salmon of particular streams by their general
appearance, which is incredible. The employees of the Alaska Packers’ Association
state that the red salmon taken at Uganuk are always smaller than those taken at
Karluk, both places on the north coast of Kadiak Island, Alaska; that 13 of the
former are required to make a case of canned salmon, while only 11 of the latter are
necessary. Tills seems to indicate that the salmon of the two localities are distinct,
but the larger salmon may go to Karluk, not because they have been hatched in
Karluk Lake, but because they are larger.

In 1897 855,000 quinnat salmon fry were released in Paper-mill Greek and its
tributaries draining into Tomales Bay, California, and 2,000,000 alevins were
released in the same streams in 1898. (See “Observations on alevins artificially
reared.”) In 1900 a few salmon were seen in Paper-mill Creek, and in 1901 they were
abundant. In one haul of the seine in the tide-water portion of Paper-mill Creek,
covering a section about 150 feet long, 7 quinnat salmon were taken November 16,
1901. It is well known that quinnat salmon did not breed in Paper-mill Creek or
its tributaries previous to 1897, for which reason these streams were selected for the
experiment. Mr. Thomas Irwin reports that he saw two large salmon in Paper-mill
Creek about 1890, but with these exceptions he never saw any fishes in the stream
that might be taken for quinnats until 1900. lie lives on the banks of the creek
and knows the stream thoroughly. His statement agrees with that of other persons.

Paper-mill Creek is not suitable for quinnat salmon, being entirely too small,
but it is frequented by dog salmon and steelheads.

But there is no conclusive evidence that the fishes which were found in Paper-
mill Creek in 1900 and 1901 were the same individuals released there three or four
years previously. They may have been merely stray fishes, and their being found
there at that time only a coincidence; or their coming into Tomales Bay may have
been caused by there being an extra large number of salmon in the ocean, which
might very well be, owing to the large output of young from the hatcheries; or those
found in Paper-mill Creek in 1900 and 1901 may have been some of those released
there, in which ease it is very probable that they had never reached the ocean at all,
but remained in Tomales Bay. Paper-mill Creek would then be their only stream.

It is incredible that the salmon remember their native stream during their two
or three years of ocean life and that they consciously seek it when they desire to
return to fresh water. Probably most of them do return to the stream from which
they entered the ocean, not because it is their native stream, but because they do not
get far away from its mouth, and when ready to return to fresh water it is the first to
attract them.

The two runs of salmon. — Adult salmon may be found in the Sacramento River
at almost any time of the year. There are, however, two more or less distinct runs,
the first of which passes up the river during April, May, and June, and the latter
during August, September, and October. The former is known as the spring run,
the latter as the fall run.

The salmon of the spring run ascend the river to the headwaters, such as the
Upper Sacramento, McCloud River, and Hat Creek, and some of the earlier ones even

122

BULLETIN OF THE UNITED STATES FISH COMMISSION.

pass Pit River Falls and ascend Fall River to its source. They are not found in Pit
River above the mouth of Fall River. By the time they reach this portion of the
stream, the Upper Pit River is very low and the water impure, and the salmon all
turn into Fall River. The salmon of this, the spring run, spawn mainly in August.

The fall salmon do not ascend the river as far as the spring run, but turn into
the lower tributaries or spawn in the main river. They reach their spawning-
grounds during the latter half of October, November, and the first half of December,
and spawn soon after. The main river is very low at that time of the year, and the
portion between Tehama and Redding is an important spawning-ground. (See chart
of spawning-grounds, plate 17.)

As a matter of fact there is no definite distinction between the spring and fall
runs; that is, there is no time during the summer when there are no salmon running.
First there are a few very early salmon that begin running up the river in February,
and the number increases until May when it decreases till July; then it increases
till the first of September when it again decreases, there being a very few each month
until the next spring run.

The spawning seasons merge in the same way. The earliest salmon go farthest
upstream, and as the season advances they stop at lower points. The localities and
dates of the spawning of the earlier salmon have not been determined, except that
Superintendent Lambson, of Baird, reports having seen a pair of spawning salmon in
the McCloud at the hatchery on the 20th of Api’il, 1902, which is the earliest record
known. By the 1st of October spawning fishes are found as far downstream as
Redding, and as far as Tehama by the first of November.

Details of migration. — When the salmon enter San Francisco Bay they come in
against the ebb tide, stem the current till the tide changes, and then run out against
the flood tide, losing much of the distance gained during the ebb. How it is that
they do not lose altogether as much as they gain will be understood f rom the following
explanation: The tide runs up the bay and river as a broad, low wave, on the upper
side of which is flood tide and on the lower side ebb tide. When the crest of a wave —
that is, slack high water — is at Isleton, the trough, or slack low water, is about at the
Golden Gate. This wave is about three hours reaching Benicia and four in reaching
Collinsville. The farther up the bay and river it reaches the smaller the wave
becomes, the shorter the flood, and (as the flood and ebb combined must equal 12
hours) the longer the ebb.

The following diagram will illustrate the movements of a salmon in passing
through the bays: a, b, and c represent the tide wave at successive points as it
passes up the bay, ■*— indicates ebb tide, and » — > flood tide. Suppose that a

salmon enters the Golden Gate, GG, at the beginning of ebb tide, which would be
the most favorable time. His position on the wave wiil be at .s. If he is able to
travel up the bay as fast as the wave he will keep his position near the crest, that is
at s. But he can hardly do that, especially as the current would be very slight,

NATURAL HISTORY OF THE QUINN AT SALMON.

123

and in the broad bay hardly strong enough for his guidance. Suppose that by the
time he reaches Benicia, B , he has fallen behind the wave until he has the position
at .s'. It is then slack low water, and he can make no headway. Soon the next
wave reaches him and he is in flood tide. He will therefore swim back against the
current. As the wave is going up the bay and he is going down, he soon gets past
the crest and finds himself in the ebb tide at s". He then turns and stems the ebb
tide, and as the wave is going in the same direction he is, he goes much beyond
Benicia, B, before he again falls back to slack low water at s'", and gets into the
flood of the next tide wave.

There is no way of tracing the passage of the salmon through the bays, but from
records made at Vallejo, Benicia, and Collinsville it seems to require about a week
to reach the mouth of the river after they enter the Golden Gate.

Plate 14 indicates the catch of fish at various places from Vallejo to Sacramento
for a certain period, and is intended to show the passage of two schools between
the two places. Each division of the diagram indicates the relative amount of
salmon taken at the ten places during one day, the unit being the average daily
catch at the given place. The vertical spaces indicate tenths of the average daily
catch. By a careful study of the diagram the following points will be noted:

On Monday, April 25, there were few fish taken anywhere, the catch being less
than the average at all points. This is the more marked because the Monday
catch is on an average 25 per cent greater than that of other days, on account of
there being no fishing on Sunday. On Tuesday there was a big catch at Vallejo
(3.0 times the average), and a slight increase at Benicia (1.1), Dutton (1.2), Black
Diamond (1.2), and Collinsville (0.9). There was little or no increase at other points.

On Wednesday, the second day of the run, the catch at Vallejo had fallen off,
and by Thursday the run had entirely passed that point. The points on Suisun
Bay and along the river as far as Isleton were gained on the second and third days,
and the run reached Court land on Friday, the fourth day. There was no fishing at
some of the upper stations on Saturday, that is, Friday night, the law prohibiting
fishing from sunrise Saturday to sunset Sunday, and the record for the fifth day is
incomplete. This run was two days in passing Vallejo, and four days in going from
Vallejo to Courtland.

On Friday, April 29, another run began passing Vallejo, the catch being over
three times the average, and the next day it had increased to over six times the
average. On Monday the Vallejo catch decreased to 1.9, on Tuesday to 1.7, and on
Wednesday to 0.4, the run being five days in passing that place. This new run
was not noticed at the other points on Friday, but on Saturday, the second day, it
had reached all points up to Collinsville at the mouth of the river. By Monday, the
fourth day of the run, it had reached all points from which we have records, the
greatest increase being at the stations farther up the river. During the remainder
of this week the catch continued to fall off at the lower stations, but continued very
large at Sacramento. By Wednesday, the sixth day, it had passed Rio Vista, and
Walnut Grove by the seventh. On Friday there was still a big catch at Sacramento
(5.9) and at Courtland (3.8). The record is imperfect for Saturday as usual, but
apparently the run had passed all stations. To summarize: This run was five days
in passing Vallejo. The foremost were four days going from Vallejo to Sacramento,
and the run was five days passing Sacramento.

124

BULLETIN OP THE UNITED STATES FISH COMMISSION.

The spring run passes upstream quite rapidly, reaching their spawning-grounds
on the McCloud River in about six weeks after entering the river at Collinsville.

The fall run moves more slowly. They are about two months reaching their
spawning-grounds, which are not so far upstream. The flood and ebb tides are more
nearly equal, owing to the smaller amount of water coming from the rivers, making
the passage of the salmon through the bay a little longer. The nets of the fish-
ermen also offer a greater obstruction during the low water and in this way hold
the salmon back. In 1900 salmon were taken in abundance in Suisun Bay and
in the river as far up as Rio Vista- by the middle of August, but were not taken
at Sacramento until after the fii'st of September. The low water doubtless made the
movement slow, and the taking of from 2,000 to 10.000 daily out. of a slow run would
account for their nonappearance at Sacramento.

Upon reaching the shoals in the middle portion of the river they cease their
migration, having already found good spawning-grounds. In 1898, 1899, and 1900
the water was normally low and a large proportion of the salmon. found spawning-
places in the main river. The early high water and frequent fall rains in 1897 sent
them into the tributaries.

The latter part of September, 1901, 150 salmon were weighed and branded with
serial numbers and released in the river near Rio Vista. Three of these were taken
at the hatcheries the latter part of November, just at the close of the season. The
following is a particular account of these three specimens:

No. 8, a female, was branded September 20, when it weighed - 3,930 grams. It
was taken again at Mill Creek fishery November 23, when it weighed 10,180 grams,
having been 64 days on the road and having lost 26 per cent of its weight.

No. 91, also a female, was branded September 24, when it weighed 8,470 grams.
It was retaken at Mill Creek November 20, when it weighed 7,160 grams, its time in
passing up the river being 56 days and its loss in weight being 15 per cent. This
specimen was returned to the creek after being weighed November 20. It was
found dead on the racks 8 days later, when it had spawned all but 20 of its ova.
Its weight had decreased 1,860 grams.

No. 43, a male, was branded September 20, when it weighed 10,080 grams. It
was taken at Battle Creek November 25, when it. weighed 6,275 grams, making its
time from Rio Vista 66 days and its loss in weight 25 per cent.

This important experiment proves that the fall salmon travel very slowly, at a
rate of 4 or 5 miles a day, and require about two months to reach the spawning-
grounds from the month of the river.

The salmon of the spring run arrive at their spawning-grounds from two to six
weeks or even longer before they are ready to spawn. This time they spend lying
quietly in the pools. The fall salmon are more nearly ripe when they reach their
spawning-grounds. Indeed, it is probable that many of them cease to ascend the
streams only when they are ready to spawn.

Downstream movement. — Under ordinary conditions there is probably little or
no downstream movement, yet when the salmon meet with such obstructions as the
racks at the fish-culture stations, there is a tendency to go back downstream. At
Battle Creek fishery more salmon are taken at the lower end of the pool than at
the upper, indicating that they go as far downstream as possible under the circum-
stances. The large number of fishes in good condition that get caught on the rack

Bull. U. S. F. C. 1902. (To face page 124.)

Plate 14.

7.0 .

6.0

S.O

4.0

TV

— To

1.9
1.8
A 7
1.6
l.S
/A
1.3

Line of Average /. /

catch for Season. i °
.9
•a
.7
.6
.5
.4
.3
.2

L

Monday,

April 25.

10

WUa.

Thursday,

MayS.

tve.

TtbU&y-

Friday,

Mayo.

Saturday,
May 7.

NATURAL HISTORY OF THE QUINNAT SALMON.

125

indicates an attempt to go downstream. They may frequently be seen coming
downstream toward the rack, though I have never seen any try to get through it.
When they get close enough to the rack to feel the force of the swift current, they
always try to turn back. Eventually they become so weak that they are unable to
keep from being carried onto the rack, where many of them perish.

There are also a few fishes that drop downstream as they become exhausted
from long residence in fresh water; rarely from spawning. Such were found almost
daily on the upper rack at Battle Creek fishery in 1900. Very few of them had
spawned, though they were almost completely exhausted and hardly ever lived over
a day after coming near the rack. Such specimens usually lie in the less swift
water some 10 or 15 yards above the rack, where little effort is required to maintain
a position. As they become weaker and the current carries them back toward the
rack, they swim back and forth across the creek, their bodies set obliquely to the
current, and their tails frequently almost touch the rack. A very little of such
exertion soon exhausts them, and frequently they go but a few feet before being
carried against the rack, where they die in a few minutes.

Relation between weather and migration. — It is popularly supposed that the
movement of salmon in the rivers is largely determined by weather conditions.
Almost any fisherman can tell of a notable run of salmon that accompanied or fol-
lowed a south wind. Observations made during two years at Battle Creek fishery
show that there is no relation whatever between the direction of the wind and the
movements of salmon. A strong wind of any direction, however, does apparently
cause them to move upstream when they have been lying in a pool for some time.
The most notable movement at Battle Creek fishery in 1898 was coincident with a
strong north wind. A rain or a slight rise in the water usually causes them to run
upstream, but not always. There is apparently no relation whatever between
weather conditions and ripening of fish.

CHANGES IN SALMON AFTER ENTERING FRESH WATER.

The alimentary canal. — It is not uncommon for fishes of the salmon family to
fast during the breeding season. Such is the case with the Atlantic salmon, the
various white-fishes of the Great Lakes, and probably with other species, and it is
well known that adult Pacific salmons do not eat while in fresh water. The Sacra-
mento salmon will often snap at bright floating objects and can frequently be taken
with the spoon while on their spawning-grounds or while passing up the river.
Seventy-five specimens were taken in this way at Jelly Ferry during October and
November, 1900. They have never been known to take food, though indigestible
material, such as leaves, is sometimes found in the stomach, A 13-inch, mature,
sea-run male salmon was taken at Battle Creek fishery in October, 1898. The
stomach was contracted the same as in the ordinary adults, but contained two small
bits of ehitinous substance looking somewhat like portions of the thorax of a grass-
hopper, but may have been portions of a crustacean.

As they do not eat after leaving salt water, the digestive organs immediately
begin to shrivel up. In most of the specimens of the fall run that reach the head
of Suisun Bay the stomach and cieca have already contracted and their walls have
become firm. Only rarely are they thin and flaccid, as if food had recently been

126

BULLETIN OF THE UNITED STATER FISH COMMISSION.

digested. The longer the time since leaving salt water, the more the digestive
organs become contracted.

The figures on pages 126-128 illustrate the successive changes in the alimentary
canal as observed in specimens from Monterey Bay, head of Snisun Bay, Sacramento
River at Sacramento, and at Battle Creek fishery.

Stomach, pyloric appendages, and part of intestine of two female salmon taken in Monterey Bay, Cal., drawn to
same scale. A , July 6, 1900, stomach containing food. B , July 10, 1900, stomach empty.

The skin. — The most immediate change noticeable in the salmon after leaving
the ocean is the great increase in the amount of slime that exudes from the skin
upon removal from the water. This point is of physiological interest, but lias
not yet been studied. By the time the fish reaches the spawning-grounds the skin
in most cases has thickened considerably, and frequently the scales are entirely
embedded and invisible. In the upper figure of plate 13 the scales can be seen only
where the skin has been worn off.

NATURAL HISTORY OF THE QUINNAT SALMON.

127

Loss in weight,. — Many weights and measurements were made in 1900 for the
purpose of determining the loss in weight sustained by the salmon during their resi-
dence in fresh water, but our scales proved somewhat inaccurate, and the data can
not be used in detail. The loss was shown to be very large, about 35 per cent.

Weights were again taken in 1901 with accurate scales. The results are shown
in the tables on pages 128 and 129, which give: 0

First, the length in centimeters of the specimens weighed, the measurements
being made from the nostril to the last joint in the spinal column. The nostril was
selected as a point of measurement rather than the tip of the snout because the
snout becomes lengthened in breeding males.

Second, the average weights of specimens of various lengths delivered at the
cannery at Black Diamond September 5 to 11.

Third, similar weights and averages for specimens taken upon their arrival at
the spawning-grounds, but before they had begun spawning; also the percentage
of loss in weight in these specimens, the average weight of Black Diamond specimens

128

BULLETIN OF THE UNITED STATES FISH COMMISSION.

of the same length being taken as a basis in each case. The males were weighed at
Battle Creek October 29, the females at Mill Creek November 15.

Fourth, similar weights, averages, and losses of spent fishes, taken either before
or immediately after death. Specimens found between October 10 and November 30.

In determining the percentage of loss in the weight of males no account is taken
of the loss of milt in spawning, which is very slight, the average total weight of the
spent fishes or those just arrived at Battle Creek being compared with the average
total weight of Black Diamond specimens of the same length.

Stomach, pyloric appendages, and intestine of two salmon from upper Sacramento River drawn to scale. J?, Bat
Creek, October 20, 1900, a spent female. Jr’ Battle Creek, September 15, 1900, a male dying on spawning-grounds.

Table of average weigh ts of male salmon.

At Black Diamond.

On arrival at Battle Creek.

Spent fishes.

Length, centimeters.

N umber
of speci-
mens.

Weight.

N umber
of speci-
mens.

Weight.

Loss, per
cent.

N umber
of speci-
mens.

Weight.

Loss, per
cent.

44

1

2,200

9

1,905

13

45

1

2, 030

2

1,905

24

46

1

2,580

9

2; 375

8

47

1

2,080

4

2, 198

18

48

3

3,057

2, 439

20

49

3

3,167

3

2,700

15

1

2,890

2,770

9

50

1

3, 060

1

9

51. ..

2

3; 720

2

2,928

21

52

1

3,910

4,100

1

1,630

61

53

1

2

3, 303

17

2

5,295

6,220

2

4,025

24

60

2

1

5, 495
5,510

61

2

6, 455

1

15

63

4

7,236

2

5,533

5,980

23

20

64

3

3

7,378

1

65

8,040

1

6,970

8.870

13

75

2

12, 390

1

28

1

8,535

30

76

4

12,855

1

10,950

15

78

5

13, 192

2

10, 768
10,015

18

79

3

1

16,797

1

40

80

15,360

2

13,707

11

81

1

13; 415
16,135

1

10,770

11,250

20

82

1

33

83

1

15,670

2

14,175
15,040
l7, 665

10

84

2

17,010

1

12

89

1

18,490

1

10

Average

16

26

NATURAL HISTORY OF THE QUINNAT SALMON.

129

The percentage of loss in ripe females is determined by comparing the total
average weight with the average weight of specimens of the same length weighed
at Black Diamond, the latter being taken as a basis. The loss percentage as stated
for the spent fishes does not include the loss of ova. The percentage is determined
by comparing the weight of spent fishes plus the weight of the extruded ova with
the weight of Black Diamond fishes of the same length. The weight of the extruded
ova is determined by finding the difference in weight between the ovaries of spent
fishes and those of fishes just arrived at the spawning-grounds.

Table of average weights of female salmon.

Length, centime-
ters.

Black Diamond.

Ripe, unspawned.

Spent fishes.

Num-
ber of
speci-
mens.

Average

weight.

Num-
ber of
speci-
mens.

Average

weight.

Loss,

per

cent.

Num-
ber of
speci-
mens.

Average

weight.

Loss,

per

cent.

Total.

Ovary.

Total.

Ovary.

1

7 70S

i

5,680

67

13

06

17

8,249

2

7, 048

1,595

15

67 ..

15

8’ 134

i

4,680

80

24

68

9

8 631

2

7,760

2,018

u

i

69

6

3, 949

i

7,720

1,921

14

4, 988

140

25

70

6

8,981

]

8,390

2,057

71

4

i

6, 340

113

15

72 .

10,227

6,573

69

16

73 .

7

10, 060

2

6, 398

69

15

74

1

10,700

2

9,095

2, 192

15

2

7,450

92

11

76

3

11,508

1

9,930

2,014

14

I

6,060

90

31

77

3

12,330

1

11,600

6

78

2

12, 220

2

1 1 ' 995

80 _ _ .

g

14,763

3

11,993

2, 372

19

1

10,990

159

11

82.

14,405

1

8,730

104

24

12

19

The averages of the loss percentages are: For males upon their arrival at the
spa wning- grounds, 1G per cent; for males at time of death, including loss of milt,
2G per cent; for females upon arrival at the spawning-grounds, 12 per cent; for
females at the time of death, not including loss of ova, 19 per cent. The difference
between the loss as deterlnined in 1900 and 1901 is accounted for by there having
been many more grilse weighed at the spawning-grounds the former year. (See
“ Two forms of males,” page 130.)

Under the heading “Details of migration” (page 124) will be found an account
of three salmon that were released in the Sacramento near Rio Yista, after being
weighed and branded, and subsequently taken at the Government fisheries. One
had lost 15 per cent of its weight, another 25 percent, and the other 26 percent during
the migration.

One important point to be considered in this study of the loss in weight during
migration is the deterioration in the value of the flesh as a food. The loss of 12 or
16 or 25 per cent is entirely in nutriment. If even a very fat beef were starved two
months, or until it had lost 16 per cent of its total weight, no one would care to eat
of its flesh. But such is the condition of the fall salmon upon their arrival at the
upper portion of the river. They have eaten nothing for over two months, and nutri-
ment to the extent of about 16 per cent of their weight has been absorbed, almost
wholly from the flesh.

It is evident, therefore, that the fall salmon taken at the upstream points have
but little value as food, and their capture should be prohibited,

F. C. B. 1902—9

130

BULLETIN OF THE UNITED STATES FISH COMMISSION.

COMPARISON OF THE SEXES.

Relative changes in fresh ivater. — Before entering the fresh water the two
sexes are identical in appearance. With the fall run the sex can often be distin-
guished by the external appearance in specimens taken at the head of Suisun Bay,
and it can always be distinguished in normal specimens by the time of their arrival at
the spawning-grounds. Soon after entering fresh water several cartilaginous teeth
appear in the front of the jaws of the males, and at the same time the jaws begin to
grow longer. By the time the males reach the spawning-grounds the .jaws are much
prolonged and hooked, and the teeth have grown to be large and solid canines; the
body becomes deep and slab-sided, and the color usually more or less reddish.

The principal change in the females lies in the diminution in the muscular tissue
of the back and sides and in the distension of the abdominal walls on account of the
development of the ova. Their color is usually more or less olive. After spawning
the female is as thin as the male, but the jaws are not prolonged.

The following illustrations indicate the changes better than descriptions :

salmon, known as adult and grilse, that are found in the headwaters. Those here
shown were nearly of the same length, though it is very rare to find as small a speci-
men as the upper that has the adult form, and the lower was a rather large grilse:
(See also plate 13, photographs of these same specimens.)

The differences are obvious. The grilse simply fails to develop the character-
istics of the breeding male, viz, the prolonged and hooked jaws, the large, hooked
teeth, the deep, slab-sided body, and red color, and retains its salt-water appearance
except in the loss of flesh. Grilse weigh from a half pound up, and intergrade with
the adults both in weight and appearance; specimens with a length of 90 centimeters
(35.5 inches) are occasionally found. I have seen two, which, from their olive color,
could be distinguished as sea-run fishes, that were only 13 inches long.

At Battle Creek fishery in 1900 the males were nearly all grilse, though there
were almost as many of the adult males as there were females. The great prepon-

Head of male salmon taken at Sacramento, September 5, 1900,
showing the beginning of the jaw prolongation and canine
teeth while yet cartilaginous.

Head of female salmon taken at Sacramento
September 5, 1900. This is also the head of
the male in salt water.

Two forms of males. — The illustrations on page 131 show the two forms of male

NATURAL HISTORY OF THE QUINNAT SALMON.

131

derance of the grilse over the adult males and females is due to their being too small
to be taken by the nets of lawful mesh.

The cause or causes that lead to the production of the grilse form are not known.
Mention is made in another jilace in this report of the sexual maturing of the male
parrs that remain in fresh water during their first summer. It it possible that this
stunts them and causes the production of grilse. If grilse were simply young indi-
viduals that followed the adults into fresh water, we would expect to find females
among them. The two forms can not be distinguished except in breeding fishes.

The dwarfed form of the female is practically unknown. Among many thousand
specimens handled at Battle Creek fishery only one dwarfed specimen was found.
This was 1G inches in total length and weighed 2 pounds. Its ova were mature.

Two spent male salmon, adult and grilse forms, Sfound dead on the rack at Battle Creek fishery October 28, 1900.
Length from hinder edg'e of eye to base of tail, larger specimen 500 mm., smaller 455 mm.; weight, larger speci-
men, 3,100 grams, smaller 1,200 grams.

Comparative statemen t of ■ weight and length of the. snout in adult and grilse salmon of the same

body length.

Length from
hinder edge
of eye to base
of caudal fin
(centi-
meters) .

Weight in grams.

Length of snout
from hinder
edge of eye
(millimeters).

Length from
hinder edge
of eye to base
of caudal fin
(centi-
meters).

Weight, in grams.

Length of snout
from hinder
edge of eye
(millimeters).

Adult.

Grilse.

Excess
of adult.

Adult.

Grilse.

Adult.

Grilse.

Excess
of adult.

Adult.

Grilse.

40

1,125

1,025

100

68

49

2, 475

2,375

UK)

77

70

41

1,650

1,000

650

75

57

49

2,700

1,425

1,275

79

67

43

1,725

1,575

150

66

66

50

2,975

1,725

1 225

92

71

43

2,500

1.600

900

66

65

51

2, 775

1,875

900

99

75

45

2,350

1,075

1,275

72

65

54

3,700

2, 425

1,275

90

71

40...

1 , 875

1,675

200

. 72

62

57

3,950

2,775

1,175

95

82

40

2,450

1,525

925

87

65

57

4,000

2, 400

1.600

100

72

47

1,875

1,800

75

75

64

77..

11,600

7.050

4,550

137

128

47

2,825

1,725

1 , 100

90

60

79

10, 800

8,550

2,250

145

102

47

2, 175

1,300

875

71

64

132

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Hermaphrodites. — I am indebted to Mr. J. P. Babcock, of the California Fish
Commission, and to Mr. Chamberlain for two hermaphrodite salmon. Mr. Cham-
berlain’s specimen was discovered by the spawning crew at Battle Creek hatchery in
December, 1900; the other specimen was discovered by Mr. F. A. Coles while cleaning
fish for the cannery at Black Diamond, and was taken in that vicinity in May, 1901.
The accompanying illustration represents the Black Diamond specimen.

The genital organs of the two specimens are essentially the same in structure.
There is but one pair, as in ordinary individuals, but each organ is developed partly
as testis and partly as ovary. One organ in the Black Diamond specimen has
about 3 inches of the anterior portion wel l developed as a testis, and nearly mature.

Immediately behind this are a few ova that
are about as large as ova ordinarily are in
salmon taken in this part of the river. The
next 4 or 5 inches of the organ consists
merely of the supporting membrane and
seminal duct, with half a dozen ova devel-
oped in one place. Then follows a portion
about 2 inches long developed as testis.
The usual seminal duct leads posteriorly.
The other organ of this specimen also has
the anterior portion developed as testis,
while all of the posterior portion is ovarian.
While the ova are of normal size, the local-
ity being considered, their number is not
over one-fourth as great as would be pro-
duced by a similar portion of a normal
ovary.

The Battle Creek specimen is similarly
developed. Some of the ova are attached,
while others are free, as if the fish had been
only partly ripe. All are variable in size,
but none of normal appearance are as large
as the average ova taken at Battle Creek.
Many of the ova evidently were dead at the
time the fish was taken, and some of these
were abnormally large. I understand that
some of the free ova were spermatized with
milt from the testis portion of the same organs, but none of them lived. It is not
known whether they were fertilized.

The genital organs only of these specimens were sent me, and I do not know
the condition of the cloacal openings.

Relative number of males and females. — In measuring and weighing salmon at
Black Diamond, on August 20 and 21, 1900, it was noticed that the females were
more numerous than the males. To determine whether this was merely a peculiarity
of the catch of those two days, Mr. F. V. Hubler, the weigher for the Black Diamond
cannery, was employed to make notes of the relative number of males and females
during the season, with the following result.

NATURAL HISTORY OF THE QUINN AT SALMON.

133

Table, showing relative number of male ancl female salmon taken at mouth of river.

Date.

Males.

Females.

Date.

Males.

Females.

Aug. 21

74

158

Sept,. 1

163

214

113

139

1 (53

Aug. 25.

4

8

Sept. 4

ioi

96

Id

22

124-

1(53

Aug. 28--.

73

219

Sept, 6

137

195

Aug. 28

135

2(51

Sept. 7

163

14(5

Aug. 30

121

202

Sept. 8.- . . ..

130

148

Ang. 31

7(5

121

Sept. 10 -.

241

316

Total

1,781)

2,545

Percent.

41

59

During October and November, 1900, Mr. Arthur Sergison caught 70 male and
32 female salmon on spoon hooks at Jelly Ferry.

The following is a statement of the relative number of males and females taken
in the seine at Battle Creek fishery in 1900:

Table showing relative number of salmon taken at Battle Creek fishery.

Date.

Males.

Females.

Date.

Males.

Females.

Oct. 12

9

3

Oct. 23

236

22

86

14

98

18

Oct. 14

187

20

73

4

Oct. 20

45

10

Nov. 1 ( river )

12

6

28

4

Nov. 19

5

2

20

22

6

Total

1,107

156

175

34

Per cent . .

88

12

Oct. 23

111

11

Relative weight. — The average weight of all salmon taken at Black Diamond
cannery from August 20 to September 10, 1900, varied daily from 21.3 pounds to
23.8 pounds. As the largest specimens taken were invariably males, it is probable
that the average weight of the males was greater than that of the females. The
reverse was true in the upper river, as will be seen from the following statement
of weights of salmon taken by Mr. Sergison at Jelly Ferry:

Table showing relative weight of male and female salmon taken at Jelly Ferry.

Date.

Weight , pounds .

Date.

Weight, pounds.

Date.

W eight, pounds.

Males:

Females.

Males.

Females.

Males.

Females.

Sept. 28

4. 5

11.5

Oct. 7

10.0

Oct, 24

10.0

Oct. 1 . . .

9. 0

8.0

24.0

10 0

31.0

11.5

Oct. 9

3 o

8.0

37 0

Oct. 2

3.0

13.0

4.0

Oct. 25...

4.0

(i.O

8.0

12.0

7.0

10.0

Oct. 10

14.0

10 0

8.0

13.5

Oct. 12

3.0

7.0

Oct. 26

4.5

J

11.0

4.0

12.0

Oct. 27 .

8.0

12.0

5.0

15.0

4.0

14.0

8.0

32.0

Oct. 28-.-

4.(1

24. 0

Oct. 3-..

1.5

9.0

Oct. 13.

6.0

18.0

10 0

3.0

12.0

Oct. 14

4.0

22.0

Oct. 29

14.1)

3.0

14.0

24. 0

14.5

3.0

23. 0

Oct. 16

3.0

Oct. 31

38. 0

3. 5

26.0

4.0

4.5

3.5

34.0

12.5

4.0

6.0

Nov. 3...

8.0

4.0

7.0

4.0

7.0

18. 0

4.5

Oct. 17...

6.0

11.5

6.0

10 5

5. 0

Oct. 18...

23 0

13.0

6. 0

6.0

Nov. 12.-.

11.5

8.0

8.0

Nov. 13

10. 0

12.0

9.0

38. 0

24. 0

12.0

Oct,. 22

3. 5

32.5

21. 0

34 0

Oct. 23. . .

14 0

Oct. 7

4.5

12.0

29. 0

Average.

9. 1

16.8

7.0

Oct. 24

4.0

134

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The average weight of the Battle Creek specimens could not be determined,
owing to the selection of the larger males for spawning, but it was certainly less
than that given in the Jelly Ferry record.

It will thus be seen that throughout the fall season of 1900 there was a greater
proportion of female salmon taken by fishermen in Suisun Bay and the lower river.
On the other hand, the small males, being too small to be taken in the regulation net
of the market fishermen, were greatly predominant in the headwaters. The evidence
here given does not indicate that more of one sex is produced than of the other.

This point should be considered in making laws governing salmon fishing. The
small males are not desirable for propagation, either natural or artificial, and on
account of their great number they are a nuisance at the Government fisheries.
They are simply so much valuable food wasted. The present law prohibits the use
of nets that would catch them, and it should be amended. As there are no small
females, the small-mesh net would not affect the supply of breeding females. A
law prohibiting the taking of small fishes is of value only when the small fishes are
growing fishes. But the small salmon that come in from the ocean are not growing
fishes. None of the salmon ever return to salt water. Their sole value lies in adding
so many pounds to the market supply or in reproducing their kind on the spawning-
beds. A large fish is worth more on the markets than a small fish; but so are large
cattle worth more on the market than small cattle, yet a stock- raiser would never
think of selling his fine cattle and keeping only the runts to breed from. It would
be better for the salmon as a species, and therefore better for the salmon industry,
if the present minimum net-mesh were made the maximum. A small-meshed net
does not catch so many large fishes, which would allow the larger individuals to
reach the spawning-grounds. The salmon will certainly deteriorate in size if the
medium and larger sizes are taken for the markets and only the smaller with a few
of the medium allowed to breed.

NATURAL PROPAGATION.

Spawning habits. — Salmon in spawning usually take a position at the upper end
of a riffle where the current is strong and where there are gravel and cobblestones
among which the eggs may lodge. After selecting the place the female extrudes a
few eggs and then moves away. The male immediately takes her exact position, or
perhaps a point one or two feet downstream from it, and extrudes a small quantity
of milt. In about five minutes the process is repeated, the female always taking the
position first occupied. This they continue day and night for over a week, usually
nearly two weeks. I have observed salmon spawning at night, but have never been
able to watch one pair until spawning was completed. Branded salmon No. 91, pre-
viously referred to, was only eight days in spawning, although some eggs had been
extruded before it was taken. Two weeks is the spawning time usually assigned by
persons living in the vicinity of salmon streams, which is probably about right.

On account of the difficulty in seeing eggs under water, it has been impossible
to determine the rate at which ova are deposited. The motions of the fish show just
when ova are being extruded, but observation at a distance of 5 feet, with the aid
of a field glass, has failed to disclose the eggs.

The female at irregular intervals turns over on her side and digs her tail into
the gravel. If the gravel is fine there is often a considerable hillock thrown up,
leaving a hole 6 or 8 inches deep and 2 feet across. This digging is probably not

NATURAL HISTORY OF THE QUINNAT SALMON.

135

for the purpose of covering the eggs, nor to make a space for them to lie in, but by
the violent exercise to loosen the eggs from the ovaries. If the purpose were to
cover the eggs it would be repeated every time any were deposited. Gravel does
not drift so far as the eggs, and if such were the purpose it would not be accom-
plished. Besides, it is almost impossible to cover eggs with gravel; the eggs, being
almost as light as the water, slide away from the gravel. More than that, a covering
of over an inch of even fine gravel kills them. The hillock, by forming an eddy at
the bottom of the stream, prevent s many eggs from floating away and being devoured
by other fishes, but such are liable to be covered too deeply and killed in that way.
Some of the fine sediment, however, may settle on the eggs and tend to make them
invisible to egg-eating fishes. The “nest” can hardly be made as a place for the
eggs to lie in, for the current always carries them below it.

The presence of the other sex is not necessary to excite either to spawning
efforts. I have seen the female spawning alone at Battle Creek fishery, and other
persons have reported similar observations from other places. In September, 1900, I
saw a male spawning alone near Sims, the female having been killed b}^ a sportsman
in order to get front bait. Like observations have been reported by other persons.

Percentage of fertilization. — As one pair of salmon deposits an average of 6,000
eggs the increase would be enormous unless there was great loss at some period. It
is usually supposed that the greater part of this loss is due to a lack of fertilization
of the ova. The great care necessary to secure perfect fertilization artificially has
led fish-culturists to suppose that the percentage of fertilization under natural con-
ditions must necessarily be very low. In artificial fertilization the ova and milt ai’e
mixed together in a vessel, insuring a coating of milt or spermatized water over each
ovum. In natural spawning the ova are caught in the eddies among the rocks,
either near the nest or within a few yards below it. A few seconds after the ova are
spawned a small quantity of milt is disseminated in the current to be carried against
them. It seems very unlikely that a large percentage could be fertilized under such
conditions. The following experiments throw some light on the question:

To determine the percentage of fertilization under as nearly natural conditions
as possible a box was built 4 feet wide, 14 feet long, and 15 inches deep, and a strong
current of water turned through it,. About 5 inches of gravel was put in the upper
three-fifths. A pair of salmon were placed in the box October 28, 1897. A female
not quite ripe was selected, in order to allow a few days to become accustomed to
the place. Pickets nailed to the side prevented the fishes from jumping out. By
November 2 they seemed to be at home in the box, and their actions indicated that
they wei’e ready to spawn. A few eggs were deposited the next day. On the 4th
the male died, having become almost entirely covered with fungus in the one week.
Another was put in immediately, but the spawning was interrupted, as it required a
day or two to get used to the place. The female died November 12, having deposited
but few eggs. No cause of death could be ascertained. Of the 512 eggs deposited,
343 were killed while being deposited. Of the remaining 169, 129 or 76 per cent were
fertilized. In the second attempt 82 per cent were fertilized.

In 1898 a pair of salmon were put in a ravine with simply a rack to prevent
their going downstream. No eggs were deposited.

So far as the number of eggs killed is concerned this experiment is not a fair
test. The level floor of the box made few eddies, and the eggs were washed into the
corners and killed The percentage of fertilization would certainly be no greater

136

BULLETIN OF THE UNITED STATES FISH COMMISSION.

tlian under entirely natural conditions. We would expect the death of the male
and the introduction of a new one to cause some eggs to he left unfertilized.

November 18, 1897, I dug up live or six nests in comparatively still water where
fishes had been seen spawning for a month. The sand and gravel were thrown into a
screen and carefully sifted, but no eggs were found; but on stirring the gravel and
cobblestones in a strong current and setting the screen below so as to catch floating
objects, 13 eggs were secured; 11 were alive and all fertilized, 4 about 3 days old and
the others about 28 days. I could see no indication of fertilization in the dead eggs.

In November, 1898, in order to obtain eggs naturally spawned, I placed a screen
obliquely in the Avater 2 or 3 feet below AArhere salmon were spawning. The screen
was covered with small cobblestones, that the eggs might lodge among them and
be protected from spawn-eating fishes. The first trial was unsuccessful. The second
secured 48 eggs; 30 were dead when found. All of the live eggs were fertilized, the
others could not be tested.

The experiment with the screen is not a fair test for the number killed, as the
screen caught much gravel Avhich pressed the eggs against the wires, and without
doubt killed many more than would have been killed under natural conditions. The
1897 experiment in securing eggs from the stream, Avhen 15 per cent \vere dead, Avas
a fair test, but the number of eggs Avas too small to warrant definite conclusions.
In both experiments all live eggs Avere fertilized.

In November, 1900, 39 eggs were secured from natural spawning-beds in Battle
Creek, 25 of them evidently fertilized, and 1 certainly not fertilized. The condition
of the others could not be satisfactorily determined, as they were killed in securing.

At another time a fish Avas artificially spawned in the creek on natural spawning-
beds, a screen being placed beloAV to catch the eggs. A male Avas held in the same
position immediately afterwards and milt expressed. The test Avas not quite fair,
for although there Avas probably a larger quantity of milt than is discharged at once
naturally, yet there was also a larger number of eggs. The eggs being caught by the
screen and thereby remaining closer to Avhere the male was stripped was of no
advantage, as the life of milt in water is ample to allow it to come to rest. If they
had been farther away it would have given time for the milt to become more thoroughly
disseminated through the water. The eddies caused by our standing in the water and
holding the fish prevented the eggs and milt from floating off in a natural manner.
In two trials, 35 per cent Avere fertilized in the first, and 50 per cent in the second.

These various experimentsindicate a high percentage in natural fertilization, prob-
ably over 80 per cent. It is significant that all live eggs found that had been spawned
naturally, excepting one, Avere fertilized. The fertilization of dead eggs could not be
determined, though they Avere no more liable to be unfertilized than live ones.

Mortality among ova. — These experiments also point to a high but indefinite
mortality from being covered by the gravel. The greatest loss, however, is probably
due to spawn-eating fishes. In the. middle portion of the river, as at Battle Creek,
when salmon are spaAvning, great numbers of other fishes, mostly the split-tail
( Pogomclithys ), gather about to feed on the spawn. Fifty or more split-tails may
often be seen lying a few feet downstream from a spawning salmon, and although
each fish may eat but few eggs, all together they probably destroy a large per-
centage of the eggs spawned in the middle portion of the river.

Trout have been taken near the spawning platform at Battle Creek station Avith
the stomach and throat gorged with eggs, the waste from artificial spawning, and 3

NATURAL HISTORY OR THE QUINNAT SALMON.

137

ova were found in the stomach of a trout taken in Mill Creek December 1, 1901, several
days after any artificial spawning had been done at the station. As there were several
salmon spawning in the creek at that time, there is little doubt that the eggs were
secured from natural spawning-beds. Trout are adapted to catching floating objects
and are doubtless very destructive to salmon spawn where the salmon breed naturally.

Other fishes, such as the hitch ( Lavinia ), hardhead ( Mylopharoclo n ) , and sucker
( Catostomus ), have not been found to eat salmon spawn, though they probably do;
the black-fish ( Orthodon) has not even been found near the spawning-beds. A large
Sacramento pike ( Ptychocheilus ) that had secured spawn from natural spawning-
beds was taken in the river near the mouth of Battle Creek in 1900.

Natural versus artificial propagation. — Probably the most important problem
yet remaining unsolved in connection with the natural history of the salmon is the
efficiency of natural propagation. If we could segregate a certain number of fishes
in a small stream, then put a fine screen across below where they are to spawn, and
later catch all the alevins and fry produced, we could solve the problem. But a
small stream, such as could be experimented with, is liable not to have an average
number of fishes to prey upon the spawn and alevins, and the conditions would not
be entirely natural. The following statement represents approximately the com-
parative value of natural and artificial propagation :

Percentage of loss in

Items.

propagation.

Natural.

Artificial,

Not spawned ...

1

«1

Unfertilized

15

2

Killed before hatching

^70

8

Alevins killed

ft 13

c2

Total loss..

99

13

« 10 to 20 per cent if unspawned eggs are not removed by abdominal section.
b No definite data

c At least 50 per cent if alevins aie planted.

From the foregoing it will be seen that the heavy loss in artificial propagation
has been in not spawning all the eggs and in planting alevins, both of which can be
remedied, as is elsewhere shown in this report. The total loss in artificial propaga-
tion should not be 15 per cent.

There is a much greater loss when alevins are planted artificially than when
they hatch out naturally.

(а) From a given number of ova, as those produced by one fish, which is the
basis of the percentage, there are more alevins to be destroyed in the case of artifi-
cial propagation. In natural propagation they have been already largely destroyed
before they become alevins, and there are not 87 per cent left to be destroyed, as in
the case of artificial propagation.

(б) But even with a given number of alevins the percentage is greater in artifi-
cial planting It is not possible to scatter them as well in artificial planting as in
natural propagation. No amount, of care will prevent their collecting in bunches,
which has not been seen in natural propagation.

Something of the value of artificial propagation can be learned from an experi-
ment tried at Clackamas hatchery, Oregon. In March, 1896, 5,000 salmon fry 2.5
inches long were marked by cutting off the adipose fin. The eggs from which the
fry were hatched were spawned .at Baird hatchery in September, 1895. Mr. Hub-

138

BULLETIN OF THE UNITED STATES FISH COMMISSION.

bard, superintendent of Clackamas hatchery, who tried the experiment, reported
that 375 of the marked Ashes were taken in 1898. The smallest weighed 10 pounds,
the largest 57 pounds, and the average was 27.7 pounds. Besides these, 5 were
taken in the Sacramento River in 1898. A few more were taken both in the Colum-
bia and in the Sacramento in 1899, and also in 1900. The 1900 specimens, however,
may have been of those marked in the Sacramento in 1898. From those 5,000 fry 2.5
inches long, costing less than a dollar to produce, Ash weighing over 5 tons were
taken. That means that for every female Ash stripped at the hatchery the Ashermen
should catch about 5 tons three years later. About 400 of the 5,000 marked Ashes
were reported taken. We have no means of knowing how many came back to fresh
water and escaped the nets, or how many were caught but not noticed.

INJURIES AND DISEASES.

General effects of spawning . — Notwithstanding their long journey from the ocean,
the salmon reach their spawning-grounds in good condition. They are not nearly
so fat as when they left the ocean, but all their bruises are received after arrival at
the spawning-grounds. This fact has already been noted by Evermann (Bulletin
United States Fish Commission 1896, p. 191).

Spawned-out female. Battle Creek, October 20, 1900.

As spawning progresses the abdominal walls of the female contract and she
becomes as thin as the male. Her caudal An is worn off to a mere stub. All A ns of
both sexes become more or less frayed, the skin wears off the sides of the tail and
the prominent portions of the body, such as the edges of the jaws and bases of the
Ans. Fungus nearly always attacks the gills and the various bruised places and
frequently destroys one or both eyes.

It has been supposed that the exertions of spawning completely exhaust the
female and that she dies immediately upon its completion. It would seem rather
strange if there were just enough energy to spawn all the ova, and that with the
extrusion of the last one the Ash should die at once. Observations indicate that
the female has considerable energy left after spawning all of the ova, and that she
continues on the spawning-beds for some time thereafter. The injuries are
received only after most of the ova have been spawned. She probably does not
know when the ova have all been extruded, and her instinct compels her, when once
spawning has begun, to continue the spawning efforts as long as energy lasts. The
complete extrusion of the ova, since it is not noticed, is merely incidental.

In 1900, 14 spent females were taken alive on natural spawning-beds; 7 of them
had extruded all ova, in one specimen there was 1 ovum yet unspawned, in two

Bull. U. S. F. C. 1 902. (To face page 138)

Plate 1 5.

SPENT AND DISEASED SALMON.

A, Female that had spawned all but about 500 ova; showing that the injuries are received while spawning the last few ova, or after all
lave been spawned.

li, Male, apparently exhausted from long residence in fresh water, but not from being on spawning beds; typical condition of late
iummer-run males, Battle Creek, September 15, 1900.

C, Female, with ova but half developed, Battle Creek, September 15, 1900: died from long residence in fresh water.

B, E. Two males, grilse and adult, showing extreme cases of fungous growth, October 22, 1900.

NATURAL HISTORY OF THE QUINNAT SALMON.

139

there were 2 each, and in the others there were 3, 58, 92, and 107, respectively.
If they died immediately after spawning the last ovum we would not have found
such a large proportion of live specimens completely spawned out; and if spawning
completely exhausted them we would not have found them on riffles hut in the
more quiet water. In one instance when only one female was taken she was
entirely spawned out and had been seen spawning just before the seine was hauled.
Of course, it is possible that the last ova were spawned just before hauling the
seine, but in any case the fish was far from being in a dying condition. (See plate
16 and fig. A, plate 15.)

Spent salmon. — A few sample field notes on the condition of spent salmon found
dead at Battle Creek in 1900 are here given. Several hundred similar specimens
were examined during the season.

The following notes refer to spent females :

September 30. Nearly spawned out. Numerous parasites (copepods) and a small patch of
fungus on each gill. Top of head without fungus, but with skin worn off. No fungus on body.
Fins and skin in good condition.

October 10. But two eggs left in body cavity. Gills about one-fourth covered with fungus;
several gill parasites. Skin worn off caudal fin and the rays about half worn off.

October 15. All but 10 eggs spawned. Died in shallow water. Caudal fin entirely worn off,
but fish otherwise in good condition. But little fungus. Gills but slightly injured.

October 26. Entirely spawned out, except 2 eggs yet attached to ovary. Half of caudal fin
worn off evenly; other fins in good condition. Gills one-third destroyed. Small patches of fungus
in various places on body. One eye blinded.

N ovember 1 . Specimens Nos. 2 and 3 from the river were of the same length. No. 2 had spawned
all but 92 eggs, and No. 3 all but 442. No. 3 weighed 900 grams more than No. 2. The skin was
entirely worn off the caudal fin of No. 2, and the rays half worn off. Caudal fin of No. 3 in good
condition.

The following notes refer to spent males:

November 5. Badly scarred, one eye blinded, skin worn off edges of fins and jaws. Another
specimen, not badly scarred, blind in both eyes, skin worn off snout and edges of fins.

November 6. Skin and flesh worn off in several places behind dorsal and on tail nearly to back-
bone; skin worn off edges of fins, jaws, and the whole snout; both eyes blinded; gill filaments
half destroyed by fungus and parasites.

November 10. One eye blinded; much scarred; little fungus. Another specimen, blind in both
eyes; skin worn off jaws and edges of fins; skin dead all over tail and caudal fin; nearly every gill
filament with one or more parasitic copepods, and many sloughed off for one-third their length.

Diseases of intestine. — The intestine of the spawning salmon is frequently
inhabited by tapeworms, which sometimes completely fill it and extend into the cmca,
but I have never found them in the stomach. They were much more abundant in
1898 than in 1900. In addition to the tapeworms the intestine, especially posteriorly,
is filled with a viscid, greenish yellow fluid. No examination of this has been made,
but it is probably formed by the disintegration of the lining of the intestine, a
catarrhal desquamation such as has been found in the Scottish salmon.

Fungus. — Fungus as related to salmon deserves special investigation. Nearly
all the salmon that reach the vicinity of Battle Creek fishery during September and
October become affected with fungus, which grows in velvety patches on various parts
of the body. The points most commonly affected are the top of the head, the gills,
fins, and eyes. Of 31 specimens noted on the racks at Battle Creek fishery during
October 16, 17, and 18, 1900, 5 were blind in both eyes and 14 others blind in one ejre,
as a result of fungus.

140

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The following extracts from my field notebook indicate the rapidity of growth
of fungus, and show a condition rather worse than the average, though by no means
exceptional. The two descriptions refer to the same specimen; first, on September
30, when it was caught and tagged and returned to the creek; and second, on October
4, when it was found dead against the rack:

September 30.— Male, ripe, weight 2,800 grams. A notch in left pectoral, a slit in dorsal;
caudal with a few small dead spots, one worn through; 3 parasites on left gill, 7 on right; whole top
of head and upper edge of pectoral fins covered with fungus; skin partly worn off sides of tail.

October 4. — Fungus covering following portions of fish; whole top and sides of head to below
eyes, lower jaw, back in front of dorsal, edges and bases of pectoral fins, upper side of ventrals, a
spot behind right pectoral and one on back before adipose fin, half of adipose fin, spot behind left
pectoral, left side below dorsal nearly to lateral line and half way to adipose fin, base of anal on
left side, and belly behind ventrals. Left gill with seven streaks of two or three dead filaments
each; a little fungus on each dead portion; a small patch of fungus at tip of gill matting together
the filaments of all the arches. Right gill with a patch of fungus at tip matting together filaments
from all the arches, and another anteriorly on the inner arches. Skin of tail and most of caudal
fin dead; some of caudal rays gone.

There were worse cases of fungus than the one here described, but this shows
what can grow in four days. Another specimen that was in good, condition when
tagged November 1 was “half covered with fungus” when seen last on the 8th.

The iiest almost disappeared in December. Figures D and E of plate 15 show
the extent to which salmon are sometimes affected with fungus.

Grill parasites. — Another common pest of the salmon in fresh water is a parasitic
copepod which attaches itself to the gill filaments. There are usually not very many
on one fish, but sometimes the gills are almost destroyed by them. Plate 15 shows
an extreme example. The gills sometimes decay without being affected with fungus
or parasites, as Avas found in a specimen at Battle Creek fishery, October 7, 1900, in
which one-third of the gill filaments were dead. (See also plate 13.)

Diseased ova. — In all of the females found dying during September and October,
1900, the oati Avere more or less diseased. Sometimes there Avere only a feAv addled
and misshapen ova crowded into the interstices of the healthy ova, but sometimes
almost all were addled. Occasionally there were a feAv abnormally large ova, half an
inch or more in diameter, and in a specimen taken October 12, 1900, nearly all Avere
in this condition. In another taken about the same time half Avere of this charac-
ter, Avliile the normal ova had been spawned. In another a third of the ova were
addled, and the others had absorbed water and were turgid. See fig. C, plate 15.

Length of life of fall salmon after reaching spawning -grounds. — September 30,
1900, numbered metallic tags Avere attached to 3 male salmon, Avhich were then
released in the pool between the racks at Battle Creek fishery; 1 of these Avas found
dead October 5, having survived 5 days. October 22, 36 others were tagged and
released in the pool; 27 of these Avere seen at various times, some of them cpiite fre-
quently, up to November 1, and 5 of them Avere found dead within that time, the
maximum time being 10 days. On October 25 36 Avere tagged and released in the
creek below the racks; 8 of these Avere found dead on the racks up to November 10,
a period of 16 days. Four Avere tagged and released in the mouth of the creek,
about 2 miles beloAv the fishery, on November 4. One of these Avas seen on the 5th
and again on the 8tlx, when it was almost dead, a period of 4 days. November 9, 39
Avere tagged and released in the river below the mouth of Battle Creek; 3 were seen
at the fishery on the 16tli, 7 days afterwards.

■»» *• .#

Bull. U. S. F. C. 1902. ■ (To face page 140.)

Plate 1 6.

TAILS OF FEMALE SALMON FROM SPAWNING BEDS.

A, B, Tails of spawned-out salmon; C, tail of branded specimen No. 91, the tail being perfect eight days before photograph was taken;

D tail of salmon with about 500 ova yet remaining.

NATURAL HISTORY OF THE QUINNAT SALMON.

141

Altogether 12 tagged fishes were seen after dying, and the average time that
they lived after tagging was 11 days. The longest time was 16 days; some had
probably been in the creek a few days when tagged, though the freshest were selected.
Two weeks is a very fair estimate of the length of life after reaching the spawning-
grounds. Branded specimen No. 91, a female, lived but 8 days after reaching the
spawning-grounds.

DEATH.

The salmon of the genus Oncorhynchus apparently has no instinct whatever to
return to salt water after spawning. Worn-out specimens are sometimes seen drift-
ing down stream and have been found as far down as Sacramento, though it is by
no means certain that such have been on spawning-beds. In such cases they are
simply too weak to stem the current and, according to a Sacramento fisherman,
“not fit to look at. ” Dead salmon rarely float, though the current sometimes washes
them along the bottom a short distance I have seen dead salmon lie for several
days in rapids and have seen them in all stages of decay in strong currents. Of the
200 or more dead salmon that were marked and thrown over the upper rack at
Battle Creek fishery in 1900, only 2 were carried to the lower rack, which was a half
mile further down stream. In small streams the water is often greatly contaminated
by the dead fish, and the stench is a great nuisance to people living in the vicinity.

The great variation in size of spawning salmon, together with the occasional pres-
ence of certain scars, such as a broken nose, has led many people to doubt whether
they all die after spawning once. The variation in size amounts to nothing as an
argument, when we know that with about 60 marked fishes known to be of the same
age, taken in the Columbia River in 1898, the variation in size was from 10 to 57
pounds. The broken nose could be received at many other times than when spawning.

It is sometimes thought that if a spawned-out salmon would float down stream to
salt water it would revive, but such is not the case. Humpback and dog salmon
often spawn in small creeks and brooks that empty directly into the ocean, yet they
die like other species. They have been seen dying and dead in brackish water.
The investigation of the blueback salmon or redfish in Idaho in 1895 (see Bulletin
United States Fish Commission 1896, p. 192), when a net was placed across the
mouth of a small stream containing about a thousand salmon, proved that that spe-
cies has no tendency to return to salt water after spawning. Lake Karluk, Kadiak
Island, Alaska, is but about 20 miles from the ocean and is a great spawning-place for
the blueback salmon. The outlet is shallow near the mouth, and if the salmon
ever went back the Indians would be sure to see them, but they do not. In June,
1897, the shore of the lake for miles was lined with the bones of the salmon that
had died six to eight months previously.

The fact that all salmon of the genus Oncorhynchus die very shortly after
spawning once can not be questioned.

Plate 17.

REID F-ECRUY

W.FTtt. o»4 C v-f K.

■

SACRAMENTO RIVER BETWEEN REDDING AND TEHAMA. SPAWNING BEDS OF FALL SALMON INDICATED BY CLUSTERS -F DOTS.

I

s

■ i« • • ' Mj/ :

n...»iiilu.,im'»L» ..

B

Plate ia

CHART OF SACRAMENTO

OBSERVATION STATIONS INDICATED THUS O.

NOTES

ON

FISHES FROM STREAMS AND LAKES OF NORTHEASTERN CALIFORNIA
NOT TRIBUTARY TO THE SACRAMENTO BASIN.

By CLOUDSLBY RUTTER,

Naturalist , United States Fish Commission Steamer Albatross.

143

NOTES ON FISHES FROM STREAMS AND LAKES OF NORTHEASTERN
CALIFORNIA NOT TRIBUTARY TO THE SACRAMENTO BASIN.

By CLOUDSUEY RUTTER,

Naturalist, United States Fish Commission Steamer Albatross.

The fishes forming the basis of the following report were collected in 1898 and
1899 while studying the distribution of the fishes of the Sacramento Basin. The
collection was studied at Leland Stanford Junior University, where special facilities
for study and comparison were afforded by the ichthyological museum.

The localities from which the collection was obtained represent four basins, now
distinct, though at one time probably tributary to Lake Lahonton. These basins
are Grasshopper Lake, Eagle Lake, Honey Lake, and Truckee River.

Grasshopper Lake is an alkaline pond, with no outlet, at the southern end of
Grasshopper Plains, in Lassen County, and contains no fishes. A species of Agosia
was found to be abundant in a spring emptying into the lake.

Eagle Lake also has no outlet, the lowest point in the surrounding watershed
being over 50 feet above the surface of the lake. Its water is slightly alkaline,
though very clear, and near the shore supports thick aquatic vegetation. Only two
species of fishes were obtained, the Eagle Lake white-fish ( Rutihis olivaceus) and
a sucker ( Chasmistes chamberlaini) here described as new. A trout is known to
inhabit the lake, but none was obtained.

Two streams were fished in Honey Lake Basin — Willow Greek and Susan River.
The former rises at the lowest point in the Eagle Lake watershed, and its upper
part is a rough mountain stream. It was fished about 15 miles north of Susanville,
where it passes through an extensive meadow. Susan River rises on the eastern
slope of Lassen Butte, and above Susanville is a mountain torrent. Its lower part
lies in the plains adjacent to Honey Lake and is dry during part of the year.

Collections were made in three streams of Truckee Basin — Little Truckee River,
Sage Hen Creek, and Prosser Creek. The former, a considerable stream, 15 to 30
feet wide, with very rocky bottom, drains Webber Lake and Independence Lake,
and was fished a short distance below the outlet of the latter. Sage Hen Creek is
tributary to Little Truckee River. It is but a small stream, flowing through a nar-
row wooded valley. Prosser Creek is tributary to Truckee River and drains the
table-land north of Truckee. It was fished near Prosser Bar, where it is a meadow
stream from 6 to 10 feet wide and 6 inches to 6 feet deep.

The fish fauna of these waters is very limited. Three species are described as
new. The collection consists of 9 native and 2 introduced species, distributed as
follows: One species, Agosia robusta, is common to three of the four basins and is
probably to be found in the other basin, Eagle Lake. Another, Rutilus olivaceus, is
found in all the basins except that of Grasshopper Plains. Chasmistes chamberlain i ,

145

F. C.B.1902 — 10

146

BULLETIN OF THE UNITED STATES FISH COMMISSION.

of Eagle Lake, is not found elsewhere. Honey Lake and Truckee River basins have
5 native species in common — Pantosteus lahonton , Catostomus talioensis, Rutilus
olivaceus, Agosiarobusta, and Coitus beldingii. Salmo irideus is found in both basins,
but it has been introduced into Truckee Basin, and possibly into Honey Lake Basin.
Leuciscus egregius is also known from Honey Lake Basin, and Coregonus williamsoni
and Salmo lienshawi from Truckee Basin. Salvelinus fontinalis has been intro-
duced into the latter basin.

Pantosteus lahonton Rutter, new species.

Head 4.5 in length, depth 5.5; eye G in head; D. 10 or 11; A. 7; scales 17-81 to 96-12, 47 to 50
before dorsal. Body terete, caudal peduncle hut little compressed; interorbital slightly convex, or
flat, width of bone 2.8 in head; eye posterior, 3 in snout, 2.5 in interorbital space, 1.5 in distance
between eye and upper end of gill-opening; snont equal to half of head, broadly rounded both ver-
tically and horizontally, projecting beyond the large mouth; 4 rows of papillae on upper lip, 4
rows across symphysis of lower lip, 10 papillae in an oblique row from corner of mouth to inner corner
of lobe of lower lip; isthmus broader than interorbital, equal to distance between pupils; fontanelle
present, but less than half width of pupil in a 6 -inch specimen; dorsal inserted from 49 to 52
hundredths of body from tip of snout; ventrals inserted under ninth ray of dorsal, halfway
between tip of snont and tip of middle caudal rays; caudal 1.5 in head, deeply emarginate, not
forked; pectoral 1.3 in head; height of dorsal about 1.4 in head, the base equal to snout, margin
slightly emarginate; ventrals 1.7 to 1.8 in head. Very dark, almost black above, abruptly paler
below, lower fins slightly dusky. Maximum length, about 6 inches.

Closely related to Pantosteus generosus, but with the following differences, determined by
comparison with specimens of that species of the same size from Provo, Utah. The Provo speci-
mens have the dorsal 10 or 11 instead of 9 or 10, as described by Jordan & Evermann.

Species.

Interor-

bital

space.

Scales in
lateral
line.

Scales

before

dorsal.

Distance of ven-
trals from
snout equals
distance from
their inser-
tion—

Caudal
fin con-
tained in
head.

Width of
lower jaw
(cartilage) in
head.

Bows of
papillae
across
symphysis
of lower lip.

Papillae on
lower lip in
oblique row
from corner
of mouth.

P. lahonton..-

Plat ....

81 to 96

45 to 50

To tip of middle
caudal rays.

1.2

3. 5 to 4, mar-
gin convex.

4

10

P. generosus ..

Convex.

77 to 87

41 to 45

To middle of
middle caudal
rays.

1

3 to 3. 7, mar-
gin nearly
straight.

2

7

Found in abundance in Susan River, and also in Little Truckee River and Prosser Creek.
Types (No. 50587, U. S. Nat. Mus.) from Susan River, collected by Rutter and Chamberlain.

FISHES FROM NORTHEASTERN CALIFORNIA.

147

Catostomus tahoensis Gill & Jordan.

Head 4.4 in length, 5 to tip of middle caudal rays; depth 4.8 in length; width of head through
opercles equal to its depth; eye 5.5 in head, 2.5 in snout, 2.7 in interorbital space, 1.7 in distance from
eye to upper end of gill-opening (by eye is meant the orbital opening, not the eye-ball nor socket) ;
interorbital (bone) 2.3 in top of head; width of isthmus 3.8 in head, a little less than distance
between eye and gill-opening, equal to width of opercle, and also equal to distance between
corners of mouth. (Measurements made on a 7.3 inch specimen.) D. 11; A. 7; scales 17-89 to
105-16. Body rather slender, profile steep; snout blunt; mouth large, with full lips, covered
with rather coarse papillae which do not become much smaller toward margin of lips; upper lip
with about six rows of much-crowded papillae: lower lip with two rows across symphysis, and
about 8 papillae in a longitudinal row through lobes; posterior margin of lower lip reaching ver-
tical through posterior nostril. Orbital rim but little developed, middle ridge of skull broad, the
interorbital space rather high and rounded. Insertion of dorsal in middle of body, its length 1.4
in its height. Insertion of ventrals under fifth ray of dorsal. Length of caudal 1 to 1.1 in head,
rather deeply forked, middle ray 1.6 in longest. Anal reaching past base of caudal, its height
equaling length of caudal; length of ventrals equals height of dorsal; pectoral a little shorter than
caudal. Least depth of caudal peduncle 2.6 in head. Lateral line complete, straight. Peritoneum
black. Color nearly black above, slightly mottled with pale yellowish below.

Taken in Willow Creek, Susan River, Little Truckee River, and Prosser Creek. Description
based on specimens from Susan River.

Chasmistes chamberlaini Rutter, new species.

One young example, 1.7 inches long, and a dried head 3.3 inches long, which can not be referred
to any hitherto described species, were obtained at Eagle Lake. Eye 7 in head, 3 in snout, 2.8 in
interorbital bone, 2 in distance from eye to upper end of gill-opening. Premaxillary spines form-
ing a prominent hump, maxillary inclined about 40°,. falling far short of anterior nostril, its length
from free end to tip of snout just equal to snout in front of nostril, 3.2 in head; lower jaw 3.5 in
head. Interorbital (bone) 2 in head, considerably arched transversely; a low, sharp longitudinal
ridge along middle suture, showing even in the young example. Nasal spines very prominent;
fontanelle closed, covered by a thin bone. Mucus canals prominent, but probably intensified in
dried specimen. Lips thin, two rows of papillae on upper; lower incised to base, lobes small, with
scattered papillae (The above data from the dried head.)

Cross series of scales 93; D. 10; A. 7. Origin of dorsal in middle of body, ventrals inserted
under sixth or seventh ray of dorsal. Pectoral broad, reaching two-thirds distance to ventrals.
Ventrals scarcely reaching vent, the outer two rays longest. Anal low, when depressed reaching
halfway to caudal. Caudal peduncle long and slender.

Has smaller scales than any other species of the genus. The dorsal and anal are the same as
in C. copei, but it differs from that species in the broad interorbital and the papillose lips, in addi-
tion to the small scales. The sharp ridge on interorbital also seems to be a distinctive character.

Named for Mr. F. M. Chamberlain, of the U. S. Fish Commission steamer Albatross.

Type (No. 50588 U. S. Nat. Mus.). Collected in Eagle Lake by Rutter and Chamberlain.
Leuciscus egregius (Girard).

The specimens here noted are not quite so deep as specimens from Winnemucca, Nev., but
otherwise can not be distinguished. They have two red stripes along side, with a darker stripe
between. Lower part of cheek yellowish, with some yellow along edge of belly. Scales in lateral
line 55 to 63. D, 8 or 9; A. 8 or 9. Common in Willow Creek and Susan River.

Rutilus olivaeeus (Cope). Eagle Lake White-fish. — 5' '• ^ i

This species was met with in Eagle Lake and Willow Creek, where it attains a length of 8 inches.
Plead 3.3 to 3.7 in body; depth 3.7 to 4.5; eye 4.4 to 5 in head; insertion of dorsal 0.53 to 0.57 of
body from snout. Scales 15-58 to 64-8; D. 8; (A. 8;. teeth 5-4 or 5-5. Body elongate, little com-
pressed, little elevated, regularly curved from occiput to dorsal, highest over tip of pectoral. Head
long; mouth oblique; jaws even, the lower forming a distinct though very obtuse angle with lower
profile. Premaxillary on level with, lower half of pupil. Top of head slightly concave. Lateral
line but little decurved. Tip of depressed dorsal over front of anal. Caudal peduncle long, but
little tapering, its length from anal equal to head behind front of eye, its thickness over end of anal
equal to snout. This species differs from Rutilus bicolor in the finer scales and in having the same
number. of rays in the anal that it has in the dorsal, R. bicolor having one fewer in the anal.

148

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Agosia robusta Rutter, new species.

Body heavy, highest above insertion of pectorals; the ventral outline curved almost as much
as the dorsal. Head 3.8 to 4 in body; snout blunt, but little overlapping the premaxillary and
never extending beyond it; mouth oblique, barbels usually absent, present on 10 to 50 per cent of
specimens from any one locality. Fins small; D. 8; A. 7; pectoral about equal to head behind nostril,
variable; caudal moderately forked, middle rays two-thirds length of longest; rudimentary caudal
rays forming prominent keels along upper and lower edges of tail; margin of anal slightly rounded,
the anterior rays not all produced, not extending beyond posterior rays when fin is depressed.
Lateral line nearly always incomplete, but with scattered pores frequently extending to base of
caudal; scales 56 to 77, varying about 12 in any one locality. Usually two dusky lateral stripes, the
upper extending from snout to caudal, the lower branching off from the upper behind the head
and ending along base of anal; cheek abruptly silvery below lateral stripe; tinged with orange
about lower jaw, upper end of gill-opening, and at base of lower fins.

Type (No. 50589 U. S. Nat. Mus. ). Collected in Prosser Creek by Rutter and Atkinson.

Taken in Spring Creek, Willow Creek, Susan River, Little Truckee River, and Prosser Creek.

Coregonus williamsoni Girard.

Abundant, in streams tributary to Truckee River. A “native white-fish,” probably this species,
is reported from Bigler (Tahoe) and Donner lakes in the California Fish Commission Report for
1883-84.

Salmo henshawi (Gill & Jordan). Lake Tahoe Trout.

Occurs in only the Truckee Basin, and taken in Little Truckee River, Sage Hen Creek, and
Prosser Creek. The black spots of sides much larger and fewer than in Salmo irideus.

Salmo irideus Gibbons. Rainbow Trout.

Readily distinguished from the above by the very small and numerous black spots, as well as
by the absence of the red blotch on inner edge of mandibles. Introduced into Truckee Basin, and
possibly also into Honey Lake Basin. It was observed in Susan River and Prosser Creek.
Salvelinus fontinalis (Mitchill). Brook Trout.

This species has been introduced into Prosser Creek, where specimens were taken.

Cottus beldingii Eigemnann & Eigenmann. Blob.

Palatine teeth wanting; no prickles on skin; lateral line broken under posterior rays of dorsal,
sometimes a few pores on caudal pedunele, usually none. Top of head covered with minute
pimples. Dorsal spines 6 to 8, dorsal rays 15 to 19, anal rays 11 to 14.

Found in Susan River, Little Truckee River, Sage Hen Creek, and Prosser Creek.

BREEDING HABITS OE THE YELLOW CAT-FISH.

By HUGH M. SMITH and L. G. HARRON.

149

BREEDING HABITS OE THE YELLOW CAT-FISH.

By HUGH M. SMITH and L. G. HARRON

In view of the paucity of information in regard to the spawning habits of cat-
fishes, and owing to the possible inauguration of cat-fish culture in response to a wide-
spread demand, we think it worth while to present these observations on one of the
most important members of the family. The notes may be taken in conjunction
with Dr. Albert C. Eycleshymer’s “ Observations on the breeding habits of Ameiurus
nebulosus ,” published in the American Naturalist for November, 1901.

On duly 3, 1902, it was observed that among a lot of yellow cat-fish ( Ameiurus
nebulosus) from the Potomac River near Washington, which had been in the Fish
Commission aquarium since May 19, 1902, two had paired and exhibited breeding
tendencies. They had withdrawn to one end of the aquarium tank and maintained
themselves there, the male driving away any others which approached. The other
fish were thereupon removed and the two in question left unmolested. They were
kept under daily observation, and their behavior furnished the principal data on
which this paper is based. In the latter part of July another pair of fish in the same lot
showed an inclination to spawn and afforded additional information, as did also a lot
of eggs of the same species found in a pool in the Fish Commission grounds on June 16;
these eggs, which were about ready to hatch, were removed to an aquarium, where
two-thirds hatched the same night, the others being dead the next morning.

Nest-making. — The aquarium in which the fish were held was 5 feet long and 16
inches wide on the bottom and 18 inches high, the posterior wall inclining obliquely
backward so that at the surface of the water the tank was 2 feet I inches wide. The
front was of solid glass, and the sides, bottom, and back were of slate. The bottom
was covered with gravel and a little sand to the depth of 11 or 2 inches.

The nest-making, as modified by the artificial conditions of the aquarium, con-
sisted in removing all the stones and sand from one end and keeping the slate bottom
scrupulously clean from all foreign objects, even the smallest particles of food, sedi-
ment, etc. In moving the pebbles, which were mostly from one-half to three-fourths
of an inch in diameter, the fish took a vertical or slightly oblique position and
sucked a pebble into the mouth, usually beyond the lips and out of sight, then swam
toward the other end of the tank and dropped it by an expulsive or blowing effort.
Sometimes the gravels were carried only a few inches and sometimes the entire
length of the aquarium. Usually the fish swam horizontally near the bottom when
carrying a stone, but sometimes turned obliquely upward and dropped it from near
the surface. Both fish participated in this operation. The removal of finer sediment
was effected by a quick lateral movement of the body which caused a whirl that lifted
and floated the particles beyond the limits of the nest.

The pair of fish more particularly under consideration, during the first night they
were in the aquarium, removed all the gravel from over a space nearly 2 feet long and
li feet wide, upward of a gallon of stones being transferred as described. After

151

152

BULLETIN OF THE UNITED STATES FISH COMMISSION.

the second pair of fish had cleared a similar space, a pint or more of gravel was
scattered. on the nest; the fish immediately began to remove the stones, and in a few
minutes had completely freed the nest from gravel. The gravel- regarded by bass
and other fishes as desirable material for the bottom of nests— may be removed by the
cat-fish for two reasons: (1) To have a clean place for the eggs and young, so that they
may be better guarded and agitated as hereafter described; (2) to provide a smooth
place on which to rest and against which to rub the abdomen.

Upward of twenty years ago the yellow cat-fish was much more abundant in
the Potomac River than at present. The marshes in Piscataway Creek were a favorite
place for the fish to spawn, and large quantities were there taken each season in
spring and summer, mostly by colored people living near the river. The fish at that
time of year were found in shallow water occupying depressions in the muddy bottom,
with most or all of their bodies concealed in an excavation extending laterally from
the rounded depression. The fishermen easily made large catches by wading and
thrusting their hands into the depressions. An old colored man whom we knew used
to refer to a marsh as his “meat market,” and would often bring ashore a sackful
of yellow cat-fish caught in this way. We are inclined to believe that these fish were
brooding, but we have had no opportunity of late years to examine them critically.

Behavior of adult fish before spawning. — Two days intervened between the
beginning of the nest-making and the laying of the eggs. As soon as the nest was
made ready, the fish became very quiet. During most of the time they rested on
the bottom, with practically no body or fin movement, except at intervals. The fish
lay close together, often parallel, with their abdomens just clear of the bottom, their
weight being borne on the anal and ventral fins. At frequent intervals the female
compressed her distended abdomen against the smooth slate bottom with a quivering
or convulsive movement, the male often accompanying or following the female in
this action, which is obviously for the purpose of loosening the eggs.

The second pair remained on the nest from July 18 to September 30, when they
were removed to make room for other species, as it was evident no eggs would be laid.
During this time they behaved in the same way as the other pair and their failure to
spawn can not be positively accounted for, though such an outcome has been the
rule among fish retained in the Fish Commission aquarium. The enlargement of the
abdomen and ripening of the eggs go on to a point when spawning seems imminent;
the actions of the fish suggest the arrival of breeding time; but no eggs are laid.
After a few weeks the enlargement of the abdomen subsides, and dissection has some-
times shown a liquefaction of the egg mass. It has been suggested that the presence
of alum in the circulating water has an injurious astringent action on the mucous
membrane of the vent, and it is a significant fact that the change from an open to a
closed circulation, with consequent elimination of the alum filter, was soon followed
by the spawning of the cat-fish first mentioned — an unprecedented occurrence at
the Fish Commission aquarium. The second pair of fish had been in the alum-
filtered water for a few days, some time before the spawning season.

Number , character , and incubation of eggs. — On July 5, between 10 and 11 a. m.,
the eggs were deposited in four separate agglutinated masses on the clean slate bottom.
Unfortunately, the fish were not under observation at this time, although they were
watched for about fifteen minutes after the extrusion of the first two lots of eggs,
when it was supposed the spawning had been completed. The masses of eggs were
of nearly uniform size, about 4 inches long, 21 inches wide, and half an inch thick.
The newly laid eggs are one-eighth of an inch in diameter, nearly transparent, and of

BREEDING HABITS OF THE YELLOW CAT-FISH.

153

a pale yellow color. The number of eggs deposited was estimated at 2,000. The
incubatory period was 5 days in a mean water temperature of 77 F. , the lowest
temperature being 75° and the highest 80°. About 12 hours intervened between the
hatching of the first and last eggs. Active movement was observed in the embryos
40 hours after the eggs were laid. Fully 99 per cent of the eggs hatched into normal
fry, a few weak and deformed fry and a few unfertilized or dead eggs being noticed.

Growth of young. — When the fry first emerged from the egg they were about
one-quarter of an inch in length, and of a yellowish, transparent color. By the sec-
ond day the skin of the back had begun to darken, and by the end of the fourth day
the entire upper parts were uniformly bluish black and the under side had become
whitish. On the third day the barbels at the angles of the mouth and the pectoral
and dorsal spines were clearly visible through the glass front of the aquarium.

Until 6 days old they remained on the bottom in densely packed, wriggling
masses, the largest lot in the nest and several smaller lots in other parts of the
aquarium. On the sixth day they began to rise vertically a few inches above the
bottom, at first falling back at once, but gradually remaining longer above the bot-
tom. By the end of the seventh day they were swimming actively, and practically
all collected in a school just beneath the surface, where they remained for two days.
They then began to scatter, and subsequently did not school.

The relatively large yolk-sac had nearly disappeared by the sixth day, when
they began to eat finely ground beef liver, and they were feeding ravenously by the
eighth day. Between feeding times, they passed much of the time on the bottom of
the aquarium in search of food, which they ate in an almost vertical position, head
downward; they also browsed on the sunny side of the aquarium, where there Avas
a short growth of algae. The early growth Avas rapid, but not uniform; on the
eleventh day their length varied from one-half to three-fourths of an inch. At the
age of 2 months the average length was 2 inches; but after that time the growth was
very slight, and in January, 1903, six months after hatching, the length of the sur-
vivors was only 2| to 2^ inches. The slow growth Avas undoubtedly due to the fact
that the fry Avere retained in small troughs where the conditions were unnatural.

Care of eggs and young. — During the entire hatching period both parents were
incessant in their efforts to prevent the smothering of the eggs, to keep them clean,
and to guard against intruders. The eggs were kept constantly agitated and aerated
by a gentle fanning motion of the lower fins, and foreign particles, either on the
bottom of the nest or floating near the eggs, were removed in the mouth or by the
fins. The most striking act in the care of the eggs was the sucking of the egg masses
into the mouth and the blowing of them out, this being repeated several times Avith
each cluster before another lot was treated.

The male was particularly active in watching for intruders, and savagely attacked
the hands of the attendant who brought food, and also rushed at sticks or other
objects introduced into the aquarium. Practically the entire work of defense was
assumed by the male, although the female occasionally participated.

During the time the fry were on the bottom the attentions of the parents were
unrelaxed, and, in fact, were increased, for the tendency of the different lots to become
scattered had to be corrected, and the dense packing of the young in the corners
seemed to occasion much concern. The masses of fry were constantly stirred as the
eggs had been by a flirt of the tins, which often sent dozens of them 3 or 4 inches
upward, to fall back on the pile.

154

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The very young1 fry were also taken into the mouths of the parents and blown
out, especially those which became separated from the main lot and were found in
the sand and sediment. The old fish would take in a mouthful of fry and foreign
particles, retain them for a moment, and expel them with some force. After the
young began to swim and became scattered, the parents continued to suck them in
and mouth them, and, as subsequently developed, did not always blow them out.

An interesting habit of the parents, more especially the male, observed during
the first few days after hatching was the mixing and stirring of the masses of young
by means of the barbels. With their chin on the bottom, the old fish approached the
corners where the fry were banked, and with the barbels all directed forward and
Hexed where they touched the bottom, thoroughly agitated the mass of fry, bringing
the deepest individuals to the surface. This act was usually repeated several times
in quick succession. The care of the young may lie said to have ceased when they
began to swim freely, although both parents continued to show solicitude when the
attendant approached the aquarium from the rear.

When 12 days old, about 1,500 of the fry were removed from the aquarium
to relieve crowding, and placed in a hatching-trough such as is employed for salmon
and trout. For some unknown cause, about 1,000 of these died during the first three
days. The others survived with little or no loss, and are still on hand.

The fry which were left with their parents continued healthy, but their number
steadily decreased. There being no way for them to escape, and a closely woven wire
screen preventing inroads from the exterior, it was suspected that the old fish were eat-
ing their young, though they were liberally fed at suitable intervals. They were kept
under close observation during the day, and were seen to be fond of mouthing the
fry, more especially the weaker ones — a habit which at this stage seemed unnecessary.
They were frequently seen to follow leisurely a fry, suck it in their mouth, retain it
for a while, and then expel it, sometimes only to capture it again. There was no
active pursuit of the fry, and the tendency seemed to be to spit them out. In one or
two instances, however, it appeared that fry taken into the mouth were not liberated,
the feeding instinct becoming paramount to the parental instinct. After all the fry
which had been left with their parents had disappeared- in about 6 weeks after
hatching — 18 fry from the trough were placed in the aquarium one evening, and only
2 of these had survived on the following morning.

During the entire period covered by these observations liver and beef were fed
regularly to the brood fishes, and at no time did their appetites fail. There was
apparently no interference with deglutition, or closure of the oesophagus, such as
has been observed in some other cat-fishes, as half-inch cubes of meat were readily
ingested during the entire time the fish were under observation.

External sexual characters of adults. — Besides the fullness of abdomen which the
mass of eggs gives to the female, there was in both pairs of fish under consideration
another external feature by which the sexes could be distinguished. This was the
shape of the snout and interorbital region, which in the males were noticeably flatter
and broader than in the females. The males in both these cases were about 12 inches
long and were an inch longer than their partners.

THE DESTRUCTION OE TROUT FRY BY HYDRA.

By A. E. BEARDSLEY,

Professor of Biology , Colorado State Normal School.

THE DESTRUCTION OF TROUT FRY BY HYDRA.

By A. E. BEARDSLEY,

Professor of Biology, Colorado Stale Normal School.

The following observations were made during an investigation at the United
States Fish Commission hatchery, Leadville, Colo., in August, 1902. On August 4,
some eggs of the black-spotted trout in a number of the hatching- troughs were just
hatching, while in others the young fish were several days old. Each trough was
separated by screens into three divisions. The first division — that into which the
water enters from the supply pipes — contained no eggs, these having all been removed
several days before on account of the great mortality of the young fishes hatched in
this division of the troughs. In the second, or middle, division, the newly hatched
fry were dying in considerable numbers, some before leaving the egg trays. In the
third division of these troughs, as well as in the troughs not directly fed from the
supply pipes, the death rate was merely nominal.

These facts clearly indicated that the cause of the mortality was directly connected
with the water supply, which was found to be derived chiefly from two sources.
The main supply pipes were fed from Rock Creek, and an auxiliary set of pipes led
from a spring near the hatchery. Connected with the main pipes was a branch
leading from the third or lowest of the Evergreen lakes. This was closed at the
time, only a small quantity of water coming through leaks around the gate at the
head of the pipe.

The water from the main pipes was clear, containing very little sediment, and with
a temperature of 48° F. ; that from the spring was very clear and pure, without
sediment, its temperature being 43° F. There was very little sediment in the
hatching-troughs. In this, however, microscopical examination disclosed the pres-
ence of great numbers of a very transparent hydra, which had been discovered by
the attendants at the hatchery a few days before, when the sun’s rays, just before
sunset, had fallen obliquely into one of the troughs. In the dim light of the
hatchery this hydra was quite invisible, but by placing a large mirror outside of
the building so as to throw a beam of sunlight through the window, with a hand
mirror reflecting this beam so as to throw it into the trough, the hydras could be
plainly seen as slender, whitish threads, 1 to 2 centimeters in length and 0.15 to 0.30
millimeter in diameter, fixed by one end to the bottom or to the side of the trough,
and bearing a crown of 5 or 6 long tentacles around the mouth at the free end. The

157

158

BULLETIN OF THE UNITED STATES FISH COMMISSION.

hydras were found quite equally distributed through the first division of all troughs
supplied directly from the main pipes. A careful count of the number on several
square inches in different troughs gave an average of 131 hydras per square inch
(20+ per square centimeter). Comparatively few were found in the middle division
of the troughs, most of them having fixed themselves before reaching the first screen.
Very little animal life other than hydra was found in the sediment of the troughs.

Since no other cause for the mortality of the young fishes could be discovered,
and as the hydras were exceedingly abundant and are well known to be armed with
great numbers of dart cells or nettling cells which secrete a fluid that quickly causes
paralysis in small crustaceans and other minute forms of animal life, it appeared
that the injury was probably due to the hydras. In so far as the writer is aware, no
injury to fishes by hydra has heretofore been known. The following experiment
was therefore instituted to determine what injury, if any, was to be attributed to
this cause:

Five beakers, each of 250 cubic centimeters capacity, were filled with water from
the supply pipes; in each of the first four of these was placed the sediment from 21
square centimeters of the bottom of the first division of one of the hatching-troughs,
containing about 430 hydras; the fifth beaker was intended as a control, and contained
water only. Five trout newly hatched and apparently in good health were taken
from the hatching-trays and placed in each beaker. Nos. 1 and 2 were filled with
water from the spring and were placed in running water, so that the temperature was
nearly constant; Nos. 3, 4, and 5 were filled from the main supply pipes, No. 4 having-
been kept over night in the office, and all three were set on a shelf in the hatching-
room. At the end of the experiment, Nos. 1 and 2 were at nearly the same tem-
perature as at the beginning, while Nos. 3, 4, and 5 had acquired the temperature of
the hatching-room.

The following table shows the result of this experiment:

1.

2.

3.

4.

5.

Temperature at beginning of experi-
ment.

Hour of beginning

43° F

43° F

48° F

5S° F

48° F.

9.23 a. m.
0 dead.

0 dead.

0 dead.

0 dead.
55° F.

9.13 a. m

9.16 a. m

1 dead

9.20 a. m

2 nearly dead.

3 dead a

1 dead

4 dead

4 dead

4 dead

4 dead

5 dead

5 dead

Temperature at end of experiment. .

44° F

44c F

55° F

55° F

a One of these had burst the yolk-sac in its struggles.

In this experiment 25 per cent of the trout were killed by hydras in less than 30
minutes, 60 per cent in 45 -minutes, 80 per cent in 60 minutes, and 100 per cent in 75
minutes; those trout which were least active in the beginning of the experiment were
the ones that survived longest, probably because they came in contact with a smaller
number of stinging ceils of the hydra. With the aid of a lens, the hydra| could be
seen with their mouths closely applied to the surface of the fish, particularly on the
yolk-sac; in some cases more than a dozen hydras were seen attached to a single fish.
Soon after the fishes were placed in the beakers most of them were seen to struggle
violently, one of them bursting its yolk-sac in its struggles and dying immediately;
these struggles recurred at intervals, but with diminishing intensity, until death

THE DESTRUCTION OF TROUT FRY BY HYDRA.

159

supervened. The five trout in the beaker without hydras were kept in the beaker
until the next day and were then found to be all alive and in good health.

No other cause of injury having been discovered after the most careful search,
the destructive effects of the hydras upon the fishes in the foregoing experiment were
taken as conclusive evidence that these were the cause of the unusual mortality of the
trout fry. This fact being demonstrated, a careful examination of all the sources of
water supply to the hatchery was made. The lower of the three lakes was first
visited. This lake is quite shallow, being about 12 feet (1 meters) in depth in the
deepest part; along the borders there is considerable aquatic vegetation, consisting
of sedges and cat-tails; here the hydras were found in immense numbers, clinging to
the submerged stems and leaves as well as to the green filamentous alga which was
growing abundantly on the bowlders which are scattered over the bottom. The other
lakes, and Rock Creek for a distance of about half a mile above the head of the supply
pipe, as well as the spring, were examined in turn, but although very careful search
was made, no hydras were found in either of these waters.

The temperature of the water in each of the three lakes at 1 foot below the
surface was taken August 8, about 2 p. m., and was found to be as follows: Upper
Evergreen Lake, 60° E. ; Middle Evergreen Lake, 64°; Lower Evergreen Lake, 65°.

The leaks about the head of the pipe leading from the lower lake were immediately
stopped and no water from the lake is now entering the hatchery.

The natural causes which control the development of the different species of hydra,
favoring or retarding it, are as yet but little understood. At one period hydras may
be very abundant at a given point, and soon afterwards entirely disappear without any
apparent cause. They have been found in the vicinity of Greeley, Colo., during all
months of the }'ear, sometimes in great abundance; sometimes, however, a whole year
has passed without a single one being seen, although searched for most diligently.
They occur in lakes, ponds, and marshes, usually in clear water. Warm water (60°
to 80° F.) appears to favor their rapid multiplication, since they are usually most
abundant in summer and earl}' autumn; cold water does not apparently injure them,
however, as the writer has frequently taken vigorous individuals in the winter,
through holes in the ice. Hydras reproduce at certain times by eggs, which settle
to the bottom and probably remain dormant through the winter, but the usual and
most rapid mode of multiplication is by budding. Little buds arise from the side of
the parent, soon acquire a mouth and tentacles like the parent, and after a time break
loose and lead an independent existence. In the lake most of the hydras examined
were bearing from two to six buds, showing that the conditions there were favorable
to their rapid multiplication. In the hatchery troughs, on the contrary, very few
were found bearing buds, and these were probably recent arrivals. The conditions
within the hatchery do not, therefore, appear favorable for their increase, and it
only remains to rid the troughs of them in the most practicable manner.

As the hydra is very tenacious of life and may even be cut into several pieces
without serious injury, each piece developing the lost parts and becoming, in a few
days, a complete hydra, it is not probable that it can be destroyed in the troughs
without injury to the fish eggs or young fry. By removing all eggs and fry, briskly
scrubbing the bottom and sides of the trough with a stiff brush so as to cause the
hydras to loosen their hold, then quickly flushing the trough into the waste-pipe,
most of them can be removed.

160

BULLETIN OF THE UNITED STATES FISH COMMISSION.

When the lower divisions of the trough contain hatching eggs and fry that can
not readily be removed without injury, the first or upper division (in which nearly
all the hydras will have fixed themselves) may be cleaned by shutting off the supply
pipes, placing a temporary partition between the upper and lower divisions, and,
after a brisk scrubbing, quickly siphoning off the water and floating hydras from the
upper division of the trough.

As the water now entering the hatchery is taken from Rock Creek and from the
spring, both of which are free from hydras, it is probable that loss from this cause
will cease as soon as all the hjulras now in the hatching troughs and supply pipes can
be removed.

It was impossible to find characters other than those of color and size by which
to differentiate this hydra specifically from the well-known II. fusca Linnaeus. It
differs from that common form only in being of larger size and in the entire absence
of coloration. Among the large number of individuals observed, both in the troughs
of the hatchery and at the lake, not one showed a trace of fuscous coloration. These
differences appear to be constant, and I propose the n^mo, pallida for the new species,
in allusion to its lack of color. It may be described as follows:

Hydra pallida Beardsley, new species.

Characters. — Body cylindrical, 1 to 2 cm. in length and .15 to .30 mm. in diameter; tentacles 5 or 6,
when fully extended two or three times as long as the body; color, milk-white in reflected light,
whitish and translucent in transmitted light.

Differs from typical Hydra fusca in being somewhat larger in average size and in the entire absence
of fuscous coloration.

Type locality. — United States fish-cultural station, Leaclville, Colorado.

DESCRIPTIONS

TV GENERA AND SPECIES OF FISHES FROM
THE HAWAIIAN ISLANDS.

By DAVID STARR JORDAN AND BARTON WARREN EVERMANN.

F. C. B. 1902—11

161

DESCRIPTIONS OF NEW GENERA AND SPECIES OF FISHES FROM THE

HAWAIIAN ISLANDS.

By DAVID STARR JORDAN and BARTON WARREN EVERMANN.

During the investigations of the aquatic resources of the Hawaiian Islands carried
on by us in 1901 under the direction of the Hon. George M. Bowers, United States
Commissioner of Fish and Fisheries, very large collections of the fishes and other
animals occurring in the waters of those islands were made.

A detailed report, covering the entire aquatic fauna of that group of islands, is
now in preparation, which, it is hoped, will be ready for publication within the year.

Among the fishes collected are many species which appear to be new. Descrip-
tions of 57 of these are given in the present paper. Illustrations of these new
species, together with more extended notes regarding their abundance, distribution,
habits, and commercial value, will be given in the general report to follow.

The types of all the new species have been deposited in the United States National
Museum, and, when possible, one or more cotypes have been donated to each of the
following museums and institutions: Museum of Leland Stanford Junior University
(L. S. Jr. Univ. Mus.), U. S. National Museum (U. S. N. M.), Reserve series of the
U. S. Fish Commission (U. S. F. C.), Museum of Comparative Zoology at Cambridge,
Mass. (M. C. Z.), American Museum of Natural History, New York City (Am. Mus.
Nat. Hist.), Academy of Natural Sciences of Philadelphia (Ac. Nat. Sci. Phila.),
University of Indiana (Mus. Ind. Univ.), Field Columbian Museum, Chicago (Field
Col. Mus.), California Academy of Sciences, San Francisco (Cal. Ac. Sci.), and the
Bernice Pauahi Bishop Museum at Honolulu (Bishop Mus.). When possible, we have
given in this paper the numbers which the types and cotypes bear on the records of
the various museums to which they have been assigned.

The majority of specimens here described were obtained by us in the market or
directly from the fishermen at Honolulu, Oahu Island. Others were obtained in the
market or from the fishermen at Hilo, island of Hawaii; others at Kailua, island
of Hawaii; others on the reef at Waikiki, near Honolulu, and one at Heeia, Oahu
Island.

Family CARCHARIIDA. The Sand Sharks.

1. Carclxarias phorcys Jordan & Evermann, new species.

Head 4.8 in length; depth 6.5; width of head 1.75 in its length; depth of head 1.8; snout about
2.2 in head; interorbital space 2.2; space between tip of snout and front of mouth 2.5; width of
mouth 2.5; eye 6 in interorbital space; internasal space 1.8; least depth of caudal peduncle a little
’over 4.8; caudal 3.5 in body; pectoral 5.75.

Body elongate, rather robust, the tail compressed; head elongate, somewhat narrow and
depressed; snout long and narrowly pointed when viewed above, the tip rounded; eyes small, their

163

164

BULLETIN OF THE UNITED STATES FISH COMMISSION.

posterior margins about midway between tip of snout and first gill-opening; nictitating membrane
well developed; mouth large, very convex, the anterior margin of mandible below front rim of orbit;
teeth in upper jaw narrow, with broad bases, compressed, serrate, not notched, and with 4 or 5 basal
cusps behind; teeth in mandible rather long, pointed, and not serrate, the edges smooth; nostrils
without flap, inferior, and nearer eye than tip of snout; interorbital space broad and convex, the
upper profile of the head rising gradually in a nearly straight line to back of head; gill-openings of
moderate length, the posterior over the base of the pectoral; peritoneum white or pale; body very
finely roughened when stroked forward; height of first dorsal less than depth of body, its origin a
little nearer tip of snout than origin of second dorsal; origin of second dorsal nearer origin of first
dorsal than tip of caudal, the fin small and about over the anal so that the origins of the two fins are
opposite; caudal long, with a notch at its tip, deep, the lower lobe 2.25 in the length of the fin;
pectoral with margin of fin slightly concave; ventrals small, their origins a little nearer base of lower
caudal lobe than origin of pectoral; back convexly ridged, broader between the dorsals; base of caudal
with a pit above and below.

Color in alcohol, pale brown, the lower parts pale or whitish with a brown streak the color of the
back along side from gill-opening to origin of ventral; tips of dorsals, edge of caudal, and tip of pec-
toral blackish.

This description is based upon the type, No. 50612, IT. S. N. M. (field No. 03747), a specimen 27.5
inches long, obtained by us at Honolulu. The collection contains 4 other examples, all from Hono-
lulu, which we take as cotypes. They are: No. 12715, L. S. Jr. Univ. Mus. (field No. 03745); No.
12715, L. S. Jr. Univ. Mus. (field No. 03746); No. — , M. C. Z. (field No. 03748); and No. 1685, Bishop
Mus. (field No. 03749).

We have examined 2 other specimens, each about 29 inches long, obtained by Dr. Jenkins in 1889,
and one foetus obtained by us at Honolulu.

Family OPHICHTHYIDTE. The Snake Eels.

2. Microdonophis fowleri Jordan & Evermann, new species.

Head about 5 in trunk measured from tip of snout to vent; tail shorter than head and trunk by
the length of the former; eye nearly 1.6 in snout or 1.5 in interorbital space; snout 6 in head; inter-
orbital space about 6.75; mouth 2.75; pectoral a little over 4.25 in head.

Body elongate, cylindrical, the tail tapering gradually to a conical horny point; head cylindrical
and pointed; snout moderately long and pointed, slightly flattened above, projecting over and beyond
the mandible; eye elongate, small, anterior and superior, about midway in length of mouth; mouth
lather large; lips somewhat fringed; teeth large and canine-like in front of jaws, and on vomer in a
single row; tongue small, adnate to floor of mouth; anterior nostrils in short tubes near tip of snout,
the posterior with broad flaps on the lips and opening downward; interorbital space concave, each
supraocular ridge slightly elevated; peritoneum silvery; skin perfectly smooth; head with mucous
pores, a series of which encircle head above and about midway in its length; lateral line well developed,
pores about 140; origin of dorsal slightly in advance of gill-opening or base of pectoral; pectoral small,
the rays just above the middle the longest, fin rounded; dorsal fin long and low, its height about equal
to length of snout; anal similar to dorsal, its height a trifle less.

General color, when fresh, white, rendered somewhat shaded on upper portions by very minute
points (seen only with a good lens) of gray; back and upper surface with numerous round brown spots
and about 17 indistinct transverse dark brown crossbands which do not extend over the dorsal; the
interspaces between the spots on the head yellow, the pectoral bright lemon yellow; end of tail for
about 1 inch from point bright yellow; spots on margin of dorsal brown, with yellow borders; a band
of yellow runs from under one eye backward, upward, across the top of head, and down under the
other eye; the transverse series of pores which encircles the head above and about midway in its length,
with black margins, and also a similar series over head along the margin of mouth, and then up, back
of eye, over head; pores of lateral line without black margins.

This species is based upon a single specimen, the type, No. 50613, U. S. N. M. (field No. 03431),
an example 23 inches long obtained by us in the market at Honolulu, July 21.

FISHES FROM THE HAWAIIAN ISLANDS.

165

Family MUR/ENID7E. The Morays.

3. Mursena kailuse Jordan & Evermann, new species.

Head 7 in total length; depth 11.5; eye 14 in head; snout 6; interorbital 12; gape 2.75.

Body short, stout, and moderately compressed; distance from tip of snout to vent less than that from
vent to tip of tail by a distance equal to two-thirds length of head; head very small and pointed; snout
long, quadrate, the jaws equal, the lower curved so that the mouth does not completely close; lips
thin, the teeth showing; each side of upper jaw with a single series of unequal, sharpish canine-like
teeth, inside of which is a single depressible fang-like tooth near middle of side; front of median line
with 2 long, sharp, fang-like, depressible teeth; shaft of vomer with a single series of short, movable
teeth; each side of lower jaw with a single series of unequal, sharp canines, those in front largest; eye
small, midway between angle of mouth and tip of snout; anterior nostrils each in a pointed filament,
whose length is about half that of eye, situated at tip of snout just above lip; posterior nostrils each
with a long filament, equal to snout in length, and situated just above anterior edge of eye; inter-
orbital space very narrow and flat; gill-opening small, nearly circular; dorsal fin very low anteriorly,
increasing much in height on tail; anal low.

Ground color in life, dark brown, with fine yellow and blackish spots and reticulating lines, the
yellow predominating on anterior part of body; end of tail dark purplish brown; edge of dor'sal and
anal dull dark red, with short pale bands bordered with darker and with small pale spots interspersed;
ground color of cheek and throat yellow, with pale spots bordered with black; jaw orange red, with
pale black-edged bars; tips of jaws bright coral red; tips of nostril filaments bright red.

Color in alcohol, body with a ground color of light grayish brown, marked with fine whitish lines
or specks, and profusely covered with numerous small, round, white spots, each ocellated with black;
among these are scattered larger black spots and blotches; white spots smallest on back and largest on
belly, where some are as large as eye; a broad, dark brown bar over nape, extending on side to level of
eye; top of head and snout with fine white spots; side of snout with a well-defined vertical white bar
about midway between eye and tip; a short white line downward to mouth from front of eye, and a
similar longer one downward and backward from posterior lower angle of eye; lower jaw crossed by
3 V-shaped white bars opening forward and bordered by darker; tip of jaw with 2 oblique white bars
separated by a narrow brown line; last V-shaped white bar extending across angle of mouth and form-
ing a large white area at base of upper jaw, behind which the angle of the mouth is dark brown; inside
of mouth mottled brown and white; nasal filaments mottled with brown and white; throat light brown,
with large white spots, some of which unite to form oblong spots or lines; gill-opening not surrounded
by dark; anal fin dark brown, crossed by about 28 short white bars; posterior portion of tail crossed
by about 12 distinct but somewhat irregular vertical white bars, which extend upon dorsal and anal
fins; tip of tail brownish black, with 1 or 2 whitish specks. Only one specimen known.

Type, No. 50614, IT. S. N. M. (field No. 03709), a specimen 19 inches long, obtained August 9,
1901, by Messrs. Goldsborough and Sindo at Kailua, Hawaii.

4. Gymnothorax vinolentus Jordan & Evermann, new species.

Head 7.2 in total length, 3.6 in distance from tip of snout to vent; depth 14.5 in total length; eye
14 in head; snout 6.4; gape 2; interorbital 8.6; vent a little nearer tip of snout than tip of tail.

Body long, but stout and not greatly compressed; tail moderately stout and compressed; head
much swollen above; snout long and slender, the anterior profile ascending somewhat abruptly from
interorbital region; mouth large, extending beyond eye a distance equal to eye and snout; lower jaw
projecting, strongly curved, so that the mouth does not completely close; eye small, over anterior half
of gape; interorbital narrow, about half greater than diameter of orbit; anterior nostril in a tube whose
length is 1.6 times eye, situated near tip of snout; posterior nostril slightly anterior to vertical at front
of orbit, oval, surrounded by a narrow, raised, flattened flap whose diameter is two-thirds that of
orbit; lips rather thin, not covering the teeth; gill-opening small, its length less than diameter of
orbit. Teeth in a single series on each side of upper jaw, the posterior ones short, sharp, and close-
set; the anterior ones, about 12 in number, slender, sharp canines of unequal length; inside of these is
a series of 5 or 6 long, slender, depressible canines; median line of roof of mouth with 2 long, sharp,
depressible canines in front, and a third somewhat farther back; vomer with a single series of short,
blunt teeth; lower jaw with a single series of rather close-set, short, backwardly directed canines,
somewhat compressed, inside of which anteriorly are 3 or 4 much longer depressible canines on each

166

BULLETIN OF THE UNITED STATES FISH COMMISSION.

side. Origin of dorsal midway between gill-opening and angle of mouth; height of dorsal 2 in
distance from tip of snout to posterior edge of orbit; anal much lower than dorsal.

Color in alcohol, rich purplish brown or wine-color, almost uniform over entire body and head;
side of head with about 7 shallow longitudinal grooves which are darker than ground-color; under side
of lower jaw yellowish white, blotched with brown; throat blotched with yellowish white and brown;
gill-opening rather paler than surrounding parts; body everywhere with numerous, but very obscure,
dark points, posteriorly with numerous narrow vertical dark lines appearing as shallow grooves in the
skin; dorsal and anal fins uniform dark brown, not white-edged; tip of tail not white.

The only specimen of this species which we have is the type, No. 50615, U. S. N. M. (field
No. 03726), 29 inches long, obtained by Messrs. Goldsborough and Sindo, at Kailua, Hawaii.

5. Gymnothorax steindachneri Jordan & Evermann, new species.

Head 7.3 in length; depth 9.5; eye 9.5 in head; snout 5; interorbital 7.2; gape 2; distance from
from tip of snout to vent less than distance from vent to tip of tail by more than half length of head.

Body moderately long and slender, much compressed; head small; snout small and pointed, the
anterior dorsal profile concave above the eyes; the nape and sides of head much swollen; gape long,
extending far behind eye; lower jaw shorter than the upper, curved so that the mouth does not quite
completely close; lips moderately thick, entirely covering the teeth in the closed mouth; eye small,
about midway between tip of snout and angle of mouth; teeth on sides of upper jaw in a single series,
rather close-set, short, compressed, triangular canines, those in front scarcely enlarged; vomer with a
single row of bluntly rounded teeth; each side of lower jaw with a single series of rather strong, back-
wardly directed canines, the anterior ones somewhat enlarged, those on tip of jaw movable'; anterior
nostril in a long tube, its length about half diameter of eye, situated near tip of snout just above lip;
posterior nostril without tube, just above anterior edge of eye; pores on sides of jaws inconspicuous.
Origin of dorsal fin about midway between gill-opening and angle of mouth, its height about equal to
length of snout; anal similar to soft dorsal, but much lower; tail moderately slender and pointed; aseries
of inconspicuous pores along middle of side; gill-opening a long oval slit exceeding diameter of orbit.

Color in alcohol, pale brown or w'hitish, sprinkled with ragged or dendritic brown spots formed
more or less into irregular vertical blotches or crossbands; margins of fins narrowly creamy w'hite or
yellowish, that of the anal much wider; corner of mouth and space about gill-opening deep blackish-
brown; about 5 longitudinal blackish-brown grooves on lower side of head; under side of lower jaw
with 2 blackish longitudinal lines which meet at an acute angle under chin; throat and belly creamy
white, with few scattered brownish markings; sides and top of head whitish, with small, sparingly
scattered, irregular brownish spots most numerous around and between the eyes.

This species is related to G. kidako (Schlegel), from which it differs much in coloration, the
present species being much paler and less reticulated, the angle of the mouth with more black, the
gill -opening being surrounded by a broad black area (nearly or quite absent in kidako) , and in having
the w'hite border to the dorsal fin much more distinct.

This species is known only from Laysan (whence Dr. Steindaehner had 2 examples) and from
Honolulu, w'here the Albatross obtained 1 specimen in 1891 and the Fish Commission 3 examples in 1901.

The specimens from Laysan which Dr. Steindaehner identified w'ith Murxna flavomarginata
Riippell, and of which he gives a good figure, evidently belong to this species. As suspected by Dr.
Steindaehner, the species is quite different from G.flavomarginatus, of which species we have examined
several specimens from Pedang, on the west coast of Sumatra. The present species is therefore known
from the 2 examples which Dr. Steindaehner had from Laysan, one specimen obtained by the Albatross
at Honolulu in 1891, and 3 specimens secured by us at Honolulu in 1901.

Field No.

Length.

Locality.

Final disposition of specimen.

03775

04904

04905
1318

Inches.

24

14

8

17

Honolulu

do

do

Honolulu (Albatross)

Type, No. 50016, TJ. S. N. M.

Cotype, No. 7447, L. S. Jr. Univ. Mus.
Cotype, No. 2697, U. S. F. C.

Murxna flavomarginata var., Steindaehner, Dents. At. Wiss. Wien, lxx, 1900, 514, pi. vi, fig. 3 (Laysan); not of Ruppell.

FISHES FROM THE HAWAIIAN ISLANDS.

167

6. Gymnothorax g-oldsboroughi Jordan & Evermann, new species.

Head nearly 3 in trunk (exclusive of head and tail), or 9 in total length; head and trunk about
1.5 in tail; eye 1.75 in snout, 1.2 in interorbital space; snout 5 in head; interorbital space 7.5; mouth 2.

Body rather compressed, the tail gradually tapering narrowly behind; head compressed, swollen
above; snout pointed, the tip blunt and the sides compressed; eye rather small, a trifle nearer tip
of snout than corner of mouth; mouth large, snout slightly projecting beyond mandible; lips rather
fleshy and concealing the teeth when the mouth is closed; teeth in a single series in jaws, anteriorly
large and canine-like, and the vomer with a single large, depressible fang; anterior nostrils at tip of
snout in small tubes; posterior nostrils directly above eye in front; interorbital space more or less
flattened like top of snout; gill-opening about equal to eye; skin smooth; head with a number of
mucous pores; origin of dorsal a little nearer corner of mouth than gill-opening; caudal small.

Color in alcohol, brown, covered all over body except anal fin with round or roundish white spots,
those on anterior part of body small, very small and numerous on head, becoming larger on trunk,
and finally increasing very much in size on tail where they are scattered and rather far apart; reticu-
lations around the light spots blackish brown upon posterior part of dorsal fin, same color as base of
anal; margins of anal and dorsal fins whitish; gill-opening and anus bordered with blackish brown.
General color of body in life, brown, rather pale olivaceous anteriorly, and covered all over with small
white spots wrhich are close-set and small on head where the dark color forms a network; spots sparse
and irregular on posterior parts, and also much larger; vent and gill-opening dusky; dorsal colored
like the body, with a broad white edge, growing broader behind; anal dark brown, unspotted, and
with a broad pale border.

This species is known only from the type, No. 50617, U. S. N. M. (field No. 03392), a specimen
21 inches long, obtained by us at Honolulu.

7. Gymnothorax hilonis Jordan & Evermann, new species.

Head 8.2 in length; depth 16; eye 7 in head; snout 6; interorbital 6; gape 2.4; distance from
tip of snout to vent 1.2 in distance from vent to tip of tail.

Body rather short, moderately compressed, the tail more compressed and bluntly pointed;
head short, the nape swollen; interorbital space broad; a distinct median groove from near the tip
of snout to origin of dorsal; angle of mouth posterior to eye a distance equal to eye’s diameter;
lower jaw but slightly curved, shorter than the upper; front of upper jaw with 3 short, bluntly
pointed, movable teeth; side of upper jaw with a single series of short, pointed canines directed
backward; shaft of vomer with short, blunt teeth; lower jaw on each side with a single series of
rather long, pointed canines, longest in front and curved backward; anterior nostril in a long tube,
about 2 in eye, near tip of snout just above lip; posterior nostril small, round, without tube,
situated just above anterior part of eye; gill-opening small, its direction obliquely forward toward
nape; a series of 4 pores on each side of upper jaw; similar pores on lower jaw. Origin of dorsal
fin on nape midway between gill-opening and middle of eye; dorsal fin well developed, its greatest
height somewhat exceeding length of snout; anal similar to dorsal, but lower.

Color in alcohol, rich, velvety black above, paler below where it is marbled and reticulated with
narrow white lines; series of pores on side of upper jaw and those on tip of lower, white; cheek with
a few irregular white spots; gill-opening whitish ; side of body anteriorly with some small white specks
and irregular whitish markings; lower jaw with larger, oblong, white cross-lines; dorsal fin rich
brownish black, the edge posteriorly with a narrow, irregular, white border, sometimes interrupted
by black; anal brown, with a narrow white edge from which extend narrow intrusions of white, some
reaching base of fin; end of tail with a few small white spots, the tip narrowly white.

The only known example of this species is the type, No. 50618, IT. S. N. M. (field No. 04902), a
specimen 9.5 inches long, obtained by us at Hilo, Hawaii.

8. Echidna zonophaea Jordan & Evermann, new species.

Head 3 in trunk, or 6.5 in total; tail longer than head and trunk by a little more than the snout;
eye 2 in snout, 1.5 in interorbital space; snout 6; interorbital space 7.75; mouth 2.8.

Body compressed, the tail tapering rather narrowly posteriorly; head deep and compressed,
pointed in front; snout rather long and pointed, the tip obtusely rounded and projecting consid-
erably beyond the mandible; eye rather small, midway between tip of mandible and corner of

168

BULLETIN OF THE UNITED STATES FISH COMMISSION.

mouth; mandible shutting completely, arched below so that only the anterior teeth touch the
front of the jaw above, though the thick fleshy lips conceal them all; teeth molar, those in front
of jaws pointed; anterior nostrils in short tubes, the posterior pair above the eye with a slightly
elevated margin; interorbital space convex; top of head more or less swollen or convex in profile;
gill-opening 1.67 in eye; skin smooth; head with a few pores; origin of dorsal beginning at last
fourth of space between corner of mouth and gill-opening; caudal small.

Color in alcohol, grayish white, the body and tail crossed by about 25 broad rich brown bands,
extending upon the dorsal and anal fins; dark bands anteriorly broadest above and not meeting across
belly, their width about equal to the distance from tip of snout to middle of eye; first brown band
through eye, second across nape, the fourth across gill-opening; gray bands of ground color anteri-
orly broad, and widening much upon belly; posteriorly the gray bands are narrower and better
defined, especially on the fins, their width scarcely greater than half that of the brown bands; tip of tail
very narrowly white; body anteriorly, especially within the gray bands, profusely covered with numer-
ous small, roundish, black specks, less numerous and more scattered posteriorly; no black spots on head;
angle of mouth black, with a small white blotch immediately in front on lowrer jaw, continued across
under jaw as a broad whitish band; side of head with about 4 or 5 narrow blackish lines between mouth
and gill-opening; region of gill-opening marbled with dark brown and whitish, the opening dark.

One example (No. 03545) had much yellow on the head and between the brown zones.

This species is known from the type and 3 cotypes, all obtained by us at Honolulu.

Field

No.

Length.

Local it y.

Final disposition of specimen.

04899

Inches.

21

Honolulu

Type, No. 50621, U. S. N. M. AJ >
Cotype, No. 2698,. U. S. F. C.

03361

it

03545

17

do

Cotype, No. 7448, L. S. Jr. Univ. Mils.

04900

15

do

Cotype, No. 3965, Field Col. Mus.

Family MYCTOPHID/E. The Tar. tern -fishes.

9. Bbinoscopelus oceanicus Jordan & Evermann, new species.

Head 3.5 in length; depth 4.1; eye 2.5 in head; snout very short, about 6; interorbital 3.5;
D. about 12; A. about 18; scales 2-35-3.

Body strongly compressed, particularly posteriorly, where it tapers into the long, slender caudal
peduncle; head exceeding depth of body; mouth large, somewhat oblique, the jaws equal, the max-
illary reaching beyond the orbit, its posterior end club-shaped; eye large;' anterior profile rather
evenly convex from tip of snout to nape; teeth difficult to make out, but a single row of minute ones
can be seen on the edge of each jaw, the exterior granular or short, villiform stripe, if it exists, being
invisible even with the aid of a good lens; teeth on vomer and edges of palatines more distinct than
those on jaws, and forming a broader line as if there were 2 or 3 rows; no granular patches visible on
disk of palatine bone; an elevated acute mesial line separating one nasal prominence from the other;
interorbital space convex, rounded; preopercle nearly vertical, sloping slightly backward from above
downward; scales large, undulated and very irregularly and sparingly toothed or crenate, and having
about 3 basal furrows; scales of lateral line conspicuous and more persistent; 7 photophores along base
of anal, 5 along lower edge of caudal peduncle, 2 at base of caudal, 1 on middle of side above last anal
photophore, 4 on each side of belly between ventrals and origin of anal fin, 5 between base of ventral
and gill-opening, 1 on side above base of ventral, a row of 3 upward and backward from front of anal,
1 above and 1 below base of pectoral, and 1 on lower anterior portion of opercle; origin of dorsal
somewhat behind base of ventrals, the posterior rays, together with those of anal, divided to the base;
no spine at base of caudal.

Color in alcohol, uniform brownish, the scales, especially on middle of side, metallic steel blue;
top of head brownish; side of head bluish; photophores black with silvery center; fins dusky whitish.

This species was recorded by Fowler from “near the Sandwich Islands,” as Rhinoscopelus corus-
cans (Richardson), the record being based upon 4 specimens (Nos. 7972 to 7975) collected by Dr.W. H.
Jones, and now in the Philadelphia Academy. During the Agassiz South Pacific expedition of the
Albatross in 1899-1900, 2 examples of this species were taken in the surface towing net at 8 p. m., Sep-
tember 8, 1899, at latitude 10° 57' N., longitude 137° 3-r/ W., southeast of the Hawaiian Islands. These

FISHES FROM THE HAWAIIAN ISLANDS.

169

2 specimens are doubtless identical with those recorded by Mr. Fowler, and are apparently distinct
from li. corusccms, the tyjje of which came from between St. Helena and Ascension Islands and others
from between Australia and New Zealand. They are near li. an dr ex Liitken, from which they seem
to differ in the blunter snout, the more slender tail, and in having the posterolateral photophore
somewhat before the adipose fin.

Type, No. 50622, U. S. N. M. (field No. 05805), 1.3 inches long, collected by the Albatross at
8 p. m., September 8, 1899, at the surface at 137° 35 / W., 10° 57' N. ; cotype, No. 2736, U. S. F. C., same
size, collected at same time and place.

Bhinoscopelus corusccms, Fowler, Proc. Ac. Nat. Sci. Phila. 1900, 498 (near the Sandwich Islands); not of Richardson.

Family SYNGNATHIDTE. The Pipe-fishes and Sea-horses.

10. Hippocampus fisheri Jordan & Evermann, new species.

Eye 2.8 in snout; snout 2 in head; D. 18, on 4 rings; A. 4; P. 15; rings 12 + 34.

Tail longer than head and trunk; trunk rather deep, compressed, its width 1.7 in depth; eye
small, equal to interorbital width; interorbital space concave; gill-opening small, high; spines on head
and body rather high, sharp; 2 rings on trunk between each pair of larger spines; tail with 3 rings
between each pair of larger spines; coronet well developed, with 5 spines; spines over eye blunt; base
of dorsal about equal to snout; anal small, long; pectoral broad, rays rather long.

Color in life, trunk below middle row of rings yellowish golden, above middle row blackish brown
on orange ground; knobs orange; lower portion of knobs on 8 to 11 rings spotted with dark brown;
side and top of tail same as back of trunk; ventral side pale dirty orange; head, crown and snout dirty
dark brown; an orange band across snout and one before eyes; pale brownish golden over gills; chin
orange; iris yellowish golden with 8 reddish streaks radiating from pupil; fins. pale; a red spot before
each eye at each side of preorbital spine.

Color in alcohol, pale brown, upper surface with dark brown marblings; side with small roundish
dark spots.

The above, description is from the type, No. 50625, U. S. N. M. (field No. 03835), a specimen 2.6
inches long, obtained at Kailua, Hawaii, where the species was new to the natives. We have 5 other
examples, each about 3 inches long, taken from the stomach of a dolphin (CorypJuma sp.) which
was captured at Hilo, July 18, 1901.

When fresh, No. 03507. a male, was pink or pale cardinal along and near the keels; plates on back
and above middle row of knobs on side mottled blackish on pale red ground; plates belowr middle row
of knobs and on belly porcelain wdiit.e; egg-pouch uniform pale cardinal-red, paler than rest of body;
tail same pink or pale cardinal, mottled with blackish blotches; top of head and crown blackish on
pale red; cheek, jaw, and snout pink. Some examples had ventral side of tail and portion behind
fourth prominent spine of tail uniform pale cardinal-red.

This species is named for Mr. Walter Y. Fisher, of Stanford University.

We have the following specimens:

Field

No.

Length.

Locality.

Final disposition of specimen.

03835

03507

Indies.
2. 6

Kailua

Hilo

do ....

Type, No. 50625, U. S. N. M.

Cotype, No. 7450, L. S. .Tr. Univ. Ivins.
Cotype, No. 2700, U. S. F. C.

Cotype, No. 3946, Field Col. Mus.
Cotype, No. 1687, Bishop Mus.

Cotype, No. , M. C. Z.

do ....

do

11. Hippocampus hilonis Jordan & Evermann, new species.

Eye about 4 in snout; snout 2 in head; D. 16, on 3 rings; rings 12 + 35. Tail a little longer than
head and trunk; trunk rather deep, compressed, its width 2 in depth; eye small, equal to interorbital
width, which is concave, broader posteriorly; gill-opening high, rather large; spines on head and body
very blunt, rounded, or obsolete, though forming knobs of more or less equal size along tail; coronet
with rounded knobs, before which is a short keel or trenchant ridge; base of dorsal about 1.35 in snout.

Color in alcohol, dark or blackish brown, more or less uniform.

170

BULLETIN OF THE UNITED STATES FISH COMMISSION.

This species is known to us only from the example described above. It is closely related to the
Japanese Hippocampus aterrimus Jordan & Snyder, but on comparison with the type of that species
was found to differ, in the presence of the keel on the top of the head and in other minor characters.
It is also close to II. ringens.

Type, No. 50626, U. S. N. M. (field No. 03832), a specimen 6 inches long, presented to us by Mr.
A. M. Wilson, of Hilo, Hawaii, where he obtained the specimen.

Family ATHERIN1DAE The Silversides.

12. Atherina insularum Jordan & Evermann, new species.

Head 4 in length; depth 4.75; eye 3 in head; snout 4; interorbital 2.8; maxillary 2.5; mandible 2.2;
D. vi-i, 11; A. 17; scales 46, 6 rows from anterior base of anal upward and forward to spinous dorsal.

Body oblong, compressed; head triangular, the sides compressed, top flat; mouth large, oblique,
maxillary reaching front of pupil, lower jaw included; teeth in rather broad villiform bands on jaws,
vomer, and palatines; interorbital space very broad and flat; snout broad, truncate; origin of spinous
dorsal slightly posterior to vertical at vent, slightly nearer tip of snout than base of caudal; longest
dorsal spine about 2.4 in head, reaching nearly to vertical at front of anal; distance between spinous
and soft dorsals equal to distance from tip of snout to middle of pupil; edge of soft dorsal concave,
anterior rays somewhat produced, their length 1.9 in head; last dorsal ray about one-half longer than
one preceding; base of soft dorsal 1.8 in head; origin of anal considerably in advance of that of soft
dorsal, the fins similar, anterior rays about 1.7 in head, base of anal 1.3 in head; caudal widely forked,
the lobes equal; ventral short, barely reaching vent; pectoral short, broad, and slightly falcate, its
length about 1.4 in head. Scales large, thin, and deep, 19 in front of spinous dorsal, 6 rows between
the dorsals and 9 on median line of caudal peduncle.

Color when fresh, clear olive green with darker edges to scales; lateral stripe steel blue above,
fading into the silvery belly; fins uncolored.

Color in alcohol, olivaceous above, silvery on sides and below; scales of back and upper part of
side with numerous small round coffee-brown specks, disposed chiefly on the edges; median line of
back with a darkish stripe; middle of side with a broad silvery band, plumbeous above, especially
anteriorly, more silvery below; top of head and snout with numerous dark brownish or black specks;
side of head silvery, opercle somewhat dusky, sides and tip of lower jaw dusky; dorsals and caudal
somewhat dusky, other fins pale; pectoral without dark tip.

This small fish is common inside the reef in shallow bays everywhere in the Hawaiian Islands.
Many individuals were seen off the wharf at Lahaina on Maui. Our collections of 1901 contain 20
specimens from Kailua, from 1.5 to 3.5 inches long; 43 from Hilo, 1.5 to 2.25 inches long; and 1 from
Honolulu, 2.25 inches in length. Numerous specimens were obtained by the Albatross at Honolulu in
1902, 1 of which is taken as our type and 3 others as cotypes.

Type, No. 50819, U. S. N. M., 4.25 inches long, obtained by the Albatross at Honolulu. Cotypes
No. 2741, U. S. F. C., 3.9 inches long; No. 2302, Am. Mus. Nat. Hist., 3.9 inches long; and No. 4063,
Field Col. Mus., 3.5 inches long, all collected at Honolulu by the Albatross.

Family II 0 L0 C E NT R I DTE. The Squirrel-fishes.

13. Myripristis berndti Jordan & Evermann, new species.

Head 2.8 in length; depth 2.4; eye 2.7 in head; snout 4.7; maxillary 1.7; interorbital 4.9; D. x-i,
16; A. iv, 14; P. i, 14; V. i, 7; scales 4-32-7.

Body elongate, deep, compressed, its greatest depth at base of ventral; head large, compressed, its
depth less than its length; snout short, blunt, convex, its width about twice its length; upper profile
of head straight from above nostril to occiput; eye large, high, its diameter a little less than posterior
part of head, and its upper rim hardly impinging upon upper profile of head; mouth very large,
oblique; mandible slightly projecting, the maxillary not reaching posterior margin of eye; distal
expanded extremity of maxillary 1.7 in eye; several enlarged, blunt teeth on outer front edges of jaw
and sides of mandible; teeth in jaws fine, in broad bands, also on vomer and palatines; tongue thick,
pointed, and free in front; suborbital rim narrow, finely serrate; lower posterior margin of maxillary
with blunt denticulations ; lips rather thick and fleshy; nostrils close together, posterior very large,
close to front rim of orbit; bones of head all finely serrate; opercle with well-developed spine; gill-

FISHES FROM THE HAWAIIAN ISLANDS.

171

opening large, filaments rather large; gillrakers long, fine, longest longer than longest gill-filaments;
pseudobranchise very large, outer portions free for half their length; dorsal spines slender, first 3.4 in
head, second 2.6, third 2.2, fourth 2.2, tenth 6.4, and last 3.5; anterior dorsal rays elevated, produced
into a point, first 1.8, second 1.7, and last 8; first and second anal spines short, third 2.6, and fourth
2.8; soft anal similar to soft dorsal, anterior ray 1.75, third 1.8, and last 6; caudal forked, lobes pointed,
1.2; pectoral rather small, pointed, 1.4; ventral 1.6, reaching .65 distance to anus; caudal peduncle
elongate, compressed, 2.2 in head, its depth 3.25; scales large, ctenoid, deep; lateral line slightly
arched, running obliquely down on side along upper part of caudal peduncle; 4 rudimentary caudal
rays above and below, slender, sharp-pointed, and graduated.

Color in life (No. 03370), deep red, with silvery luster; no stripes on side, a blood-red band across
gill-opening and base of pectoral; fins deep red, without white edgings, distal half of spinous dorsal
shading into orange.

Color in alcohol, pale straw-color, fins plain and paler; upper margin of opercle blackish, and axil
of pectoral black; anterior margins of soft dorsal and anal Whitish.

We take pleasure in naming this species for Mr. Louis E. Berndt, superintendent of the Honolulu
market. Described from an example (No. 03346) taken at Honolulu, where the species is rather
common. Our collections contain the following specimens:

Field

No.

Length.

Locality.

Final disposition of
specimen.

Field

No.

Length.

Locality.

Final disposition of
specimen.

Inches.

Inches.

'• 03346

9

Honolulu.

Type, No. 50627, U. S. N. M.

04870

9

Honolulu.

Cotype, No. 3947, Field

03370

8

do

Cotype, No. 16818, Bishop Mus.

Col. Mus.

04834

8

do....

Cotype, No. , M. C. Z.

04872

8

do ....

Cotype, No. 24212, Ac.

04836

7

do —

Cotype, No. 2701, U.S. F.C..

Nat. Sci. Phila

04849

7

do

Cotype, No. 2282, Am. Mus,

04873

8.5

do

Cotype, No. 7451, L. S.Jr.*

Nat. Hist.

Univ. Mus.

04850

7

do

Cotype, No. 1489, Cal. Ac. Sci.

04874

7

do

04856

7

do

Cotype, No. 50628, U. S. N. M.

04878

7

do

04857

8.5

do

Cotype, No. 9801, Mus. Ind.

04882

7

do

Univ.

04885

8

do....

04891

9

do ....

14. Myripristis chryseres Jordan & Evermann, new species.

Head 2.75 in length; depth 2.5; eye 2.4 in head; snout 5.5; maxillary 1.9; mandible 1.8; inter-
orbital 5; D. x-i, 14; A. iv, 12; scales 4-34-6.

Body short, stout, and compressed; dorsal profile evenly convex from tip of snout to origin of soft
dorsal; ventral outline nearly straight to origin of anal whose base is equally oblique with that of
soft dorsal; caudal peduncle short but slender, and not greatly compressed, its length from base of last
dorsal ray to first short spinous caudal ray 1.3 in eye, its least width about 3 in its least depth which is
1.8 in eye; head heavy, short; mouth moderately large, the gape in closed mouth reaching vertical of
middle of eye; maxillary very broad, triangular, reaching nearly to vertical of posterior line of eye,
with a broad, curved supplemental border; surface of maxillary roughly striated, anterior edge near
the angle strongly dentate; lower jaw strong, somewhat projecting, the tip with 2 rounded rough
prominences fitting into a distinct notch in upper jaw; teeth short, in narrow villiform bands in jaws
and on palatines, a small patch on vomer, none on tongue; eye very large, orbit exceeding postocular
part of head; lower edge of eye on level with axis of body; snout short, 2 in orbit; interorbital space
nearly flat, strongly rugose; 2 long ridges from preorbitals to nape; outside of these a short ridge
beginning above front of pupil, extending backward and branching upon nape; supraocular ridge
spinescent posteriorly; suborbital narrow, strongly dentate below, upper edge in front somewhat
roughened; opercular bones all strongly toothed; opercular spine short and obscure (stronger in most
of the cotypes); dorsal spines slender, fifth longest and strongest, its length 2.5 in head; first dorsal
spine somewhat posterior to base of pectoral, its length 2 in eye, spines gradually shorter from fifth;
space between dorsals very short, about equal to length of tenth spine; dorsal rays long, length of
longest a little greater than orbit, last equal to pupil; first anal spine very short, second short and
triangular, its length about 1.5 in pupil; third anal spine long, strong and straight, longer than fourth,
its length equal to diameter of orbit; fourth anal spine slender, its length 1.3 in orbit; anal rays longer
than those of dorsal; caudal widely forked, lobes equal, their length 1.5 in head; pectoral long and
narrow, its length 1.4 in head, the tip reaching past tips of ventrals; ventrals slender, pointed, nearly
reaching vent and nearly as long as pectoral.

172

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Scales smaller than in M. murdjan, number in lateral line .34 in type, 35 to 38 in some of the cotypes;
scales strongly dentate, and striate near the edges; a strongly dentate humeral scale.

Color in life, bright scarlet, centers of the scales paler; a blackish-red bar behind, and on edge of,
opercle, continued as red (not black) into the axil; first dorsal golden, with red basal blotches on mem-
branes; second dorsal golden, with crimson at base, spine and first ray white; caudal golden, first ray
white above and below; anal golden, the siiines and first ray white; all the vertical fins narrowly edged
with red; ventrals mostly pink, with golden wash on first rays; pectoral plain crimson; axil light red.

Color in alcohol, yellowish or orange white, the edges of the scales paler; some of the scales with
small brownish dustings on the edges; edge of opercle black; opercle and cheek somewhat silvery; fins
all pale yellowish, without dark edges. In some individuals the general color is more silvery, and in
one example (No. 04860) the axil of the pectoral is somewhat dusky. In life the color is more scarlet
than in M. murdjan and the fins yellow, not red as in M. murdjan and all other Hawaiian species.

Mi/ripristis chrysere s is related to M. murdjan, from which it differs in the smaller scales, larger eye,
less black in the axil, and the absence of bla'ck edges to the dorsal and anal fins as in the life colors
already noted, the yellow fins being the most conspicuous character in life. It, reaches a length of 9 or
10 inches and appears to be moderately abundant at Honolulu and Hilo.

Type, No. 50629, U. S. N. M. (field No. 03463), a specimen 8 inches long, obtained at Hilo, Hawaii.
The numerous cotypes and the museums in which they have been deposited are indicated in the fol-
lowing tabular list of specimens:

Field.

No.

Length.

Locality.

Final disposition of specimen.

Field

No.

Length.

Locality.

Final disposition of specimen.

Inches.

Inches.

2558

4

Honolulu.

04862

9

Hilo

Cotype, No. , M.C.Z.

03463

8

Hilo

Type, No. 50629, U. S. N. M.

04863

9

do

Cotype, No. 9802, Ind. Univ.

04823

8.5

do

Cotype, No. 7452, L. S. .Tr.

Mus.

Univ. Mus.

04867

Si

Honolulu.

Cotype, No. 2702, U. S. F. C.

04827

8

do

Do.

04868

7

do

Cotvpe, No. 1689, Bishop Mus.

04833

0

Honolulu.

Cotvpe, No. 2283, Am. Mus.

04869

8

do

Cotype, No. 1490, Cal. Ac. Sci.

Nat. Hist.

04887

8

do

04839

7

do

Cotype, No. 24273, Ac. Nat. Sci.

04889

7

do

Phila.

04890

8

do ....

04860

9.5

Hilo

Cotvpe, No. 50630, U. S. N. M.

04892

8

do —

04861

9

do

Cotype, No. 3948, Field Col.

Mus.

15. Myripristis argyromus Jordan & Evermann, new species.

Head 3.5 in length; depth 2.75; eye 2.4 in head; snout 5; maxillary 1.8; mandible 1.6; interor-
bital 3.75; D. x-i, 15; A. iv, 13; scales 4-33-5.

Body rather long and compressed, dorsal and ventral outlines about equally and evenly convex
from snout to origins of anal and soft dorsal fins; head rather large but short; mouth moderate, max-
illary reaching vertical at posterior edge of pupil, the exposed portion broad, triangular, the upper
edge concave, the end rounded and the anterior edge with short blunt teeth, strongest at angle; tip of
upper jaw with a shallow notch roughened at its outer edges; jaws equal, lower fitting into the notch
of upper and with 2 patches of strong blunt tooth-like tubercles at its tip; eye large, its middle above
level of tip of upper jaw; interorbital space wide and slightly convex; 2 low, nearly parallel median
ridges from tip of snout to nape, diverging slightly at their middle, another low ridge from above orbit
backward to nape, and another backward around orbit; ridges on nape divergent; suborbital narrow,
dentate on both edges; opercular bones all striate and dentate at the edges; opercle with a short, flat,
triangular spine; scales large, rough, striate near the edges which are finely toothed; a series of 4 or
5 large modified scales across nape, and a series of triangular scales along bases of dorsal and anal;
about 10 scales in front of dorsal; origin of dorsal about over lower base oi pectoral; dorsal spines
slender, the first 3.2 in head, third and fourth longest, about equal to orbit; interval between dorsals
very short; anterior dorsal rays somewhat produced, their length equal to snout and eye; edge of fin
concave, last rays nearly 3, or equal to pupil; anal spines graduated, the first very small, second short
but stout, third much longer and stoutest, its length 1.3 in eye, fourth still longer and more slender;
anterior anal rays produced, their length about equal to that of longest dorsal rays, free edge of fin
concave; caudal evenly forked, the lobes equal to length of head; peetoral long and pointed, reaching
beyond tips of ventrals, about 1.3 in head; ventrals shorter, 1.6 in head, their tips equally distant
between their bases and that of first anal ray.

FISHES FROM THE HAWAIIAN ISLANDS.

178

Color in alcohol, pale yellowish-white, brightest above, more silvery on side and belly; opercular
bones with fine round brownish specks; edge of opercle not black, scarcely dusky; axil dusky inside
but not showing above fin; fins pale yellowish-white without any dark on edges.

Type, No. 50631, U. S. N. M. (field No. 04829), a fine specimen 9.5 inches long, obtained by us at
Hilo, Hawaii.

M. argyromus is related to M. berndti, but is distinguished by the more slender body, the absence
of black on the opercle, and the paler axil. It does not appear to be abundant and is represented
in our collections by only 8 specimens. All the other specimens are taken as cotypes. The data
regarding each will be found in the following list:

Field

No.

Length.

Locality.

Final disposition of specimen.

01829

Inches.

9.5

Ililo

Type, No. 50631, U. S. N. M.

04830

9.0

Honolulu. .

Cotype, No. 7453, L. S. . Jr.Univ. Mus.

04835

6.5

do

Cotype, No. 1690, Bishop Mus.

048-40

6.5

do

Cotype, No. 2284, Am. Mus. Nat. Hist.

04877

9.0

Honolulu. .

Cotype, No. 2703, U. S. F. C.

04879

8.0

do

Cotype, No. , M. C. Z.

01880

7.5

do

Cotype, No. 9803, Ind. Univ. Mus.

04881

9.0

do

Cotype, No. 3949, Field Col. Mus.

16. Myripristis symmetricus Jordan & Evermann, new species.

Head 3.2 in length; depth 2.4; eye 2.2 in head; snout 5; interorbital 3.8; D. x-i, 15; A. iv, 14;
P. i, 14; V. i, 7; scales 4-36-6.

Body elongate, deep, compressed, greatest depth about midway between origin of ventrals and anal;
upper and lower profiles evenly convex; head compressed, as long as deep, its width 1.7 in its length;
snout short, broad, blunt, and steep; upper profile of head straight from above nostril to occiput; eye
very large, high, hardly infringing upon the upper profile of head, its diameter greater than postocular
region; mouth very large, oblique; mandible slightly projecting, and reaching posteriorly to below
posterior rim of pupil; distal expanded extremity of maxillary 2.35 in eye; several enlarged blunt
teeth on outer front edges of mandible; teeth in jaws, on vomer, and palatines very fine, in bands;
tongue thick, pointed, free; suborbital rim narrow, finely serrate; lower posterior margin of maxillary
smooth; lips rather thick and fleshy; nostrils close together, posterior very large, close to front rim of
orbit; bones of head all finely serrate; opercle with well-developed spine; gill-opening large, filaments
large; gillrakers long, fine, longest longer than longest gill-filament; pseudobranchiae very large; dorsal
spines slender, sharp, first 2.75 in head, second 2.1, third 2, fourth 1.9, tenth 6, and last 3.6; soft
dorsal with anterior rays elevated, produced into a point which projects beyond tip of posterior rays
when fin is depressed, first ray 1.4 in head, third 1.35, and last 3.75; anal spines graduated to last,
third enlarged, 2.5 in head, fourth 2.9; soft anal similar to soft dorsal, anterior rays produced, first 1.4,
third 1.3, and last 4.6; caudal elongate, deeply forked, the lobes pointed, 1.2 in head, and reaching
slightly behind tips of ventrals; ventrals sharp-pointed, 1.4 in head, spine 2.2; caudal peduncle
elongate, compressed, its length 1.8 and its depth 3.2; scales large, finely ctenoid, deep on middle of
side; lateral line running obliquely back, slightly curved at first, and posteriorly along upper side of
caudal peduncle; 4 rudimentary, slender, sharp-pointed, graduated rays along upper and lower edges
of caudal; scales narrowly imbricated along middle of side.

Color in alcohol, pale straw-color; fins paler, except the anterior dorsal and anal rays, which are
grayish; margin of opercle above blackish; axil of pectoral black.

This species was found both at Honolulu and Hilo, but does not appear to be abundant. Only 4
specimens are in our collections:

Field

No.

Length.

Locality.

Final disposition of specimen.

04866

04864

04865
04924

Inches.

5.5

5.5

5.5

5.0

Hilo

do

do

Honolulu

Type, No. 50632, U. S. N. M.

Cotype, No. 7454, L. S. Jr. Univ. Mus.
Cotype, No. 2704, U. S. F. C.

Cotype, No. 3950, Field Col. Mus.

174

BULLETIN OF THE UNITED STATES FISH COMMISSION.

17. Flammeo scythrops Jordan & Evermann, new species.

Head (measured to end of flap) 2.75 in length; depth 3; eye 3 in head; snout 4; maxillary 2.1;
mandible 1.8; interorbital 5; D. xi, 13; A. iv, 9; scales 5-48-7, 5 rows on cheek; Br. 7.

Body oblong, rather slender; dorsal outline gently and rather evenly curved from tip of snout to
origin of soft dorsal, more nearly straight from tip of snout to nape; ventral outline less convex; head
long; snout long and pointed; maxillary broad, with a strong supplemental bone whose lower edge
forms a broad angle; end of maxillary slightly concave; lower jaw long, much projecting, tip promi-
nent; mouth large, not greatly oblique; maxillary nearly reaching vertical at posterior line of pupil;
lips broad, rounded, and soft; eye large, lower edge of pupil on axis of body; interorbital space with a
broad, shallow groove between low ridges, one on each side; space between ridge and eye with short,
curved ridges; nape on each side with a group of 8 or 10 short, sharp ridges, diverging backward and
ending in short, sharp spines; posterior part of supraocular with a patch of short spines; suborbital
dentate on its lower edge; preorbital with 2 blunt prominences in front, a strong, recurved spine
below, ridges and spines on its upper surface; opercular bones all strongly striate, the striae ending
in short spines; entire surface of interopercle seriate; opercle with 2 strong spines, the lower the
stronger, its length 1.6 in orbit.; preopercle with a 'very strong spine at angle, its length nearly
equaling diameter of orbit, its surface striate, and its base with a series of small spines; under surface
of dentary somewhat roughened; surface of articular bone much rougher; jaws each with a broad
band of villiform teeth, the outer series on upper jaw stronger; a narrow series on each palatine- and
a patch on vomer; scales moderate, the surfaces usually nearly smooth, the edges finely toothed; a
series of strongly striate scales across nape, and a strong, striated plate at shoulder; lateral line well
developed, little arched, with about 45 pores; bases of soft dorsal and anal each with series of
modified triangular scales; caudal with small scales on base and fine scales on membranes, extending
well toward tips of outer rays; origin of spinous dorsal in advance of base of pectoral or over middle
of upper opercular spine; dorsal spines in a broad, deep groove, moderately strong, middle one
longest, 2.3 in head, first a little shorter than snout, tenth more than half eye; dorsal rays longer
than spines, longest 2.2 in head; first anal spine very short, second about 3 times as long; third
anal spine very long and strong, but little curved, reaching past base of anal, its length 1.5 in head;
fourth anal spine shorter and more slender, its length 2.25 in head and equaling longest anal rays; last
anal ray much shorter, 1.6 in eye; pectoral long and slender, 1.2 in head, the tip nearly reaching vent;
ventrals shorter, equal to snout and eye; caudal forked, the lobes equal, not strongly divergent, their
length about equaling that of third anal spine; rudimentary caudal spines 5 above, 4 below, strong
and sharp.

Color in life, head red above, paler on sides, nearly white below; tips of jaws rich red; side of
body with about 10 or 12 narrow' yellow stripes separated by red or rosy stripes of about same width,
those below paler and somewhat purplish; under parts purplish or pinkish white; the stripes begin-
ning at edge of opercle and ceasing at base of caudal peduncle, which is rich red above, becoming
paler on side and below; membranes between the first and third dorsal spines rich blood-red, those
between other spines white at base, each with distal portion lemon-yellow in front and red behind,
last 2 or 3 membranes with little or no yellow'; dorsal spines pale rosy, nearly white; soft dorsal, anal,
pectoral, and ventral with rays rosy, membranes pale; ventral with a little yellow at base; anal spines
somewhat dusky; caudal rich blood-red, paler distally; eye red, a narrow yellow ring around pupil.

Another example (No. 03041), much faded, was bright red; stripes on side equally bright golden;
fins red; edges of dorsal membranes pale; no markings evident on fins.

Color in life of another example (No. 03451), side with 10 or 11 longitudinal golden or yellow
bands; spinous dorsal more or less white; membranes between first and third dorsal spines more or
less deep vermilion, except the upper marginal portion behind second spine, which is white; a red
blotch along margin of membranes just before each of the other dorsal spines.

Color when fresh of another specimen (No. 03490), violet-rose with 10 stripes of bright golden on
side; dorsal red, mottled with golden, the first, two spines deep red; soft dorsal and other fins rather
light red without edgings, and scarcely darker behind third anal spine; pectoral and ventrals pink; a
red dash across cheek, space above and below whitish; temporal region deep red; iris red.

All these colors fade in alcohol and the fish becomes a pale yellowish white, the longitudinal lines
on side showing faintly as duller and brighter stripes of yellowish white; fins all whitish or yellowish
white, membranes of spinous dorsal w'hiter.

FISHES FROM THE HAWAIIAN ISLANDS.

175

The above description from the type, No. 50633, U.S.N.M. (field No. 03188), a specimen 9 inches
long obtained by us at Honolulu. An examination of our large series of cotypes shows but slight vari-
ations, the characters appearing quite stable. In some examples the upper opercular spine is the
larger, in others the 2 are equal; in 2 examples we find 3 opercular spines.

This species has been several times called Holocentrum argenteum. The species described under
that name by Quoy & Gaimard from New Guinea resembles this in the slender body and general
coloration, but differs in having the lower jaw included, eye much smaller, mouth smaller, and the
preopercular spine weaker. It was probably intended for some species with the lower jaw included.

Holocentrus Here Lesson, from Tahiti, is more likely to be the present fish. It is figured as elon-
gate, with the spinous dorsal low and the opercular spines equal. The plate is, however, too rough to
permit certain identification and approaches almost as closely to Holocentrus diploxiphus as to Flammeo.

This is one of the most abundant species in the markets at Honolulu and Hilo. It reaches a
length of 8 to 10 inches. Our collections contain the following specimens, all of which, except the
first, are taken as co types:

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Inches.

Inches.

03488

9. 25

Honolulu.

Type, No. 50633, U. S. N. M.

04965

8.50

Honolulu.

Cotvpe, No. 7456, L. S. Jr.

03041

8.00

do

Cotype.

Univ. Mus.

03451

8. 50

do

Cotvpe, No. — , M. C. Z.

04966

8.50

do

Cotype, No. 2705, U.S.F.C.

03490

8.50

Hilo

Cotype, No. 7455, L. S. Jr.

04967

7.50

do ....

Cotvpe, No. 2705, U. S. F.C.

Univ. Mus.

04968

8.25

do ....

Cotype, No. 24214, Ac.

04925

8.25

do

Do.

Nat. Sci., Phila.

04920

10. 00

do ....

Do.

04969

7. 50

do....

Cotype, No. 2285, Am.

04927

9.00

do

Cotype, No. 1491, Cal. Ac. Sci.

Mus. Nat. Hist.

04928

9.25

do....

Cotype.

04970

8.50

do....

Cotype, No. 7456, L. S. Jr.

04949

6.25

Honolulu.

Cotype, No. 3967,Field Col. Mus.

Univ. Mus.

04954

8.75

do

Cotype, No. 7456, L. S. Jr.

04971

7.75

do

Do.

Univ. Mus.

04972

7. 75

do....

Do.

04956

8. 75

do....

Cotype, No. 9804, Ind. Univ.

04973

8.75

do....

Do.

Mus.

04974

8. 75

do....

Do.

04957

7.25

do

Cotype.

04975

8. 00

do

Do.

04958

8.50

do

Cotype, No. 1691, Bishop Mus.

04976

5. 00

do....

Do.

04959

7.50

do....

Cotype.

04977

7. 00

do ....

Do.

04964

9.00

do ....

Cotype, No. 50634, U. S. N. M.

04988

9.00

do

Do.

Holocentrum argenteum, Steindachner, Denks. Ak. Wiss. Wien, lxx, 1900, 492 (Honolulu and Laysan); not of Cuvier &
Valenciennes.

18. Holocentrus xantherythrus Jordan & Evermann, new species.

Head 2.8 in length; depth 3; eye 3 in head; snout 4; maxillary 2.7; interorbital 5; D xi-14;
A. iv, 10; scales 4-47-8.

Body elongate, compressed, greatest depth about base of ventral; upper profile steep; lower
profile nearly horizontal; head compressed, its depth about 1.2 in length, width 2.25; eye large, high,
impinging upon upper profile in front, anterior, and a little less than postocular region; snout short,
pointed, its upper profile obliquely straight; jaws rather large, subequal; maxillary reaching beyond
front margin of pupil or to first third of eye, its distal expanded extremity 2.7 in eye; supplemental
maxillary large; lips rather thick, fleshy; teeth small, short, in rather broad bands in jaws and on
vomer and palatines; tongue elongate, pointed, free in front; nostrils close together, posterior, a deep
cavity in front of middle of eye; interorbital space broad, very slightly concave; preorbital with a
large spine in front, its margins serrate; suborbital narrow, with finely serrate margin; preopercle
with a large dagger-like spine at lower angle; opercle with 2 similar spines on upper margin, upper
one much the larger; bones of head with serrate margins; gill-opening rather large, filaments and
pseudobranchioe well developed; gillrakers short, compressed, few, and much shorter than longest
filaments; fleshy axillary flap small; dorsal spines sharp-pointed, first 3.2 in head, second 2.8, third
1.9, last 7; anterior dorsal rays high, second 2.4 in head, third 2.2, last 6.5; third anal spine very
large, not reaching beyond soft rays, 1.7 in head, fourth 2.25; anterior anal rays longest, first 1.75 in
head, second 1.9, last 6; caudal rather small, deeply forked; pectoral small, 1.6 in head; ventral
sharp-pointed, 1.4, spine 2; caudal peduncle .elongate, compressed, its length 2.1 in head, depth 4;
scales rather large, ctenoid; lateral line nearly straight, running obliquely down along upper side of
caudal peduncle.

176

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Color in life (No. 02989), bright red, belly more or less silvery; about 10 narrow longitudinal
silvery stripes; uppermost pinkish; side of head silvery with pinkish shades; a white stripe from
preorbital to base of preopercular spine; spinous dorsal deep red without streaks or black marking, a
white spot behind first and second spines at base, tips of third to seventh spines whitish; soft dorsal,
anal, caudal, and pectoral plain pink; anal with membrane of third spine and first soft ray deep red;
ventral pink, spine and first soft ray white, second soft ray deep red anteriorly, posteriorly whitish.

Another example (No. 03161), was rose red when fresh, with about 10 very faint light rosy streaks
along rows of scales, these much less distinct than in other species; cheek rosy with one broad
oblique white band; dorsal plain red, the membranes fading to white, no light stripes on dark areas;
other fins plain light red; membrane of fourth anal spine not darker; iris pink.

Another example (field No. 03467), deep crimson when fresh, with 10 narrow, sharply defined,
white stripes along rows of scales; an oblique white stripe below eye from snout to base of preopercular
spine; dorsal clear deep red, clouded with darker; soft dorsal, caudal, and anal light bright red;
membrane between third and fourth anal spines blood-red; pectoral deep red; veDtrals red, spines
white, their membranes blood red.

Color in alcohol, pale brown or brownish white, washed more or less with silvery or brassy
white; side with 9 or 10 longitudinal white stripes; fins pale.

This species is related to Holocentrus en&ifer, differing mainly in the presence of two well-developed
spines on the upper margin of the opercle. It is one of the most abundant of the family in Hawaiian
waters. It is represented in our collections by 40 examples, as follows:

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Inches.

Inches.

05999

6

Honolulu.

Type, No. 50635, U. S. N. M.

04227

6.5

Honolulu .

Cotype, No. 9805, Ind. Univ.

02934

6.5

do....

Cotype) No. 50636, U. S. N. ML

Mus.

02933

5.5

do

Cotype', No. 2706, U. S. F. C.

04936

5. 75

do —

Cotype, No. 1492, Cal. Ac. Sci.

02989

4. 5

do

Cotype, No. 3951, Field Col.

04938

6.25

do

Mus.

04939

6.25

do

03160

5. 25

do

Cotype, No. 7457, L. S. Jr. Univ.

04960

5.5

do ....

Mus.

04961

5.5

do

03161

6.25

do....

Do.

04940

5. 5

do

03294

5. 75

do

Cotype, No. 1692, Bishop Mus.

04980

5. 75

Kailua . . .

03295

5.75

do ....

Cotype, No. , M. C. Z.

04234

5.25

Honolulu.

03203

5

do

Cotype, No. 22S6. Am. Mus.

04230

5.5

do....

Nat. Hist.

04244

3.25

Kailua . . .

03467

5. 75

do

Cotype, No. 24215, Ac. Nat. Sci.

04978

3. 75

do —

Phila.

04226

5.75

Honolulu .

03468

G. 5

do

Cotype, No. 2733, U. S. F. C.

05905

5.6

do ....

And 15 other examples from Honolulu, ranging in length from 4 to 6.5 inches.

19. Holocentrus ensifer Jordan & Evermann, new species.

Head 3 in length; depth 2.7; eye 3 in head; snout 3.5; maxillary 2.25; interorbital 5; D. xi, 15;
A. iv, 11; P. i, 14; V. i, 8; scales 4-47-8.

Body elongate, compressed, greatest depth at ventral fin; upper profile decidedly more convex
than lower; head compressed, much longer than deep, pointed, its width a little more than half its
length; eye moderate, about 1.2 in postocular part of head, and slightly impinging upon upper profile;
snout pointed; mouth moderate, oblique; maxillary broad, with large supplemental bone distally,
equal to half diameter of eye; lips thick, fleshy; teeth minute, in broad bands in jaws, and on vomer
and palatines; tongue pointed, free in front; nostrils close together, posterior a large cavity with sev-
eral small spines projecting over; preorbital with 2 large strong spines and about 6 strong serrations on
its margin; suborbital rim narrow; bones of head all more or less finely serrate, the opercle above and
preopercle below each with a long, strong, dagger-like spine; interorbital space broad, very slightly
concave; a fleshy axillary flap; gill-opening large, filaments moderately long, much longer than gill-
rakers which are compressed and not very numerous; pseudobranch ite large; spinous dorsal long,
membrane between spines not much incised, first 2.2, second 2.1, third 2, last 4.2; anterior dorsal rays
longest, fourth 1.8, last 7.5; third anal spine largest, 1.75, fourth 2.3; soft anal similar to soft dorsal,
third spine not reaching beyond rays; caudal rather small, forked; pectoral 1.3; ventral 1.4, spine 2;
caudal peduncle compressed, its length 2.2, depth 4; scales rather large, ctenoid; lateral line arched
a little at first and running down obliquely on upper side of caudal peduncle.

FISHES FROM THE HAWAIIAN ISLANDS.

177

Color in life, bright red; side with about 8 yellow longitudinal bands; spinous dorsal vermilion
tinged with yellow; soft dorsal rosy with front margin white and behind this above, red; anal whitish
with red between third spine and first ray; caudal red, margined above, and along the emargination
with whitish; pectoral whitish with red lines; ventral rosy with front margin white.

Another example (field No. 03454), in life had yellow and red longitudinal bands above and yel-
low and white below; spinous dorsal vermilion, other fins red with white borders. Another (field
No. 03472), was brilliant scarlet red with 11 golden streaks along rows of scales, upper 4 broadest, and
third and fourth most distinct and oblique; a white or golden streak across cheek; fins plain scarlet
without dark patches. Color, when fresh, of another specimen (field No. 03494), bright red verging to
scarlet; side red, with 4 golden stripes along back and 6 silver stripes below these, golden and silver,
very bright; head crimson; a white band on cheek; spinous dorsal deep scarlet with crimson edge;
soft dorsal light crimson with a white, then a dark crimson edge; caudal blood red, edged above and
below with white, posterior part of fin abruptly pale; anal with pale spines, then blood red, then
pinkish; ventral with white spine, then dark red, then pink; pectoral light red, axil deep red.

Color in alcohol, pale brown or brownish white, the longitudinal bands on sides, together with
scales on cheeks and opercle, silvery; fins pale.

This species was obtained by us at Honolulu and Kailua, and appears to be common at the former
place. The collections contain the following specimens:

Field

No.

Length,

Locality.

Final disposition of specimen.

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Inches.

Inches.

03448

6

Honolulu.

Type, No. 50637, IT. S. N. M.

01951

9.5

Honolulu.

Cotvpe, No. 24216, Ac.

03472

6

Kailua ...

Cotype, No. 7458, L. S. Jr.

Nat. Sci. Phila.

Univ. Mus.

03454

9. 75

do....

Cotvpe, No. 9806, Ind.

04929

8. 75

Honolulu.

Cotype, No. 2708, U. S. F. C.

Univ. Mus.

04930

8.5

do....

Cotvpe, No. 1693, Bishop Mus.

03494

8.5

do....

Cotype, No. 3952, Field

04931

8

do

Cotype, No. , M. C. Z.

Col. Mus.

04950

8.75

do —

Cotype, No. 2287, Am. Mus.

Nat. Hist.

Family CARANGID7E. The Pampanos.

20. Carangus elacate Jordan & Evermann, new species.

Head 3.6 in length; depth 3.4: eye 4.5 in head; snout 3.8; interorbital 3.8 in snout; maxillary 2.1;
preorbital 8.5; mandible 1.9; D. vii-i, 19; A. ii-i, 16; scutes 28.

Body slender, compressed, not greatly elevated; snout rather short, profile ascending to nape in a
gentle curve, slightly trenchant; mouth large, slightly oblique; lower jaw somewhat projecting; max-
illary reaching posterior edge of orbit, its width at tip 1.5 in orbit; supplemental maxillary well
developed, its width 3.25 in entire width; gape reaching vertical of posterior edge of pupil; villiform
teeth on vomer, palatines and tongue, those on jaws in a single row, small and somewhat canine-like;
eye large, anterior; adipose eyelid strongly developed behind; supraocular region with two ridges,
extending to humeral region, the lower the stronger; posterior half of body, beginning at origin of
soft dorsal, long and gently tapering to caudal peduncle; caudal peduncle much depressed, its least
depth scarcely half its least width; distance from base of last dorsal ray to origin of caudal fin equal
to snout and pupil; fins small; origin of spinous dorsal posterior to base of pectoral by a distance
equal to eye; longest dorsal spine slightly greater than snout; anterior rays of soft dorsal somewhat
produced, about 1.8 in head; anal similar to soft dorsal, its origin under eighth soft dorsal ray, anterior
ray produced, but scarcely equaling longest soft dorsal rays; caudal widely forked, lobes apparently
equal; pectoral long and falcate, reaching past origin of anal, exceeding head in length by 0.65 diameter
of eye; ventrals short, 2.4 in head; scales rather large, a low sheath at base of soft dorsal and anal
anteriorly; breast entirely scaled; lateral line strongly arched above pectoral, joining straight portion
under sixth dorsal ray, chord of arched portion 1.6 in straight part.

Color in alcohol, rusty olivaceous above, paler on side below lateral line, belly white; top of head
dark olive, side and lower jaw lighter, with strong brassy tinge on postocular and on lower portions
of opercle; lower jaw profusely covered with fine brown points; a black spot at upper end of opercular
opening; axil black; vertical fins all more or less dark; produced part of soft dorsal almost black, low

F. C. B. 1902—12

178

BULLETIN OF THE UNITED STATES FISH COMMISSION.

part of soft dorsal black at base, then lighter, narrowly tipped with dark; anal dark brown, with a
subterminal stripe of yellowish white along edge of fin; pectoral and ventrals pale.

The above description based upon the type, No. 50638, U. S. N. M. (field No. 01452), a large
example 27 inches long, from Honolulu.

This species somewhat resembles Carangus rnarginatus, from which it differs in the much more
slender body, larger eye, and dark anal fin. The type is the only example obtained

Family SERRANID+L The Sea-basses.

21. Pikea aurora Jordan & Evermann, new species.

Head 2.5 in length; depth 3; eye 5 in head; snout 4; interorbital 6.2; maxillary 2.25; D. vm,
13; A. iii, 8; scales 5-55-22; Br. 7; gillrakers short and rather weak, about 9+5.

Body moderately stout, the back slightly elevated, head rather long and pointed; snout depressed,
the anterior profile nearly straight from tip of snout to Occiput; mouth large, maxillary reaching
posterior margin of pupil, supplemental bone not developed, the tip broad, 1.5 in orbit; mouth some-
what oblique, the lower jaw strongly projecting; teeth in broad villiform bands on jaws, vomer, and
palatines; tongue naked; eye moderate, high up, chiefly above axis of body; anterior nostril in a short
tube at edge of prenasal; posterior nostril small, round, near upper anterior edge of orbit; edge of
preopercle slightly dentate, especially on lower arm; opercle ending in a broad flap with a weak, flat
spine; pseudobranchise rather small; interorbital low, very little convex; caudal peduncle stout, com-
pressed, and very deep, the depth equaling snout and eye; fins rather small; origin of dorsal posterior to
that of pectoral, slightly nearer base of last ray than tip of snout; dorsal spines low and weak, the third
longest, 3.6 in head; soft portion of dorsal somewhat elevated and pointed, with longest ray 1.9 in head;
anal similar to soft dorsal but smaller and somewhat posterior, fifth ray 2 in head ; caudal truncate or
slightly lunate; ventrals short, not nearly reaching vent, their length 1.75 in head; pectoral rather long
and slender, reaching origin of anal, its length about 1.4 in head; scales rather small, finely ciliate,
somewhat loose; entire head, except interorbital, snout, and under parts, scaled; lateral line well devel-
oped, complete, with a strong arch above the pectoral and distinctly decurved under last dorsal ray.

Color in life (field No. 03342), top of head, upper half of anterior part of body, and whole posterior
half of body pale rosy; lower part, of head, and lower parts of anterior half of body white with faint
rosy wash; top of head and back in front of dorsal vermiculated with greenish yellow lines; middle
portion of upper jaw yellow with a broad sulphur-yellow stripe from it to eye, then back of eye to
opercular opening; a narrow sulphur stripe on posterior edge of maxillary and continued interruptedly
downward and backward across cheek to opercle; a few small yellow spots across cheek between
the two stripes; tip of lower jaw yellow; yellow of back in about 6 indefinite lines; dorsal pale rosy,
spinous part greenish yellow at base, this extending toward tip posteriorly and forming a submarginal
yellow stripe on soft part, narrowly bordered above by rosy; rest of fin rosy; caudal dark rosy, paler
toward tip, then with blackish red edge, a greenish yellow stripe along upper and lower margins nar-
rowly edged with rosy; anal yellow anteriorly, rest of fin pale rosy; pectoral and ventrals pale rosy;
yellow of lower jaw bounded by rosy, rest of jaw and chin whitish; some examples with posterior
half of side with scattered small greenish yellow spots, these extending on caudal; eye with a broad
brown bar through the middle, white above and below.

Color in alcohol, pale yellowish white, lighter below; body, especially posteriorly, caudal, and soft
dorsal fins with numerous small distinct brown spots; head pale, a white line extending along upper
edge of maxillary and across cheek to opercular opening, a similar but less distinct white line from eye
to upper edge of gill-opening; between these 2 a few white specks; all the fins except caudal and soft
dorsal plain yellowish white.

Four specimens of this interesting and handsome species were obtained by us, 2 at Honolulu and
2 at Hilo. Four others are in the collection made at Honolulu in 1898 by Dr. Wood.

Field

No.

Length.

Locality.

Final disposition of specimen.

Field

No.

Length.

Locality.

05232

Inches.

6.2

Hilo

Type, No. 50675, U. S. N. M.

O. P.J.

403

Inches.

4.5

Honolulu.

05233

4.8

do ....

Cotype, No. 3971, Field Col. Mus.

682

5. 25

Do.

05231

5. 75

Honolulu.

Cotype, No. 7484, L. S. Jr. Univ. Mus.

687

6

Do.

03312

5.2

do

Cot.ype, No. 2734, U. S. F. C.

6074

6

Do.

FISHES FROM THE HAWAIIAN ISLANDS.

179

22. Anthias kellog'g'i Jordan & Evermann, new species.

Head 2.5 in length; depth 2.5; eye 4.5 in head; snout 3.6; maxillary 2; interorbital 5.4; D. xi, 15;
A. in, 7; P. 15; scales 4-36-10; gillrakers 16 + 4.

Body short, deep, and compressed; dorsal outline greatly arched, profile from origin of spinous
dorsal to tip of snout nearly straight, being gently concave over interorbital space; ventral outline
nearly straight; caudal peduncle compressed, its greatest depth 3 in head; head longer than deep;
snout bluntly pointed, lower jaw prominent, slightly the longer; mouth large, nearly horizontal; a
narrow band of small, sharp, conic teeth on palatines, a small patch on vomer, a band of cardiform
teeth on upper jaw, a narrower band in lower jaw; several large canine teeth in each jaw anteriorly,
3 of these close together on middle of each side of lower jaw, these hooked backward; 6 or 8 large
pores on lower side of mandible and several on upper part of snout; maxillary reaching to posterior
edge of orbit, its greatest width 1.5 in eye; edge of preopercle above angle and edge of opercle below
the upper middle of base of pectoral denticulate; 2 broad opercular spines, the upper the larger; eye
anterior, its lower edge on line with upper base of pectoral; fins large, the second soft dorsal ray and
upper rays of upper caudal lobe being produced each as a filament, the dorsal filament being produced
half its length beyond rest of fin; dorsal spines stout and strong, the first spine 2.3 in third, the fifth
being the highest, 2.5 in head; base of spinous dorsal 1.15 in head; base of soft dorsal 2.3 in head, its
fourth ray 3.5 in head, the last ray 1.4 in fourth; caudal truncate, the lower rays produced slightly as
a filament, but not nearly so long as the upper lobe; second anal spine longest, 2.5 in head; second
soft ray longest, 2 in head; pectoral very long and large, reaching to origin of soft anal, the eighth
and ninth rays from the top the longest, 1.4 in head; scales large, finely ctenoid, in regular series;
entire body and head scaled; basal portion of all fins except spinous dorsal with small scales; lateral
line strongly convex, not concurrent with the dorsal profile, becoming straight on middle of caudal
peduncle; one row of scales behind tip of last dorsal ray.

Color in alcohol, pale brown, the fins lighter; in life, red.

Only three specimens of this species were obtained, all having been taken with the hook in deep
water off Kailua, in southwestern Hawaii. It is allied to Anthias japonicus Steindachner & Doderlein.

Named for Dr. Vernon Lyman Kellogg, professor of entomology in Stanford University.

Field

No.

Length.

Locality.

Final disposition of specimen.

05278

Inches.
7. 75

Off Kailua

Type, No. 50642, U. S. N. M.

03703

8

flo

Cotvpe, No. 7460, L. S. Jr. Univ. Mus.

05277

8.5

do

Cotype, No. 2711, U. S. F. C.

Family APOGONIDvE. The King of the Mullets.

23. Apogonichthys waikiki Jordan & Evermann, new species.

Head 2.4 in length; depth 3; eye 3.2 in head; snout 4.6; interorbital 6; maxillary 2; D. vii-i, 8;
A. n, 7 ; scales 2-24-5.

Body short, stout, and compressed; dorsal outline strongly arched from tip of snout to posterior
base of soft dorsal; ventral outline comparatively straight from tip of mandible to origin of anal; vent
immediately in front of origin of anal; caudal peduncle deep and compressed; head rather large;
mouth large, slightly oblique, jaws equal, maxillary reaching posterior edge of pupil; eye rather small,
slightly above axis of body; interorbital space narrow, little convex; opercular and preorbital bones
entire; a band of small villiform teeth in each jaw, and on vomer and palatines; fins moderate, origin
of spinous dorsal nearer base of last soft ray than tip of snout; first dorsal spine very short, second
about half length of third, which is equal to eye and snout; base of soft dorsal equal to depth of
caudal peduncle; longest dorsal rays 2.25 in head; caudal rounded, its length 1.75 in head; origin of
anal slightly posterior to that of soft dorsal, its longest rays 2.4 in head; pectoral slender, reaching
past origin of anal, its length 1.5 in head; ventrals short, barely reaching origin of anal, their length
nearly 2 in head; scales large, weakly ctenoid, firm and somewhat deeper than long; lateral line
strongly developed, following outline of back until under last dorsal ray, where it curves downward,
following middle line of caudal peduncle to base of caudal fin.

180

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Color in alcohol,- head and body rather dark brownish, a lighter crossband around body at nape
and across opercles; another light band surrounding body between the 2 dorsal fins; 3 dark brown
lines radiating from the eye, the first downward across cheek to tip of maxillary, the second
backward across cheek toward base of pectoral, the third upward and backward to origin of lateral
line; spinous dorsal blackish, especially on last spine; soft dorsal, anal, and caudal dusky, narrowly
edged with white; pectoral pale, crossed by about 6 obscure brownish crossbars; ventrals black or
very dark brown, the outer rays somewhat paler.

The above description is based upon the type, No. 50639, U. S. N. M. (field No. 20), a specimen 1.5
inches long, obtained from the coral rocks in front of Waikiki, near Honolulu, August 22, 1901.

This species is distinctly related to A. alutus of the coast of Florida, from which it differs mark-
edly in color and in the more slender body. Only one specimen was obtained.

Genus FOWLERIA Jordan & Evermann, new genus.

Fowleria Jordan & Evermann, new genus of Apogonidx ( aurita ).

This genus differs from Apogonichthys only in the character of the lateral line, which is developed
only on the anterior part of the body.

Several species occur in crevices of coral rock in the South Seas. All of them are of very small
size and some are brightly colored.

This genus is named for Mr. Henry Weed Fowler, of the Academy of Natural Sciences of
Philadelphia.

24. Apogon snyderi Jordan & Evermann, new species.

*

Head 2.7 in length; depth 3.1; eye 3.7 in head; snout 3.7; interorbital 4.5; maxillary 2.2; mandible
2; gape 3; I). vii-i, 9; A. ii, 8; C. 17; P. 10; scales 2-25-5; Br. 6.

Body short and stout, moderately compressed, the dorsal and ventral outlines about equally
curved; head rather large, conic; snout conic, the anterior profile very slightly curved from tip of
snout to origin of spinous dorsal; mouth oblique, jaws subequal, the lower slightly included; maxillary
long, reaching not quite to posterior edge of pupil, its width at tip 2 in eye, supplemental bone well
developed; interorbital space rather broad, slightly convex, preorbital narrow, least width 3 in eye;
teeth on vomer and jaws, the latter in villiform bands; none on palatines; gillrakers slender, 10 on lower
limb of first arch; caudal peduncle compressed and deep, the least width about 4 in its depth; scales
large, deep, closely imbricated, strongly ctenoid and loose; lateral line beginning at upper end of gill-
opening, nearly straight to base of caudal fin, 4 scales in front of spinous dorsal; nape with a striated
shield; edge of opercle thin and smooth; both margins of preopercle and edge of interopercle serrate,
teeth strongest at angles; a series of moderately strong teeth along lower edge of orbit ; origin of spinous
dorsal nearer snout than base of last dorsal ray; first dorsal spine very short, fourth longest, about 2
in head, second 2 in the fourth, seventh 2 in second; first soft rays longest, 1.8 in head; caudal deeply
emarginate, longest rays about 1.6 in head; anal similar to soft dorsal; somewhat smaller, its origin
under last rays of soft dorsal; ventrals pointed, scarcely reaching vent, 1.9 in head; pectoral reaching
vertical of vent, 1.7 in head.

Color in alcohol, pale yellowish brown, darkest above; a darker brownish band extending from
upper edge of opercle along side, just above lateral line, to posterior edge of soft dorsal; another
broader, more distinct brown band from tip of snout through eye and along middle of side to base of
caudal fin, covering lateral line on caudal peduncle; caudal peduncle at base of caudal fin with a
broad dusky crossbar, usually darkest on upper half, sometimes obscure, sometimes with a darker
blotch or spot in the upper portion; upper parts of head covered with fine dark brown punctulations;
lower jaw similar, but somewhat paler; membranes of anterior 2 or 3 dorsal spines black, others
finely punctulate; soft dorsal pale at base, above which is a broad indistinct dark crossband, the
color confined chiefly to the interorbital membranes, this color extending to near tip of last rays; outer
part of soft dorsal pale; anal similar to soft dorsal, the black bar narrower and nearer base of fin, rest
of fin white; caudal dusky on membranes of outer 1 or 2 rays, the fin otherwise white, with a few fine
punctulations on the interradial membranes; ventrals pale; distal parts of the first and second rays
and their connecting membrane black; pectoral pale; axil and base of pectoral somewhat dusky.

Color in life (field No. 198, O. P. J.), pale red; 2 longitudinal pearly lines on body; first dorsal
with a dusky olivaceous anterior border; white lines along fourth, fifth, sixth, and seventh spines, the

FISHES FROM THE HAWAIIAN ISLANDS

181

membrane olivaceous; second dorsal with many white and some olivaceous spots; anal with a dusky
line along base, the distal part red; base of caudal dusky, rest of fin pale red; ventral with a white
spot near tip; pectoral pink; iris yellow.

Another example (field No. 03499) was coppery brown when fresh, with trace of dusky band along
side; a faint black bar at base of caudal, forming a black spot above end of lateral line; some dusky on
opercle; first dorsal dusky; second dorsal brownish red with some dark; anal same with a basal flesh-
colored bar below it; caudal reddish brown; ventrals same, with first ray pinkish and dusky behind
it; some dusky on opercle.

This species reaches a length of about 6 inches. It was obtained by Garrett in the Hawaiian,
Society, and Paumotu islands. Our collections contain numerous specimens from Honolulu and Hilo.
We have examined also 12 specimens in the collection made by Dr. O. P. Jenkins.

This species closely resembles Apogon menesemus, from which it differs chiefly in coloration; the
black caudal crescent, which is such an excellent distinguishing mark in A. menesemus, is wholly absent
in this species; moreover, the 2 silvery lateral bands, which become dark brown in spirits, are not
found in A. menesemus; and the black on the anal and soft dorsal is less conspicuous in A. snyderi. It
belongs to the subgenus Pristiapogon of Klunzinger, having both limbs of the preopercle serrate.

This species is figured by Bleeker, Day, and Gunther, the figures of Bleeker and Gunther being
colored. The best figure is that of Gunther in Fische der Siidsee, who calls it Apogon frenatus, but the
species originally thus named seems to be quite different, as Bleeker has already noticed.

Named for Mr. John 0. Snyder, assistant professor of zoology in Stanford University.

Our collection contains the following specimens of this species:

Field

No.

Length.

Locality.

Final disposition of specimen.

Field

No.

Length.

Locality.

Final disposition of specimen.

Inches.

Inches.

03072

5.25

Honolulu.

Type, No. 50640, U. S. N. M.

13

3.4

Honolulu.

03065

3.8

do

Cotvpe, No. 50641, TJ. S. N. M.

14

3

do

03067

4.8

do ....

Cotype, No. 7459, L.S. Jr. Univ.

21

2

Hilo

Mus.

16

3.25

Honolulu.

Cotype, No. 9807, Mus. Ind.

02941

4.5

do ....

Do.

Univ.

03079

4.4

do

Cotype, No. 2709, U. S. F. C.

05145

4.8

Hilo

Cotype, No. 9808, Mus. Ind.

03215

4.6

do

Cotype, No. 2710, U. S. F. C.

Univ.

03216

4.5

do

Cotvpe, No. , M. C. Z.

05146

4.8

do

Cotvpe, No. 3956, Field Col.

03217

4.25

do

Cotvpe, No. , M. C. Z.

Mus.

03218

4.8

.„...do

Cotype, No. 2288, Am. Mus.

05147

4.6

do

Cotype, No. 3957, Field Col.

Nat. Hist.

M us.

03219

4.75

do

Cotype, No. 2289, Am. Mus.

05148

4.75

do

Cotype, No. 1493, Cal. Ac. Sci.

Nat. Hist.

05149

4.4

do

Cotvpe, No. 1494, Cal. Ac. Sci.

02499

5

do

Cotype, No. 24217, Ac. Nat.

05150

4.75

do

Cotype, No. 1694, Bishop Mus.

Sci. Phila.

05151

4.5

do ....

Cotype, No. 1695, Bishop Mus.

1

4.8

do

Cotype , No. 24218, Ac. Nat.

O. P. J.

Sci. Phila.

198

4. 75

Honolulu.

2

4.75

do

088

5

do ....

3

4.75

do

411

5

do ....

4

4.5

do

409

5.4

do

5

4.5

do

406

4. 75

do ....

6

4.25

do

408

4.5

do

7

4.5

do

417

4.4

do

8

4

— do

« 428

4.25

do

9

4

do

« 680

4.25

do

10

3.5

do

6 430

4

do

11

3.8

....do ....

6 431

3.75

do ....

12

3.5

....do ....

b 673

5.5

do

a Dr. Wood. b Jordan & Snyder.

Apogon frenatus , Gunther, Fische der Siidsee, I, 19, taf. 19, fig. A, 1873 (Hawaiian, Society, and Paumotu islands); Stein-
dachner, Denies. Ak. Wiss. Wien, lxx, 1900, 484 ( Honolulu); not Apogon framatus Valenciennes, Nouv. Ann. Mus.
Hist. Nat. 1832, 57, pi. 4, fig. 4, nor of Klunzinger.

Family PRIACANTHID7E. The Catalufas.

25. Priacanthus alalaua Jordan & Evermann, new species. “Alalaua.”

Plead 3.2 in length; depth 2.65; eye 2.4 in head; snout 3.6; maxillary 2; interorbital 3.8; D. x,
14; A. in, 15; scales 13-85 to 90-45, 70 pores; Br. 6; gillrakers, about 22 on lower arm.

Body short, deep, compressed, ovate; upper profile of head nearly straight; snout very blunt;
mandible prominent, produced ; mouth very oblique; teeth small, sharp, in bands on jaws, vomer and

182

BULLETIN OE THE UNITED STATES FISH COMMISSION.

palatines; tongue rounded, free in front; maxillary reaching almost to front margin of pupil, its greatest
width 2 in eye; edge of preopercle finely serrate, with a sharp, flat, serrated spine directed backward
at angle; margins of interopercle, subopercle, and opercle entire; opercle with an obscure flat spine;
interorbital space slightly convex; eye very large, its lower edge a little above base of pectoral and in
line with axis of body; nostrils small, close together, the anterior with elevated rim; posterior nostril
oblong, with broad flap; gillrakers rather slender, about 22 on longer arm of first arch, longest about
3 in eye; origin of spinous dorsal over upper base of pectoral; dorsal spines rather uniform, the
longest about equal to orbit; soft portion of dorsal somewhat elevated, rounded, fourth ray 1.7 in
head; anal spines rather stronger than those of dorsal, third the longest, 1.1 in orbit; soft portion of
anal similar to that of soft dorsal, rays of about equal length; caudal truncate, the middle rays slightly
greater than orbit; pectoral short, bluntly pointed, not reaching tip of ventral, length 1.4 in head;
ventrals longer, just reaching base of second anal spine, their length 1.2 in head; ventral spine about
1.25 in longest ray, or 1.7 in head; scales small, firm and rugose, those of lateral line somewhat
enlarged; entire head, as well as body, densely scaled; lateral line rising abruptly for 6 or 7 pores
from gill-opening, thence concurrent with back to caudal peduncle.

Color in life, silvery, light olive above, somewhat flushed with red in irregular blotches; chin red;
spinous dorsal olive-yellowish, especially on edge; ventrals black, rays whitish; fins unspotted.

Young of 4 inches in length are dirty gray, browner above, with no trace of red in life; some brown
spots along lateral line; fins dusky, anal and ventral darkest; iris a little brownish red.

Color in alcohol, plain yellowish-white; spinous dorsal and anal somewhat dusky; ventral mem-
branes black, the rays white, other fins pale yellowish-white. In some examples the color is much
more flushed with red, especially above; the red paler and more evanescent than in the other species;
fins red, unspotted; the spinous dorsal edged with golden; upper lip golden; ventral membrane
black, pectoral pale.

There seems to be but little variation in this species; the younger individuals appear to be more
brightly colored or with more evident wash of red than was shown in the type.

We have examined the following specimens:

Field

No.

Length.

Locality.

Final disposition of specimen.

01170

Inches.

8.25

Honolulu

Type, No. 50643, U. S. N. M.

03395

7

do

Cotype, No. 7461, L. S. Jr. Univ. Mus.

03420

6

do

Cotype, No. 2712, U. S. F. C.

Family LUTIANIDdi. The Snappers.

Genus BOWERSIA Jordan & Evermann, new genus.

Bowersia Jordan & Evermann, new genus of Lutianidx (violescens).

Body long, rather slender and moderately compressed; top of head evenly rounded, the supra-
occipital crest extending forward on cranium; jaws equal, lower not projecting; bands of villiform
teeth on both jaws, the outer series sonrewhat enlarged and canine-like; villiform teeth on vomer,
palatines, and tongue; maxillary slipping for its entire length under the rather broad preorbital; eye
large; opercle entire, ending in 2 flat, obscure spines, the space between them deeply emarginate,
but filled by soft membrane; preopercle scarcely dentate; dorsal fin continuous, the last ray produced
nearly twice length of preceding one.

This genus is related to Apsilus, with which it agrees in the presence of villiform teeth on the
vomer and palatines, but from which it differs in having well-developed teeth on the tongue, and in
the produced last dorsal and anal ray. Two species are known.

We take much pleasure in naming this new genus for the Hon. George M. Bowers, United States
Commissioner of Fish and Fisheries, in recognition of his active and intelligent interest in promoting
scientific work, especially the investigation of the aquatic resources of the Hawaiian Islands.

a. Scales rather large, about 60 in lateral line: preorbital broad, 7.75 in head violescens.

aa. Scales smaller, about 68 in lateral line; preorbital narrow, 10 in head ulaula.

FISHES FROM THE HAWAIIAN ISLANDS.

183

26. Bowersia violescens Jordan & Evermann, new species. “Opakapaka.”

Head 3.25 in length; depth 3.5; eye 4.4 in head; snout 3; maxillary 2.6; mandible 2; inter-
orbital 3; preorbital 7.75; scales 8-60-15; D. x, 10; A. in, 8; Br. 7; gillrakers 5 + 14.

Body long, rather slender, moderately compressed, tapering gradually into the rather long caudal
peduncle; head large, longer than deep; snout moderate, rather bluntly conic; mouth large, maxillary
reaching anterior third of pupil, slipping for its entire length under the thin edge of the rather broad
preorbital, the width of its tip 2 in eye; mandible strong, but not projecting; broad bands of villiform
teeth on jaws, vomer, palatines, and tongue, the outer series in the jaws slightly enlarged and canine-
like; eye large, its lower edge in line with axis of body; interorbital broad, gently convex; anterior
profile but slightly curved from tip of snout to nape, thence more strongly arched to origin of dorsal,
descending in a long, low curve to caudal peduncle; ventral outline but slightly convex; caudal
peduncle rather long, 2 in head, its least width about 1.6 of its least depth, which is 1.8 in its length,
measured from base of last dorsal ray to base of supporting caudal rays; gillrakers few, rather strong
and short, the longest about 2.6 in eye; opercle smooth, ending in 2 flat, obscure spines (more strongly
developed in each of the cotypes).; preopercle obscurely serrate at the angle (more distinctly so in the
cotypes) ; fins moderately developed, the dorsal fin continuous, without notch, its origin over base of
pectoral and equally distant from tip of snout and base of fourth ray, length of entire base of fin and
to tip of last ray twice length of head; first dorsal spine moderately short, closely bound to the second,
whose length exceeds it by about one-half; seventh dorsal spine longest, its length equal to that of
snout; last dorsal ray produced, its length about 1.7 times that of the preceding; anal similar to soft
dorsal, its origin under base of third or fourth dorsal ray; first anal spine very short, third longest and
strongest, its length equaling diameter of eye; last anal ray produced, its length equaling that of produced
dorsal ray; caudal rather widely forked, lobes about equal, their length, measured from base of first sup-
porting ray, equaling head; ventrals pointed, their tips not reaching vent, length 1.4 in head; pectoral
long, slightly falcate, the tip about reaching tips of ventrals, its length about 1.2 in head; scales large,
deeper than long and rather loose; cheek and opercles scaled, 5 rows on cheek; a large bony humeral
scale, from which extends to nape a series of somewhat modified scales, in front of which is a patch
of ordinary scales; lateral line complete and well developed, beginning at lower edge of humeral scale
and following curvature of back to base of middle caudal rays; the pores little or not at all branched.

Color in life (field No. 03404), light rosy olive, with violet shades, pale below; center of each scale
of hack shining violet; dorsal reddish flesh-color, its base anteriorly yellowish olive; caudal flesh-
color, rosy along the edges; anal similar, its edge light lavender gray; ventrals pale, shaded with light
orange; pectoral flesh-color, violaceous at base; snout violet, iris light yellow. A flesh-colored vio-
laceous fish without color markings anywhere. Another specimen (field No. 03417) freshly dead, had
the body, head, and caudal light rosy; ventrals white; outer margin of spinous dorsal golden, the
membranes with irregular golden areas; pectoral and anal not distinctly colored; iris yellow.

Color in alcohol of type (field No. 03018), above dusky silvery, bases of scales brown; sides and
under parts silvery, with pale greenish-yellow tinge; top of head somewhat olivaceous, sides rusty
silvery; axil of pectoral dusky; fins all pale or yellowish-white.

This species reaches a length of about 2 feet and is an important food-fish.

Only 4 specimens were secured:

Field

No.

Length.

Locality.

Final disposition of specimen.

03018

05040

Inches.

24

Honolulu

do

Type, No. 50660, U. S. N. M.

Cotype, No. 7473, L. S. Jr. Univ. Mus.
Cotype, No. 2721, U. S. F. C.

Cotype, No. 9813, Ind. Univ. Mus.

03404

do

03417

do

1 '

27. Bowersia ulaula Jordan & Evermann, new species. “Ulaula.”

Head 3.6 in length; depth 3.8; eye 3.8 in head; snout 3.8; maxillary 2.9; mandible 2.4; inter-
orbital 3.6; preorbital 10; scales 8-68-14; D. x, 11; A. in, 8; Br. 7; gillrakers 21 + 5.

Body long and slender, the dorsal outline in a low, gentle curve from tip of snout to base of caudal,
the ventral outline but gently convex; head moderate, bluntly conic; snout rather short; mouth mod-
erate, somewhat oblique, the jaws equal; maxillary moderate, slipping for its entire length under the

184

BULLETIN OF THE UNITED STATES FISH COMMISSION.

narrow, thin preorbital, its width at tip 2.8 in eye; bands of villiform teeth on vomer, palatines, tongue
and jaws, those of outer series in the latter scarcely enlarged; opercle ending in 2 obscure, flat spines,
the space between them deeply emarginate but filled by membrane; preopercle rather distinctly ser-
rate, the teeth very short; eye rather large, its lower border in line with axis of body; preorbital very
narrow, much narrower than in B. violescens; interorbital space narrower than in the preceding species,
slightly convex; caudal peduncle long, its length from base of last dorsal ray to first supporting rays
of caudal 1.7 in head, its least width about 2.1 in its least depth, which is 2.1 in its length; gillrakers
rather numerous, close-set, the longest about 2.2 in eye; fins moderately developed, the dorsal contin-
uous, without notch, its origin slightly behind base of pectoral and equally distant between tip of snout
to base of fifth or sixth dorsal ray; head 2 in distance from origin of anal to middle of last dorsal ray;
first dorsal spine rather short, about 1.9 in length of second; fifth dorsal spine longest, its length equal
to distance from tip of snout to pupil; last dorsal ray produced, its length about 1.8 times that of the
preceding; anal similar to soft dorsal, its origin under base of third dorsal ray; first anal spine very
short, the third longest, its length 1.2 in diameter of eye; soft anal similar to soft dorsal, the last ray
produced and of equal length with that of dorsal; caudal densely scaled and widely forked, lobes equal,
their length, measured from base of first supporting rays equaling that of head; ventrals not pointed,
their tips not reaching vent, their length 1.6 in head; pectoral long, slightly falcate, its tip reaching
vent and much beyond that of ventral, its length equaling that of head; scales rather small, closely
imbricated, deeper than long, their edges finely ciliated; cheek and opercles scaled, 6 rows on cheek;
a large bony humeral scale from which extends a series of modified scales to nape, and in front of
which is a patch of ordinary scales; lateral line complete and well developed, beginning at lower edge
of humeral scale and following contour of back to base of middle caudal rays, the tubes little branched.

Color in alcohol, brownish or purplish olivaceous above, paler on side; under parts nearly plain
white; each scale of back and upper part of side with a darker brown spot, these forming indistinct rows,
about 6 above lateral line; side below lateral line with less distinct horizontal lines; upper parts of head
olivaceous brown, lower parts paler, spines of dorsal fin purplish, the membranes white, purplish at tips;
soft dorsal with rays whitish, membranes purplish; caudal slightly dusky, other fins plain whitish.

This species is related to B. violescens, from which it differs chiefly in the shorter snout, larger
eye, shorter maxillary, shorter mandible, narrower interorbital space, decidedly smaller scales, more
numerous gillrakers, and more posterior insertion of dorsal fin. Only one specimen known, type
No. 50661, U. S. N. M. (field No. 01104), 14.25 inches long, from Hilo, Hawaii Island.

28. Etelis evurus Jordan & Evermann, new' species.

Head 3.2 in length; depth 3.6; eye 3 in head; snout 3.9; maxillary 2.2; interorbital 3.6; D. x, 11;
A. in, 8; scales 5-50-11; Br. 6; gillrakers 15 -f- 6, longest about 2 in eye.

Body rather long, tapering, moderately compressed; dorsal outline slightly convex, ventral out-
line nearly straight; head considerably longer than deep, compressed, subconic; snout bluntly pointed,
less than eye, equal to portion of eye anterior to posterior edge of pupil; mouth large, oblique; small
bands of villiform teeth on vomer, palatines, and anterior part of each jaw; a single row of small,
wide-set, slender canine teeth on the outer edge of each jaw, those in upper jaw slightly larger and
more wide-set; a single larger canine tooth on the side of each jaw in front, those in the upper jaw7 the
larger; maxillary extending to middle of pupil; eye very large, its lower edge slightly below axis of
body; preopercle finely serrate; opercle with 2 broad, flat spines, not produced, the upper rather
obscure; fins moderately developed; origin of spinous dorsal slightly posterior to base of pectoral, its
distance from tip of snout equaling that to base of sixth dorsal ray; dorsal fin deeply notched, almost
divided; first dorsal spine short, its length but slightly greater than diameter of pupil; third dorsal
spine longest, 2.1 in head; ninth spine short, its length 2.75 in third; soft dorsal not elevated, the rays
about equal, the last 1.75 in third spine; anal similar to soft dorsal, the first spine very short, the third
about 1.8 in third dorsal spine, last anal ray about equal to last dorsal ray; caudal deeply notched, the
lobes much produced, the upper the longer, its rays greatly exceeding length of head, or about 2.4 in
body; ventrals long, but not reaching vent by a distance equaling half diameter of pupil, their length
1.5 in head; pectoral long, reaching vent, the upper rays somewhat produced, their length 1.2 in head;
scales moderate, firm, covering body, nape, opercles, and breast; a large humeral scale; lateral line
beginning at lower edge of humeral scale and following contour of back to base of caudal fin.

Color in life, of a specimen (field No. 03481) 14 inches long, brilliant rose-red, the side from level
of eye abruptly silver, with rosy shades; snout, jaws, eye, and inside of mouth red; fins all rose-color,

FISHES FROM THE HAWAIIAN ISLANDS.

185

the dorsal and caudal bright; ventrals and anal pale, the former washed with red on center; axil
pale pink; pectoral pale rosy.

Color in alcohol, uniform yellowish white, paler below; fins all pale yellowish white, the caudal
lobes somewhat dark.

This species is related to Etelis oculatus of the West Indies, from which it differs in the somewhat
larger scales, much longer caudal lobes (9.5 times length of middle rays instead of 4 times, as in E.
oculatus), and larger eye. From E. carbunculus Cuvier & Valenciennes, from Isle de France, it seems
to differ in having only 16 instead of 20 scales in a transverse series and in the coloration.

This species, one of the handsomest of all Hawaiian fishes, is thus far known only from Hilo,
Hawaii, in the market of which we obtained 13 fine examples, measurements of which are given in
the following table:

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Inches.

Inches.

03482

12. 5

Hilo....

Type, No. 50662, U. S. N. M.

04086

13

Hilo....

Cotype, No. 24224, Ac. Nat.

04082

13. 25

. ..do

Cotype, No. 2722, U. S. F. C.

Sci., Phila.

0:3481

14

... do

Cotype, No. 50663, U.S. N. M.

04087

12.75

...do ....

Cotype, No. 3958, Field Col.

04080

15

...do ....

Cotype, No. 7474, L. S. Jr.

Mus.

Univ. Mas.

05287

13

...do ....

Cotvpe, No. 9814, Ind. Univ.

04081

15. 25

. . .do

Do.

Mus.

04083

15.5

. . .do

Cotype, No.1701 Bishop Mus.

05288

11.6

...do ....

Cotype, No. 1498, Cal. Ae.Sei.

04084

12. 75

...do ....

Cotype, No. , M. C. Z.

05289

12 6

. . .do

04085

16.5

...do

Cotype, No. 2296, Am. Mus.

Nat.. Hist.

Family KYPHOSID^E. The Rudder-fishes.

29. Sectator azureus Jordan & Evermann, new species.

Head 4 in length; depth 3; eye 5 in head; snout 3.65; maxillary 4; interorbital 2.4; D. xi, 15;
A. hi, 13; scales 14-81-20.

Body elongate, ovoid, greatest depth about at tip of pectoral; head slightly longer than deep,
compressed; snout very bluntly convex; jaws about equal, maxillary not reaching front of eye; mouth
small, horizontal; teeth very small, compressed, in a single series in each jaw; minute villiform teeth
on vomer, palatines and tongue; tongue broad, rounded and free in front; preopercle entire, posterior
edge very oblique; lower edge of eye on a line with upper base of pectoral, posterior margin well in
front of middle of head; interorbital broad, strongly convex, a deep groove in front of eye to nostril;
caudal peduncle rather long, 1.9 in head; origin of spinous dorsal slightly in front of base of ventrals,
well behind pectoral, its distance from tip of snout slightly greater than depth of body; longest dorsal
spine 3 in head, last dorsal ray elongate, being one-fourth longer than other rays, its length 3.4 in head;
third anal spine longest, 4.9 in head; first anal ray longest, 3.4 in head; base of anal 1.8 in baseof dorsal;
caudal deeply forked, lower lobe the longer, 3.5 in body; pectoral short, slightly longer than ventrals,
1.8 in head, the spine more than half length of longest ray; scales cycloid, present on head except on
jaws and in front of eye, very minute on all the fins except ventrals; lateral line concurrent with dorsal
outline; peritoneum dark gray.

Color in life, dark steel-blue, becoming paler below; a definite deep blue stripe from snout below
eye widening on opercle, and thence straight to center of base of caudal; below it a narrow bright
golden stripe from angle of mouth to lower part of caudal, and then a fainter blue stripe below this;
a blue stripe from eye to upper part of gill-opening, interspace shaded with green; a deep blue stripe
from upper part of eye along each side of back to base of upper caudal lobe; upper fins dusky golden
or olivaceous; ventrals yellow; anal and lower lobe of caudal dirty golden; pectoral translucent.

Color in alcohol, deep steel gray, brown above, each scale with a very pale spot, the edge pale;
lower surface whitish silvery; a pale streak of gray behind eye to edge of opercle; dorsal fin gray
brown like the back; caudal and pectoral whitish; inside of ventrals dusky orange; ventrals and anal
dusky; inside of pectoral blackish brown.

Type, No. 50664, U. S. N. M. (field No. 03363), a specimen 15.25 inches long, taken off the shore
near Heeia, Oahu Island.

This species must be very rare, being unknown to the fishermen and only the single specimen
having been obtained by us.

186

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Family MULUDtE. The Surmullets.

30. Mulloides flammeus Jordan & Evermann, new species.

Head 3.6 in length; depth 4; eye 4.3 in head; snout 2.25; interorbital 3.5; maxillary 2.6;
mandible 2.1; shortest distance from eye to upper edge of maxillary 1 in eye; D. vii-9, longest dorsal
spine 1.75 in head, longest dorsal ray 2.6; A. 7, longest ray 2.7; scales 3-41-6; pectoral 1.5; ventral 1.4.

Body oblong, not much compressed; head heavy, broad, the interorbital space broad and slightly
convex; snout rather long and pointed, not abruptly decurved; mouth rather large, somewhat
oblique, the lower jaw but slightly included; maxillary broad, slipping for most of its length under
the thin preorbital, its tip not reaching orbit by diameter of pupil; eye rather large, high, slightly
posterior; gillrakers 18 4- 7, the longest about 2 in eye, serrate; opercular spine obscure in adult, more
plainly developed in the young;* origin of dorsal a little nearer posterior base of soft dorsal than tip of
snout; distance between dorsals considerably less than snout, about 2.6 in head; anal similar to soft
dorsal, its origin somewhat more posterior; ventrals rather long, reaching slightly beyond tip of
pectoral; caudal deeply forked, the lobes equal, about 1.2 in head.

Color in life (field No. 03459), bright rose-red, with 5 broad crossbands of darker clear rose,
which vanishes very soon after death; a very faint yellow lateral streak, with yellow shades on scales
below; lower side of head rose, snout and lips very red; 2 wavy golden streaks from below eye to
angle of mouth, lower conspicuous; first dorsal clear red; second dorsal deep red on the lower half,
fading above; caudal deep red at base, fading outward; anal pink, pectoral light yellow; ventral
creamy red; barbels red, paler toward tip; iris silvery.

A color note on specimens, field Nos. 03054 and 03055, says that they were rosy in life.

Color in alcohol, pale dirty olivaceous above, yellowish white on sides and belly; head yellowish
olive above, pale on cheek and below; a yellowish band from snout under eye; fins all colorless, the
spinous dorsal slightly dusky, all with slight yellowish tinge; ventrals with the middle membranes
blackish. Smaller examples show considerable rosy on the sides, indicating that the fish in life was
probably red or rosy in color.

This species somewhat resembles Mulloides auriflamma, from which it differs in the smaller eye,
larger, more oblique mouth, longer maxillary, the longer, less decurved, more pointed snout, and fewer
gillrakers. It bears some resemblance to M. pflugeri, but has the eye larger and the snout longer and
more pointed. Compared with M. samoensis, it has a much larger and more oblique mouth, and a
considerably longer maxillary, as well as a different coloration. It does not agree with any of the
plates of Day, Gunther, or Bleeker, nor with any current descriptions. In life its banded coloration
gives it a very handsome appearance. It is found in deeper water than most of the other species.

M. flammeus seems to be fairly abundant and is represented in our collections by the following 9
specimens, the first of which is taken as the type and all the others as cotypes:

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Field

No.

Length.

Locality.

Final disposition of speci-
men.

03740

Inches.

9.5

Kailua . . .

Type, No. 50665, U. S. N.M.

03995

Inches.

11.25

Hilo

Cotype, No. 24225, Ac.

03029

03054

6

6.5

Honolulu.
do

Cotype, No.1702, Bishop Mus.
Cotvpe, No. 7475, L. S. Jr.

03934

9

do

Nat. Sci. Phila.
Cotype, No. 9815, Ind.

03055

6

do ....

Univ. Mus.

Cotvpe, No. , M. C. Z.

03937

12

do ....

Lfniv. Mus.

Cotvpe, No. 3953, Field

03778

03459

9. 75
10

do

Hilo

Cotype, No. 2723, U. S.F. C.
Cotype, No. 2297, Am. Mus.
Nat. Hist.

Col. Mus.

31. Pseudupeneus chrysonemus Jordan & Evermann, new species.

Head 2.8 in length; depth 3.4; eye 5.3 in head; snout 1.7; interorbital 3.5; maxillary 2.3; D.
vm-9; A. i, 7; scales 3-30-7.

Body slender, not greatly compressed, the back gently and rather uniformly elevated from tip of
snout to dorsal; ventral outline slightly convex; head moderate; snout long; bluntly pointed; mouth
moderate, slightly oblique, the lower jaw included; maxillary broad at tip, falling short of vertical of
orbit by diameter of pupil; interorbital space convex; eye small, in posterior half of head; teeth rather
large, in a single band in each jaw; barbels long, 1.2 in head, reaching nearly to base of ventrals;
opercular spine small; fins rather large; third dorsal spine longest, 1.5 in head, or equal to distance

FISHES FROM THE HAWAIIAN ISLANDS.

187

from tip of snout to middle of pupil, third ray longest, 3.2 in head; base of spinous dorsal 1.4 in third
spine; base of soft dorsal 1.4 in longest spine; origin of spinous dorsal nearer last dorsal ray than tip of
snout by longitudinal diameter of pupil; distance between dorsals 1.5 in eye; length of caudal peduncle
1.5 in head; pectoral long, pointed, slightly falcate, 1.4; ventrals slightly longer, 1.3; last anal ray 2.9,
equal to base of tin; caudal shallowly forked, lobes 1.3 in head, middle rays 2.75 in upper lobe; scales
finely ctenoid and obscurely dendritic; lateral line concurrent with the back, the pores with few
branches, the number usually not exceeding 5 or 6; 2 scales between the dorsals, 8 on dorsal side of
caudal peduncle; peritoneum somewhat silvery.

Color when fresh, deep scarlet red, especially a shade from snout through eye toward tail; first
dorsal plain scarlet, second paler golden with oblique stripes of scarlet and yellow edge; caudal orange;
reddish at base, yellowish at tip; anal like second dorsal; pectoral pale orange; ventrals deep red;
barbels bright yellow; iris red. In life, a pale streak backward from eye to middle of side parallel
with back; side with 2 blotches of deep red; a row of dark spots along bases of both dorsals; young of
3 inches, from the rock pools, in life, dark olive green above with a dark olive streak along lateral line
and 3 dark shades under first dorsal, second dorsal, and back of caudal peduncle; tip of first dorsal
cherry-red, edged with white; second dorsal and caudal translucent, scarcely reddish; ventrals and
anal bright cherry-red, former mesially dusky; barbels golden.

Color in alcohol, pale yellowish; each scale below dorsal with brownish edgings, generally most
distinct in young and often entirely disappearing with age; a series of smaller obscure spots along
median line from opercle to tip of pectoral; sides and under parts with faint traces of rosy.

This species may be known by the series of dusky blotches along each side of the dorsal fin and
by the simple structure of the lateral line. In life it is at once known by its golden barbels.

The above description based upon a specimen (field No. 03929), 8 inches long, obtained at Honolulu,
in 1898, by Dr. Wood. We have examined 4 other specimens of approximately the same size obtained
at the same time, and numerous examples collected by us at Honolulu and Hilo.

The following is our list of specimens of this species:

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Inches.

Inches.

03476

6.75

Hilo

Type, No. 50666, U. S. N. M.

03989

<S. 25

Hilo ...

03920

4.5

Honolulu.

Mus. Nat. Hist.

03928

7

Hilo

03990

7.5

03930

8.5

do

Nat. Sci. Phila.

03931

6.5

do

03991

8

do

Cotvpe, No. 9816, Ind.

03932

7

do....

Univ. Mus.

03933

8.5

do....

03992

8

do

Cotype, No. 3955, Field

03935

8.5

do

Col. Mus.

03936

8.5

do

03993

8.5

do

Cotvpe, No. 1499, Cal. Ac.

03938

5.25

do

Sci.

03939

4.75

do

03994

7. 75

do —

Cotype, No. 60676, U. S.

03983

5

do

Cotype, No. 7476, L. S. Jr.

N. M.

Univ. Mus.

04005

4

Honolulu.

03984

5

do ....

Cotvpe, No. 7476, L. S. Jr.

04006

5

do....

Univ. Mus.

O. I'. J.

03985

5. 25

do

Cotype, No.2724,U. S. F. C.

03929

a 8

do

03986

6

do....

Cotype, No.2725,U. S. F. C.

a 5. 5

do....

03987

6. 75

do ....

Cotype, No. 1703, Bishop

«5. 5

do

Mus.

a 5. 5

do....

03988

6.5

do ....

Cotype, No. , M. C. Z.

65. 75

do

a Collected t>y Dr. Wood, 1898. 6 Collected by Jordan & Snyder in 1900.

32. Upeneus arge Jordan & Evermann, new species. “Weke” or “ We Ice Puco.”

Head 3.75 in length; depth 4.1; eye 5 in head; snout 2.25; interorbital 3; maxillary 2.3; shortest
distance between maxillary and eye 1.25 in longitudinal diameter of eye; D. vm-9, second spine 1.5
in head; A. ii, 6, longest anal ray 1.9 in head; pectoral 1.5; ventrals 1.45; scales 3-40-7.

Body oblong, compressed, deepest through the anterior base of the spinous dorsal; head moder-
ate, compressed, profile arched from origin of the spinous dorsal to tip of snout, steepest on snout;
snout bluntly rounded; lower jaw included; mouth moderate, slightly oblique; tongue short, rounded
anteriorly, not broad nor thick, and not free; teeth in villiform bands on each jaw and on vomer
and palatines; maxillary moderate, reaching anterior edge of eye, moderately broad and sheathed
for more than half of its length; eye rather small, high, median, adipose eyelid well developed;
barbels not reaching edge of gill-opening; pseudobranchise well developed; gillrakers 16+6, finely

188

BULLETIN OF THE UNITED STATES FISH COMMISSION.

serrate, last 5 or 6 on longer limb very blunt and short, pupil of eye contained 1.5 in longest; spinous
dorsal 1.5 in depth, first 2 spines even, longer than the others and longer than base; distance from
snout to origin of spinous dorsal one-third distance from snout to last scale on caudal; distance between
dorsals slightly less than base of soft dorsal; soft dorsal slightly concave; caudal deeply forked, upper
lobe longer; anal similar to soft dorsal, inserted slightly behind the latter; ventrals reaching slightly
beyond pectoral, rays of pectoral slightly the longer; lateral line concurrent with dorsal outline; scales
large, finely ctenoid; entire body and head scaly.

Color in life, pale green, changing to white below; edges of scales on back and down to lateral
line purplish brown, giving the appearance of 3 rather distinct stripes of purplish brown, with
greenish centers on the scales; side with 2 broad yellow stripes, the upper beginning on opercle at
level of eye and running to caudal just above lateral line, the latter being crossed under soft dorsal;
second beginning on base of pectoral and running to base of caudal just below lateral line, this stripe
less distinct and narrowing posteriorly; opercle bright rosy; top of head dusky; cheek white with
some rosy; lower jaw white; barbels yellow; dorsal fins pale, each crossed by 2 or 3 brownish rosy
bars; caudal white, upper lobe with 4 broad brownish red bars running downward and backward, 1
at base narrow; lower lobe with similar but much broader black bars running upward and backward,
2 of them more distinct than the others; 2 longish dark spots on inner rays; anal, ventrals, and
pectoral pale, ventrals rather pale yellowish; iris yellowish, pink above.

Color in alcohol, above, bluish olivaceous, the side becoming lighter, almost white on belly;
borders of scales dusky ; first dorsal spine with 3 or 4 dark spots, and the upper posterior edge of
membranes with dark spots; soft dorsal with 3 dark spots on anterior edge and similar spots on
upper part of fin; caudal fin with dark bands, upper lobe with about 6, those on lower lobe 4, much
broader; other fins pale.

This species resembles Upeneus vitlatus (Forskal), described from Djidda, Arabia, but the latter
has the belly abruptly deep yellow in life.

This is an abundant and important food fish at Honolulu, where we obtained 10 specimens and
where 4 others were collected by Dr. Jenkins in 1889. It is equally common at Hilo and in Pearl
Harbor. It lives in shallow water along quiet shores, and is known as “ Weke” or “ Weke Puco.”

The following is our list of specimens:

Field

No.

Length.

Locality.

Final disposition of
specimen.

Field

No.

Length.

Locality.

Final disposition of
specimen.

Inches.

Inches.

02999

8.5

Honolulu.

Type, No. 50667, U.S.N.M.

03795

10.25

Honolulu.

Cotype, No. 3954, Field Col .

03019

10

do

Cotype, No. 7477, L. S. Jr.

Mus.

Univ. Mus.

03796

10

do

Cotype, No.9817,Ind. Univ.

0314S

10.5

do

Cotype, No. 2726, U.S.F.C.

Mus.

03288

8

do

Cotype, No. 1704, Bishop

03797

8.5

do

Cotype, No. 1500, Cal. Ac.

Mus.

Sci.

03791

10

do

Cotype, No. M. C. Z.

O.F.J.

03793

9. 25

do —

Cotype, No. 2299, Am. Mus.

135

9

do

8

do

03794

12.5

do

Cotype, No. 24227, Ac. Nat.

8

do —

Sci. Phila.

10.5

do —

XJpeneoides vittatus, Streets, Bull. U. S. Nat. Mus., No. 7, 71, 1S77 (Honolulu); not of Forskal.

Family POMACENTRIDAi. The Demoiselles.

33. Glyphisodon sindonis Jordan & Evermann, new' species.

Head 3.5 in length; depth 1.75; eye 3.4 in head; snout 3.5; maxillary 3.4; interorbital 2.8; D.
xii, 19; A. ii, 15; scales 4-28-9, 22 pores.

Body short and deep, dorsal outline evenly arched from tip of snout to soft dorsal; head deeper
than long, compressed; snout short and conic; mouth small, horizontal, lower jaw slightly shorter;
maxillary reaching to anterior edge of orbit; a single row of small, rather blunt, slightly compressed
teeth on each jaw; preopercle entire, opercle ending in 2 small flat spines, upper very small and
obscure; eye anterior, high, its lower edge above upper base of pectoral; interorbital broad, steep and
convex; fins large, origin of dorsal over base of ventrals, its distance from tip of snout equal to dis-
tance from base of last ray to tip of upper caudal lobe; spines strong and long, first 0.7 of fourth, which
is 1.9 in head and of same length as following spines; middle dorsal rays produced, longest ray 1.25 in

FISHES FROM THE HAWAIIAN ISLANDS.

189

head; anal similar to soft dorsal, longest ray 1.25 in head, second spine longest, 2 in head; caudal
forked, upper lobe the longer; ventrals reaching past vent, outer rays longest, about equal to head;
pectoral broad, upper rays longest, equal to head; scales large, ctenoid, covering entire body and
head except lower jaw and snout anterior to eye; lower limb of preopercle scaled; large scales cover-
ing nearly all of dorsal spines, smaller scales covering as much of soft dorsal and anal and nearly all
of caudal; very minute scales on base of pectoral, none on I'ays of ventrals; lateral line concurrent with
dorsal outline, on 22 scales, ending 3 rows of scales short of posterior base of dorsal, then dropping
3 rows of scales and continuing obscurely on middle of caudal peduncle to base of caudal fin.

Color in alcohol, uniform very dark brown, nearly black; 2 narrow wavy bands of white on side,
first beginning about under fourth dorsal spine and extending under about middle of pectoral, thence
curving slightly backward toward vent, rather indistinct below pectoral; second band beginning under
last dorsal spine and first ray, extending toward middle of anal, rather obscure, indistinct for 2 or 3
scales before reaching anal; fins all black, pectoral slightly lighter than others; a large black ocellated
spot with a narrow white border on back and lower part of soft dorsal, larger than eye, just back of
last white bar.

This species agrees with typical Glyphisodon in all respects except that none of the teeth appears
to be emarginate. It agrees with Chrysiptera in the entire preopercle and preorbital and naked snout,
but differs from the type of that genus in having the teeth in a single series.

The above description based on the type, No. 50669, U. S. N. M. (field No. 04524), a specimen 3.75
inches long, from Honolulu. One other specimen obtained. It is taken as a cotype and is No. 2727,
U. S. F. C. reserve series (field No. 03732), a specimen 2.75 inches long, from Kailua, where it was
first discovered by Michitaro Sindo, for whom the species is named.

34. Pomacentrus jenkinsi Jordan & Evermann, new species.

Head 3.4 in length; depth 1.8; eye 3.3 in head; snout 4; maxillary 3.2; interorbital 2.75; D. xm,
16; A. ii, 13; scales 4-29-11; Br. 4.

Body ovate, deep, compressed, dorsal outline rather steep, evenly curved from tip of snout to soft
dorsal, following edge of scales on spinous dorsal; head deeper than long, compressed subconic; snout
bluntly conic, jaws equal; maxillary reaching anterior edge of eye; mouth small, horizontal; a single
row of close-set, incisor teeth in each jaw; posterior edge of preopercle roughly serrate; opercle ending
in 2 short flat spines, the upper very obscure; interorbital wide, strongly convex; fins rather large;
origin of dorsal over ventral, origin of each equally distant from tip of snout; first 2 or 3 dorsal spines
shorter than others; others about of equal length, shorter than the longest dorsal rays, the median rays
being longest, 1.5 in head; caudal forked, lobes rounded, upper the longer; anal rounded, longest ray
1.5 in head, second spine rather stout and strong, 2.2 in head; ventrals long, reaching vent, 1.1 in head;
pectoral broad, upper rays the longer, 1.2 in head; scales large, finely ctenoid; body and head, except
lower jaw and snout, scaled, scales on top of head small; bases of all the fins except ventrals well
covered with fine scales, those on spinous dorsal larger; lateral line concurrent with dorsal outline to a
line under base of third or fourth dorsal ray, where it drops 3 rows of scales to middle of caudal
peduncle, whence it continues to base of caudal fin, the detached portion little developed.

Color in life, ground color dark drab; central portion of scales olivaceous, each one with black on
lower part of posterior edge forming vertical bands on body; axil black; outer border of dorsal fin,
above scaled part, black; pectoral dusky olivaceous, black at base; ventral and anal black; caudal
dusky with posterior border lighter; iris bright yellow.

Color in alcohol, dark brown, edges of scales darker; a dark stripe on upper edge of membranes
of spinous dorsal, broadest and most distinct anteriorly; rest of dorsal, and caudal and pectoral dark
brownish; ventrals and anal dark, almost black; a black blotch at upper base of pectoral, continuous
with the black axil.

This is a very abundant species among the Hawaiian Islands. Numerous specimens were obtained
at Honolulu in 1889 by Dr. Jenkins, and others by Dr. Wood in 1898 and Dr. Jordan in 1900. Our
own collections, made in 1901, contain numerous SDecimens, the localities represented being Honolulu,
Hilo, and Kailua.

The above description is based chiefly upon a specimen (field No. 04526) 4.8 inches long, obtained
by us at Honolulu.

190

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The field numbers and lengths of a few of our specimens are given in the following table:

Field

No.

Length.

Locality.

Final disposition of specimen.

03331

Inches.
3. 75

Honolulu

No. 7479, L. S. Jr. Univ. Mus.

04516

4. 75

Hilo

No. 7480, L. 8. Jr. Univ. Mus.

04517

4.5

do

No. 3959, Field Col. Mus.

—04518

4.2

do

-No. 50671, U. -S. N. M.

04519

4.2

do

No. 2728, U. S. F. C.

04520

4

do

No. 2729, U. S. F. C.

04521

3.8

do

No. 1705, Bishop Mus.

04522

5.25

Kailua

No. . M. C. Z.

04523

4.8

do

No. 24228, Ac. Nat. Sci. Phila.

-^G4.rV2fi

4.8

4.5

L~No. 50670, IT. S. N. M.

No. 2300, Am. Mus. Nat. Hist.

04527

do

Pomacentrus nigricans , Quoy & Gaimard, Voyage Uranie, Zoo)., 399, 1824 (Sandwich Islands); Cuvier & Valenciennes,
Hist. Nat. Poiss., v, 425, 1830; Gunther, Cat , iv, 34, 1862 (Sandwich Islands) ; not Holocentrus nigricans Laciipede,
Hist. Nat. Poiss., IV, 332 and 367, 1803, locality unknown, collected by Commerson.

Eupomacentras marginatus Jenkins, Bull. U. S. Pish Comm, for 1899 (June 8, 1901), 391, fig. 5, Honolulu (Type, No. 49700,
U. S. N. M., Coll. O. P. Jenkins); not Pomacentrus marginatus Euppell.

Family LABRIDTL The Wrasse-fishes.

35. Lepidoplois strophodes Jordan & Evermann, new species.

Head 2.75 in length; depth 2.75; eye 4.65 in head; snout 3.25; mouth 3.1; interorbital 4; D. xii,
10; A. hi, 12; scales 7-34-13.

Body oblong, compressed; head longer than deep; upper and lower profiles evenly and slightly
convex; snout long, pointed, rounded above; jaws produced, pointed, about equal; mouth large,
maxillary reaching beyond front of eye; teeth strong, forming a sharp cutting edge on sides of jaws,
front of each jaw with 4 large canines; eye rather large, anterior, high in head; posterior margin of
preopercle very finely emarginate; interorbital space rather broad, convex; nostrils small, anterior in
short tube; dorsal spines pungent, longest 3 in head, last 3.5; third anal spine longest, 2.8; third anal ray
1.9inhead; pectoral rounded, 1.7; ventrals pointed, 1.4; caudal broad at base, truncate; caudal peduncle
broad, compressed, its depth 2; scales large, thin, those on front of dorsal, along its base and that of
anal, small ; lateral line concurrent with back, sloping down at caudal, then running straight to its base.

Color in life, pale rosy white; upper parts of the snout, nape, and side to base of about ninth
dorsal spine, lemon-yellow, extending down on side to level of upper edge of pupil; side of head
very pale rosy, 2 irregular broken lines of wine-colored spots across snout and through eye to
posterior edge of opercle, a similar row of 4 oblong spots from angle of mouth downward and
backward to edge of opercle; cheek and side of lower jaw with numerous small irregularly placed
orange spots; side with about 16 brighter rosy longitudinal lines, those above less distinct on
account of the deeper rosy ground color, those below more distinct, the ground color being more
white; side between anal and soft dorsal fins with a broad sooty black spot extending irregularly
upon both fins and fading out upon body anteriorly, the posterior edge being nearly vertical and
well defined; caudal peduncle and base of the caudal fin whitish, with a slight tinge of rosy, a pale
rosy band separating this from the black lateral area; region in front and below the pectoral with
about 4 series of small reddish brown spots; pectoral region and the under parts somewhat bluish;
dorsal fin rich lemon-yellow, the tips of the soft rays whitish, and a small, round, black spot on
middle of membrane of second spine; base of soft rays and last dorsal spines rosy from intrusion
of the rosy wash on side of body; last dorsal rays sooty black at the base from extension of the
black spot on the side; caudal pale lemon-yellow; anal pale rosy in center, lemon on spines and
along tip of fin, base of fin sooty black from intrusion of black spot on side of the body, the black
extending farthest down on the interradial membranes; pectoral very pale rosy; ventrals pale
rosy, the membranes bluish, the tip of second ray blackish.

Color in alcohol (field No. 04291), gray-brown, gradually darker posteriorly; space between soft
dorsal and anal abruptly black, the color extending forward in darker streaks along the rows of scales
and forming a large black blotch on soft dorsal and anal; top of head and space before dorsal abruptly

FISHES FROM THE HAWAIIAN ISLANDS.

191

pale; posterior part of caudal peduncle also abruptly pale; a black blotch on dorsal between second
and third spines, not involving third and fourth, as in L. bilunulatus; dorsal and caudal otherwise
pale; a pale blotch at base of posterior dorsal rays; side with narrow dark brown longitudinal lines,
coalescing posteriorly with the black blotch; 2 narrow brown streaks from lip to front of eye, then
back across side of head above; edged with narrow, darker, wavy lines; a wavy streak from corner
of mouth toward base of pectoral, lower side of head with small brown spots or blotches; ventral fin
mostly dusky.

This species is very close to Lepidoplois bilunulatus, differing chiefly in the dark zone on posterior
part of body and in the smaller size of the dorsal spot. Our specimens are all young, but we have the
young of L. bilunulatus scarcely larger and showing the markings of the adult.

Our collection contains the following specimens of this species, all from Honolulu, where it is not
uncommon :

Field

No.

Length.

Locality.

Final disposition of specimens.

04291

Inches.

4.7

Honolulu

Type, No. 50672, U. S. N. M.

03520

3.8

do

Cotype, No. 2730, U. S. F. C.

03532

4.5

do

Cotype, No. 7481, L. S. .Ir. Univ. Mus.

04292

3.75

04293

3. 75

do

Cotype, No. 3963, Field Col. Mus.

VERRICULUS Jordan & Evermann, new genus.

Verriculus Jordan & Evermann, new genus of Labridi e ( sanguineus ).

Body elongate, subfusiform, compressed, with rather long, pointed snout; snout rather large, with
anterior canines strong, f- to f; posterior canines present; lateral teeth short, confluent in a serrated
cutting edge; cheek and opercle scaly; preopercle entire, both limbs more or less scaly; scales moder-
ate, about 40 in lateral line; lateral line continuous; D. xii, 10; A. in, 12; dorsal spines low, pungent;
soft dorsal and anal not elevated, their bases without scales; caudal subtruncate; pectoral short.

This genus is allied to Verro and Nesiotes. From its nearest relative, Nesiotes, it differs in the
presence of a posterior canine tooth. The single known species is brilliantly colored.

36. Verriculus sanguineus Jordan & Evermann, new species.

Head 2.9 in length; depth 3.5; eye 6.2 in head; snout 3.1; mouth 2.8; interorbital 4.75; D. xii,
10; A. in, 12; scales 5-40-13.

Body elongate, compressed oblong; head long, pointed, conic, its depth 1.7 in its length; eye
small, its posterior margin in middle of length of head; snout long, pointed, rounded; jaws produced,
equal; mouth large, nearly horizontal, corner reaching below front rim of eye; lips thick, fleshy; teeth
strong, those on sides short, close-set, forming a sharp cutting edge on side of jaw; 5 canines in front
of upper jaw, 4 in front of lower, a posterior canine on each side of upper jaw; tongue long, pointed,
free in front; preopercle not serrate; interorbital space broad, convex; nostrils small, anterior in short
tube; dorsal spines strong, sharp-pointed, longest in middle and posteriorly; last dorsal spine 4 in head;
anal spines strong, last spine longest, 3.75; seventh anal ray 3; caudal rounded; dorsal and anal fins
scaled at base; pectoral rounded, 1.9 in head; ventrals short, spine strong, pointed, two-thirds longest
ray, which is 2 in head; caudal peduncle broad, deep, 2.2 in head; scales small, thin, cycloid; head
with very small thin cycloid scales on occiput, cheek, greater part of opercle, behind eye, and on
opercles, otherwise naked; lateral line slightly curved in front, then obliquely down to base of caudal.

Color in life, deep red, edge of upper jaw and lower tip golden; a long stripe from eye along back
to base of caudal golden, with a red shade, a vertical black bar edged with golden above, on opercular
region; a long blackish area covering it from eye to above pectoral, with some blackish before, behind
and above; a black spot at base of caudal; dorsal and caudal golden, first dorsal edged with violet and
with the lower half violet; anal entirely deep blood-red ; ventrals golden ; pectoral reddish, golden at base.

Color in alcohol, very pale brown; a dusky band from snout across back of head and on side,
fading out indistinctly posteriorly; a blackish spot at middle of base of caudal; opercle posteriorly
with black vertical blotch; fins all pale or light brown.

192

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Described from the type, No. 50677, U. S. N. M. (field No. 03489), an example 7.5 inches long,
taken at Hilo, with hook and line, in deep water1 with Elelis evurus, Eteliscus marshi, Erythrichthys
ichlegeli, Antigonia steindarhneri, and Anthicis fuscipinnis.

We have examined only one example, the one described above.

37. Pseudocheilinus evanidus Jordan & Evermann, new species.

Head 3 in length; depth 3.8; eye 4.5 in head; snout 3; preorbital 6.2, interorbital 5.5; D. ix, 11;
A. hi, 9; scales 2-25-6.

Body short, deep and compressed; head long, conic; snout long, sharply conic; anterior profile
rising in a relatively straight line from tip of snout to nape, thence gently convex to base of caudal
peduncle; ventral outline less convex; mouth large, horizontal, below axis of body, gape reaching
anterior line of orbit; upper jaw with 3 pairs of anterior canines, outer strongest, curved outward
and backward; lower jaw with a single pair at tip, similar to inner above; jaws laterally with a single
series of smaller conic teeth; preorbital narrow, oblique; eye high up, its lower border on axis of
body; interorbital space rather broad and flat; depth of caudal peduncle about 2 in head; scales large,
surfaces finely striate; head, nape, and breast with large scales; lateral line following contour of
back until under base of sixth dorsal ray, where it is interrupted, reappearing 2 rows farther down and
continuing on 6 or 7 scales to base of caudal fin; fins rather large; dorsal spines somewhat greater than
eye in length, spines with a sheath of large scales reaching nearly to their tips; soft dorsal and anal
with a lower sheath; soft dorsal elevated, rays equal to snout and eye; anal similar to soft dorsal,
second spine- strongest, nearly as long as snout; anal rays somewhat longer, equaling those of soft
dorsal; caudal rounded, its length 1.3 in head, its base covered with very large, thin scales.

Color in life, according to Mr. Sindo, body dull brick-red; belly and base of anal pale purplish;
about 17 thin, thread-like longitudinal yellowish streaks along side anteriorly; dark greenish blotches
above eye and on snout; a bluish horizontal bar on cheek, below which is a yellow bar; median line
of throat and tip of snout brick-red; edges of opercle and preopercle bright purple; a purple stripe
with reddish edges through middle of dorsal fin, below which the color is dull brick-red, like that of
body, and above which the spinous dorsal is orange-yellow, the margin of the membranes bright
cardinal-red; above the purple streak in the soft dorsal is a bright yellow streak, above Avhich the fin
is cardinal-red, fading gradually upward; dorsal rays purplish; tip of soft dorsal somewhat red; caudal
rays purple, the membranes immediately next to the rays yellow, middle part dull brick-red; anal
> same as caudal; ventrals pale purplish; pectoral pale; iris scarlet-red.

The same specimen after having been in spirits more than a year has the body light brownish-
blue; a pale streak along each row of scales, but no trace of the narrow yellowish streaks above noted;
top of head and upper part of cheek dusky blue; opercle and edge of preopercle rich blue; dorsal, anal,
and caudal fins bright blue, the soft dorsal pale on outer two-thirds, dorsal rays bright blue; ventrals and
pectoral light blue, latter darker blue at base. The color of this specimen in spirits is wholly different
from that which it possessed in life, and it would be difficult to believe that such changes had taken
place except that the specimen was carefully tagged in the field when the color note in life was taken.

Since writing the above, we have noticed similar changes in the Samoan species, P. hexatsenia.
The blue shades are permanent in spirits, while the pink or crimson wash soon vanishes in spirits.

The 17 thread-like streaks, mentioned in Mr. Sindo’ s field notes above, have vanished entirely in
the original type. A number of specimens taken by Mr. Snyder at Laysan, while on the Albatross,
retain these traits, the streaks being almost white, like white threads, covering most of the side
anteriorly. This is a very peculiar color mark, which should well distinguish the species in life.

A single specimen, type, No. 50678, U. S. N. M. (field No. 05757), was taken by Mr. Sindo in
Henshaw’s pool near Hilo, a deep tide-pool in the lava rocks.

38. Hemipteronotus baldwini Jordan & Evermann, new species.

Head 3.25 in length; depth 3; eye 5.75 in head; snout 1.75; maxillary 3; preorbital 2.2; inter-
orbital 4.8; D. ii-vni, 13; A. in, 13; scales 3-27-9.

Body moderately short and deep, greatly compressed; head slightly deeper than long; anterior
profile nearly vertical from mouth to front of eye, sharply cultrate; dorsal outline gently convex, sloping
to the deep caudal peduncle; ventral outline less convex; caudal peduncle very narrow, the depth
2.25 in head; mouth small, horizontal, the maxillary nearly reaching vertical of orbit; the jaws equal,

FISHES FKOM THE HAWAIIAN ISLANDS.

193

each provided anteriorly with a pair of strong curved canines and laterally with a single row of short
close-set conic teeth; lower jaw strong, its outline very convex; preorbital nearly vertical and very
deep; preopercle and opercle smooth, with membranous edges, the latter produced somewhat in a
broad rounded flap; origin of dorsal but little posterior to orbit, far in advance of base of ventrals;
first 2 dorsal spines somewhat removed from third but connected to it by a low membrane, their length
scarcely greater than the gajie of mouth ; remaining dorsal spines short and weak, scarcely equaling
gape; soft dorsal low, the rays slightly longer than the spines; anal similar to soft dorsal, rays some-
what longer; caudal slightly convex, rays 2 in head; outer ray of each ventral somewhat produced,
not reaching vent, the length about 1.9 in head; pectoral broad, the longest rays 1.7 in head; scales
large, thin, smooth, firmly attached, those on breast somewhat reduced; head naked, except about 4
series of small scales extending from eye downward to level of mouth; lateral line curving abruptly
upward from upper end of gill-opening, following contour of back to the scale under third dorsal ray
from last where it drops 3 rows and continues to base of caudal, the pores simple, unbranched.

Color in life (field No. 03123), pale, yellowish white over head and body; a diffuse lemon-vellow
blotch underand above pectoral; a large brownish-black blotch on lateral line under seventh to tenth
dorsal spines; dorsal fin yellowish-white, tip of detached part with a jet-black crescent ( this marking
variable in position, it sometimes being farther posterior), rest of fin faintly mottled with yellowish
and olive, the latter in narrow oblique lines; caudal yellowish white; anal yellowish white, with
narrow, wavy, pale-blue lines, and a large jet-black spot bordered with blue on membrane of last 5
rays; iris whitish.

Color in alcohol, creamy yellowish white; head somewhat orange on cheek and opercles; faint
rosy lines downward from eye to mouth and on preopercle; median line of anterior profile bluish;
middle of back with a large black or brownish black blotch lying on lateral line, beneath which is a
large white blotch under and above pectoral fin ; anterior part of spinous dorsal blackish at edge, the
color ocellated, rest of dorsal yellowish white with narrow purplish cross-lines; anal similar, with a
large jet-black spot on last 4 rays; caudal color of soft dorsal; pectoral and ventrals yellowish white.

Color in alcohol, of one of the female cotypes (No. 03372), pale olivaceous, the general color that
of the male; dorsal with black spots on membranes of second, third to fourth, and eighth spines, the
latter ocellated; a series of about a dozen small black spots back of the dorsal blotch on side above
lateral line; no black spot on anal.

The above description based upon the type, No. 50644, U. S. N. M. (field No. 03414), a male
example, 8.5 inches long, obtained at Honolulu.

Another specimen, also a male (field No. 03371), was in life, livid gray; each scale posteriorly
with a vertical spot of violet; anterior line of profile bright violet; a violet line downward from eye
with a whitish area behind it on cheek; an oblique violet line downward and backward from
opercular flap to behind axil; behind this a vague yellow area, behind which is an ovate white spot,
each scale around which has a vertical bar of bright violet; above this a large black blotch washed
with brick-red; dorsal bluish-gray, the rays posteriorly with an increasing amount of orange, where
the blue is reduced to oblique crossbands, an intermarginal line of violet, a small black spot on last
ray; membranes of second to fourth dorsal spines with a terminal black ocellus; anal pale golden,
with oblique bluish stripes, a large jet-black ocellus bordered with blue on last rays; caudal pale
orange, crossed by bluish lines; ventrals and pectoral pale.

Still another male example (field No. 03004) was described as follows: General color very pale
smoky white, edges of scales pale bluish, beneath seventh to ninth dorsal spines a large blotch, brick-red
above, pale rosy below, all irregularly overlaid with black or brown; median line from tip of snout
to base of first dorsal spine bright blue; a narrow bright blue line downward from anterior part of eye
to angle of mouth; region above pectoral pale lemon yellow, a short oblique pale blue line above base
of pectoral; dorsal pale flesh-color, with short vertical bluish lines, with 3 jet-black spots at tips of first,
second, and fourth spines; anal pale yellowish, a black spot on distal half of last 3 rays; caudal pale,
with obscure bluish cross-lines; pectoral and ventrals white; iris yellowish, red at lower posterior angle.

Another example, a female (field No. 03372), 7.5 inches long, from Honolulu, which is taken as
a cotype, differed in life coloration from the male in lacking the black ocellus on the anal and in
having more violet on the white lateral spot, also more golden before it; violet lines and spots obscure,
but present; 3 to 8 small blackish points above lateral line behind black dorsal blotch; a small black
ocellus on second to third dorsal spines and one on seventh dorsal spine, these wanting in some
females; fins otherwise colored as in the males, but the blue fainter and the orange of dorsal brighter.

F. C. £. 1902—13

194

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Another female example (field No. 03005) differed in color from field No. 03004 only in the
absence of black on the dorsal and anal fins, the paler blue lines on head, the paler caudal fin, and in
having black spots on the back. Another female example (field No. 03271), 7.5 inches long, in life
had the head and body smoky white; a large bluish white spot under tip of pectoral; snout bluish
around border and surrounded by a broad pale yellow space involving nearly all of anterior half of
side below level of eye; a large black spot under fifth to sixth dorsal spines, crossing lateral line and
penetrating yellow of side, nearly reaching white spot; back of this a series of about a dozen small
black specks, scattered along side above lateral line to near end of dorsal fin; median line of snout
and head blue; dorsal pale, with wavy yellow cross-lines, pinkish toward margin; caudal pale; anal
pale, with about a dozen pale yellow crossbars; pectoral and ventrals pale; iris yellow and red.

This beautiful and abundant species is represented in our collection by 41 specimens, 19 of which
are males and 22 females. The differences in coloration of the two sexes are very marked. The
male, in all specimens examined, lias the jet-black spot upon the last rays of the anal, a marking
which is not present in any of the females examined. The female always has a series of small black
specks on the side above lateral line posterior to the large lateral blotch. These markings, the small
black spots on the side of the female and the large jet-black spot on the anal of the male, would
apparently always serve to distinguish the two sexes.

The extent of variation in color among individuals of the same sex is indicated in the color
descriptions given above. We should have added that occasionally there is a small jet-black spot
upon the last rays of the dorsal.

This is one of the most abundant and beautiful species found among the Hawaiian Islands. It
appears to be related to H. melanopus of Bleeker, but differs from it markedly in the presence of the
large black lateral blotch and in the absence of the large red lateral blotch shown in Bleeker’ s figure.

This species is named for Mr. Albertus H. Baldwin in recognition of his paintings of American
and Hawaiian fishes.

Field

No.

Length.

Sex.

Locality.

Final disposition o£ specimens.

Field

No.

Length.

Sex.

Locality.

i

03414

Inches.

8.5

Male ..

Honolulu. . .

Type, No. 50644, U. S. N. M.

05603

Inches.

8.5

Male . . .

Honolulu.

03123

8.8

Male ..

do

Cotype, No. 2713, U. S. F. C.

05604

8.5

do ...

Do.

03372

7.5

Fern . . .

do

Cotype, No. 2714, U. S. F. C.

05605

8.5

do ...

Do.

03371

7.5

Male ..

do

Cotype, No. 7462, L. S. .Tr. Univ. Mus.

05606

7.75

— do ...

Do.

03004

7

do

Cotype, No.3961, Field Col. Mus.
Cotype, No. 3962, Field Col. Mus.
Cotype, No. 7462, L. S. Jr. Univ. Mus.

05609

6. 75

Fem

Do.

03005

6. 25

Fern . . .

do

05627

9.2

Male ...

Do.

03271

7.25

Fem . . .

do

05628

7.6

Fem

Do.

-4)3084

6.5

Fem . . .

do

Cotype, No. 50645, U. S. N. M.

05629

8.2

Male ...

Do.

05438

05589

8.25

7.25

Male ..
Fem . . .

Hilo

05630

8.5

Do

Honolulu...

Cotvpe, No. , M. C. Z.

05631

8. 5

do ...

Do.

05439

7.5

Male . .

Hilo

Cotype, No. 2290, Am. Mus. Nat. Hist.

05632

8.6

Fem

Do.

05591

6.2

Fem . . .

Honolulu. . .

Cotype, No. 2291, Am. Mus. Nat. Hist.

05633

7

do ...

Do.

05465

9.25

Male ..

do

Cotype, No. 24219, Ac. Nat. Sci. Pliila.

06634

7

do ...

Do.

05599

6.8

Fem . . .

do

Cotvpe, No. 24220, Ac. Nat. Sci. Phila.

05635

7

do ...

Do.

05586

8.25

Male ..

do

Cotype, No. 9809, Mus. Ind. Univ.

05636

6.25

do ...

Do.

05600

6.5

Fem . . .

do

Cotype, No. 9810, Mus. Ind. Univ.
Cotvpe, No. 1495, Cal. Ac. Sci.

05637

6.5

do ...

Do.

05587

8.6

Male ..

do

05639

6.25

do ...

Do.

05607

7.4

Fem . . .

do

Cotype, No. 1496, Cal. Ac-. Sci.

05640

6.25

do ...

Do.

05588

8

Male . .

Cotype, No. 1696, Bishop Mus.
Cotype. No. 1697, Bishop Mus.

0564 1

6.2

do . . .

Do.

05608

7

Fem . . .

do

05642

5.75

do ...

Do.

03124

8.75

Male ..

do

O. P.J.

a 624

6.5

Male . . .

Honolulu.

a Collected by Jordan & Snyder.

39. Xyrichthys niveilatus Jordan & Evermann, new species.

Head 3.3 in length; depth 2.4; eye 6.2 in head; snout 1.8; preorbital 2; maxillary 3; interorbital
4.7; I). ii-vii, 12; A. in, 12; scales 3-28-8.

Body short, deep, and very greatly compressed; anterior profile nearly vertical from tip of upper
jaw to front of eye, thence in a parabolic curve to dorsal fin; anterior dorsal outline very trenchant;
body tapering rather evenly from head to caudal peduncle, which is greatly compressed and very
deep, depth at middle equaling preorbital; head short; snout very short and blunt; mouth small, hor-
izontal, the maxillary nearly reaching anterior edge of orbit; jaws equal, each with a pair of strong
curved canines in front, and a single series of smaller, conic teeth laterally, the canines of lower jaw
most prominent and extending in front of upper jaw; eye small, high up; the interorbital space narrow
and trenchant; opercles smooth, without spines or serrations, ending in thin flexible edges; preorbital

FISHES FROM THE HAWAIIAN ISLANDS.

195

vertical and very deep; origin of dorsal fin above posterior line of orbit, far in advance of base of ven-
trals; first 2 dorsal spines somewhat removed but not detached from third, the membrane between
second and third spines moderately notched, length of second spine about 2.7 in head, remaining dorsal
spines subequal, weak, about equal to gape; dorsal rays low, the last few somewhat produced, their
length 3 in head; anal similar to soft dorsal; caudal short, slightly convex, rays about equal to preorb-
ital; outer ray of ventral somewhat produced, not reaching vent, its length equaling depth of preorbital;
pectoral broad, its tip reaching vent, its length equaling distance from snout to edge of preopertie.

Scales large, thin, and with membranous edges, those on breast somewhat smaller; head entirely
naked, except for a few small scales below the eye; lateral line beginning at upper end of gill-opening
following closely the curvature of back to the scale under the last dorsal ray but 2, where it drops 3
scales and continues to base of caudal, the pores simple, rarely branched.

Color in life, grayish; each scale of posterior half of body with a large violet spot, more narrow
and brighter near middle of body, the edge of each scale broadly golden-olive; a large golden area,
anteriorly deep orange, above pectoral and on edge of opercle; behind this a large quadrate pure white
area extending to tip of pectoral; a few scales in golden area with bright violet markings; head shaded
with violet, a bright violet stripe downward from eye to behind angle of mouth; a lunate black area
shaded with red just below front of soft dorsal; spinous dorsal violet-gray, edged with reddish; soft
dorsal golden, with violet vermiculations at base, its edge orange; anal golden, with bluish vermicula-
tions; caudal similar, with the bluish markings; pectoral faintly reddish; ventrals dirty white.

One of the cotypes, a male (field No. 03373) , agreed in life coloration with the type except that
behind the opercle is a golden area with the bright violet stripes across anterior basal part; behind
this a large milk-white patch beyond tip of pectoral; a violet border was around the white and
blackish above the yellow.

Color in alcohol, dirty yellowish white, dusky above; head with some purplish reflections; a thin
purplish line downward from anterior edge of orbit to tip of maxillary, a similar but less distinct line
from humeral region downward to subopercle; a yellowish white blotch on side above base of pectoral,
in the base of which are 2 or 3 small purplish spots; a large white area on middle of side under and
above tip of pectoral, separated from the yellowish blotch by purplish brown on 2 or 3 scales; a black
spot covering the larger part of 3 scales on side above lateral line under base of first 3 dorsal rays, back
at base of last dorsal rays somewhat dusky; anterior portion of dorsal fin dusky olivaceous, soft dorsal,
anal and caudal pale yellowish crossed by narrow, wavy, pale purplish lines; ventrals and pectoral
plain yellowish white.

This handsome fish is rather common about Honolulu.

Field

No.

Length.

Locality.

Final disposition of specimen.

03345

Inches.

9.5

Honolulu

Type, No. 50646, U. S. N. M.
Cotype, No. 3960, Field Col. Mus.

05464

9. 75

do

03373

9.25

do

Cotvpe, No. 2715, U. S. F. C.

05597

8.2

do

Cotype, No. 7463, L. S. Jr. Univ. Mus.

05598

7.4

do

Cotvpe, No. , M. C. Z.

05590

6. 5

do

Cotype, No. 2292, Am. Mus. Nat. Hist.

Family SCAR1D9:. The Parrot-fishes.

40. Scarus jenkinsi Jordan & Evermann, new species.

Head 3 in length; depth 2.5; eye 6.5 in head; snout 2.6; preorbital 4.7; interorbital 3; D. ix, 10;
A. in, 9; P. 13; scales 2-24-7.

Body short, very deep and greatly compressed; head short, nearly as deep as long, snout short
and blunt; mouth small; each jaw with 1 or 2 blunt canines; dorsal and ventral outlines about equally
convex; anterior profile rising rather irregularly from tip of snout to origin of dorsal; caudal peduncle
deep, its least depth 2 in head. Scales large, deeper than long; 2 rows of large scales on cheek and 1
row on subopercle; a row of thin modified scales at base of dorsal and anal; a few very large, thin
scales on base of caudal; lateral line ceasing under last dorsal ray, reappearing 2 rows lower down and
continuing to base of caudal, the pores with 2 or 3 irregular branches; dorsal rays soft and flexible,
not pungent; dorsal spines somewhat elevated posteriorly, longest a little more than 2 in head; first
ventral spine obscure, the others soft and flexible; anal rays somewhat shorter than those of doisal;

196

BULLETIN OF THE UNITED STATES FISH COMMISSION.

caudal shallowly lunate, the outer rays not greatly produced; ventrals moderate, 1.6 in head, not
reaching to origin of anal by a distance equal to two-fifths their length; pectoral broad, 1.2 in head.

Color of a nearly fresh specimen, bright blue-green, brightest on posterior half of body, each
scale broadly edged with reddish brown; lower anterior part of body reddish brown, with traces of
blue-green; top of head brownish red or coppery, a broad deep blue-green band on the upper lip,
extending on side of head to below eye; lower lip with a narrow brighter blue-green band connecting
at angle of mouth writh the one from upper lip; chin with a broad coppery-red bar, followed by a
broader bright blue-green one; caudal green, median part pale, banded with green spots; dorsal
bright green at base and tip, the middle pale greenish, translucent; anal similar, the distal band
broader; pectorals and ventrals deep vitriol-green with whitish markings.

Color in alcohol, dirty greenish, side with about 8 longitudinal series of greenish blotches; head
olivaceous above, paler on cheeks; upper lip broadly pea-green at edge, this color continued backward
to under eye; edge of lower lip pale green, continued around angle of mouth uniting with the same
color from upper lip; chin with a broad, pale crossbar behind which is a broader, pale green one
which extends up on cheek nearly to orbit; back of this is a still broader, white crossbar interrupted
in the middle by greenish; subopercle and lower edge of preopercle with a large, irregular, green patch;
a median green line on breast to base of ventrals; dorsal green at base and along edge, the middle
portion paler; anal similar to dorsal, the green border broader; caudal bright [tea-green on the outer
rays, the inner ones pale with 4 or 5 cross-series of green spots, tips of rays darker; ventrals pale
green, the edges dark pea-green; pectoral pale green, darker green on the upper rays.

This species is related to Scarus gilberti from which it differs in the greater depth and the somewhat
different coloration. It is also related to Scarus lauia, but differs in the much greater depth, the
less produced caudal lobes, the greater width of the green head markings, and the color of the fins.

Only one specimen was obtained, type, No. 50647, U. S. N. M. (field No. 02944), 14 inches long,
obtained at Honolulu, June 6. Named for Dr. Oliver P. Jenkins.

41. Scarus lauia Jordan & Evermann, new species. “Lauia.”

Head 2.8 in length; depth 2.7; eye 6.75 in head; snout 2.6; preorbital 4.8; interorbital 2.8; D. ix,
10; A. iii, 9; P. 13 on one side, 14 on other; scales 2-25-6.

Body short, stout and compressed; head heavy; snout rather short, bluntly rounded; dorsal and
ventral outlines about equally arched, anterior profile slightly concave before the eyes; nape strongly
convex; mouth small, nearly horizontal, in axis of body; upper jaw with 1 or 2 moderately strong,
backwardly directed canines; a similar but smaller canine sometimes present on lower jaw; cutting
edge of upper jaw fitting outside that of lower; teeth white; eye small, entirely above axis of body;
opercle with a broad short flap. Scales large, their surface with fine lines and granulations; nape and
breast with large scales; cheek with 2 rows of large scales, about 7 scales in each; subopercle and
lower limb of preopercle each with a row of scales; opercle with large scales; lateral line broken under
last dorsal ray, reappearing one row lower down and continuing to caudal fin, the pores writh 2 to 4
branches; a series of these oblong scales along base of dorsal and anal; base of caudal with 3 or 4 very
long, thin scales. Dorsal spines soft and flexible, not pungent, the longest about 2.7 in head; soft
portion of dorsal somewhat higher, especially posteriorly where the rays are about 2.4 in head; anal
spines soft and flexible, the first obscure, the third about 4.3 in head; anal rays higher, the last but one
longest, 3 in head; caudal deeply lunate, the 3 or 4 outer rays above and below produced, length of
middle rays 2.3 in head, or 2 in outer rays; ventrals moderate, not reaching vent, 1.9 in head; pectoral
broad, the free margin oblique, length of longest rays 1.3 in head.

Color in life, head brownish yellow before eyes, the jaws lighter yellow; cheek washed with brown-
ish and blue, throat greenish; nuchal and opercular regions brownish orange; body salmon-color
above, the belly lighter yellow, most of the scales with an edging of greenish blue; a deep blue line
from nostril before and behind upper part of eye; upper lip deep blue, the streak forming an inter-
rupted line before eye; lower jaw with 2 blue cross-lines, 1 marginal; a dark blue spot behind angle
of mouth; deep blue blotches on interopercle; dorsal deep blue with a peculiar jagged stripe of light
brownish yellow; anal with blue spots at base, then light yellow, then deep blue, then green with blue
edge; caudal brownish yellow, with bright blue edgings and a median area of bright golden green;
ventrals golden, trimmed with bright blue; pectoral golden with deep blue above and greenish blue
on lower rays, a salmon streak across base with greenish blue behind it.

Color in life of another example (No. 03040, 10 inches long), pale coppery rosy, darker on first 3
rows of scales; the center of each scale in the first 5 rows greenish blue; under parts pale rosy, with

FISHES FROM THE HAWAIIAN ISLANDS.

197

orange wash; head pale rosy, a small postocular blue spot, a short blue line forward from eye, and a
second of same color on upper lip and across cheek to eye, where it has a slight break, then continues
under eye as a greenish-blue bar; under lip with narrow blue edge; chin faded salmon with a double
blue crescent; space from chin to isthmus bright blue; an oblong bright blue spot on subopercle,
behind -which is a smaller irregular one bordered above by a broad greenish-blue space; dorsal green-
ish blue, with a broad submedian orange band, the lower greenish-blue band made up of large, scarcely
connected, bluish spots, the upper half continuous with a narrow bright blue border; a small orange
blotch on base of last dorsal ray; caudal pale rosy at base, then with a greenish bar, followed by a
broad rosy bar, then by a broad terminal greenish -blue bar, dark blue in front, greenish in middle
and pale blue on outer third ; upper and lower edges of caudal blue, below which is a broad rosy
orange stripe; anal greenish blue at base, then a broad orange stripe, the outer half greenish blue with
narrow bright blue edge; pectoral orange anteriorly, pale bluish behind, the anterior border blue; ven-
trals orange, anterior edge and tip blue; iris pale orange.

Color in spirits, light dirty grayish white, lighter below; a narrow pea-green stripe on edge of
upper lip, breaking up into irregular spots from angle of mouth to lower edge of orbit, a similar stripe
from nostril to eye and slightly beyond upper posterior border of eye; these lines sometimes continu-
ous and unbroken; lower jaw edged with green, a broader pea-green cross-stripe at anterior edge of
branchiostegal opening; subopercles each with a broad green stripe; line of union of gill-membranes
broadly green; dorsal with a series of large olive-green spots at base and a broad band of similar color
on distal half, these separated by a paler band and cut by intrusions from it both above and below;
dorsal fin with a very narrow paler border; and with a series of greenish spots at base, then a broad
pale yellowish white line, bounded distally by an indefinite, wavy, black line shading off into the
greenish of the distal half; last ray of anal dusky on its outer third; caudal greenish-olive at base and
on produced outer rays, edges of fin above and below green; middle rays with a broad lunate area of
pale green, scalloped proximally by dark green, separated from the lighter green base by a broad
whitish interspace, the upper and lower edge also darker green; ventrals creamy white, the outer edge
pale greenish; pectoral whitish, the upper edge dusky.

There is some variation in the width of the green markings on the head, sometimes the stripes on
the lower part of the head being very broad.

This species is related to Scarus gilberti, from which it differs in the more strongly produced caudal
lobes, in the narrower lines on the snout, the broader green lines on the throat, the absence of a green
median line on the breast, and in the very different coloration of the fins.

Field

No.

Length.

Locality.

Final disposition of specimen.

03485

04152

03040

04353

04352

Inches.

14

14

Hilo

Honolulu

do

do

do

Type, No. 50648, U. S. N. M.

Cotype, No. 7464, L. S. Jr. Univ. Mus.
Cotype, No. 2716, U. S. F. C.

Cotype, No. 9811, Ind. Uniy. Mus.
Cotype, No. 3964, Field Col. Mus.

Also one specimen (No. 12046), 9 inches long, obtained by Dr. Wood at Honolulu.

42. Scarus barborus Jordan & Evermann, new species.

Head 3.2 in length; depth 3.2; eye 6.6 in head; snout 2.9; interorbital 2.9; preorbital 4; D. ix,
10; A. hi, 9; P. 14; scales 2-25-6.

Body oblong, not very deep nor greatly compressed; head about as long as deep, conic, com-
pressed; snout short, blunt and rounded; upper jaw produced, its lip double, covering entire dental
plate; lower lip covering half of dental plate; no canine teeth; eye anterior, high, its lower border con-
siderably above upper base of pectoral; caudal peduncle short and deep, its depth 2 in head.

Origin of dorsal over upper base of pectoral, spines flexible, short, not quite as long as rays; long-
est ray 2.1 in head; longest anal ray 2.2 in head; caudal truncate; ventrals 1.9 in head, not reaching
vent by half their length; pectoral 1.5 in head. Scales large and thin, very slightly roughened by
radiating lines of granulations extending to margins of scales; lateral line interrupted, the pores being
on 18 scales, then dropping 2 rows to row of scales under posterior base of dorsal, and continuing to base
of caudal on middle of caudal peduncle, 7 pores in the shorter part, which begins on the row following
the row on which the upper part ends, there not being 2 pores in the same row; the scales extend well

198

BULLETIN OP THE UNITED STATES FISH COMMISSION.

out on the caudal, the last scale of lateral line, very large and thin, being the largest scale on the fish;
4 scales in median line before dorsal; 2 rows of scales on cheek, 5 scales in upper row and 2 to 4 in
lower, sometimes only 2 on posterior part; 2 rows on opercle, and 1 on lower margin.

Color in alcohol, grayish leaden brown, lighter below; no markings on fins different from corre-
sponding parts of body evident.

The above description is based on the type, No. 50649, U. S. N. M. (field No. 04316), a specimen
7.75 inches long, from Honolulu; cotype, No. 2735, T7.8. F. C. (field No. 0^354), 7.5 inches long, from
Honolulu; cotype, No. 7465, L. S. Jr. Univ. Mus. (held No. 650), 5.5 inches long, from Honolulu.

Family TEUTHIDID7E. The Tangs.

43. Teuthis atrimentatus Jordan & Evermann, new species.

Head 3.8 in length; depth 1.9; eye 4.2 in head; snout 1.2; interorbital 3; D. ix, 27; A. in, 25.

Body deep, compressed, ovoid, the upper profile steeper than lower, evenly convex; jaws low, not
produced, lower inferior; mouth small, inferior; teeth broad, compressed, edges crenulate; nostrils
close together, anterior larger, with small fleshy flap; anterior dorsal spines graduated to posterior, the
longest 1.5 in head; fourth dorsal ray 1.4; third anal spine longest, 1.9; first anal ray 1.5; caudal large,
emarginate, upper and lower rays produced in sharp angular points, upper much longer than lower;
pectoral about 3.5 in body; ventrals sharp-pointed, 3.6 in body, spine half the length of fin; caudal
peduncle compressed, 2 in head; caudal spine large, depressible in a groove, 3.1 in head; scales very
small, ctenoid, few, and very minute on vertical fins; lateral line high, arched, at first descending
under fifth dorsal spine, then straight to below middle of soft dorsal, finally falling down and running
along side of caudal peduncle to tail.

Color in life (No. 02996), coppery brown, crossed by numerous, very narrow, pale blue lines, those
above axis of body running somewhat upward and backward, and with short broken lines of same
interspersed, those below more regular but less distinct; cheek brassy, with about 5 narrow pale blue
lines from eye to snout; a conspicuous jet-black spot on caudal peduncle at base of last dorsal ray, each
of these extending slightly upon pale rusty, each with 5 or 6 narrow brassy lines parallel with margin;
edge of each blackish; last rays of dorsal and anal more brassy; caudal dark, blackest on outer part of
middle rays; pectoral pale lemon; ventrals dusky, blacker toward tips; iris brownish, white on
posterior part. Another example (No. 03474) was dull olive-gray, unmarked, save a faint whitish
band across nape and back part of head; fins plain dusky gray.

Color in alcohol, very dark chocolate brown; side with about 40 narrow irregular or incomplete
series of indistinct dark slaty longitudinal lines; cheek with similarly colored lines running obliquely
downward; fins, except pectoral, all more or less blackish or dusky; dorsal with about 5 blackish
longitudinal bands; anal with several similar indistinct blackish bands; base and axils of last dorsal
and anal rays blackish; pectoral brown.

This common species is well distinguished from Teuthis dussumieri and other streaked species by
the black ink-like spot in the axil of the dorsal and anal fin. It has several times been recorded
under the erroneous name of Acanthurus lineolatus, but the species originally called by that name must
be something else. Numerous specimens were obtained by us at Honolulu, where it was also secured
by Dr. Jenkins and Dr. Wood. We have examined the following examples:

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Field

No.

Length.

Locality.

Final disposition of speci-
men.

Inches.

Inches.

05481

9.5

Honolulu.

Type, No. 50673, 1T.S. N. M.

05378

6.2

Honolulu.

Cotype, No. 1501, Cal. Ac.Sci.

02996

0.8

do....

Cotype, No. 7482, L. S. Jr.

C. 1

do

Cotype.

Univ. Mus.

05491

5.2

do

Cotype.

03146

5.7

do

Cotype, No. 2731, U. S. F. C.

05486

5.75

do

Cotype.

03205

5. 75

do....

Cotype, No. , M. C. Z.

05484

0.7

do

Cotvpe.

03474

4.5

do....

Cot.ype, No. 2301, Am. Mus.

05488

5.5

do....

Cotvpe.

.8

do

Cotype.

03729

0.75

do....

Cotype, No. 24229, Ac. Nat.

Dr.Wood.

Sci. Phila.

5.3

do

Cotvpe.

05018

7

do....

Cotype, No. 3965, Field Col.

6. 6

do

Cotype.

Mus.

O.P.J.

05020

5.2

do ... .

Cotype, No. 9818, Ind. TTniv.

do

Cotype.

Mus.

141

4.4

do

Cotvpe.

05305

5.6

do....

Cotype, No. 1707, Bishop Mus.

4.6

do ....

Cotype.

Acanthurus lineolatus, Gunther, Fisc.he der Stidsee, I, 112, taf. i.xxnr, fig. A, 1873 (Society Islands); Steindachner, Denks
Ak. Wiss. Wien, lxx, 1900, 493 (Honolulu); not of Cuvier & Valenciennes.

FISHES FBOM THE HAWAIIAN ISLANDS.

199

Family BALISTIDiF. The Trigger-fishes.

44. Pachynathus nycteris Jordan & Evermann, new species.

Head 3.5 in length; depth 1.9; eye 5 in head; snoutl.25; interorbital 2. 6; preorbital 1.5. D. in — 33;
A. 29; scales about 80. Body short, stout, deep, and greatly compressed; head short; dorsal and ventral
profiles about equally curved; caudal peduncle short, compressed, its least depth about twice diameter of
eye, its least width about equal to diameter of eye; a short horizontal groove in front of eye below nostrils;
nostrils small, close together in front of upper part of eye; teeth broad, close-set, forming a continuous
plate, the teeth, however, not united; lips thin; mouth small, horizontal, in axis of body; lower jaw very
slightly the longer; gill-opening short, nearly vertical; a group of bony scutes under pectoral back of gill-
opening, one of these considerably enlarged; scales regularly arranged in rows, their surfaces granular;
lateral line beginning at posterior edge of eye, ascending to within 7 scales of spinous dorsal and con-
tinuing to near origin of soft dorsal, where it disappears; scales on posterior portion of body and on
caudal peduncle each with a slightly raised crest at its center, these forming series of ridges along the
side. First dorsal spine strong, blunt, and rough, its length about 2 in head; second dorsal spine
shorter and much weaker, its length scarcely more than one-third that of first; third dorsal spine
remote from the second and very short, not extending above dorsal groove; soft dorsal gently rounded,
its rays of approximately equal length, the longest equaling the distance from tip of snout to posterior
edge of eye; base of soft dorsal slightly greater than distance from tip of snout to posterior base of first
dorsal spine or equaling distance from tip of snout to lower base of pectoral axil; anal similar to soft
dorsal, the rays somewhat longer than those of soft dorsal, the base somewhat shorter; caudal short
and rounded, the rays about 1.75 in head; pectoral short, the upper rays longest, about 3 in head.

Color in alcohol, rich brownish or velvety black; spinous dorsal black; soft dorsal pale yellowish
or whitish, margined with black, the lower half crossed by 4 narrow parallel black lines; anal similar
to soft dorsal, but with only 2 narrow black lines on its basal half; caudal dusky, yellowish at tip;
pectoral yellowish.

Only one specimen. Type, No. 50821, U. S. N. M. (field No. 05089), 6.25 inches long, Honolulu.

Family TETRODONTID/E. The Puffers.

45. Lagocephalus oceanicus Jordan & Evermann, new species.

Head 2.8 in length; depth 3.6; eye 4.5 in head; snout 2.4; interorbital 3.2; depth of caudal
peduncle 6; D. 12; A. 12; C. 10; P. 14.

Body rather elongate, moderately compressed, greatest depth at vertical of pectoral; head long;
snout long, blunt at tip, the sides flattened; anterior profile from tip of snout to vertical of pectoral in
a long, low, even curve; ventral outline little convex when not inflated; mouth small; teeth pointed at
median line, the cutting edge sharp; nostrils separate, not in tubes, the anterior somewhat the larger,
their distance from eye about half their distance from snout or about half the interorbital space; gill-
opening vertical, 1.2 in eye, extending a little above base of pectoral, inner flap entirely hidden by
outer; eye rather large, wholly above axis of body; interorbital space very little convex; cheek long;
caudal peduncle nearly round, tapering, its length from anal fin equaling snout; back, upper parts of
sides and entire head entirely smooth, no spines or prickles evident; belly covered with small 4-rooted
spines, most prominent when belly is inflated, spiniferous area not extending on throat anterior to eye,
nor on side above base of pectoral, but in front of anal extending upward to level of lateral fold; a line
of very small mucous pores curving above eye on interorbital space; a strong cutaneous fold on lower
part of side of caudal peduncle from above anterior base of anal to lower base of caudal fin; no dermal
fold on head or anterior part of body; mucous pores inconspicuous; dorsal fin somewhat anterior to
anal, pointed, anterior rays produced, their length equal to that of snout; anal similar to dorsal, its
rays somewhat longer; caudal lunate, outer rays about 2 in head; pectoral broad, its length a little
greater than snout, 2.3 in head.

Color in life, back blackish, fading into deep steel-blue on side; side and below from level of
upper edge of eye abruptly silvery-blue; sides of belly white, with round black spots about as large as
pupil, these most distinct about pectoral, before, below, and behind the fin; upper fins dusky; caudal

200

BULLETIN OF THE UNITED STATES FISH COMMISSION.

mottled black, tipped with white; pectoral black above and behind, pale below; anal pale, broadly
tipped with blackish.

Color in alcohol, bluish black above; side from upper level of eye abruptly bluish silvery; back
crossed by 7 or 8 narrow darker cross-streaks; belly white, with a series of about 9 to 12 small
roundish black spots, chiefly below the pectoral; cheek dusky; pectoral, dorsal, and caudal dusky,
tips of the latter paler; anal whitish,- a little dusky at tip. A somewhat smaller example (4.5 inches
long) has larger dark spots along middle of side above level of pectoral.

This species is known to us from 2 small examples obtained in the market of Honolulu. It is
related to Lagocephalus stellatus ( Donovan) of Europe ( Tetrodon lagocephalus of . Gunther, not of
Linnseus) , but differs in the much shorter pectoral, more conspicuous spots, and rather greater extension
of the prickly region of the breast. The types of Tetrodon lagocephalus Linnseus are reputed to have
come from India. According to Linnaeus, this species had 10 dorsal and 8 anal rays. It may have
been based on Lagocephalus sceleratus or some other East Indian species, but there seems to be no
evidence that it was identical with the European Lagocephalus stellatus. In any event, the Hawaiian
form seems different from any other yet known.

Type, No. 50820, U. S. N. M. (field No. 03379), 5 inches long, obtained at Honolulu; cotype, No.
7784, L. S. Jr. Univ. Mus. (field No. 534, paper tag), 4.5 inches long, also from Honolulu.

Family 0STRACIID7E. The Trunk-fishes.

46. Ostracion oahuensis Jordan & Evermann, new species. ‘ ‘ Morno Awaa .”

Head 3.9 in length; depth 2.9; eye 2.9 in head; snout 1.2; preorbital 1.6; interorbital 1; D. 9;
A. 9; P. 10; C. 10.

Body 4-sided; dorsal side of carapace evenly convex, its greatest width one-fourth greater than
head; lateral dorsal angles not trenchant, slightly convex anteriorly, then evenly convex; snout blunt,
the anterior profile ascending abruptly then strongly convex in front of eyes; interorbital space nearly
flat; cheek fiat; side of body concave, its width about equal to head; ventral keel prominent, evenb'
convex; ventral surface nearly flat posteriorly, but little convex anteriorly, its greatest width 1.4 time
length of head, its length just twice its width; gill-opening short, not exceeding two- thirds diamett i
of eye; least width of anterior opening of carapace 1.75 in interorbital, or 1.5 times diameter of orbit,
the depth nearly twice orbit; mouth small; teeth rich brown; least depth of posterior opening of
carapace much less than width of anterior opening, equaling distance from lower edge of preorbital to
pupil; length of caudal peduncle less than that of head, its depth 2.2 in its length; no spines anywhere.
Dorsal fin high, its edge obliquely rounded, its length 1.3 in head; anal similar to dorsal, the edge
rounded, its length 1.2 in dorsal; caudal slightly rounded, its rays nearly equal to head; pectoral with
its free edge oblique, the rays successively shorter, length of fin equal to height of dorsal.

Color in life, dark brown with blue tinges; interorbital space showing more or less golden; small
whitish spots profusely covering entire dorsal surface; no spots on side of body or on face; no spots on
ventral surface except a faint one of a slightly darker color than general gray color of surface; one
longitudinal row of golden spots on each side of upper part of caudal peduncle from carapace to base
of caudal fin; pectoral, anal, and dorsal fins with transverse rows of faint spots; caudal bluish black at
base, white on posterior half; a broad light or yellowish area below eye; iris golden.

Color in alcohol, rich brown above, the sides darker, and the ventral surface paler, brownish
about margins, dusky yellowish within; entire back with numerous small, roundish, bluish-white spots;
upper half of caudal peduncle with similar but larger spots; forehead and snout dark brown; lips
brownish black; cheek dirty yellowish; sides and ventral surface wholly unspotted; base of caudal
blackish, paler distally, the dark extending farthest on outer rays; other fins dusky, with some obscure
brownish spots.

This species is related to 0. camurum Jenkins, from which it differs in the smaller, more numerous
spots on back, the entire absence of spots on side, the smaller size of the spots on the caudal peduncle,
and the brighter yellow of the suborbital region. Only 2 specimens known, both from Honolulu.

Type, No. 50668, U. S. N. M. (field No. 03443), a specimen 5.6 inches long, obtained by us at
Honolulu, July 25, 1901. Cotype, No. 7478, L. S. Jr. Univ. Mus. (field No. 2156), an example 5.25
inches long, collected at Honolulu, in 1898, by Dr. Wood.

FISHES FROM THE HAWAIIAN ISLANDS.

201

Family SCORP^ENIDd?. The Rockfishes.

47. Pterois sphex Jordan & Evermann, new species.

Head 2.4 in length; depth 2.65; eye 3.8 in head; snout 3.2; interorbital 5.2; maxillary 2 35;
mandible 2; D. xiii, 11; A. in, 7; P. 16; V. i, 5; scales 10-56-13.

Body elongate, compressed, greatest depth at first dorsal spines; back only slightly elev^sed;
snout rather short, rounded; mouth large, oblique; maxillary reaching below anterior rim of orbit,
its distal expanded extremity 1.75 in eye; teeth fine, in bands in jaws and on vomer; lips rather
thin, fleshy; tongue pointed, compressed and free in front; jaws nearly equal; eighth dorsal spine
longest, equal to head; penultimate spine 4; fifth dorsal ray 1.75; third anal spine longest, 2.2; third
anal ray longest, 1.5; caudal rounded, elongate, 1.4; pectoral long, the rays more or less free for
at least half their length; ventral 1.3 in head, reaching beyond origin of anal; ventral spine 2.1;
caudal peduncle compressed, its depth 3.75; nasal spine very small; preocular, supraocular and post-
ocular spines present, the upper bony ridge over eye being serrate; tympanic, coronal, parietal,
and nuchal spines present, coronal very small and close together and parietal with 4 serrations; a
finely serrate ridge from behind eye over opercle to suprascapula; a finely serrated ridge over
preorbital and cheek to margin of preopercle, ending in a strong spine, below this 2 other spines;
preorbital with a strong spine over maxillary posteriorly, and with fine serrations above; scales
ctenoid, present on top of head, cheeks, and opercles, head otherwise naked; tubes of lateral line
single, in straight line to base of caudal; several fleshy flajis on head, 1 above eye, 1 from lower
preorbital spine, and 2 from along margin of preopercle.

Color in alcohol, very pale brown, whitish beneath; side with 9 broad, deep brown bands alternat-
ing with narrow brown bands on trunk and posterior portion of head; narrow brown bars from below
penultimate dorsal spine with a narrower brown line on each side above lateral line; lower surface of
head whitish, without crossbands; spinous and soft dorsal and caudal each with 4 dusky brown cross-
bands; base of anal with 2 broad similar bands, and soft portion of anal with 3 series of irregular
crossbands; axil of pectoral above with white blotch; pectoral whitish with 10 blackish crossbands; a
brown band in front of base of pectoral extending on lower pectoral rays; ventral with dusky blotch
at base, outer portion with about 5 dusky crossbands.

The only example we have seen of this species is the type, No. 50650, U. S. N. M. (field No. 05030),
6 inches long, obtained by us at Honolulu.

48. Scorpsenopsis catocala Jordan & Evermann, new species.

Head 2.1 in length; depth 2.75; eye 7.25 in head; snout 3.1; interorbital 4.3; maxillary 1.8; D.
xn, 10; A. hi, 5; P. 18; V. i, 5; scales 9-42-22.

Body elongate, greatest depth at first dorsal spines; back elevated, swollen, or convex, below first
dorsal spines; snout rather long, with an elevated prominence; mouth large, oblique; maxillary large,
expanded extremity broad, 6.5 in head; teeth in broad villiform bands in jaws, those on vomer small;
no teeth on palatines; tongue small, pointed, free in front; lips rather thick, fleshy; eye small, a little
in front of middle of length of head; a deep pit below eye; top of head with deep square pit just behind
interorbital space; anterior nostril with broad fleshy flap; posterior large, without flap; four spines on
side of snout above anterior nostril; preocular, supraocular, postocular, tympanic, parietal, and nuchal
spines present; a series of spines running across cheek below eye; several large spines on lower part of
preopercle; several spines on opercle; side of head above with many small spines; suprascapular with
several small spines; dorsal spines rather strong, third longest, 3.75 in head; last dorsal spine 3.8;
second dorsal ray 2.7; second anal spine enlarged, a little ionger than the third, 3.4 in head; first anal
ray longest, 2.4 in head; caudal rounded, 2 in head; pectoral large, lower rays thick, fleshy, curved
inward; sixth pectoral ray 1.7 in head, lowest 3.7; base of pectoral broad, 2.25; ventral spine strong,
3.1 in head, second ray longest, 1.9; the innermost ray joined by a broad membrane to belly; caudal
peduncle compressed, its depth 4 in head; head and body with many fringed fleshy flaps; scales mod-
erately large, ctenoid.

Color in life (field No. 03382), excessively mottled, streaked, and spotted; body dark purplish
brown or claret shaded, the spaces gray tinged with sulphury yellow; head all dull brown, flaps colored
like the space about; belly to axillary region whitish with reticulations and irregular marks of yellowish

202

BULLETIN OP THE UNITED STATES FISH COMMISSION.

olive; axillary region wine-brown, finely mottled with yellowish white in streaks and spots; a few
round black spots behind and in axil; inside of pectoral with a large jet-black blotch at upper part of
base, bordered with orange; around this a large yellow area, then 6 oblong black spots on the mem-
branes of upper rays above middle, then a broad rose-red band, fading into violet below, the rim gray;
ventrals bright brown and gray, red shaded on inner face; inside of branchiostegals salmon-color,
striped with white, the membranes yellow; membranes of upper jaw salmon-color mottled with light
yellow; tip of upper jaw orange with a golden ridge dividing a triangular spot of iiuligo-blue between
vomer and premaxillary; a golden line on each side in front of palatines; tip of tongue light yellow;
a triangular indigo-colored spot behind teeth of tip of lower jaw; a golden streak behind it on membrane
before tongue; lower lip salmon-color especially behind where hidden.

Color in alcohol, dark purplish, beautifully mottled with dusky and darker; head mottled above
with dusky; fins with many fine dusky and brown wavy lines; base of pectoral both outside and
inside brownish, the latter variegated with white and blackish brown; outer portion of inside of
pectoral covering first 5 rays with a series of broad blackish spots; ventrals more or less brownish
variegated with gray and whitish; body whitish, mottled with pale brown; edges of buccal folds,
inside of mouth, deep yellow; a deep blue blotch direct!)' behind teeth in front of each jaw.

This species is related to Scorpsena gibbosa (well figured by Gunther in Fische der Siidsee),
from which it differs in the much rougher and less depressed head, much larger flaps on opercles and
mandible, and the presence of a very large fringed flap on the anterior nostril, this being obsolete in
S. gibbosa. We have compared our specimens with examples of S. gibbosa from Apia.

This species was obtained both at Honolulu and Hilo, and appears to be not uncommon. Our
collections contain 8 excellent examples, as follows:

Field

No.

Length.

Locality.

Final disposition of specimen.

05298

Inches.

9.5

Honolulu

Type, No. 50651, U. S. N. M.

03382

9.3

do

Cotype, No. 7466, L. S. Jr. Univ.Mus.

03521

6.8

do

Cotype, No. 2717, U. S. F. C.

05294

7. 75

do

Cotype, No. 1698, Bishop Mus.

05295

6

do

Cotype, No. , M. C. '/.

05296

8.2

Hilo

Cotvpe, No. 3966, Field Col. Mus.

05299

7.5

Honolulu

Cotype, No. 2293, Am. Mus. Nat. Hist.

G. 3

Hilo

Cotype, No. 24221, Ac. Nat. Sci., Phila.

49. Dendrochirus hudsoni Jordan & Evermann, new species.

Head 2.5 in length; depth 2.5; eye 3.4 in head; snout 3.3; interorbital 5; maxillary 2.1; mandible
1.8; D. xm, 10; A. in, 6; P. 18; V. i, 5; scales 8-52-13.

Body elongate, compressed, rather deep, the greatest depth at fifth dorsal spine; profiles of trunk
above and below more or less even; head compressed; snout short, rounded; mouth large, maxillary
nearly reaching below middle of eye, its distal expanded extremity equal to half eye; minute teeth in
bands in jaws and on vomer; lips thin; tongue pointed, compressed, free in front; jaws nearly equal;
anterior nostrils each with a small fleshy flap; interorbital space deeply concave; fifth dorsal spine
longest, 1.25 in head; penultimate spine 5.2; second anal spine longest, 2.1 ; third anal ray longest, 1.3;
caudal rounded, 1.25; pectoral 2.4 in trunk, reaching below middle of base of soft dorsal, rounded, and
only membranes between lower rays slightly incised; ventral rounded, reaching base of first anal ray;
caudal peduncle compressed, its least depth 3.5 in head; nasal spines very small; preocular, postocular,
tympanic and coronal spines present; parietal and nuchal spines forming a single ridge; a ridge of
spines behind eye above opercle; a ridge of spines below eye, ending in a spine on margin of pre-
opercle; 2 spines below this also on margin of preopercle; no opercular spines; margin of preopercle
with spine projecting down and back; skinny flap above eye equal to its diameter, and another from
preorbital spine; scales small, ctenoid; head naked except some scales on opercle, cheek, and side
above; lateral line running obliquely down to base of caudal.

Color in alcohol, pale brown or whitish; side with 3 pairs of deep brown vertical bands, first on
posterior part of head preceded by a deep brown streak from below eye, second on middle and posterior
part of spinous dorsal, and third extending out on soft anal and basal portion of soft dorsal; soft
dorsal, caudal, and anal pale or whitish; membranes of dorsal spines deeply incised in front, each

FISHES FROM THE HAWAIIAN ISLANDS.

203

spine with 3 brown crossbands; pectoral grayish with a blackish brown basal blotch and 5 blackish
crossbands; ventral blackish with 2 whitish or grayish blotches.

This species is especially characterized by the unspotted soft dorsal, anal, and caudal. From
Dendrochirus barberi Steindachner, it is distinguished by the longer pectoral which reaches to below
the posterior dorsal rays.

Named for Capt. C. B. Hudson, in recognition of the excellence of his paintings of Hawaiian
fishes.

We have examined 5 specimens of this species, as follows:

Field

No.

Length.

Locality.

Final disposition of specimen.

03547

Indies.

1.8

Waikiki

Type, No. 50652, U. S. N. M.

651

1.9

Reef near Honolulu

Cotvpe, No. 7407, L. S. Jr. Unlv. Mils.

052

1.9

Honolulu

Cotype, No. 2718, U. S. F. C.

O.P. J.
301

4

do

3.5

do

Family GOBIIDiE. The Gobies.

QinSQJJILIUS Jordan & Evermann, new genus.

Quisquilius Jordan & Evermann, new genus of Gobiidx ( eugenius ).

Allied to Asterropteryx. Body robust, covered with large, ctenoid scales; snout blunt; mouth large,
very oblique, with 2 series of sharp teeth in jaws, the inner depressible; side of head with several
series of short papillary fringes; ventrals separate, their rays i, 5, joined at base by a narrow frenum;
dorsals short, the first with 6 spines, the second with 12 short rays.

The genus is distinguished from other small Eleotrids by the papillary fringes on preorbital, jaws,
and opercles.

50. Quisquilius eugenius Jordan & Evermann, new species.

Head 2.8 in length; depth 3.8; eye 3.25 in head; snout 4.25; width of mouth 2.4; interorbital 2
in eye; D. vi-12; A. 10; V. i, 5; scales 25,-12.

Body robust, compressed, greatest depth about middle of belly; head large, elongate, broad, depth
1.4 in its length, width 1.25; snout short, blunt, rounded above; jaws large, lower projecting; mouth
large, very oblique, its posterior margin reaching below front of eye; upper jaw with 2 series of
teeth, sharp-pointed, outer larger, the inner depressible; mandible with teeth similar to those in
upper jaw; no teeth on vomer and palatines; tongue truncate, front margin not notched; eye large,
high, anterior; nostrils separated, anterior in small tube, posterior close to upper front margin of eye;
interorbital space narrow, very deeply furrowed; a series of fringe-like papillae running from preorbital
along upper margin of maxillary down behind corner of mouth where it joins another series running
along under surface of mandible, and continued back and upward on margin of preopercle; anterior
margin of opercle with a small vertical series of papillae, each papilla a little shorter than diameter of eye;
gill-opening large, continued forward till nearly below posterior margin of eye; spinous dorsal rather
small, spines flexible, with tips produced in short filaments; soft dorsal high, median rays rather
longer than others; anal more or less similar to soft dorsal, posterior rays very long; caudal rather
large, round; pectoral broad, round, equal to head; ventrals small, 1.25 in head, sharp-pointed, and
joined at base of inner rays by a narrow frenum; caudal peduncle compressed, its length 1.6 in head,
depth 2.4; scales large, ctenoid, those on upper part of head very small; snout, interorbital space,
jaws, and lower surface of head naked; no lateral line.

Color in life (field No. 03554), body with transverse bands of dark brown with olivaceous tinge
alternating with dirty white; edges of scales in dark brown portions lighter; dorsal, anal, and caudal
dark brown, edged in part with white; pectoral light reddish brown.

Color in alcohol, brown; 12 dark brown crossbands on side, the last 6 very broad, much broader
than the pale interspaces; vertical fins dark slaty; pectoral pale slaty; ventral pale on outer posterior
portion, blackish slaty on inner.

204

BULLETIN OF THE UNITED STATES FISH COMMISSION.

We have examined the following examples:

Field

No.

Length.

Locality.

Disposition of specimens.

03564

Inches.

1.4

.8

1.2

1.2

.9

1.1

.9

Waikiki

Honolulu

do

do

do

do

do

Type, No. 50674, W. S. N. M
Cotype, No. 7483, L. S. Jr Univ. Mus.
Cotype, No. 2732 U S. F. C.

Cotype, No. 1708, Bishop Mus.

Cotype, No. , M C Z.

Cotype, No. 3970, Field Col. Mus.
Cotype, No. 24230, Ac. Nat Sci. Phila.

51. " Gnatholepis knig\hti Jordan & Evermann, new species.

Head 3.5 in length; depth 4.25; eye 3..8 in head; snout 3.6; width of mouth 2.5; interorbital
2.25 in eye; D. vi-12; A. 12; P. 16; V. 5.5; scales 32,-9.

Body elongate, compressed, not depressed in front, greatest depth at the middle of belly; head
elongate, its depth 1.25 in its length, its width 1.5; snout oblique, blunt, broad; upper profile of the
head obtuse, with a prominence over eye in front; mouth rather broad, the maxillary not reaching
posteriorly to below front rim of orbit; lips rather thin; teeth small, sharp, in narrow bands in
jaws with an outer enlarged series; no teeth on vomer or palatines; interorbital space very narrow,
Wei; nostrils small, close together in front of eye, anterior with flap of very short, fleshy cirri;
eye high, small, a little anterior; gill-opening restricted to side, nearly vertical, its length 2.25 in
head; scales large, finely ctenoid, and becoming much larger on posterior side of trunk; scales
small on belly in front of ventrals, cycloid; scales moderately large, cycloid on the upper part and
side of head, head otherwise naked; dorsal fins well separated, spines flexible and with extremities
of most free and filamentous; first 1.6 in head, fifth 1.7, last 2.7; soft dorsal long, last rays longest,
first 1.7, last 1.25; anal similar to the dorsal, but lower, first ray 2.8, last 1.25; caudal rounded, the
median rays very long, a little longer than head; pectoral with upper median rays longest, all
rather fine, about equal to length of caudal; ventrals rather large, frenum uniting in front, rather
broad, length equal to pectoral; caudal peduncle compressed, length 1.2 in head, depth 2.25.

Color in life, pale flesh-color, upper parts with dark brownish spots and blotches; a series of about
8 brownish blotches along middle of. side; a small dark spot on base of pectoral; opercle dusky; fins
all pale, spinous dorsal with brown edge; iris bluish white.

Color in alcohol, pale brown, side with numerous small dark brown spots and 7 large dark brown
blotches; a dark brown streak below eye, and another across opercle; spinous dorsal Very pale brown
with about 3 blackish brown cross-lines, very distinct on first spines, running somewhat obliquely,
and becoming indistinct posteriorly; soft dorsal with the spines pale or whitish brown and membranes
between blackish brown; anal more or less dark gray brown; caudal very pale brown or whitish,
spotted in cross-series with brown; pectoral pale brown; ventrals dark brown, paler along edges.

Color when fresh, of example from Hilo, olive-green, rather pale, and with 7 blackish crossbands;
caudal spot small and inconspicuous; black bar below eye, narrow and very distinct; back crossbarred
with many spots of dusky olive; side with longitudinal streaks of dark brown spots along rows of
scales, these irregular and variable, mixed, especially behind, with spots of pale sky-blue; dorsal,
anal, and caudal dotted finely with dark olive; pectoral pale olive; ventrals blackish; anal plain
blackish, paler at base. In most examples examined the head was finely dotted with bright pale blue
on cheeks and opercles.

This small but interesting species is generally common in brackish water about Hilo and Honolulu.
Our collections contain a total of 123 specimens; 15 of these have been tagged and their measurements
are given in the table; 101 other specimens from Hilo range in length from 1.1 to 2.5 inches, the
average length being 1.81 inches. From Waianae we. have 5 specimens, 1.3 to 1.8 inches in length, the
average being 1.62 inches. From the pond at the Moana Hotel at Waikiki, we have 2 examples, each
0.8 of an inch long. The average length of our 123 specimens is 1.81 inches.

The species is named for Master Knight Starr Jordan, who first noticed it in the pond at the Moana
Hotel at Waikiki Beach near Honolulu.

a The genus Gnatholepis Bleeker seems to be equivalent to Ilazeus of Jordan & Snyder.

FISHES FROM THE HAWAIIAN ISLANDS.

205

The following are some of the specimens examined:

Field

No.

Length.

Locality.

Final disposition of specimen.

•2150

Inches.

2.25

Hilo

Type, No. 50653, U. S. N. M.

778

1.75

do —

Cotype, No. 7468, L. S. Jr. Univ. Mus.

783

2.5

do

Do.

784

2.2

do —

Cotype, No. 2719, IT. S. F. C.

785

2.2

do —

Cotype, No. , M. C. Z.

788

2.25

do

Cotype, No. 2294, Am. Mus. Nat. Hist.

789

2.2

do —

Cotype, No. 24222, Ac. Nat. Sci. Phila.

790

2.2

do

Cotype, No. 9812, Ind. Univ. Mus.

792

2.1

do

Cotype, No. 3969, Field Col. Mus.

793

1.9

do

Cotype, No. .

794

2

do —

Cotype, No. 1497, Cal. Ac. Sci.

795

1.8

do

Cotype, No. 1699, Bishop Museum.

797

1.75

do

799

1.7

do

800

1.5

do

Acentrogobius ophthalmotamia, Streets, Bull. U. S. Nat. Mus., No. 7, 60, 1877 (coral reefs at Oahu); not of Bleeker.

52. Gobiopterus farcimen. Jordan & Evermann, new species.

Head 3.25 in length; depth 3.5; eye 3.2 in head; snout 3.5; D. vi-11; A. 9; scales 28 (27 to 29)-10.

Body rather robust, compressed, greatest depth at gill-opening; head rather large, depth 1.25 in
length, width 1.4; upper profile of head evenly convex from tip of snout to origin of dorsal; jaws
large, mandible very large, slightly produced; mouth large, very oblique, maxillary extending beyond
front margin of eye; teeth in jaws uniserial, rather large, somewhat canine-like; two small depressible
canines on posterior part of bone behind anterior series; lips large, thick, fleshy; tongue not emargi-
nate, large, thick, rounded; nostrils close together, posterior very large, in front of upper margin of
orbit with elevated rim; interorbital space very narrow, concave; scales large, ctenoid; a large pore
behind and above base of pectoral; gill-opening large, continued forward below; spinous dorsal small,
flexible, spines ending in filaments, beginning behind base of pectoral; soft dorsal high, rays of nearly
uniform length; anal with posterior rays elongate, much longer than anterior; caudal elongate,
rounded; pectoral broad, round, equal to head; ventrals long, equal to head, broad, without any
frenum in front; caudal peduncle compressed, its length 1.5 in head, depth 2.25.

Color in alcohol, pale brown, trunk covered all over with very pale minute brown dots; fins very pale
brown, dorsals dusky, especially the spinous; 3 vertical pairs of pale brown cross-lines over side of head.

Described from an example 1.1 inches long, taken at Hilo. Type, No. 50654, U. S. N. M.

VITRARIA Jordan & Evermann, new genus.

Vitraria Jordan & Evermann, new genus of Gobiidx, subfamily Luciogobinix (clarescens) .

Body elongate, translucent, covered with very small thin scales; mouth small, oblique; teeth
minute; gill-opening rather narrow; dorsals small, the rays v-11 ; pectoral rather long; ventrals small,
united in a circular disk. Small gobies of the coral reefs, allied to the Japanese genus Clariger, but
with the first dorsal of 7 small spines instead of 3.

53. Vitraria clarescens Jordan & Evermann, new species.

Head 4.6 in length; depth 6.7; eye 3.5 in head; snout 4.5; D. viii-11; A. i, 10.

Body elongate, slender, compressed, greatest depth between dorsal fins; head elongate; pointed,
conic, depth 1.75 in its length, width 2; snout rather long, rounded; jaws prominent, upper slightly
produced; mouth oblique, maxillary reaching a little beyond anterior margin of eye; teeth not evident;
tongue broad, truncate; snout above, interorbital space, and top of head more or less flattened; nostrils
well separated, anterior nearly midway in length of snout, posterior close to front of eye; eye rather
large, anterior; gill-opening restricted to side, rather small; scales very small; dorsal spines flexible,
first dorsal small, the last three spines very small (minute stubs, broken in the type) the fin beginning
behind tip of ventrals; soft dorsal beginning a little nearer base of caudal than tip of snout, about

206

BULLETIN OF THE UNITED STATES FISH COMMISSION.

over insertion of anal, and anterior rays of both fins longest, those of anal gradually smaller behind,
the last 2 minute and close together; caudal emarginate, lobes rounded; pectoral rather long, lower
rays longest; ventrals small, united to form a small round disk whose diameter is 2.25 in head;
caudal peduncle compressed, elongate, its length equal to head, its depth 2.25 in head.

Color in alcohol, very pale translucent brown, 7 V-shaped pale brown markings on upper side of
body united over back; fins whitish.

We have examined 7 examples obtained at Hilo, each about 1.2 inches in length:

Locality.

Final disposition of specimen.

Locality.

Final disposition of specimen.

Hilo

Do....

Do....

Do....

Type, No. 50655, U. S. N. M.

Cotype, No. 7469, L. S. Jr. Univ. Mus.
Cotype, No. 2720, U. S. F. C. .

Cotype, No. , M. C. Z.

Hilo

Do

Do

Cotype, No. 2295, Am. Mus. Nat. Hist.
Cotype, No. 24223, Ac. Nat. Sci. Phila.
Cotype, No. 1700, Bishop Museum.

Family PTEROPSARIDAi.

OSURUS Jordan & Evermann, new genus.

Osurus Jordan & Evermann, new genus of Pteropsaridx ( Parapercis schauinslandi Steindachner).

This genus is allied to Parapercis, from which it differs in having the caudal fin deeply forked
instead of truncate.

Family FIERASFERIM.

54. Fierasfer umbratilis Jordan & Evermann, new species.

Head 10.2 in length; depth 15.2; eye 5 in head; snout 4.8; mouth 2.6; interorbital 4.5.

Body very elongate, compressed; tail very long and tapering gradually in a long point; head
elongate, conic, its depth 2 in length, width 2.25; snout rather broad, conic, and produced beyond
mandible; mandible broad, flattened below; mouth nearly horizontal, broad, the gape reaching below
posterior margin of eye; eye rather small, anterior, without eyelid, and placed about first quarter of
head; nostrils well separated, anterior with elevated rim, posterior a short, crescent-like slit; inter-
orbital space rather broad, convex; gill-opening low, inferior, rather long; gill-membrane free from
isthmus, its angle nearly an eye diameter distant from posterior margin of eye; dorsal fin almost
rudimentary, very low and thin; anal rather broad, in middle its height is about 0.75 in eye, from
which point it gradually decreases to tip of tail, where it is rudimentary, like dorsal; tail ending in
a fleshy point, caudal fin apparently absent; pectoral small but relatively large, 3.1 in head, rays very
minute; lateral line distinct, running down along middle of side on posterior half of tail; no scales.

Color when fresh (field No. 03506), pale olivaceous, with pale greenish spots; a pale bluish streak
in each spot over lateral line; pale purplish oblong spots on lower half of body; head greenish-olive,
with pale green spots closely set on cheek and jaw; pale purplish dots on upper part of cheek and
behind eye; first dorsal same as body, but the spots yellowish; a black spot behind first and second
rays, tips pale; rays of second dorsal checked alternately with yellowish-green and white; caudal
same as second dorsal, but margin yellowish; anal, yellowish-olive; tip blackish; pectoral and ventrals
pale; iris greenish-vellow; dull red streaks radiating from pupil.

Color in alcohol, brown; head and end of tail dark sooty or blackish brown, the color formed of
dark points; greater part of anal fin, lower surface of body anteriorly and pectoral and branchiostegal
membranes, pale straw color; lower surface of trunk more or less blotched with pale brown.

Our collection contains but 2 specimens of this species, both obtained at Hilo. Type, No. 50656,
U. S. N. M. (field No. 03506), a specimen 7.6 inches long; cotype, No. 7470, L. S. Jr. Univ. Mus. (field
No. 528), an example, 6.4 inches long.

Fierasfer umbratilis occurs also in the South Seas, and is readily distinguished from most related
species by its dark, non-translucent coloration.

F. boraborensis from Borabora, briefly described by Kemp, has the pectoral 6 to 7 times in head.

FISHES FROM THE HAWAIIAN ISLANDS.

207

Family PLEURONECTM. The Flounders.

55. Eng-yprosopon hawaiiensis Jordan & Evennann, new species.

Head 3.8 in length; depth 1.75; eye 3.25 in head; snout 4.25; interorbital 6.3; maxillary 2.8; D. 79,
A. 56; P. i, 10; V. i, 5; scales 14-46-15.

Body elongate, deep, rather ovoid, greatest depth about end of pectoral; head very deep, its length
0.7 ip depth; upper profile very convex in front, steep; snout short, obtuse; jaws small, produced a
little, the mandible slightly projecting; lips rather thin; mouth curved a little, very oblique, the small
maxillary reaching a little beyond front margin of eye; teeth in jaws very small, sharp-pointed; eyes
well separated, lower anterior, placed in first third of head, the upper about two-fifths an eye diameter
posterior; nostrils close together, with elevated rims: interorbital space a little more than half an eye
diameter in width, deeply concave; gill-opening small; gillrakers rather short; scales large, finely
ctenoid, very small on rays of vertical fins; lateral line strongly arched at first for first two-ninths its
length, then straight to base of caudal; dorsal beginning on snout, the anterior rays free for only a
short portion of their extremities, first 5 in head, fiftieth 2.1, this the highest region of the fin; anal
more or less similar, first 3.25, thirtieth 2; caudal rounded, middle rays longest, 1.1; pectoral short,
pointed, 1.4; ventrals rather broad, base of left 3, first and last rays about equal; right ventral smaller;
caudal peduncle compressed, its depth 1.9.

Color in alcohol, dark olivaceous brown, fins dark gray-brown, each ray finely specked with oliv-
aceous brown; left pectoral specked with. dark brown, right pectoral dull creamy or brownish white
like the right side of body.

Type, No. 50657, U. S. N. M., taken at Hilo, the only example we have seen, 3 inches long.

56. Engyprosopon arenicola Jordan & Evermann, new species.

Head 3.6 in length; depth 1.9; eye 4.3 in head; maxillary 3; D. 78; A. 57; P. i, 11; V. 5; scales
14-36-17. Body elongate, very deep, rather ovoid, the greatest depth at tip of pectoral; head much
deeper than long, the upper profile steep, strongly convex; snout obtuse; jaws very oblique, mandible
slightly projecting; maxillary very oblique, reaching below anterior margin of eye; lips rather thin,
fleshy, fringed along margins; teeth in jaws minute, sharp-pointed; eyes close together, lower anterior
placed about first third of length of head; upper eye about one-third an eye diameter posterior;
nostrils well separated, with raised fleshy rims forming a flap; interorbital space very narrow, concave;
gill-opening rather small, restricted to side; gillrakers small, short, few; scales large, finely ctenoid;
lateral line strongly arched for anterior fourth of its length, then straight to base of caudal; anterior
dorsal rays free distally for one-half their length, first ray 3 in head, forty-fifth 1.8, which is the
highest region of the fin; anal similar to dorsal, but anterior rays not free for half their length; first
ray 3.5, thirtieth 1.8; caudal elongate, median rays longest, equal to head; pectoral short, pointed, 1.5;
ventrals rather large, the left with its base 5 in head, first ray 3.6, last 2.6, almost entirely in front of
the right, which is much •smaller; caudal peduncle broad, compressed, its depth 2.2 in head.

Color in alcohol, very pale brown; side marked with many large incomplete rings of blackish or
dusky and with a number of dusky spots in between; fins whitish, the vertical or unpaired with large
blackish spots on membranes between rays and similar small ones scattered about, those of caudal
forming about 4 crossbands; several dusky spots at base of pectoral; right side whitish.

We have seen but 2 examples, both taken at Hilo: Type, No. 50658, U. S. N. M., 2.5 inches long.
Cotype, No. 7471, L. S. Jr. Univ. Mus., 1.9 inches long.

Family ANTENNARIIM.

57. Antennarius drombus Jordan & Evermann, new species.

Head (to end of opercle) 2.5 in length; depth 1.75; eye 5 in head; snout 4; width of mouth 2;
D. i-i-12; A. 7; P. 12; V. 5.

Body very deep, compressed, back elevated; head deep, with blunt conic profile in front, some-
what oblique above; snout broad, obtuse, surface uneven; mouth broad, large, nearly vertical; maxil-
lary concealed under skin, reaching below anterior part of eye; lips fleshy; teeth in jaws minute, in
narrow bands; teeth on palatines rather large, sharp-pointed, none on vomer; tongue broad, thick;

208

BULLETIN OE THE UNITED STATES FISH COMMISSION.

mandible large, with fleshy knob at symphysis, projecting; nostrils circular, well separated, with
rounded fleshy rims; interorbital space convex, roughened; top of head with rather large concave pit;
eye high, anterior; bait rather short, only reaching a little beyond first spine, with fleshy caruncle at
extremity; dorsal spines short, first free, rough, depressible in pit on top of head; second dorsal spine
twice length of first, equal to width of mouth, depressible, and united with skin of back to its tip;
posterior dorsal rays longest, and the last, like that of anal, united to caudal peduncle by a membrane;
anal similar, rounded, elongate, 1.5 in head; pectoral broad; ventral small, rounded; caudal peduncle
small, compressed, its depth equal to interorbital space; body rather rough, mucous pores on head and
in lateral line with excrescences; side of body with many pointed cutaneous flaps; second dorsal spine
and first dorsal ray very rough, also with cutaneous flaps; lateral line very convex, running down
toward middle of base of anal.

Color in alcohol, pale plumbeous gray, more or less spotted or mottled with darker; belly and lower
surface rather pale, the spots distinct; fins all more or less pale with dark spots, some at basal por-
tions of dorsal and anal darker; iris blackish with radiating lines of golden.

The above description is from the type, No. 50659, U. S. N. M. (field No. 541 ), taken at Waikiki,
near Honolulu.

Another example (field No. 539) shows some differences: Head (to end of opercle) 2.5 in length;
depth 1.7; eye 3 in head; maxillary 1.8; width of mouth 1.7; interorbital 3.7; P. i— i— 12; A. 7; P. 12; V. 5.

Body very deep, compressed, back elevated; head deep, gibbous, with blunt conic profile in front,
somewhat oblique above; snout broad, obtuse, short, surface uneven; mouth large, obliquely vertical;
maxillary large, reaching a little beyond front portion of eye; lips fleshy; teeth in jaws minute, sharp,
in bands; teeth on roof of mouth large, sharp-pointed; tongue large, broad, thick; mandible large>
with knob at symphysis, projecting; nostrils well separated, close to end of snout, each with elevated
fleshy rims, the anterior higher; interorbital space broad, elevated, uneven; top of head with rather
large pit; eye high, anterior; bait short, reaching tip of first dorsal spine, with caruncle at extremity;
dorsal spines short, depressible; first dorsal spine half length of second, free, depressible in pit on top
of head; second dorsal spine large, joined by skin to its tip; dorsal rays of about equal height, seventh
1.3 in head, and the last, like lower portion of last anal ray, adnate to caudal peduncle by a membrane;
anal rounded; caudal elongate, rounded; pectoral broad; ventral small; body rather rough, mucous
pores on head and lateral line with excrescences; along the lateral line and anterior region of dorsal
are many cutaneous flaps; lateral line convex, running down to above middle of anal.

Color in alcohol, dark gray -brown; edges of vertical fins whitish, the pale border rather broad
and very distinct along posterior, dorsal, anal, and caudal rays; side with about 6 large round
blackish spots; caudal with some pale or indistinct mottlings; pectoral and ventral with rather broad
margins, median portion dusky; iris more or less silvery.

A. drombus seems nearest related to A. nummifer Cuvier & Valenciennes, originally described from
Malabar. Probably the specimens from the South Seas referred to the latter belong rather to A.
drombus. A. nurnmifer is said to be red in color with dark spots, and, as figured by Dr. Day, differs in
several respects from A. drombus. Both these species differ from A. commersoni and its numerous
allies or variants {A. niger A. leprosus, A. rubrofuscus, and A. sandvicensis from Hawaii) in the shortness
of the first dorsal spine or fishing rod. This is scarcely longer than the second spine in A. drombus,
but in A. commersoni it is twice as long.

Our collections contain but 2 examples of this species, the type, No. 50659, U. S. N. M; (field
No. 541), and cotype, No. 7472, L. S. Jr. Univ. Mus. (field No. 539), both taken on the reef at
Waikiki, near Honolulu.

DESCRIPTIONS OF A NEW GENUS AND TWO NEW SPECIES OF FISHES
FROM THE HAWAIIAN ISLANDS.

By DAVID STARR JORDAN and BARTON WARREN EVERMANN.

Since the publication of our recent paper® on new species of fishes from the
Hawaiian Islands, further studies of our large collections have resulted in the dis-
covery of an interesting new species of Tropidichthys and a remarkable new genus
of Scorpamida}. These are described in the present paper. Illustrations of both
species will be given in our final report.

Tropidichthys psegma Jordan & Everinann, new species.

Head 3 in length; depth 2; eye 4.5 in head; snout 1.5; interorbital 2.3; D. 11 ; A. 11; 0. 8; P. 16.

Body short, stout, moderately compressed; snout long, conic; anterior dorsal profile rising evenly to
region above gill-opening, at which point the body is deepest; interorbital flat; gill-opening nearly ver-
tical, short, its length less than diameter of eye; mouth low, below axis of body; teeth strong, convex,
cutting edge sharp; eye small, supraorbital rim not prominent; caudal peduncle deep, its least depth
about 2 in head, its least width, 4 in its least depth; length of caudal peduncle from dorsal fin to base
of caudal tin 1.3 in head; from base of anal tin 2 in head; dorsal prominence equally distant between
tip of snout and posterior base of caudal; base of dorsal 1.5 in height of fin, which latter is 2 in head ; anal
similar to dorsal, its edge rounded; caudal truncate, or very slightly convex, 1.2 in head; pectoral
broad, its base 2.6 in head, free edge oblique, posterior rays 1.5 in anterior ones; body mostly smooth;
interorbital space and snout above and on side with small prickles; belly with a few prickles; a scat-
tered patch also on side above pectoral.

Color in alcohol, dark brown above, paler below ; 3 or 4 short black lines running forward from
orbit, and same number backward; lower part of side, especially posteriorly, and lower part of
caudal peduncle, with small roundish black spots; snout and interorbital space crossed by about 12
narrow black lines, these extending down on side of snout; side of snout with 3 or 4 narrow black
lines from chin toward eye, separated by paler lines; posterior to these small irregular black spots
covering entire cheek, dotted over with fine white specks; ends of spines, pectoral, dorsal, and anal
pale whitish, their bases largely brownish black; caudal dark brownish or black.

This species is known to us only from the type, No. 50885, U. S. N. M. (field No. 2561) 3.75 inches
long, obtained by us at Honolulu in 1901.

We have compared this specimen with examples of T. coronaius Vaillant & Sauvage, obtained by
the Albatross in 1902, and find them quite distinct.

IRACUNDUS Jordan & Everinann, new genus.

Iracundus Jordan & Ever maun, new genus of Scorpxnidic ( signifer ).

Allied to TIdicolenus and Pontinus. Body rather elongate, compressed, covered with small, weakly
ctenoid scales; fins not scaly; head not depressed; formed as in Sebastodes, the spines moderately
developed; head and body with dermal flaps; teeth on jaws and vomer, none on palatines; dorsal fin
deeply divided, the spines 11 in number, the fourth much elongate; pectoral rays undivided; anal
rays in, 5; ventral rays i, 5; caudal rounded; vent at base of first anal spine; air bladder obsolete.

a Descriptions of new genera and species of fishes from the Hawaiian Islands. <Bull. U. S. Fish Comm, 1902 (April 11,
1903), pp. 161-208.

F. C. B. 1902—14

209

210

BULLETIN OE THE UNITED STATES FISH COMMISSION.

Iracundus signifer, Jordan & Evermann, new species.

Head 2.4 in length; depth 3.2; eye 4 in head; maxillary 2; D. x, i, 9; A. in, 5; P. 17; V. i, 5;
scales about 9-55-30, about 45 pores.

Body rather elongate, moderately compressed, the head conic, not depressed; mouth large,
oblique, the lower jaw slightly projecting, the maxillary reaching to opposite posterior margin of
pupil; teeth in moderate bands in the jaws, the inner teeth in the upper jaw slightly largest; vomer
with small teeth; palatines toothless; interorbital area deeply concave, little wider than pupil; spines
on top of head low and rather sharp, much as in Sebastodcs; preocular, supraocular, postocular, tym-
panic, occipital, and nuchal spines present; a ridge with 2 spines outside the tympanic spine; preor-
bital moderate, about as wide as eye, with a sharp spine turned forward and a blunt spine turned
backward; suborbital stay a narrow, simple ridge, reaching base of preopercular spine, which is
straight and very short; 3 lower preopercular spines reduced to blunt points; opercle with 2 slender
diverging spines, the upper the larger, their points not reaching edge of membrane; head with numer-
ous, broad, fleshy flaps; a fringed flap at the nostril, 2 on edge of preorbital, 2 on lower limb of
preopercle, and a high fringed flap above eye, about as long as pupil; small simple flaps on the cheek,
the end of the maxillary, and elsewhere on head; large pores on lower jaw, under suborbital stay, and
elsewhere; opercle and upper part of cheek with rudimentary, embedded scales; jaws naked; top of
head scaleless, occiput covered with thin skin and scarcely depressed; gillrakers very short, thickish,
and feeble, all but about six reduced to mere rudiments; no slit behind last gill; body covered with
small, close-set scales, which are slightly ctenoid; scales on nape small, on breast minute; lateral line
conspicuous, provided with dermal flaps; numerous scattered flaps on sides of body.

Dorsal fin very deeply notched, the spines rather slender, pungent, the first a little longer than
eye, the second and third subequal, about half longer, the fourth greatly elevated, 1.5 in head, almost
twice height of third and fifth, which are subequal; sixth, seventh, and eighth slightly longer than
fifth, tenth very short, eleventh half length of fourth; soft dorsal high, the longest rays nearly half
head; rays of all the fins scaleless; caudal long, rounded, 1.4 in head; anal high, the spines graduated,
the third a little longer than second, which is 2.6 in head; longest soft rays 1.8 in head; pectoral with
the rays all simple, the longest 1.2 in head, lowest rays shortened and thickened; ventral fins inserted
below axis of pectoral, rather long, 1.6 in head, not. quite reaching anal, inner rays well free.

Color, pale in alcohol, doubtless vermilion red in life, the flaps on body pinkish; a single jet-black-
spot about half diameter of pupil near tip of membrane between second and third spines of dorsal.

The only example known is the type, No. 50886, U. S. N. M. (field No. 635), a specimen 4.2 inches
long, taken by us on the coral reef at Honolulu.

THE FRESH WATER FISHES OF WESTERN CUBA.

By C. I I. EIGENMANN,

Professor of Zoology , University of Indiana.

211

*

V

Bull. U. S. F. C. 1902.

Plate 1 9

SAN JUAN RIVER, LOOKING UPSTREAM FROM THE FIRST BEND ABOVE THE UNITED-
HA BA N A RAILROAD BRIDGE

SAN JUAN RIVER FROM ABOVE THE UNITED HABANA RAILROAD BRIDGE.

THE FRESH-WATER FISHES OF WESTERN CUBA.

By C. H. EIGENMANN,
Professor of Zoology , University of Indiana.

During March, 1902, the writer, accompanied by one of his students, Mr. Oscar
Riddle, as assistant and interpreter, made a series of collections in the fresh waters of
western Cuba, in the streams accessible by the Western Railway and the United
Habana Railways. Attempts to reach waters remote from the railways were aban-
doned on account of the expense, both in time and money. Sumidero was reached
by horse from Pinar del Rio, and the caves about Canason foot and by volante. The
original and chief object of the visit to Cuba was to secure material for a study of the
eyes of the blind fishes, Stygicola and Lucifuga . In this I was successful. The
fresh-water fishes proved also of considerable interest. As might have been expected,
many of the more abundant and larger species had been previously described by
Poey. Nothing, however, was known about the distribution of fresh-water fishes,
and there were found a surprising number of new species.

I wish here to express my thanks to Mr. A. P. Livesey and Mr. J. E. Wolfe,
the managers of the Western and United Habana Railways, who did all in their power
to make the available time profitable from a scientific standpoint. 1 am also under
obligations to Mr. Philip Hammond, the chief engineer of the United Habana Railroads,
for suggestions and various favors. Mr. Pascual Ferreiro, of the Cuban railway
postal department, kindly acted as guide to the Pedregales caves, and Mr. Francisco
Martinez and his brother to the caves about Canas. The success of the expedition
was largely due to my companion, Mr. Oscar Riddle, whose previous stay in Porto
Rico and trip to Trinidad and the Orinoco had familiarized him with the language of
the country and enabled him to deal with the natives.

The drawings illustrating this paper were made by Mr. Clarence Kennedy.

In his “ Memorias sobre la Historia Natural de la Isla de Cuba,” tomo 2, pp.
95-114, 1850, Poey describes two species of blind fishes, Lucifuga subterraneus and
Lucifuga dentatus , from caves on the southern slopes of the jurisdiction of San
Antonio, Guanajay, and San Cristobal." They were first brought to notice by the
surveyor, D. Tranquilino Sandalio de Noda. Specimens were secured for Poey by
Dubroca, Fabre, and Layunta.

« Jordan & Evermarm, in their Fishes of North and Middle America, in, p. 2501, give a number of localities for
which I can find no authority in Poey. Thus “San Antonio, Cuba (Coll. D. Tranquilino); Sandalio de Noda (Coll. D.
•Tuan Antonio Fabre)” and cave at the “ Castle of Concord.” I can not find the authority for the locality San Antonio,
Cuba; the collector given as D. Tranquilino is probably D. Tranquilino Sandalio de Noda; the second locality Sandalio
de Noda is probably the latter section of the name of the man who first called attention to them. Castle of Concord
should probably read “ Coffee plantation La Concordia.”

213

214

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The localities from which Poey secured Lucifuga subterraneus are (1) Cave of
Cajio, 5.2 miles south of La Guira de Melena (Noda, 1831); (2) cave of the coffee
plantation La Industria, halfway between Alquizar and Guanimar (Dubroca); (3)
cave of Ashton, near San Andres (Fabre); (4) cave of the Dragon, on the cattle
farm San Isidro, near Las Mangas (Fabre); (5) cave on the coffee plantation La
Concordia, 5.2 miles from Alquizar (Layunta); (6) cave near the beehouse of the
coffee plantation La Paz (Dubroca); a well near the tavern Frias (Fabre).

Lucifuga dentatus was secured in Nos. 1, 2, and 3. Those from Nos. 1 and 2
were said to be without the least vestiges of eyes; those from No 3 with vestiges.

I visited a number of the caves from which Poey secured his specimens, intending
to obtain specimens from as many of Poey’s localities as possible, but especially from
those from which he secured both species of the blind fishes. The towns Guira de
Melena, Alquizar, San Andres (now Cafias), mentioned by Poey, are successive sta-
tions along the Western Railroad, and Las Mangas is a town a short distance off the
railroad beyond Canas. We made our first stop at Alquizar, hoping to be in the
middle of the cave region. The driver we engaged at Alquizar to take us to La
Industria showed his independence and originality by taking us over a rough road
during a half day’s drive parallel to the railroad to Frias and Ashton near the station
Canas, but we were not inclined to quarrel with him, as we at once secured fishes in
Ashton, and caves were reported to us as very abundant in the whole region south of
Canas. Many of these caves were, visited by us, as well as that at Cajio (by Mr.
Riddle) and others in a widely distant part of the island. Those from which we
secured specimens 1 propose briefly to describe here. The cave on La Industria we
did not visit, and in fact, except while we were engaging our volante, no one about
Alquizar seemed to know the plantation La Industria. It is possible that the name
has been changed in recent years. Bearing in mind our experiences with the volantc-
men at Alquizar we made no attempt to find La Concordia, which is also reached
from Alquizar.

The “caves” about Canas can best be described after a few words concerning the .
country in general in which they occur. The territory about Canas is entirely
drained by underground streams. The streams rising in the lulls and mountains
forming the watershed between north and south drainage run above ground for a
distance and then disappear underground. The Ariguanabo River thus runs into a
bank at San Antonio de los Banos and disappears among fallen rocks. A few yards
away from its “sumidero” the water can lie seen running in its underground channel
through an opening in the thin roof of the channel. A few yards farther on a dry
channel leads down to the water which at the end of the channel disappears among
fallen rocks. Other rivers disappear in a similar manner. Their waters reappear,
in part at least, in a number of “ojos,” some near the coast south of San Antonio.
The region drained by underground streams is flat, with frequently no indications of
surface streams and their erosion, and extends westward to near San Cristobal,
where the first permanent surface stream is observed. At Artemisa and Candelaria
stream beds contained pools of water at the time of our visit.

From San Cristobal to Dinar del Rio there were many small but perennial
streams. Eastward from Canas the cave region has an unknown extent. Poey
limited it to the jurisdiction of Guana jay, but it certainly extends as far east as the

FRESH-WATER FISHES OF WESTERN CUBA.

215

meridian of Matanzas, and from reports probably beyond Cienfuegos. East of Rincon
there are, however, frequent river beds, dry during- the time of our visit. This
main region, belonging to the southern slope, sends a point northward from Rincon to
the Almendares River in the northern watershed. Aside from the “Ojos de agua”
along the edge of the cienegas skirting the southern coast there are two notable places
where undergound rivers find an exit. One at Yento supplies the entire city of
Iiabana with its water, the other serves to make the region about Guines a garden,
its waters being used for irrigation. Others in ail probability have a subaqueous
exit to the south.

The large spring at Yento, sufficient to supply the city of Habana with its water,
is the only one which issues on the northern slope, so far as I know. The origin
of the supply from the Veil to spring has not been traced. It issues but a few feet
from the Almendares River, and it is very remotely possible, though not at all prob-
able, that it derives its water from the upper courses of the Almendares. At the
time of our visit the water of the spring was 1 degree warmer than that of the
Almendares River. The region north of that river, across the river from Vento,
being shut out from a possible contributing source, it undoubtedly derives its water
from the system of underground streams mentioned above. An examination of the
best available map and the levels of the Western and United Habana Railroads makes
it seem quite certain that the Vento springs derive their water from the region imme-
diately south of Vento and north of Rincon and Bejucal — that traversed by the two
railroads mentioned. This region contains various sinks without surface outlets, as
well as dry sink-holes, and is the northward-projecting point of the cave region men-
tioned above. A notable sink-hole in this region is that at Aquada, on the United
Habana Railroad. This is very broad, shallow, and dry during the dry season, but
the water rises to stand over 10 feet deep on the railroad track during some of the
wet seasons.

THE CAVES VISITED.

The soil over the region under consideration is thin, the surfaces of the very irreg-
ularly corroded rocks jutting out in numerous places. This, together with the fact
that the water of the underground streams is but a few feet underground, gives the
region an entirely different aspect from the sink-hole and underground-stream region
in Indiana, Kentucky, etc. It is, in the first place, impossible to enter the under-
ground streams, and there are no funnels on the surface to indicate the location of
an underground stream or its tributary. In places the thin limestone roof of an
underground chamber has given way and enables one to get to the water, which in
all the numerous places we visited was stationary, not flowing. With one or two
exceptions the water was covered with a continuous crust of carbonate of lime, due to
the evaporation and discharge of carbon dioxide from the surface of the perfectly
quiet lime water. When the water is disturbed flakes of variable size break loose
and gradually sink to the bottom. All of the so-called caves about Canas, with one
possible exception, were sink-holes formed by the breaking of the thin roof of a
larger or smaller underground chamber. In all of the caves where stalagmites and
stalactites were noticed these extended for 3 feet or more into the water. Inasmuch
as they could not have formed under water, the latter must have risen since their

216

BULLETIN OE THE UNITED STATES FISH COMMISSION.

formation. Usually the walls of the sink-hole retreat downward and sidewise,
suggesting- that a channel til led with water extends down and out from the sink-hole.
The impression is irresistibly made that there exists here an extensive series of
drowned caves. As our visit to the caves occurred near the end of the dry season,
this excess of water could not have been due to recent excessive rains. The water in
the caves was, however, still falling and fell several inches during March.

The condition could have been brought about (1) by the rising of the southern
coast of Cuba, resulting in a changed incline in the underground streams and a con-
sequent drowning of the caives; (2) by a blocking of the outlets of the streams; (3)
by an increase of the rainfall above that provided for in the past. 1 was at first
inclined to favor the first view, since caves as widely different as those of Canas and
Alacranes showed the same characters; but a cave at Matanzas, on the northern slope,
showed exactly the same conditions. A local blocking of the outlets is scarcely
conceivable on account of the wide separation of caves showing the same conditions.
Concerning the rainfall I am not able to speak, but an}r other cause, since the caves
are above the level of the ocean, does not occur to me.

The detailed description of various caves may begin with Modesta No. 1. The
caves, for the most part, do not have distinctive names, but are called after the Finca
on which they occur. I have added the numbers 1, 2, etc., for convenience. Modesta
No. 1 is an ideal cave of which all the rest are modifications. There is nothing on
the surface to distinguish it when one is a few feet away. The cave is bell-shaped,
with an opening 10 to 15 feet at the top. A tree growing at its margin sends vertical
roots down to the bottom. On these roots notches have been cut, and the descent is
made by means of them. At the water level at the time of our visit the cave was
oval in section, 30 by 45 feet in extent. In the middle of the bell, and immediately
under the opening, there was a large pile of rocks, cemented together in places by
stalagmitic material and rising but a few inches above the water. The water, beauti-
fully clear, became rapidly deeper in all directions and could be seen to extend out
in at least two directions in deepening channels filled to their top with water. The
roots descending from the opening at the top to the island, a distance of about 15
feet, here divided suddenly into a tuft of innumerable rootlets, most of them in the
water. Such roots were found in almost all the caves, and the young blind fishes
were always found in among the rootlets; the big ones among the rocks.

Modesta No. 2, also called Hawey, is a cave of the same type, except that the
central mass of fallen rock forms an arch over the water, and that it can be reached
from one side by an inclined plane, also formed of fallen material. Rootlets were
very abundant here and small blind fishes equally so. The water was probably not
more than 10 feet below the surface and at one edge was very deep — how deep we
had no means of determining. Part of the opening had at one time been walled in
and the cave was used as a well. It is possible that this is the well mentioned by
Pocy as being near Frias and containing blind fishes. But a number of other wells
in the neighborhood reach caves.

Modesta No. 3 contained no blind fishes.

San Isidro No. 2 is a duplication of Modesta No. 1, but with different propor-
tions. The central mass is higher and holds a number of stalagmites. It is only
partially surrounded by water. The entrance is by roots.

Bull, U. S. F, C. 1902.

Plate 20.

RIO SAN DIEGO AT PASO REAL, LOOKING UPSTREAM FROM BELOW WESTERN RAILROAD

BRIDGE.

RIO DEL PINAR, LOOKING UPSTREAM FROM BRIDGE.

FRESH-WATER FISHES OF WESTERN CUBA.

217

San Isidro No. 1 is an underground chamber containing 2 to 3 feet of water in
places and muddy in the bottom. It is reached by an inclined plane from the side
and is more like a cave as ordinarily understood than the caves of the Modesta type.

La Frias No. 1 is -a cave of the Modesta type, but larger and with the roof of
one side fallen, so that the descent to the water is made by an inclined plane, and the
water forms a crescent about 15 feet wide under the retreating walls of the side
opposite to the entrance.

La Frias No. 2 is another cave of the Modesta type, 100 feet across at the bot-
tom and with the water 40 feet from the surface. The roof has fallen in at one side
and the central mass of rocks rises nearly to the level of the surrounding region, so
that the descent can be made by a winding inclined plane. Part of the roof, very
thin and worn through in places, is still standing, supported by stalagmite-stalactite
columns. The roots of trees wind about stalagmites or descend as straight and
unsupported stems 15 and more feet through the cave to the water, where there is
the usual breaking up into rootlets.

Ashton No. 1 is a cave of the Modesta type, with a large amount of fallen mate-
rial at one side, where one can descend to the water by means of steps. The diameter
of the cave is possibly 100 feet. A large tree grows in the center of the fallen
material. At the deeper part of the cave, opposite the entrance, the walls retreat
downward. Stalactites and stalagmites are present, and there is a crescent of pools
of water of variable width and depth. Channels tilled with water are seen to lead off
from the bottom of the pools. This is one of the rare places where green plants are
growing in the water. They are confined to the dextral pools as one enters.

Ashton No. 2 differs from all the other caves. It is more open; the roof has
fallen in, so that there are no retreating walls, as in the other caves. The fallen
material slopes gradually to the water, which is shallow and densely covered with
duckweed. A channel leads off from the left end of the water as one enters, and
we went into this with our canoe, but the walls soon came down to the water. The
place was sultrily hot and smelled disagreeably of decaying- plants, so that we were
very glad to get out. This was one of the few caves we visited in which there were
no fishes. This cave practically joins Ashton No. 1 , the dividing wall being but a few
yards wide.

Ashton No. 3, called Los Banos, differs from the others in that considerably
more of the roof of the original chamber remains ■ standing and the amount of
territory covered by the water is also considerably larger. The depth of the water
differs from a few inches to “overhead.” This cave is occasionally used as a bath.
At the time of our visit the water was covered with the usual crust of lime.

Cajio, Jaiguan, and La Tranquilidad were visited by Mr. Riddle.

Cajio, 6 or 8 miles southeast of La Guira do Melena, differs considerably in one
particular from all the other caves. The water lies at a much greater distance from
the entrance than in the others. The entrance is an ordinary sink-hole 10 to 12 feet
deep. Instead of finding the water at the bottom of this sink, as in the Modesta
type, it is perfectly dry. Leading from this, however, there is a dark, narrow
passageway, 100 feet long, which leads to a very large chamber with a crescent-shaped
body of water. This long channel is not an inclined plane, but runs parallel to the

218

BULLETIN OF THE UNITED STATES FISH COMMISSION.

surface. The thin roof of the main chamber has a hole blasted into it, through
which light reaches the water. The floor is of fallen rock, as in other caves

The cave of Jaiguan, 2 or 3 miles east of Cajio, is a chamber 100 feet long by
40 or 50 feet wide. It probably contains more water than any other cave visited,
with the possible exception of Pedregales. There is an opening in the center of the
roof affording considerable light. The entrance is a small sink hole at the edge of
the cave, which is provided with steps cut in the rock. The roots of a tree reach
from the central opening down to the water, a distance of 18 feet. The rootlets,
very abundant here, were found to shelter many of the young blind fishes. In this
cave is a central, rocky islet, formerly completely surrounded by water, as in Modes ta
No. 1. A narrow bridge of rocks has been built, which unites it with the entrance
to the cave.

La Tranquilidad No. 1, in the Cauas region, 3 miles west of Ashton, is of the
Modesto type. It is a large cave, entered only by means of a rope from an opening-
in the roof 4 feet in diameter. The water here presents an extensive surface, but
nowhere is it deep. Large specimens- were taken here. Side channels allow one to
follow the water farther in this cave than in any other visited. The greater part of
the cave is very dark.

La Tranquilidad No. 2 is a small cave with much light. It is in the center of
a large sink probably GOO feet in diameter. The water is not deep and is easily
accessible to cattle and swine. The bottom is formed of very soft, deep mud.

The number and species of fishes taken in these caves are listed under the head
of the various species. In many of the cases 2 blind crustaceans, both of them new
species, were found to be abundant. One of them is a very graceful Palsemonetes ,
the other, a Cirolana , is much more abundant and forms a large part of the food of
the blind fishes. a

My attention was called by Messrs. Wolfe and Hammond to the Pedregales caves
near Alacranes in Matanzas Province, 60 miles east of the easternmost of the caves
from which Pocy recorded blind fishes. They were visited more for the sake of
visiting all the caves that might possibly contain blind fishes than with the expec-
tation of finding any. When near these caves we inquired whether any of them
contained fishes and were told “.yes, but they don’t amount to anything; they don’t
have any eyes.” After this remark we felt thoroughly comfortable in a place where
certainly nothing else contributed to comfort.

Pedregales cave, about 3 miles from Alacranes, differs in some respects from the
caves of the Cabas region. The cave slopes down as steeply as can conveniently be
descended from a narrow opening, but once inside it widens out, descending continu-
ally. The floor was formed of a section of a cone that recalls the central masses in
the bell-shaped caves of Cahas. The floor and the roof were elaborately decorated
with stalagmites and stalactites, some of them united and ranging from 3 feet in diame-
ter to a fraction of an inch. These were pure white when broken, but tinted a red on
the outside l>y the coral earth. When struck they gave out a clear bell-like tone, and
the striking of various- sized columns by different members of the party produced a
pleasing chime-like effect. Among the caves that 1 have visited this is approached
in elaborateness of decoration only by the cave of the fairies in Colorado.

aSee W. B. Hay, in U. S. Nat. Mu?., xxyi, pp. 429-435.

FRESH- WATER FISHES OF WESTERN CUBA.

219

At a vertical depth, judged by the depth of a near-by well to be about 75 feet,
water was encountered in the form of the usual crescent. There was no indication
that we had reached the bottom of the cave and it is not known how deep the water
is, for roof and floor continue to slope down with the same incline and stalagmites
rise from a depth of at least 3 feet beneath the level of the water at the time of our
visit, when it was about 3 feet below its maximum height. The water was covered
with a crust of lime and no fishes were seen.

An account of the Pedregales caves will soon be issued by Mr. Pascual Ferreiro,
of the Cuban railway postal service, a member of the international copyright com-
mission.

About half a mile beyond Pedregales is the M cave, so called from the M-shaped
path that leads from the surface to the water. The descent in this cave is much
steeper than in Pedregales and the stalagmitic decorations much less elaborate. A
dim light penetrates to the water. Here, as in Pedregales, the cave floor continues
to descend for an unknown distance below the level of the water. Fishes were more
abundant here than in any other cave visited. They were all of one species.

We visited another cave in Matanzas Province, about 20 miles north of the
Pedregales, at the edge of the city of Matanzas. It was essentially like the M cave,
but contained no fishes.

THE STREAMS VISITED.

Between the western end of the island and Union, south of Matanzas, a number
of streams run by independent courses from the watershed to the sea. Those west
of San Cristobal are perennial. Those immediately .east of San Cristobal consisted at
the time of our visit of a series of independent pools. East of Artemisa the streams
run above ground only part of their course, then enter caves and continue their
course to the sea underground, or reappear as “ojos de agua” a short distance from
the ocean. In the western part of this region, from Canas to at least Guira de
Melena, there are no dry beds or other surface indications of drainage. Farther to
the west dry beds of streams, narrow and crooked, were seen, but with one exception
there is no perennial stream between San Cristobal and Union except along the coast.
Tl$i one exception is the stream formed by the large springs near G nines. East of
Union we did not go.

The Rio Sabanalamar at San Cristobal is about 20 feet wide and varied from 1 foot
to 2 inches in depth in cross-section. The water is in places very swift and shallow;
in others “over head” and in pools. The banks of the river are clay; the bottom is
grass-grown except in the deep pools and over riffles. We seined up and down from
the railroad bridge and also in an old channel of the river containing a muddy pool
entirely cut off from the river. The water of the river was clear and at 10 a. m. had
a temperature of 23° C.

The Rio Palacios at Los Palacios varied from 5 to 40 feet in width. The water
was clear, the bottom alternately gravel, mud, and weed-grown. Temperature of
the water 23° C. We seined up and down the ford at the end of the main street.
Between Los Palacios and Paso Real the country is in part swampy, with lily ponds.

The Rio San Diego at Paso Real de San Diego is 15 to 40 feet wide with steep
banks about 20 feet high. The water was clear, in pools and riffles, and 23° C. We

220

BULLETIN OF THE UNITED STATES FISH COMMISSION.

seined near the railroad bridge. There are several dry beds of tributaries in the
neighborhood. At Herradura a small creek, 5 feet across and with banks 20 feet deep,
was crossed. No fishing was attempted.

Between Las Ovas and Golpe were ponds with white water-lilies*. No fishing
was attempted here.

The Rio del Pinar is, at the town of the same name, a broad shallow stream with
alternating riffles, pools, and weedy patches. Collections were made above and
below the wagon road near the ice factory.

The Rio Cuy agnate je is the most western river of any size that flows to the
south. Collections were made near Sumidero. The river near this town tunnels
twice through rocky walls several hundred feet high, and in one instance probably
not much more than 200 feet thick, and in the other probably several times as thick.

The Almendares River is a deep and swift stream about 40 feet wide emptying
into the ocean near Mabana. It was scarcely possible to seine in the river itself.
Collections were made above and below a dam at Calabazar and in a small tributary
just above the dam at Calabazar. The temperature was not taken at the time of
seining. A few days later it was 25° at Vento.

The Ariguanabo is of special interest, inasmuch as it is one of the rivers that
disappears in a cave. Collections were made just above its entrance to the cave. It
is a clear, swift stream running through the town of San Antonio de los Banos.
Above the town a dam at the ice factory has deepened the water so that a small
steamer can run up to Laguna Ariguanabo. At the time of our visit the river was
for a long distance above the town blocked with water hyacinths and other water
plants. A much smaller though similar brook which runs through Guanajay also
disappears in the ground several miles below the town. No attempt was made at
seining.

Collections were also made in the outlet of the Yumuri at Matanzas, but we did
not succeed in ascending to fresh water, and no fresh-water fishes were secured.

We ascended the San Juan River from Matanzas to the head of tide water, where
a shallow ford occurs. Collections were made in the ford, above and below the ford,
and in pools of spring water. Immediately above the ford the surface of the stream
was covered with water hyacinths, and the stream was 4 and more feet deep. *At
the ford the water had a maximum depth of about L8 inches, and in places formed
shallow riffles. Below the ford the banks become steep and the water is too deep
for a collecting seine.

PECULIARITIES AND ORIGIN OF THE CUBAN FISH FAUNA.

There are recorded in the present paper 36 species and subspecies. These belong
to 25 genera and 13 families. A number of other species have been taken in the
same region, notably Lepisosteus tristcechus. Of the 37 species and subspecies (includ-
ing the last-named species) but 4, aside from members of the Gobiidae , are found in
fresh water elsewhere. They are the species of Lepisosteus , Symbranchus , Acjgnos-
tomus , and Anguilla. Lepisosteus tristcechus is found in the fresh waters of Mexico
and the southern United States. Symbranchus marmoratus is generally distributed
through the fresh waters of the tropics of America, Anguilla chrysypa is also found

FRESH- WATER FISHES OF WESTERN CUBA.

221

in the streams of eastern North America, and Agonostomus monticola is found in the
fresh waters of the West Indies and Mexico. Of the remaining species only those
of the genus Ileros belong to a strictly fresh-water family. The genus Ileros is gen-
erally distributed in South and Central American waters, one of its members entering
the United States. The members of the marine family of Gobiidae, are found in the
streams and brackish water of tropical America generally, and their presence in
Cuba is not so significant as their absence woidd be. A number of the species
enumerated are marine, and their presence in the rivers may be looked upon as
purely fortuitous; these are Ta/rpon atlanticus, Doryrhampkns l ineatus, Centropornus
undecimadis , Lutianus jocu , L. griseus , Eucinostomus meeki , Gobius soporator, G.
boleosoma , and Lophogobius cyprinoides.

Two species, in many ways the most interesting fishes found in the region exam-
ined, are members of the deep-sea family Erotulidse; they are the blind-fishes Stygi-
cola dentatus and Lucifuga subterraneus. These have evidently worked their way up
the underground streams and are now becoming readapted to the light in the upper
courses of the streams. No other members of the family are found in fresh water
anywhere. Atherina is a marine genus with the peculiar Cuban species as its sole
fresh- water representative. The remaining species are all members of the Poeciliidae, ,
a family inhabiting brackish water and coastwise streams. Of the Pceciliidae. 2
genera, Girardinus and Toxus , are peculiar to Cuba.

The origin of the Cuban fauna is then not far to seek. Wo have, as mentioned
above, a number of marine species, more or less regular visitors of the fresh water.
We have species widely distributed in the brackish water and coastwise streams
whose presence is predicable ( Gobiidae ), and wo have local modifications of families
with a wide distribution in the brackish and fresh waters of the tropics of America
( PoGciliidan ). We have, furthermore, local adaptations of marine species to fresh
water (Broiulidae and Atherina). The origin of all the above is simple of explana-
tion. The species whose presence is of greatest interest are the strictly fresh-water
species of Lepisosteus , evidently belonging to the North American fauna, and Syvi-
branehus and Ileros as evidently members of the South American fauna. The pres-
ence of the eel in the fresh waters of Cuba is to be expected, inasmuch as it very
probably breeds in the ocean near Cuba. The presence of Symbranc/ms , Ileros , and
Lepisosteus tristcechus and Agonostoma monticola shows that the fresh-water fauna of
Cuba has a greater affinity for that of Mexico than for that of Florida, and that these
forms probably reached Cuba by way of Yucatan.

BULLETIN OF THE UNITED STATES FISH COMMISSION.

222

SYSTEMATIC LIST OF FISHES COLLECTED, WITH DETAILS OF DISTRIBUTION.'*

SYM BRAND H1I)£.

Symbranchus marmoratus Bloch. T. Pinar del Rio.

ANGUILLID^E.

Anguilla chrysypa Rafinesque. San Juan, near its first ford; Paso Real.

ELOPIU/E.

Tarpon atlanticus (Cuvier & Valenciennes). M.

Pinar del Rio, 4 specimens, 20, 119, 182, and 192 mm., from a deep pool beneath the wagon
bridge, many miles from the sea. They are locally known as “sadina,” and we had been told that
we should find them in this spot.

PtECILlIILE.

The members of this family are everywhere abundant, especially in streams bordering the cave
region.

Key to the genera of Cuban Pceciliidx.

a. Anal fin of male similar to that of female, oviparous; intestinal canal short, little convoluted; teeth little
movable; dentary bones firmly connected; lower jaw strong and usually projecting beyond
upper.

b. Teeth all pointed, in villiform bands.

c. Air-bladder well developed; no caudal ocellus; gill-openings not restricted above; opercles free from

shoulder-girdle; dorsal and anal nearly equal; origin of dorsal in advance of anal Fundulus.

c. Air-bladder wanting; a black ocellus at root of caudal in male; dorsal smaller than anal, its origin behind

that of anal Rivulus.

b. Teeth tricuspid in one row, no villiform band of teeth; body short and deep, compressed; dorsal short, of

10 to 12 rays, first ray slender and rudimentary Cyprinodim-

aa. Anal fin in male placed well forward and modified into a sword-shaped intromittent organ.

d. Intestinal canal short, little convoluted; teeth all pointed, in bands; eye normal; jaws not produced;

dorsal short, of 6 to 10 rays, behind origin of anal; mouth wide, chin low Gambusia.

dd. Intestinal canal elongate, much convoluted.

e. Teeth compressed, entire, without lateral cusps.

/. Anal process in male very long, serrate behind near tip and with finger-like claspers (the prepuce
being modified into a pair of clasps); dorsal in both sexes behind origin of anal.

(j. Dentaries and intermaxillaries firmly united; teeth of outer row fixed, a band of minute teeth
behind them.

li. Teeth of outer row much expanded at tip, broadly spade-shaped in upper jaw, close-set, their
margins overlapping; teeth near middle of lower jaw asymmetrically expanded, lateral lobes

prolonged and ending in a point , Glandickthys.

hh. Teeth of outer row wide-set, scarcely expanded, spear-shaped, those near middle of lower jaw

in two irregular series Tuxes, I nov.

gg. Dentaries and intermaxillaries loosely joined; teeth of outer row movable, inserted on lips, a
few teeth behind them or none, those of outer row wide-set, scarcely expanded, spear-shaped;

those of middle of lower jaw in two irregular series Girardinus.

ff. Anal process comparatively short, a leaf-shaped prepuce attached to the anterior surface covers
the tip; tip without claspers; dorsal in female in advance of origin of anal,
i. Tip of anal process in male ending in a simple antrorse hook, no serrae on its posterior surface. . -Pxcilia.
ee. Teeth all pointed; origin of dorsal behind that of anal Heterandria.

Fundulus cubensis Eigenmann, new species. *

l am somewhat in doubt as to the generic position of this species. Its short intestine, double row
of teeth, unrestricted gill-openings, and position of its dorsal in relation to its anal, and similarity of
sexes (at least one of the three specimens is a male) seem to indicate that it is a species of Fundulus.
Type: No. 9667, Ind. Univ. Mus., 29 mm. long; Pinar del Rio.

Cotvpes: Two specimens, 26J mm. long; Pinar del Rio, at the ford just above wagon bridge.

Plead 3.6, about equal to depth; D. 11 or 12; A. 10 or i, 10; scales 24. Origin of dorsal very slightly
nearer tip of snout than base of middle caudal rays and over the eighth scale of lateral line; origin

«The following characters are used to indicate the general distribution of the genera and species enumerated;
t, genus peculiar to Cuba; *, species peculiar to Cuba; T, generally distributed in tropical fresh waters; M, marine species.
b To£ov=a quiver full of arrows.

FRESH- W ATElt FISHES OF WESTERN CUBA.

223

of anal under eleventh scale. Dorsal and anal high, angular behind, last ray somewhat produced, 4.3
in length to base of caudal; tip of last anal ray reaching base of caudal, tip of last dorsal ray nearly so;
ventrals reaching to origin of anal; pectorals about to middle of ventrals; mouth very small and very
oblique; eye longer than snout, 3 in head, very little less than interorbital; profile straight in front,
slightly convex toward dorsal. Teeth all conical, in at least two series, those of outer series enlarged.

Coloration brilliant. Ventrals and anal without pigment, caudal with two or more cross-series of
spots; a dark humeral spot just above origin of pectoral; dorsal dusky; a dark band through lower
part of eye forward below chin; a dark band from eye to tail, its median half much darker than edges,
bordered above and below by pigmentless bands bright orange in life, the upper one beginning below
origin of dorsal; back above upper light band thickly covered with pigment spots; below lower light
band a dusky band of equal width consisting of a series of contiguous V-shaped markings from axil of
pectoral to tail; below this another pigmentless band extends from axil of pectoral to just behind anal ;
below this is a narrower band of dusky from axil to last anal ray; below this another short pigment-
less band; edge of branchiostegal membranes and a line forward from their union to chin black;
ventral surface otherwise without pigment except a faint line between ventrals.

Gambusia punctata Poey. *

Everywhere abundant. The following specimens were collected: Ban Antonio, 3 females, the largest
92 mm. long; Modesta No. 2, 3 females, the largest 84 mm.; Modesta No. 3, 4 females 37 to 72 mm.;
San Cristobal, 50 females 31 to 84 mm.; 11 males 40 to 53 mm.; Palacios, 18 females 54 to 88 mm.; 5
males 50 to. 61 mm.; Paso Real, 2 females 50 to 55 mm.; Sumide.ro, 3 females and 1 male; Pinar del
Rio, 9 females, the largest 84 mm.; San Juan near Matanzas, 30 females, the largest 70 mm.; 5 males,
the largest 42 mm. ; Calabazar, 7 females, the largest 75 mm.

Gambusia puncticulata Poey. *

This species is shorter and deeper than Gambusia punctata , which it greatly resembles. The
dorsal in the specimens taken counts 8 to 10, not 11, as Carman found in his specimens.

San Antonio, 3 females, the largest 58 mm., D. 9; San Cristobal, 4 females, the largest 48 mm., D. 8,
9, and 10; Palacios, 1 female 39 mm., D. 9; Pinar del Rio, 4 females, the largest 47 mm., and 3 males,
the largest 35 mm.

GLARIDICHTHYS Garman

Intestinal canal elongate; males with anal tin modified into a very long intromittent organ; jaws
much more firmly united than in Girardinus, each with a series of close-set, broad-tipped, entire or
but slightly crenulate teeth, these teeth not movable, a narrow band of smaller, broad-tipped, conical
or tricuspid teeth behind them; fins small, anal in advance of dorsal in both sexes.

Most nearly allied to Gooclea and Girardinus , and differing from them in character of teeth, Goodea
having tricuspid teeth, Girardinus having movable loosely-set, and Toxus having hastate teeth.
Glaridichthys uninotatus Poey. *

Abundant. Fifty specimens preserved from San Cristobal, 47 to 84 mm. long; a number of
specimens have, in addition to the lateral spot, a spot on either side of anus and sometimes a black
streak connecting the two; in one instance the lateral spot on one side is entirely replaced try the anal
spot; males (13) 38 to 47 mm.

224

BULLETIN OF THE UNITED STATES FISH COMMISSION.

At Palacios 24 specimens were preserved, 50 to 81 mm. Fluctuations in the lateral spots of female
as in San Cristobal specimens; in one specimen only a small spot on one side, none on other.

Paso Real, 2 females; Pinar del Rio, 4 females.

Inner series of teeth in upper jaw small, with sharply triangular or tricuspid teeth in 4 or 5 rows;
inner series of lower jaw little expanded at tip. Teeth of outer row near middle of jaws irregularly
expanded, lateral lobes prolonged into a point; lateral teeth of lower and of upper jaw equally expanded.

Glaridichthys falcatus Eigenmann, new species. *

Type, No. 9664, Ind. Uni v. Mus., a female, 82 mm. long, from San Cristobal.

Cotvpes: Eight females from ahold river channel at San Cristobal, the smallest 60 mm. long, the
largest 85 mm. ; 4 females 50 to 53 mm., from Palacios, taken in a muddy pool in the river bed at the
ford; 8 females and 2 males from Rio del Pinar, the females 38 to 47 mm., the males 29 and 37 mm.
This species reaches its maximum size and is most abundant in warm, muddy pools.

Fig. 3. Glaridichthys falcatus Eigenmann, new species. Female.

Body long, slender, little compressed; head 4; depth 4 (in pregnant females 3.5); D. 9; A. 11;
scales 29; head broad, wedge-shaped in profile, with lower jaw very oblique, projecting; eye very
large, longer than snout, 2.6 in head, 1.4 in interorbital; mouth very oblique, small; interorbital
divided into 3 distinct regions by longitudinal grooves, central portion convex; origin of dorsal equi-
distant from base of middle caudal ray and origin of pectoral; dorsal and anal falcate; second rays
sickle-shaped, each extending for one-third its total height beyond tip of last ray when folded back,
little less than length of head; caudal emarginate, some of outer rays prolonged; origin of anal in
female about equidistant from base of middle caudal ray and anterior margin of eye, its seventh ray
under origin of dorsal; ventrals usually reaching to anal; pectorals about to middle of ventrals in
female, to base of anal in male.

San Cristobal specimens very pale; a dusky streak from nape along middle of back to caudal,

FRESH WATER FISHES OF WESTERN CUBA.

225

scales above lateral line faintly edged with black; a black line along middle of sides composed of a
single series of chromatophores; a black streak along ventral surface from anal to caudal; otherwise
colorless.

Palacios specimens colored like those from San Cristobal, except one in which each scale of the side
below lateral line is edged with a series of chromatophores and there is a faint hint of 8 dark spots
along median black line; region above lateral line dusky.

Pinar del Rio specimens colored like the darker Palacios specimen, sometimes a black streak on
either side of anus and forward to ventrals. Male with modified portion of the anal very long, with
retrorse spines behind and a little clasper on tip of longest rays.

Glaridiehthys torralbasi Eigenmann, new species. *

Type, No. 9662, Ind. Univ. Mus., male, 45 mm. long, from Pinar del Rio, represents apparently a
new species. In general appearance it very greatly resembles the males of Girardinus vuiallicus and
Glaridiehthys uninolatus. From the former it differs in possessing bands of teeth in jaws behind
spatulate row, first row of teeth fixed, and dentary and premaxillaries much more firmly united;
from the second it differs in coloration, having no lateral spot and a conspicuous dorsal band. There
are no other species which it resembles.

D. 9; A. 10; scales 28; head 3.6; depth 3.3; body compressed, elongate; mouth small, subterminal,
lower jaw not prominent; eye greater than snout, less than 3 in head, equal 'to interorbital; teeth in
outer series overlapping, those of lower jaw more pointed; a band of minute teeth behind at least
the front row in upper jaw; dorsal small, its first ray equidistant from base of middle caudal rays

and from eye; its highest ray nearly equal to distance of pectoral from anterior margin of eye; caudal
truncate, equal to distance of pectoral from tip of snout; anal process long, with seme behind, and
a clasper at end of anterior rays; ventrals small, not much longer than eye; pectoral reaching to anal;
dorsal with an arched band reaching from full length of first ray to base of last; caudal, ventral, and
pectoral colorless; first ray of anal process largely black; sides with about 10 dark cross-streaks; all
scales margined with black, dorsal ones not heavily so; a dark dorsal streak, a black ventral line
behind anal, a dusky band around head just in front of eyes.

I take pleasure in naming this species for Prof. Jose I. Torralbas, of the chair of zoology in the
University of Flabana.

Girardinus metallicus Poey. f *

San Cristobal, very abundant ; largest female, 77 mm. long; largest male, 45 mm. A few females
with black on anal.

Palacios; largest female, 69 mm.; largest male, 41 mm.; ventral band from chin to tail black in
one male; one male blotched with black.

San Antonio; largest female, 79 mm.

Ashton; females usually with black on anal; largest female, 51 mm.; largest male, 38 mm.

Pinar del Rio; largest female, 68 nun.; largest male, 41 mm.; a number of males with a black
streak of varying intensity and width along the ventral surface.

F. C. B. 1902—15

226

BULLETIN OF THE UNITED STATES FISH COMMISSION,

Girardinus garmani Eigenmann, new species. f *

Type, No. 9661, Ind. Univ. Mus., one male, 35 mm., Dinar del Rio. Cotypes, one male, 35 nun.,
Pinar del Rio; one male, 38 mm., Palacios.

D. 9; A. 9; scales 29; depth 3.4 to 3.6; head 3.6 to 3.8. Body compressed; head truncated, lower
jaw nearly vertical; mouth ver}r small, lips thick, teeth in a single series in each jaw, sloped as in G.
metallicus, very movable, interinaxillaries and dentaries not united; eye as long as snout, 2.5 in head,
equal to interorbital; origin of dorsal a little nearer pectoral than base of middle caudal rays; dorsal
rounded, small; dorsal, caudal, and pectoral of about equal height, equal to distance of pectoral from
eye; anal process 2.5 in length, serrate near its tip behind and with a clasper at its end; ventrals
very small, reaching to the anal; ventral surface colorless except a black line from anal to caudal;
scales of side with a dark margin of increasing width toward back; a dusky dorsal streak; head in
front of eye dark above and below; region below eye colorless; a well-defined black spot on base of
last 5 dorsal rays; first dorsal ray black ; anal process blackish on basal half, a small indistinct black
spot on distal half of last anal membrane and extending at times on neighboring regions; sides
without streaks or bars.

This species differs from G. metallicus and G. denticulatus in being entirely without, streaks or
spots or bars on the sides, in the number of anal rays, and in other characters.

TOXUS Eigenmann, new genus. f.

Toxus Eigenmann, new genus of Pieciliid; u ( riddlei ).

This genus differs from Glaridiclithys in its narrow teeth, from Girardinus in having its jaws firmly
joined.

Toxus riddlei Eigenmann, new species, f*

Type, No. 9656, Ind. Univ. Mus., a female, 59 mm. long, from San Cristobal.

Cotypes: 1 female 66 mm. long, and 2 males 33 and 34 mm. long, from San Cristobal. Head 4;
depth 3.4; D. 9; A. 10; scales 28; origin of dorsal midway between base of middle caudal rays and origin
of upper pectoral ray and over thirteenth scale of lateral line; origin of anal below eleventh scale; fins
moderate; longest dorsal ray equaling length of head without snout; eye equaling snout, little more
than 3 in head; interorbital convex, equaling snout and eye; profile slightly curved; outer row of
teeth movable, spear-shaped, not very closely set, brown-tipped, a band of minute teeth behind them;
a dark lateral band crossed by about 6 indistinct dark crossbands; scales of sides with a light center
and a narrower or broader margin of dark, forming reticulations; a dark streak extending down between
eye and angle of mouth ; rest of lower side of head and belly white; pectoral colorless; caudal faintly
dusky; anal with a faint dark band through the middle, the tips and base colorless; dorsal dusky.

Male much smaller, the color contrasts sharper; about 7 well-marked dark crossbands in the
larger specimen; dorsal tipped with dusky, a black band from base of last dorsal rays forward toward
basal third of fourth dorsal ray; in the larger specimen a series of dark spots on dorsal rays on a level

FRESH -WATER FISHES OF WESTERN CURA.

227

with tip of first one; a black band through middle of anal, most intense upon and entirely covering
last rays; first fully developed ray black in the large specimen, colorless in the smaller; origin of
anal under eighth scale.

I take pleasure in naming this species for Mr. Oscar Riddle, to whom much of the success of the
expedition to Cuba is due.

Poecilia vittata Guichenot. *

San Cristobal, over 50 females, the largest 94 mm. long; a number of them with peculiar black
blotches; the small ones with 3 yellow stripes below the lateral line and frequently with one or more
series of black dots along lower part of side; 32 males, the largest 70 mm.

San Antonio, 5 females, the largest 110 mm.; 2 males, the largest 73 mm.

Los Palacios, 70 females, the largest 100 mm., showing great variability in the intensity of color-
ation; 35 males, the largest 65 mm.

Calabazar, 9 females, the largest 102 mm. ; 1 males, the largest 70 mm.

Paso Real, 3 females, the largest 80 mm.; 1 male, 53 mm.

Pinar del Rio, 17-females, the largest 83 mm.; 9 males, the largest 63 ram.

Sumidero, 2 females.

Heterandria cubensis Eigenmann, new species. *

Type, No. 7663, Ind. Univ. Mus., a female, 59 mm., from Los Palacios.

Cotypes, one female with young, 53 mm. , from Los Palacios, and one female, 38 mm., from Pinar del
Rio. These specimens agree well with the characters of the genus Heterandria, as restricted by Gar-
man, except that the outer series of teeth are movable. Head 5 to 5.2; depth 3.3 to 3.7; D. 9; A. 10;
scales 29. Body elongate, little compressed, general shape that of Fundulus; profile regularly curved
from dorsal fin to eyes, flattened over eyes and forward; mouth small, opening upward, the lower jaw
projecting; bones of jaw loosely united; eye longer than mouth, 2.5 to 2.7 in head, very little less than
interorbital; origin of dorsal a little nearer head than base of middle caudal rays, over first third of anal;
dorsal and anal both falcate, the anterior rays extending considerably beyond tip of last when laid
back; highest dorsal ray slightly shorter than highest anal ray, about equal to length of head; caudal
a little longer than head; ventral reaching to anus; pectoral reaching to ventral.

Scales of the mid-dorsal line with their dorsal halves dusky, those of entire side margined with

228

BULLETIN OF THE UNITED STATES FISH COMMISSION.

black, most distinct above lateral line; a series of 12 narrow dark vertical bands about as wide as pupil
and as high as eye; an irregular black streak along middle of side; dorsal tipped with dusky, the first
membrane black; a black line and a few chromatophores along each anal ray; a black line from anal
to caudal; chin dusky; pectoral, ventral, and belly colorless.

SYNGNAT1IID/E.

Doryrliamplius lineatus (Valenciennes). M.

Two specimens from the Rio San .Tuan below the ford.

ATHERINI1DE.

Atherina evermanni Eigenmann, new species. *

Type No. 9657, Inch Univ. Mus., 45 mm., from San Cristobal. Cotypes, 35 specimens, 36 to
50 mm., from San Cristobal, and 5 specimens, 28 to 38 mm. long, from Pinar del Rio.

D. v— i, 9-11; A. i, 12 to 15; head 3.5 to 4; depth 4.5 to 5; scales 32; head rather pointed; mouth
oblique, the lower jaw projecting; maxillary reaching a little beyond front of eye; eye 2.75 in head,
equal to distance from tip of snout to anterior margin of pupil; little wider than interorbital; teeth
minute; spinous dorsal inserted behind tijas of ventrals, its origin equidistant from tip of snout and
middle caudal rays, or a little nearer caudal; caudal peduncle very slender, its least depth less than-

Fig. 9. Atherina evermanni Eigenmann, new species.

length of eye; caudal little less than length of head; anal inserted in advance of origin of dorsal; ven-
trals small, not reaching anal ; pectoral reaching tips of ventrals; a conspicuous lateral band most
intense on caudal peduncle, gradually fading out under pectoral; region above this in all cases thickly
peppered with black cells, most thickly so along median line; region below this in many cases
similarly but less intensely spotted; lower side of head and breast white; ventrals nearly free from
pigment; ail the other fins with pigment, cells of greater or less intensity along the rays. This species
is readily distinguishable from the other species of Atherina by the smaller number of scales.

I take pleasure in naming this species for Dr. Barton Warren Evermann, in recognition of his
valuable work on the fishes of the West Indies, especially his work on the fishes of Porto Rico.

FRESH- WATER FISHES OF WESTERN CUBA.

229

MUGILID£.

Agonostomus monticola (Bancroft). T.

Rio San Juan, near its first ford, abundant, the largest 140 nun. long; Sumidero, abundant, the
largest 170 nun. ; Pinar del Rio, abundant, the largest 160 mm.

CENTRO POM M.

Centropomus undecimalis (Bloch). M.

Rio San Juan, at its first ford, abundant, varying in length from 55 to 250 mm.

LUTIANID^E.

Lutianus jocu (Bloch & Schneider). M.

Two specimens from the Rio San Juan, just below the ford.

Lutianus griseus (Walbaum). M.

One small specimen from the Rio San Juan, just below the ford.

GERRILME.

Eucinostomus meeki Eigenmann, new species. M.

Type, No. 9660, Inch Univ. Mus., a specimen 135 mm., from San Juan River, just, below its first ford.
General appearance of Uliema lefroyi, differing from all other species of the genus Eucinostomus in
having but 2 anal spines.

Head 3.25; depth 3; 0. ix, 10; A. ii, 8; scales 4-46-9 ; eye 1 in snout, 3 in head, 1 in interorbital.
Body elongate, little compressed or elevated, the dorsal profile but little more elevated than the
ventral; snout pointed, the profile from snout to dorsal gently arched; mouth narrow, terminal, but
little above the lower margin of the eye; maxillary reaching to vertical from front of eye, 3.4 in head,
its exposed part boat-shaped, a trifle more than twice as long as wide, 5 in head; intermaxillary
groove entirely naked, its width 5 in the interorbital; preopercle and preorbital entire; dorsal spines
slender, the second longest, 6 in the length; ventrals short, reaching half way to anal; pectoral long,
3.5 in the length, reaching beyond tips of the ventrals, but not to vent. First anal spine minute, the
second equal to the length of tl le eye.

Color, ashy gray, with some metallic reflections; dusky lines along the rows of scales; sides and
back everywhere punctate with minute dots; vertical fins dusky; ventrals and pectorals lighter.

Named for Dr. Seth Eugene Meek, assistant curator of zoology, Field Columbian Museum, in
recognition of his excellent work on Mexican fishes.

230

BULLETIN OF THE UNITED STATES FISH COMMISSION.

CICHL1D/E.

Individuals of the genus Heros are as numerous in the streams of Cuba as individuals of the Cen-
trarchidn ■ are in the streams of equal size in the Ohio valley. They were found by us down to tide
water, but not in it. Only a single species has been recorded from Cuba, and nothing has been said
either concerning its distribution or its variation. No one, except possibly Poey, has before this com-
pared numbers of specimens from different places or even from the same place. Such a comparison
is therefore very desirable, and the material collected far surpasses all other collections made before.
We have altogether 236 specimens from various localities. An examination of all of these proves
either the presence of several instead of a single species on the island or a remarkable variation with
localities. A definition of the variations has proved very elusive. The numbers of fin rays and scales
are uniform, so that the differences exist in the proportions and the color. But the coloration also
has a certain underlying uniformity. There is a spot near the middle of the side, another at the base
of the caudal, and an obscure third above the gill-opening. There are numerous small spots on the
fins and on scales of the sides, especially below and on the opercles, and sometimes on the cheeks.
There is also a longitudinal streak from the eye through the lateral spot to the caudal spot, and a

definite number of crossbars, both streak and bars most conspicuous in the young and in light-
colored adult individuals. This uniformity of underlying structure makes defining of species or varie-
ties a difficult proceeding. The polymorphism is further complicated by instances like the following:
The specimens from San Antonio are readily referable to a certain form found at Calabazar, although
they differ from Calabazar specimens in quite readily distinguishable features; but one of them differs
notably from all other specimens collected at San Antonio, and would unhesitatingly be considered a
species distinct, from the other specimens from the same locality. But at Palacios the same form
branching from the Calabazar form approaches the characters of the single specimen from San Antonio.

I venture to describe here certain of the most aberrant forms as new, without, however, feeling
sure that they are really distinct varieties or species or that some of the other forms referred to II. tetra-
canthus are not also new.

Heros tetracanthus Cuvier & Valenciennes. *

Heros tetracanthus torralbasi Eigenmann, new subspecies.

(25 specimens, 60 to 181mm. long, from Calabazar.)

These specimens come from the Almendares River, and as this flows near to Habana it is very
probable that the type of Ileros tetracanthus came from the same river. Cuvier & V alenciennes say that
Poey’s drawing, on which their tetracanthus was based, resembled Ambloplites in outline, and possessed
spots in the angles of the scales. This very well describes some specimens I have (figs. 12 and 13).

FRESH -WATER FISHES OF WESTERN CURA.

231

D. xv, 11; A. iv, 9; depth greatest below lirst dorsal spine, 2.5 to 2.7; depth of caudal peduncle
2.5 to 2.25 in head; scales 27 to 29; pores 17 to 21 + 10 to 13; body heavy forward, tapering from the
shoulders to the caudal peduncle; jaws heavy, lips thick; snout 2.75 to 2.5 in head; eye 4.5 to 5 in
head (3.5 in young), 1.5 in interorbital; no pore in upper angle of gill-opening (except on one side of
one individual.); highest dorsal and anal rays reaching base of caudal; highest dorsal rav 4.5 to 5.3 in

Fig. 12. Heron tetracanthus tetracanthus Cuvier & Valenciennes. (Type B.)

— " i ( •+ tv. ; /

, v . V.

1 vVi i

n At I

l * i i > f mVa

m.u., A i

\ )j »> I ('■) t'.i t S ): f f /»; 4! »lj

iUJUX i )\ mm
IK K i X }: > .. * , X:XX

Fig. 13. I-Icron tetracanthus tetracanthus Cuvier & Valenciennes. (Type C.)

the length; last dorsal ray 3 in the longest; lateral and caudal spots conspicuous in young, which have
a series of light crossbars; two light bars usually confluent over the lateral spot; fins dusky, the verti-
cal ones lighter-edged and with some spots on their bases; no spots on head or body in the youngest;
in larger ones spots appear about the base of the pectoral, opercle, angle of preopercle and mandible-
There is great variation in the distinctness of the lateral bands.

232

BULLETIN OP THE UNITED STATES FISH COMMISSION.

I n the larger specimens! there are 2 types of coloration ; type A has more or less distinct vertical bars,
alternating light and dark; the lateral and caudal spots are distinct, the dark crossbars are darkest in
a line between the two; a dark streak extends from the eye to second dark bar, this with the darker
areas on the crossbands forming an interrupted lateral band; cheeks unspotted, operclesand mandible
with dark spots; ventral surface plain; a few scattered spots along the sides; vertical fins more or less
spotted at the base. Type 15 shows no crossbands, each scale of the side with a dark spot forming longi-
tudinal series; cheeks as well as opercles and mandible spotted or the spots -confluent into lengthwise
streaks; vertical fins more conspicuously spotted; cheeks in the young of both types unspotted; sides
of the young of type B less regularly spotted than in the adult. Figs. 11 and 12 are drawn from males,
of the same size and with reproductive organs in the same stage of development. Fig. 12, type B,
evidently represents the variety figured by Boev, and is the original tetracanthus. Fig. 1 1 , representing
type A, may be termed II. tetracanthus torralbasi Eigenmann, var. nov. (No. 9672, Ind. Univ. Mus. ).

(78 specimens, 44 to 160 mm. long, from San Cristobal. )

Most of these were taken out of a muddy lagoon near the river, and all were very pale in color, the
crossbars showing well. In these paler specimens there are no indications of a longitudinal stripe.
The largest, which also came out of the lagoon, is nearly uniform light ashy, there being but faint indi-
cation of crossbars and spots; there are faint spots on fins and opercle. Other smaller specimens are
everywhere profusely spotted. In the darker specimens from the river there is a dark lateral band.

Fig. 14. Ileroa tetracanthus ejriseus Eigenmann, new subspecies

(70 specimens, 47 to 190 mm. long:, from Los Palacios.)

These specimens are of types A and B from Calabazar, with some distinct features. The dark spots
(of B) along the rows of scales are, in some of the lighter individuals, nearly faded out; in the darkest
ones they spread nearly over entire margin of scales (fig. 13). The cheeks are spotted or streaked in some
of the largest specimens and not in others. The lateral band in some specimens of Type A is as well
developed as in some San Antonio specimens; vertical bars vary also very much in intensity. One
dark specimen resembles in almost all respects lleros griseus from San Antonio. Depth, 2.25 to 2.5.

(17 specimens, 62 to 155 mm. long, from Pinar del Rio, are of A and B from Calabazar.)

(4 specimens, 65 to 240 mm. long, from Sumidero.)

All these specimens are dark, the smaller one nearly uniform, with but faint crossbars; lateral
and caudal spots distinct. The largest one is a male, very dark above, without distinct markings, and
with black streaks and spots on cheeks, opercles, and lower sides. Depth 2; head 2.6; eye 5 in head.
These may referred to tetracanthus.

FRESH- WATER FISHES OF WESTERN CUBA.

233

Among 32 specimens from San Antonio there are 3 distinct types, one of which may simply be the
adult of one of the others. They are all elongate, the depth being 2.3 to 2.5 i . length. There are, in
the first place, 4 adults measuring 1(50 to 190 mm. in length, of type B from Calabazar. They differ from
the Calabazar specimens in having the spots along the scales larger and less regular. In two of the
specimens the cheeks are spotted, in two others the spots are confluent into vertical or longitudinal
streaks. In the largest some of the dorsal rays are prolonged, reaching to near middle of caudal. These
are probably the adult of 17 specimens from 92 to 157 mm. long. In all of these, even in the smallest,
there are spots on the cheeks, more numerous in the larger, and confluent into streaks in the largest.
Sides irregularly spotted; lateral band more or less conspicuous as the fish is lighter or darker; there
are traces of the usual light and dark bands in some individuals. All of these are evidently typical
II. teiracanthus.

There are 11 specimens, 90 to 123 mm. long, evidently modifications of II. ietracnnthus torralhasi
from Calabazar, in which, except one very dark specimen, the lateral band is very conspicuous. Ver-
tical bands quite well marked except in darkest specimens. Cheeks are unspotted except in one
individual which has faint spots; sides without small dark spots.

Fig. 15. Heron teiracanthus talus Eigenmann, new subspecies.

Heros tetracanthus griseus Eigenmann, new subspecies. *

Type: No. 9670, Ind. Univ. Mus. ; a specimen 117 mm. long, from San Antonio.

D. xv, 11; A. iv, 8; scales 27; pores 17 + 10; depth a little more than 2.5; depth of caudal
peduncle 2.5 in head; head 2.7 in the length. Shape and general characters of Heros teiracanthus
from the same place, differing in the color and the notably larger eye, as compared with specimens of
II. teiracanthus of the same size. Eye 3.7 in head (4.6 in II. tetracanthus of the same size) ; 1 in inter-
orbital (1.5); 1.25 in snout (1.7); preorbital five-sevenths of eye; snout 3 in head; highest dorsal
and anal reaching base of caudal; highest dorsal 4.7 in the length; highest anal 5; vent.ra.ls reaching
to vent.. No lateral spot; a faint caudal spot; sides ashy with irregular dark spot; a few whitish
streaks through some of the scales above lateral line; cheeks plain, a few spots on opercles; soft
portions of vertical fins spotted at base; no traces of dark crossbars.

234

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Heros tetracanthus latus Eigenmann, new subspecies. *

Type: No. 9669, Ind. Univ. Mus., a specimen 160 mm. long, from San Juan.

This is a narrow, deep fish with projecting lower jaw, a pointed snout, and a depression in the.
profile over eyes. D. xv, 11; A. iv, 9; depth 2; head 2.7; depth of caudal peduncle 2 in head; scales 28;
pores, 18 + 12; snout pointed 2.75 in head; eye 4.5 in head, 1.5 in interorbital; no foramen in
upper angle of gill-opening; maxillary reaching vertical from front of orbit; highest dorsal and anal
rays reaching to end of basal two-fifths of caudal; highest dorsal ray twice as high as last ray, equaling
longest caudal ray, equaling length of head without opercle; highest anal ray slightly shorter; ventrals
to vent.

Ashy gray, darker above to light below; each scale of lower parts of side with black spot on tip,
extending over to next scale and forming distinct series; lower part of opercle, preopercle, lower part
of cheek, and lower jaw with dark spots; vertical fins dusky; soft dorsal and anal spotted; ventrals
dark, inner ray light; pectorals pale.

Heros tetracanthus cinctus Eigenmann, new subspecies. *

'type: No. 9671, Ind. Univ. Mus., a specimen 129 mm. long, from Paso Real.

Four specimens were taken at Paso Real measuring a 68, h 72, c 129, and d 136 mm. long, respec-
tively, and differing in coloration from those taken at any other point. The two larger are very dark,
one having very distinct markings.

Fig. 16. Heros tetracunthus cinctus Eigenmann, new subspecies.

The specimens approach 7/. nigric.nns but have a normal lateral line: c (the type) is most aberrant
in its coloration, d approaches the coloration. of tetracanthus torralbasi from the Almendares, a and 5 are
indistinguishable from other young except in the band through the lateral spot.

I), xv, 11 (c); xvi, 10 (d); xv, 10 ( b ); xvi, 11; A. iv, 8; scales 28, pores 15 + 9 («); 16 + 8 (b);
15 +9 (c) ; 19+11 (d); head 2.6 to 2.7; depth 2.25 to 2.2; depth at end of opercle 2.4 to 2.43; snout
pointed, 2.75 to 3 in head; mouth horizontal, maxillary nearly concealed when the mouth is closed;
maxillary about reaching vertical from front of orbit; highest dorsal and anal raj's reaching little
beyond base of caudal 5 in length, last dorsal ray 2.3 in the longest; ventrals reaching vent; gill-opening
with a supplementary pouch above, reached by a larger or smaller foramen.

FRESH-WATER FISHES OF WESTERN CUBA.

235

Dark; a black lateral and a black caudal spot; side with 7 light crossbars; a light streak across
nape from upper angle of gill-opening, another across from behind eye, another between eyes; snout
light; sjraces between light bars form dark bars of about equal width except lirst two on body, which
are much wider at the top; a black streak from eye across upper angle of gill-opening to the second
dark bar; cheeks and opercles with black spots and streaks, ventral surface spotted; vertical fins and
ventrals dark; soft dorsal and anal with obscure spots; pectoral pale, unspotted. The dark band in
which the lateral spot is located continued to the back.

Young much lighter colored, a few dark spots along opercle and below cheek; ventral surface
unspotted, lateral and caudal spots conspicuous.

A fifth specimen, 152 mm. long, from Paso Real, is a typical Mrar.anthus.

Heros nigricans Eigenmann, new species.

Type, No. 9668, Ind. Univ. Mus., a specimen 192 mm. long, from Pinar del Rio.

One specimen was secured. It is the most prominent of the aberrant forms, and I should unhesi-
tatingly describe it as a distinct species if I had obtained more than one specimen.

D. xiv, 11; A. iv, 10; scales 28; pores about 15 H O ; head 2.6; depth a trifle less than 2 in the
length; depth at end of opercle 2 in length +£ diameter of eye; depth over middle of eye equals
length of head, less one ocular diameter; snout pointed, upper and lower profiles nearly equally
inclined to behind eye; upper profile gibbous behind eye; maxillary very little exposed when mouth

Fig. 17. Heros nujricans Eigenmann, new species.

is clos'd; eye 1.5 in interorbital. Gill-cavity, with a small supplementary pocket at its upper angle-
entered by a large foramen; lateral line irregularly developed on left side; no pores on either side o.
tail; soft dorsal falcate, fifth ray as long as caudal, 3 times as long as last ray, its tip reaching nearly to
middle of caudal, 3.6 in length, longer than head, less opercle; caudal broadly rounded; anal falcate
fourth ray longest, equaling head, less opercle; ventrals reaching to vent.

Color everywhere nearly black, with 7 lighter crossbars, 2 of which are on caudal peduncle; an
additional light streak from angle of gill-opening across nape; cheeks and opercles with lighter mark-
ing; fins nearly uniform black.

236

BULLETIN OF THE UNITED STATES FISH COMMISSION.

GOBIIDrE.

Philypnus dormitator (Lacepede). T.

Rio San Juan, and seen in Yen to springs near Havana; very abundant in the San Juan, where
specimens 46 to 256 mm. long were obtained; young with black stripe from tip of lower jaw to caudal.

Dormitator maculatus (Bloch). T.

Rio San Juan, at its mouth and at the first ford; very dark; lagoon at San Cristobal, very light.
Eleotris pisonis (Gmelin). T. Rio San Juan, at its mouth and at the first ford.

Lophog'obius cyprinoides (Pallas). M. Rio San Juan.

Gobius soporator Cuv. & Val. M. Mouth of Rio San Juan at Matanzas and at its first ford.
Gobius boleosoma Jordan & Gilbert. M. Mouth of Rio San Juan and at its first ford.

Awaous taiasica (Lichtenstein). T. Sumidero.

BROTULIDiE.

Styg-icola dentatus (Poey). f*

This blind fish was taken in the M cave near Alacranes, 20 specimens; Jaiguan, 5 specimens;
Frias, 2 specimens; Modesta, 4 specimens; and Tranquilidad, 7 specimens; caves near Cafias without
specific locality, 5 specimens. In all, 43 specimens were secured, ranging from 60 to 152 mm. long.
Poey records this species from Cajio and Ashton, in which it was not found by us. Pie also records it
from La Industria, which is said to be between Cajio and the Cafias caves.

The males of Stygicola dentatus are distinctly larger than the females. Average length of the 20
females caught is 97 + mm., the largest one 120 mm. The average size of the 23 males is 113 + mm., the
largest one being 152 mm. long. The males were in excess of the females in the ratio of 100 females
to 115 males. There is but an appreciable difference in the averages of the fins, as far as these could
be counted, the average formula for the females being D. 91.4, A. 74; and for the males D. 91.1, A. 73.6;
or the average for the two are I >. 91.2, A. 73.6.

Lucifuga subterraneus Poey. t *

This species was taken in all but one of the caves in which Stygicola dentatus was taken, and in
several others besides. The localities are Ashton, 13 specimens; Los Banos, 5 specimens; Cajio, 3 speci-
mens; Hawey, 16 specimens; San Isidro, 2 specimens; Jaiguan, 18 specimens; Las Frias, 5 specimens;
Modesta, 2 specimens; Tranquilidad, 3 specimens; Cafias, without specific locality, 9 specimens; total,
76 specimens, ranging from 24 to 94 mm. long. The females of Lucifuga subterraneus are distinctly
larger than the males. In making the average for the size of the sexes individuals less than a year old
were not considered, because the differences in the sexes would, if present, be but very slight, and
because in such young the sex could not always be determined with certainty. An examination of all
specimens makes it probable that at the end of a year after birth the young are about 50 mm. long.
In obtaining the average size of the sexes only specimens over 50 mm. were considered. The males
above this size measure 59.7 mm. on an average, with a maximum of 94 mm. The females measured
71.1 mm. on an average, with a maximum of 93 mm. Of the specimens over 50 mm. long 23 were males
and 22 females, or 100 females for every 104.5 males. The fin formula to the nearest decimal for those
of the individuals over 50 mm. which would lie counted is, males, D. 82.1, A. 67.4; females, D. 81.9,
A. 68. The average formula for those less than 50 mm. long is D. 83, A. 67.2, or for all together, I).
82.6+, A. 67.5.

While the average number of rays differs considerably in the two species, the number in each varies
so much that the numbers in individual cases overlap, individuals of Lucifuga reaching as high as 88
dorsal rays and individuals of Stygicola as low as 87. The same is true regarding the anal.

A female of this species (fig. 3, pi. 21), 65 mm. long, contained four young about 20 mm. long.

Bull. U. S. F. C. 1902.

Plate 21 .

STYGICOLA DENTATUS.

;

LUCIFUGA SUBTERRANEUS.

LUCIFUGA. A BLIND FISH CONTAINING UNBORN YOUNG WITH WELL-DEVELOPED EYES.

Contributions from the Biological Laboratory of the U. S. Fish Commission
at Woods Hole, Massachusetts.

I'HE ORGAN AND SENSE OF TASTE IN FISHES.

By C . J TJ D S O N HERRICK,

Professor of Zoology in Denison University.

237

CONTENTS.

Page.

Introduction 239

Review of literature - 241

Terminal buds and their innervation 247

Functions of terminal buds 250

Experiments on siluroid fishes 250

Experiments on gadoid fishes 257

The hake, Urophycis tenuis 258

The tomcod, Microgadus lomcod _ 262

Other fishes . 264

The sea-robin, Prionolus carolinus - 264

The king-fish , Menticirrhus saxatilis 265

The toad-fish. Opsanus tan 265

Conclusion 266

Addendum 270

Literature cited . - 272

238

Contributions from the Biological Laboratory of the U. S. Fish Commission,
Woods Hole, Massachusetts.

THE ORGAN AND SENSE OF TASTE IN FISHES.

By C. JUDSON HERRICK,
Professor of Zoology in Denison University.

INTRODUCTION.

The practical problems connected with the fisheries have been attacked (and in
large measure successfully solved) 1 >y a rough-and-ready application of the method
of trial and error, and the scientific investigator has merely to follow after and
explain why a given form of trap or method of lure is successful with one species of
fish and not with another. But there remain many unsolved problems of great
economic importance, and it is the function of scientific research to contribute to the
solution of these problems in a more orderly and economical manner, even though it
often happens that the investigator best qualified to solve the scientific problem has
not the practical knowledge of fishery matters necessary to apply his own results to
economic problems, and so his facts have to be worked over from the other point of
view before they become practically useful.

We are, in fact, profoundly ignorant of the senses and instincts of the fishes,
even those connected with their feeding habits, which are of so direct importance to
all commercial fisheries. Nearly all which one finds in the scientific literature bear-
ing on the senses of fishes is merely inference of function based on a study of the
structure of the organs — a most precarious pathway for scientific research. My
own studies on the nerve components of fishes have led me to certain inferences
regarding the functions and the distribution of the organs of taste in fishes, and the
present study is an attempt to follow out these inferences by the determination of
more exact facts regarding the pathways of gustatory stimuli as anatomically demon-
strable, together with sufficient direct physiological experiment to furnish definite
information of the function served by this system of sense organs and of their
nervous paths in the fishes.

Neurologists have always paid a great deal of attention to the conduction paths
within the central nervous system, and in recent years special efforts have been made
to isolate the various functional systems of neurones, tracing the exact path of the
sensory impulses from the peripheral organ to the primary sensory center, thence to
the various secondary centers and return reflex paths. This motive underlies the
recent studies on the nerve components and, indeed, much of the best morphological
work on the nervous system in all times.

239

240

BULLETIN OF THE UNITED STATES FISH COMMISSION.

Some years ago I formulated the following definition of such a functional system
of neurones, with special reference to the peripheral members of the system:

The sum of all the nerve fibers in the body which possess certain physiological and morphological
characters in common so that they may react in a common mode. Morphologically each system is
defined by the terminal relations of its libers, by the organs to which they are related peripherally,
and by the centers in which the libers arise or terminate. The fibers of a single system may appear
in a large number of nerves repeated more or less uniformly in a metameric way (as in the general
cutaneous system of the spinal nerves), or they may all be concentrated into a single nerve (as in the
optic nerve).

Now, if we add to this the secondary paths related to the primary central end
stations referred to above, and the chief reflex arcs directly associated therewith, we
shall have a picture of the system in its entirety.

The functional system with which we are especially concerned in the present
research is that known to comparative anatomy as the communis system, including
(1) unspecialized visceral sensory fibers ending free in the mucous surfaces of various
viscera without special sense organs — probably phylogenetically the more primitive
elements — and (2) specialized sensory fibers always ending in connection with highly
differentiated sense organs in the mouth, pharynx, lips, or outer skin, known as taste
buds, terminal buds, or end buds, and in general serving the function of taste. These
specialized elements are probably of more recent phylogenetic origin than the first
group, and the term “gustatory system” will be used to designate these organs, wher-
ever placed on the body surface, together with their nervous pathways toward and
within the brain. In other words, the gustatory system is that portion of the com-
munis system of neurones which serves the sense of taste, as distinguished from
those communis neurones which serve less highly specialized visceral sensations.

These two groups of fibers can easily be distinguished peripherally of the brain,
but centrally they have not as yet been successfully analyzed. Hence in treating of
the central gustatory path we can not be sure that we do not include the unspecialized
visceral system also. Hut since in some fishes the gustatory fibers preponderate
many fold over the unspecialized fibers of the communis s}rstem, there is no
ambiguity arising from this central confusion of the two elements so far as the
gustatory system is concerned, since the secondary paths as clearly traceable in these
fishes must be made up chiefly of gustatory fibers.

The central gustatory path is not definitely known either in man or in any other
vertebrate, so far as shown by the available literature. I have therefore studied with
some care the brains of some fishes in which this system is enormously developed, in
the hope that they would throw light on this unsolved problem of vertebrate anatomy.
And in this I have not been disappointed, though my study of the central paths is not
yet sufficiently advanced for publication.

As intimated above, sense organs belonging to the communis system and pre-
sumably serving the function of taste are found in the mouths of all fishes (“ taste
buds”). They are frequently found upon the lips, and in some cases they are found
likewise plentifully distributed over extensive areas of the outer skin of the head
and trunk. In this latter case they are commonly termed terminal buds or end buds
(End/cnospen, Becherorgane , of the Germans). They must in all cases be sharply
distinguished from the neuromasts, or organs of the lateral-line system (German,

THE ORGAN AND SENSE OF TASTE IN FISHES.

241

JSfervmhugel ), though these latter occur in the skin of fishes in a great variety of
forms, often resembling the terminal buds very closely. The innervation and
functions of the two systems of organs are, however, wholly different, and they really
have nothing to do with each other. I shall illustrate more fully in a later section
of this paper the structure of the terminal buds and the details of their innervation.
I here call attention merely to the important fact that both in structure and in nerve
supply they resemble most closely the taste buds of the mouth. From this one
naturally infers for them a gustatory function. Since, however, inferences are not
in order when facts are available, I have undertaken to determine experimentally
the function of these cutaneous sense organs of the communis system.

The experiments which 1 have made are of an exceedingly simple nature, the
attempt being to put the fish while under observation in as nearly normal conditions
as possible and to utilize the ordinary feeding and other instinctive reactions so far
as possible in the accumulation of the data. These, are the methods of the old-time
observational natural history, it is true, as contrasted with the methods of precision
of the modern physiological laboratory. They have, however, proved sufficient for
their purpose, which was merely to determine the class of stimuli to which the
terminal buds are sensitive, or the sensational modality which they serve, rather
than to contribute to the chemical physiology of taste in general.

The chief obstacle to experiments of this sort, and one which many observers
seem to have made no serious efforts to overcome, is the natural timidity or shyness
of wild creatures when kept in the confined and unnatural quarters necessary for
close observation. The role played by fear in animal behavior has been vividly
brought to our notice by Whitman (’99), and, like this observer, I find that young
animals which have been reared in captivity are much more approachable and
tractable under experimental conditions than adults which have been reared in their
natural freedom. In fact, with several species I quite failed to get the adults to
take food at, all in captivity, though they were under observation for long periods.

REVIEW OF LITERATURE.

Surprisingly little attention has been paid to the physiology of taste in fishes,
and this literature is very scanty. On the other hand, the anatomical investigation
of these sense organs has been extensively followed for nearly a century, though
often in a blind and profitless way. The history of opinion upon the significance of
these sense organs has been quite fully given by Merkel (’SO) in his great mono-
graph published in 1880, and the earlier phases of this history need not be again
reviewed further than to mention a few salient features.

In 1827 Weber observed the taste buds on the peculiar palatal organ of the carp
and correctly interpreted their function, lie also figured the brain of the carp,
illustrating the enormous vagal lobes from which these taste buds receive their inner-
vation. Leydig discovered in 1851 the terminal buds of the outer skin of fishes and
gave a detailed account of their structure, which subsequent research has shown to
be in some respects inaccurate. In 1863 F. E. Schulze gave a more accurate descrip-
tion of the “ beche'i formigen Organe ” of fishes, in which he distinguished the specific
sensory cells from the supporting cells. He also correctly inferred their function to

F. C. B. 1902—16

242

BULLETIN OF THE UNITED STATES FISH COMMISSION.

be similar to that of taste buds within the mouth, viz, the perception of chemical
stimuli.

In 1870 the same author (F. E. Schulze, ’70) made a further important contri-
bution to the problem of the terminal buds by the demonstration that they differ
structurally from all neuromasts, or organs of the lateral-line system. The neuro-
masts are commonly sunken below the skin in canals, tubes, or pits, but in some
cases they are strictly superficial and resemble in external form the terminal buds
very closely — a feature which led Lcydig (’51, ’79, ’94) and others to assume that the
two classes of organs are mere varieties of a common type. Schulze showed that
the neuromasts can in all cases be differentiated from the terminal buds by the fact
that their specific sensory cells (pear cells) extend only part way through the
sensory epithelium and fail to reach the internal limiting membrane, while in the

terminal buds both specific sensory

cells and supporting cells pass through

from external to internal limiting mem-

brane.

This distinction was confirmed by
Merkel (’80), who, with curious incon-
sistency, while recognizing the struc-
tural dissimilarity of the two classes of
organs, nevertheless, as we shall see
below, ascribes to both essentially the
same function, touch. This matter
was put to the decisive test in my
contribution on Ameiurus (’01), a type
which possesses both terminal buds
and neuromasts in great abundance
and diversity of forms. Schulze’s
contention is supported both by the
structure of the organs and by their
innervation, for I have shown that
all neuromasts of whatever form are
innervated by acustico-lateralis nerves from the tubereulum acusticuin of the brain,
while all terminal buds, whether within the mouth or in the outer skin, are inner-
vated by communis nerves related centrally to a single center within the brain.
This center is bilobed, the lobus vagi receiving most of the communis fibers from
the mouth cavity by way of the vagus and glossopharyngeus and the lobus facialis
the communis fibers from the terminal buds of the outer skin by way of the facial
nerve (cf. fig. 1).

Similar terminal buds have been found in the outer skin of many species of
Teleostomes and in Cyclostoines, but, so far as certainly known, nowhere else among
vertebrates (save on the lips of some other classes). Their distribution among the
fishes is very irregular, being most abundant among the siluroids, eyprinoids,
ganoids, and cyclostoines, in general bottom fishes of sluggish habit, often living
in mud and rarely belonging to the predaceous types which find their food chiefly
by the sense of sight. The following list of fishes which have been shown to possess

Fig. 1. — Dorsal view of the brain of the yellow cat-fish (Lap-
tops olivaris Rat). The olfactory bulbs with most of their
crura have been removed, also the membranous roof of the
fourth ventricle, exposing the facial and vagal lobes. This
v ventricle is bounded behind by a transverse ridge contain-
ing the commissura infima Halleri and the commissural
nucleus of Cajal. x2.

THE ORGAN AND SENSE OF TASTE IN FISHES.

243

terminal buds on the outer skin is by no means complete, but will serve to illustrate
the wide range of species which have acquired this peculiarity:

Fishes ■ possessing terminal buds on the outer skin.

Acerina. Oil fins and body (Merkel, ’80).

Acipenser sturio, sturgeon. On barbel (Merkel,
’80). Also other sturgeons.

Ayonus cataphractus, pogge. On the villi form
tentacles beneath the head (Bateson, ’90).

Ameiurus vielas, cat-fish, and other North Ameri-
can Siluridee. On barblets and nearly the
whole body surface (Herrick, ’01).

Amia caiea, bowfin. On skin of head and other
parts (Allis, ’97).

Anguilla vulgaris, eel. On the fins, lips, and ante-
rior nostril (Merkel, ’80; Bateson, ’90).

Aspius alburnus (Merkel, ’80).

Barbus fluviatilis. ( )n barblet (F. E. Schulze, ’63).

Branchiostoma lanceolatum = Amphioxus lanceolatus,
lancelet. On the oral cirri (Merkel, ’80).

Carassius auratus, gold-fish. On the whole body
(numerous authors; Herrick).

Cephalacanthus = Cobilis fossilis, flying gurnard
(Merkel, ’80).

Coitus scorpius, sculpin. On fins (Merkel, ’80).

Cynoscion — Corvina (Merkel, ’80).

Oyprinus carpio, carp, and other eyprinoids. On
whole body (Merkel, ’80, and others).

Dactylopterus (Merkel, ’80).

Discognathus lamta, Indian carp. Over the whole
body surface (Leydig, ’94).

Enchelyopus = Motella, four-bearded rockling. On
barblets and pelvic fins (Bateson, ’90).

Gadus callarias, cod. On lips, barbel, fins, and
body (Merkel, ’80; Herrick, ’00).

Gadus luscus, pouting. On the lips, barblet, and
pelvic fins (Bateson, ’90).

Gadus merlangus, whiting. On lips (Bateson, ’90).

Gadus pollachius, pollack. On lips (Bateson, ’90).

Gaidropsarus = Motella, three-bearded rockling.
On all the barblets and pelvic fins (Zincone,
’78; Bateson, ’90).

Gobius, goby. On fins (Merkel, ’80).

Hippocampus, seahorse (Merkel, ’80).

Leptocephalus conger, conger eel. On the outer
and inner lips (Bateson, ’90).

Leucaspius delineatus. On the body generally
(Leydig, ’94).

Leucvscus dobula (Leydig, ’57).

Lota vulgaris, ling. On barblet (Merkel, ’80).

Mullus barbatus, mullet. On barblet (Zincone, ’78;
Merkel, ’80).

Petromyzon fluviatilis, lamprey. On skin of whole
body (Merkel, ’80, and others).

Pygosteus = Gasterosleus pungitius, stickleback
( Merkel, 80) .

Rhodeus amarus. On the body generally (Leydig,
’94).

Scorpxna (Merkel, ’80).

Silurus glanis, cat-fish (Merkel, ’80).

Solea vulgaris, sole. “Contrary to the natural
presumption, the villi on the lower (left) side
of the head do not bear sense organs, though,
as Mr. Cunningham informs me, such organs
are found between the villi ” (Bateson, ’90).

Tinea vulgaris, tench. On barblet (Merkel, ’80).

As already suggested, our knowledge of the functions of all of the sense organs
of fishes is very imperfect, since speculation based upon structure has seemed more
attractive to most authors than accurate physiological research. The monograph of
Merkel (’80), with its great wealth of accurate anatomical data on the structure and
distribution of terminal buds in all classes of vertebrates, gives an excellent illustra-
tion of the dangers in the path of even so skillful an observer when he goes beyond
the bounds of observed fact and enters the field of speculation. This author recog-
nizes the close structural resemblance between these organs and the undoubted organs
of taste in the human body. He controverts, however, the clear argument of F. E.
Schulze for their gustatory function on merely theoretical grounds. His first objec-
tion is based on their innervation. Instead of being supplied by a single gustatory
nerve, the glossopharyngeus, they may be supplied, he says, by any other body
nerve. This objection has beeu totally removed by the discovery (compare especially
my own Ameiurus paper, already referred to, published in October, 1901) that all
terminal buds, no matter where located on the body and no matter from what nerve
branches their innervation seems to come, are in reality supplied by nerves of a single
physiological system, terminating in the brain in a single center — the communis
nerves.

Again, he objects to Schulze’s theory that the terminal buds serve to localize
gustatory stimuli on the various parts of the body, on the ground that an organ of
chemical sense stimulated by substances in solution in the environing fluid could not

BULLETIN OF THE UNITED STATES FISH COMMISSION.

244

receive a sufficiently circumscribed stimulation. It is unnecessary to follow the
argument in detail, for the experiments which I shall describe shortly show conclu-
sively that when the sapid substance is brought into contact with these organs or
very near to them the stimulus is accurately and very promptly localized, and in fact
some of the fishes studied habitually find their food by this very power, the gusta-
tory stimulus calling forth an immediate reflex movement toward the point stimu-
lated. It is probable that the local sign is not given by the gustatory (communis)
nerves, but by the accompanying tactile (general cutaneous) nerves of the corre-
sponding- cutaneous area (which general cutaneous nerves Merkel, curiously enough,
denies to the fishes altogether, whereas, in fact, they are plentifully supplied to all
parts of the skin), though my experiments do not decisively answer this question.0
Weak stimuli, especially when uniformly diffused through the water, are, it is true,
not at all localized; but strong stimuli are unquestionably localized by one method
or another.

In fact, Merkel agrees with Jobert that the terminal buds of the outer skin are
tactile in function. This is based largely on the erroneous belief, referred to above,
that there are no free tactile nerve endings in the skin of fishes, and also on the
observed tactile sensibility of the barblets and other parts of the body known to be j

most plentifully supplied with terminal buds. But I have shown that all of these
parts of the body receive, in addition to communis nerves for the specialized sense
organs, a most liberal general cutaneous innervation foi tactile sensibility; and the
experiments which follow go to show practically that these two functions commonly
cooperate in setting oft' the reflex of seizing food, though they may be experiment-
ally isolated.

Merkel now proceeds to carry his argument to its logical conclusion (and like-
wise to a reductio ad absurd am ) by denying the gustatory function to all terminal
buds, even those within the mouth supplied by the glossopharyngeal nerve, of all
vertebrates below the Mammalia.

He finally concludes that both the neuromasts of the lateral-line system and the
terminal buds are tactile organs, the buds being the more delicate; but if these are
deficient, then the neuromasts may be elevated to a more delicate functional value;
both of which conclusions, in the light of our present knowledge, illustrate the dangers
attending an attempt to determine function on the basis solely of observed structure,
without adequate physiological control.

The general works contain numerous references to the subject, but usually
chance observations or speculative conclusions. Gunther says, under the caption
“ Organ of taste”:

Some fishes, especially vegetable feeders, or those provided with broad molar-like teeth, masticate
their food; and it may be observed in carps and other cyprinoid fish that this process of mastication
frequently takes some time. But the majority of fish swallow their food rapidly and without mastica-
tion, and therefore we may conclude that the sense of taste can not be acute. The tongue is often
entirely absent, and even when it exists in its most distinct state it consists merely of ligamentous or
cellular substance, and is never furnished with muscles capable of producing the movements of exten-
sion or retraction, as in most higher vertebrates. A peculiar organ on the roof of the palate of cypri-
noids is perhaps an organ adapted for perception of this sense; in these fishes the palate between and
below the upper pharyngeal bones is cushioned with a thick, soft, contractile substance, richly supplied
with nerves from the Nervi vagus and glossopliaryngeus.

«On this point, see the further experiments recorded in the Addendum, pp. 270-271.

THE ORGAN AND SENSE OF TASTE IN FISHES.

245

Regarding the peculiar palatal organ of the cyprinoids, it has been known since
Weber’s account in 1827 that this is plentifully supplied with taste buds, and Weber
himself brought forward strong indirect evidence that its function is gustatory.
The following observations (and many similar ones might be cited from the literature
of sport) are taken from the section on “The Trouts of America,” by William C.
Harris, in the American Sportsman’s Library.

The angler can not resist the belief that the senses of smell and taste are well developed in trout.
They eject the artificial fly, if the hook is not fast in the flesh, at the instant they note its nonedible
nature, or when they feel the gritty impact of the hook. They will not eat impure food, and they
have the faculty of perceiving odors, and various scents attract or repel them. This has been verified
from the earliest days of our art, when ancient rodsmen used diverse and curious pastes and oils, which
were seductive to fish; in Walton’s day, and long after, this practice was followed and the records tell
us of its success. When I was a boy and the Schulkill River was swarming with the small white-bellied
cat-fish, than which no more delightful breakfast food ever came out of the water, the only bait used
to catch them was made of Lim burger cheese, mixed with a patch of cotton batting to hold it firm on
the hook. No other lure had the same attraction for them because, no doubt, of the decided odor of
the cheese.

The problems connected with the relative significance of the several sense organs
of the fishes have been treated both anatomically and experimentally in the excellent
paper of Bateson (’90). After, anatomical remarks, based largely on his own careful
studies, on the eyes, olfactory organs, and gustatory organs, he recounts a series of
admirable and well-considered experiments made to test the parts played by these
organs in the normal feeding of various kinds of fishes.

These observations are grouped under two chief heads, viz, “Senses of fishes
which seek their food by scent” and “The senses of fishes which seek their food
I >y sight.” Though the taste buds in the mouth and outer skin are described and
correctly interpreted in the anatomical part of the paper, these organs are scarcely
considered at all in the physiological part, and this is really the greatest weakness of
the paper. Since my own observations in part follow so closeljT in the footsteps of
Bateson (though completed in the main before his paper was accessible to me), and
since they are in general confirmatory of his, it will be of interest to review portions
of his paper at this time.

He gives the following list of fishes which he has observed “to show conscious-
ness of food which was unseen by them, as, as will hereafter be shown, there is
evidence that they habitually seek it without the help of their eyes”:

Proloplerus anncclens, mud-fish.

Scyllium canicnla, rough dog-fish.
Scyllium caiulus, nurse-hound.

Raja batis, skate.

Conger vulgaris, conger eel.

Anguilla, vulgaris, eel.

Motella tricirrata, three-bearded rockling.

Motella mustela, five-bearded rockling.
Nemacheilus barbatula, loach.

? Lepadogaster gouanii, sucker.

Solea vulgaris, sole.

Solea minuta, little sole.

Acipenser ruthenus, sterlet.

He says: “To this list may almost certainly be added the remainder of the Raiidse ,
together with the angel -fish (Rhina squatind) and Torpedo .” Unfortunately, how-
ever, Bateson in his list does not distinguish between those fishes in which smell
obviously plays the leading part and those in which taste or touch or both are used to
compensate for the reduction of vision, and it is this defect which it is hoped that the
present contribution may in part correct.

£46 BULLETIN OF THE UNITED STATES FiSH COMMISSION.

Most of the forms in the list above are more or less nocturnal animals, but they
differ much in this regard. The part attributed to the sense of sight and smell in
Bateson’s studies is so similar to my own conclusions in many respects that it seems
fitting to quote the greater part of his description, especially since the species
observed by us are in all cases different. He says:

None of these fishes ever start in quest of food when it is first put into the tank, but wait for an
interval, doubtless until the scent has been diffused through the water. Having perceived the scent
of food, they swim vaguely about and appear to seek it by examining the whole area pervaded by the
scent, having seemingly no sense of the direction whence it proceeds. Though some of these animals
have undoubtedly some visual perception of objects moving in the water, yet at no time was there the
slightest indication of any recognition of any food substance by sight. The process of search is equally
indirect and tentative by day and by night, whether the food is exposed or hidden in an opaque vessel,
whether a piece of actual food is in the water or the juice only, squeezed through a cloth, and, lastly,
whether (as tested in the case of the conger and the rockling) the fish be blind or not. * * * The
perceptions, then, by which these animals recognize the presence of food are clearly obtained by
means of the olfactory organs and apparently exclusively through them. I was particularly surprised
to find no indication of the possession of such a function by the sense organs of the barbels and lips or
by those of the lateral line. As has been already described, the pelvic fins and barbels of the rock-
lings ( Motella ) and the lips, etc,., of most fishes bear great numbers of sense organs closely comparable
in structure with the taste buds of other vertebrates. No one who has seen the mode of feeding of the
rockling or pouting ( Oadus hcscus) can doubt that these organs are employed for the discrimination of
food substances; but the fact already mentioned, that the rockling in which the olfactory organs had
been extirpated did not take any notice of food that was not put close to it, points to the conclusion
that they are of service only in actual contact with the food itself.

Bateson gives also a considerable list of fishes which he has observed to get their
food chiefly by the sense of sight, and he is doubtless correct in asserting that the
majority of fishes belong to this class. None of these sight-hunting fishes while living
in his tanks appeared able to see their food try night, or even in twilight. None of
the fishes which he enumerates as belonging to this class showed symptoms of interest
when the juice of food substances was put into the water, and other evidence is
brought forward to show that the sense of smell plays little or no part in helping
them to discover their food.

1 have not studied any of the species mentioned by Bateson, but for the forms
studied by me, which have an extensive supply of terminal buds on the outer skin,
1 fully confirm most of the statements quoted above, save that in determining the
part played by sight I did not blind any of my fishes and save that the statement
that in fishes of his first group “at no time was there the slightest indication of any
recognition of any food substance by sight” is strictly true of none of my fishes
except Ameiurus, though in some of the other cases it is approximately true.

The only important respect in which my observations are not in harmony with
those of Bateson is in connection with the part played by the sense of taste in some
of these types of fishes. 1 have studied the gustatory reactions of fishes closely
allied to the rockling and having the same arrangement of terminal buds on the barb-
lets and pelvic fins, and am convinced that Bateson’s failure to get clear gustatory
reactions from these organs was due to the insufficiency of his methods of experi-
ment rather than to the absence of the function. In general, it may be stated that
the part played by the gustatory reflex in the case of fishes having an extensive sup-
ply of terminal buds on the outer skin is of vastly greater importance than Bateson
appears to have recognized.

TIIE ORGAN AND SENSE OF TASTE IN FISHES.

247

The only other paper of importance dealing with the sense of taste in the fishes
experimentally which has come to my notice is the great monograph on the senses of
taste and smell by Nagel (’94). He investigated the sense of taste in the following
fishes:

(1) Fresh-water Types: Anguilla anguiUa (old and quite young); Cyprinus carpio; Barbus fluvia-

lilis; Leuciscus ceplialus; Gasterosteus aculeatus; Gobbis flimulilis; Silurus glcmis (young
specimen ) ; Cobitis fossilis.

(2) Marine Types: Pristiurus; Scyllium catulus and S. canicula; Syngnathus acus; Urcmoscopus

scaber; Loplnus piscatorius.

Nagel tested all the fresh water fishes mentioned in this list by bringing bitter,
sour, sweet, and salty solutions in contact with the skin, without getting any response
to the stimulus. Thus, the carp, wels ( Silurus ), and stickleback did not respond to
a stimulation of the skin of the body with quinine, though the last-named fish gave
an immediate response when the solution touched the lips. lie concludes:

l ii the fresh-water fishes, according to my observations, the power of taste is completely lacking
in the outer skin; or, more precisely, in no part except the head is there gustatory sensibility.

For such of these forms as possess no terminal buds on the skin of the body this
is doubtless true; but for the other fishes, including, doubtless, Silurus and Cyjprinus ,
it is certainly a mistake. In gadoid fishes I got a clear reaction against quinine
solution when it was applied to the free fin rays, which are known to be supplied
with terminal buds, but not from other parts of the skin.

Among the elasmobranch fishes Nagel found Scyllium catulus and S. canicula
to be sensitive to very dilute solutions of vanilla all over the body and fins. Bitters
were not perceived thus, nor oil of rosemary, but they are very sensitive to creosote.
He controverts Schwalbe’s argument that the terminal buds of the outer skin of fishes
probably have a gustatory function by reason of the similarity of their structure
with that of taste buds in the mouth, and concludes:

A real sense of taste, such as man and many other animals have in the mouth, appears to be
absent in the outer skin of all fishes and Amphibia.

It will appear from the following pages that this conclusion is erroneous. 1 will
merely add here that if Nagel had worked with sapid solutions, with which his fishes
were presumably already familiar, instead of with substances like sugar and vanilla,
toward which no clearly established reflexes had been established in the natural
environment of the fishes, his conclusions might have been different.

TERMINAL BUDS AND THEIR INNERVATION.

The terminal buds of the fishes tabulated above, and doubtless many others
which might be mentioned, are of the same type and presumably provided with
similar innervation by communis nerves, for cutaneous branches of the communis
root of the facial nerve are known to reach the areas provided with the buds in all
cases which have been adequately studied. These organs may therefore all be
defined morphologically as belonging to the communis system of sense organs, along
with the taste buds of the mouth cavity and as distinct from the lateral-line organs
and all other types of sense organs. In order to support this position there remains
merely the proof that the terminal buds and taste buds have a similar function.
This evidence is presented subsequently in this paper.

248

BULLETIN OF THE UNITED STATES FISH COMMISSION.

The terminal buds of fishes have been often described and figured, and I have
little to add to the classical descriptions save in the matter of distribution and inner-
vation. Those in the mouth are supplied by branches of the x, ix, and vii pairs of
cranial nerves, the first two nerves supplying those in the gill regions and the pre-
trematic branch of the glossopharyngeus also running forward to supply those on
the hyoid arch (tongue). The communis root of the facialis (=portio intermedia of
human anatomy) and its geniculate ganglion supply the taste buds on the palate by
the r. palatinus facialis (= great. superficial petrosal nerve of man), and other buds
on the lining of the cheek, on the jaws, and on the lips by other branches, some of
which are secondarily associated with branches of the trigeminus and most of which
have no homologous in mammalian anatomy, though some one or more of them
probably represent the chorda tympani.

In Ameiurus I have shown (’01) that terminal buds occur in the skin of practically

the whole body surface, most abundantly on
the bardlets and diminishing in frequency
toward the tail. These buds (see fig. 2)
rest on a low papilla of the dermis, quite
different from that figured by Merkel (’80,
plate v, lig. 1) for the terminal buds of
Silurus. Ilis figure shows a much smaller
organ, resting upon a greatly elongated
papilla in an epidermis which is apparently
thicker than in Ameiurus. Merkel states
(’SO, p. 72) that terminal buds always occur
on such a dermal papilla. While this is cer-
tainly the general rule, we find occasionally
instances where the papilla is absent, as on
the lilliform fins of the hake, where I find the
buds imbedded in the epidermis and extend-
ing only partway through it, with a layer of
unmodified epidermal cells between the bud
and the dermis.

All parts of the body of Ameiurus which are supplied with terminal buds are
reached by branches of the facial nerve from the geniculate ganglion. In other
words, the rami from the communis root of the facialis are distributed to nearly the
whole outer body surface of this fish. On the distal side of the ganglion these rami
usually join themselves to other cutaneous branches which are phylogenetically older,
belonging to the general cutaneous and lateral lino systems. Even the great recur-
rent branch into the trunk, the ramus lateralis accessorius, which passes out of the
cranium as a practically pure communis nerve, anastomoses with the spinal nerves at
their ganglia and its fibers are ultimately distributed along with the general cutaneous
fibers from these spinal ganglia. Fig. 3 illustrates the courses of the chief cutaneous
branches of the communis system in Ameiurus meins , the nerves of all other systems
being omitted from the. sketch.

Proximally of the geniculate ganglion the communis root of the facialis pursues
an uncomplicated course to the primary gustatory center within the medulla
oblongata. In most fishes this root passes back close to the floor of the fourth ven-

Fig. 2. — Section through the skin of the top of the head
of Ameiurus mclas, showing a terminal bud. x 375.
(From the Journal of Comparative Neurology, vol. xi,
. No. 3, Oct., 1901, plate xvii, fig. 11.) At rl is the dermis,
which is raised into a low papilla under the sense organ
and whose center is pierced by the nerve for the organ .

THE ORGAN AND SENSE OF TASTE IN FISHES.

249

tricle as the fasciculus communis ( = fasc. solitaries of mammals) to terminate in the
vagal lobe of the same side, and receives in its course the communis root of the glos-
sopharyngeus nerve. But in siluroids and cyprinoids, where the very abundant
terminal buds of the outer skin are all innervated from the communis loot of the
facial nerve, the consequent increase in the size of this root has resulted in a great
enlargement of the cephalic end of the gustatory center (vagal lobe) which appears
on the dorsal surface of the oblongata as the facial lobe. This structure is paired in
siluroids and was formerly called the lobus trigemini, an inadmissible term, since it
has nothing whatever to do with the trigeminus nerve. In cyprinoids it is unpaired
and is referred to in the older literature as the tuberculum impar.

The cyprinoid fishes also have long been known to have terminal buds {Beefier -
organe ) widely distributed over the outer body surface; but neither the innervation
of these organs nor the exact composition of the cranial nerves has ever been worked
out in any cyprinoid fish. A cursory examination of a series of sections prepared

Fig. 3. — A projection of the cutaneous branches of the communis root of the facial nerve in Amciuruss mclas, as seen from
the right side. The outline of the brain is indicated by the stippled area and the positions of the eye and anterior and
posterior nostrils are indicated. The projection is reconstructed from serial sections, but is not drawn accurately to
scale. More detailed reconstructions of the cranial nerves and lateral-line sense organs of this fish are given in the
Journal of Comparative Neurology, vol. xi, No. 3, plates xiv and xv (Herrick, ’01).

by the Weigert method through the entire head and body of a small gold-fish ( Caras -
sius auratus) has convinced me that the same conditions in general prevail in the
cyprinoids as in the siluroids. That is, the enormous size of the vagal lobes of
cyprinoids is explained by the fact that these are the terminal centers for the vast
numbers of nerve fibers entering the brain bv way of the ix and x nerves from the
palatal organ, this remarkable structure being crowded over its entire extent with
taste buds and probably Serving to filter food particles out of the mud taken into the
mouth.

On the other hand, the tuberculum impar, or facial lobe, receives the entire
communis root of the facial nerve. This root receives fibers from practically all
parts of the outer surface of the body, and we may infer b}T analogy with other
fishes that these fibers connect with the terminal buds in these cutaneous areas,
though we have as yet no actual demonstration of this fact. The terminal buds of
the skin of the head are supplied mainly, as in Ameiuvus , by way of the infraorbital
trunk. The terminal buds in the skin of the body of the gold-fish are not, however,

250

BULLETIN OF THE UNITED STATES FISH COMMISSION.

supplied by a ramus lateralis accessorius, or recurrent facial nerve, as in Ameiurus and
the gadoid fishes, for this nerve, as has long been known, is absent in the cyprinoids.

There is, however, in these fishes an intracranial anastomosis between the
v+vn ganglionic complex and the ix+x complex, the composition of which has
thuts far remained unknown. This proves to be the recurrent branch of the facialis,
carrying communis fibers from the geniculate ganglion into the trunk. The details
of the peripheral distribution of these fibers have not been fully worked out, but the
main path in the gold-fish is as follows:

The geniculate ganglion of the facialis is clearly separable from all other
ganglionic masses of the trigemino-facial complex and is composed of two portions,
each of large size. The more dorsal portion corresponds to the greater part of the
ganglion in other teleosts and distributes its fibers chiefly by way of the infraorbital
trunk. The more ventral portion sends cephalad a very large palatine nerve, and
caudad a still larger nerve which represents morphologically, though not topograph-
ically, the r. recurrens facialis of the siluroids, etc., or the facial root of the r. lateralis
accessorius as found in the cod.

This nerve passes back along the lateral side of the great auditory root and at
the level of the superficial origin of the ix nerve it divides into several strands, one
of which passes dorsal ly of the ix root, the others ventral ly. These latter, however,
pass upward so as to lie, farther back, dorsally of all of the vagus roots except that
of the lateralis branch of the vagus. All of these communis fibers now join them-
selves to the r. lateralis vagi and, passing through the ganglion of the latter nerve,
both components enter the body of the fish bound up in a single nerve trunk in
which the fine communis fibers are for a time completely surrounded by the coarse
lateralis fibers. The communis fibers go off in successive branches along with
lateralis fibers. The details of the distribution have not been worked out, though I
think it would not be difficult to do so with the material at hand. It is highly
probable that the communis fibers are for the terminal buds sparsely distributed
over the skin of the body and that the terminal buds of the trunk are all innervated
from these communis fibers in the r. lateralis vagi, just as the buds in the skin of
the head are innervated by other communis fibers from the geniculate ganglion of
the facialis, an arrangement substantially identical in morphological plan with that
of the siluroid fishes.

The conditions here, so far as studied, confirm essentially the conjectures to
which I was led from a study of the literature (Herrick, ’99, p. 400), and accord so
completely with the morphological interpretation there proposed that we merely
refer the reader to that passage in the Menidia paper.

FUNCTIONS OF TERMINAL BUDS.

EXPERIMENTS ON SILUROID FISHES.

The cat-fish {Ameiurus nebulosus ) upon which this series of experiments was
conducted (except a few experiments specifically designated) were hatched in the
open at Granville in the spring of 1901. In October of that same year they were
taken to the laboratory and kept through the following winter in tanks. Microscopic
examination of the skin and barblets shows that their skin and cutaneous sense orvans

THE ORGAN AND SENSE OF TASTE IN FISHES.

251

at this age are practically in the adult condition. During the winter they were fed
on various kinds of meat chopped fine, sometimes cooked, but usually raw.

In one small aquarium were kept half a dozen cat-fish, several ordinary “shiners”
(Notrojpis sp.a), and some small “spotted suckers” ( Mimjtrema melanops Kafinesque).
Casual observations made during the winter while feeding showed that the shiners
use the eyes chiefly in capturing their food. A bit of meat dropped into the water will
usually be seized instantly and devoured before it has time to sink to the bottom of
the tank. After it has fallen to the bottom it is apt to be long overlooked unless the
fish happens upon it in its aimless wanderings, or unless its attention is called to it by
the movements of other fishes which may be eating it. These fishes, when observed,
are usually swimming about in the mid-depths of the tank, not resting near the
bottom. I have observed the same behavior in Menidia and other large-eyed species.

The behavior of the suckers was totally different. These fishes lie on the bottom
most of the time unless disturbed, though if frightened they are very active, swim-
ming powerfully and leaping out of the water. When food is thrown in they never
pay the slightest attention, nor are they attracted by the sight of other fishes
struggling for the meat. They are exceedingly shy and rarely eat when under
observation. They lie quietly much of the time or swim slowly about, dragging- the
lies iff lips of the highly protrusible mouth over the bottom of the tank. If they thus
happen upon a bit of meat this is sucked into the month, worked over with the
pharyngeal teeth apparently, and then often ejected forcibly from the mouth, to be
again taken, perhaps, and the process repeated — a behavior very characteristic of the
way they take the bait, I am told by fishermen.

The cat-fish, like the suckers, keep strictly to the bottom of the tank. They are
often quiet in the darkest corners or lying under debris, but much of the time
are slowly dragging the mental and post-mental barblets along the bottom. 'The
nasal barblets are held projecting well upward, and the maxillary barblets are
directed outward and backward, their tips trailing the bottom or waving gently back
and forth. They appear never to use their eyes directly for catching food to the
slightest degree under the conditions of these experiments. No attention is paid to
particles of food thrown into the water, even though they settle down within a few
millimeters of the nose or barblet of the fish. The only case observed by me in
which the eyes seem to serve in finding food is when a large piece of meat is thrown
in and one fish begins to “worry” it. His movements may attract others until as
many fish as can reach it are all tugging at it at once. If, however, a shadow is
caused to fall upon the water, as by hovering the hand over the aquarium, the fishes
are greatly disturbed and dart wildly about. They always seek the darkest corners
of the tank and lie under dead leaves resting on the bottom of the tank for the most
part, showing that the eyes are not by any means functionless and the fishes are
strongly negatively phototactic.

If the cat-fishes in the course of their aimless movements along the floor of the
aquarium touch a bit of meat with the lips or barblets, it is instantly seized and swal-
lowed. Food in the immediate neighborhood of the fish is not discovered at once,
but after a time appears to affect the fish in some way, probably through the sense of

aNolropia lias very small tuberculum impar and vagal lobes, the latter scarcely larger than In the cod, Menidia, and
physoclistous fishes generally. From this one may safely infer that cutaneous terminal buds are not as highly developed
in this form as in the larger cyprinoids.

252

BULLETIN OF THE UNITED STATES FISH COMMISSION.

smell, as the maxillary barblets begin to wave about more actively and finally the fish
becomes restless. He does not find the food, however, unless in the course of his
movements it actually touches some part of the body.

During May and June, 1902, more systematic experiments were undertaken with
these fish, and since these experiments are typical of those subsequently performed
on other species of fishes 1 shall recount them in some detail. At first a few speci-
mens were taken out in a shallow tray and the attempt made to feed them in various
ways under close observation. They were, however, so much frightened by the
exposure to bright daylight and by the proximity of the observer, in spite of all pre-
cautions, that no reactions could be obtained which were at all satisfactory. A bit of
fresh meat on a long-handled needle could be thrust slowly toward the fish as he lay
quietly on the bottom, rubbed over his body or on the barblets, and even over the
lips, without evoking a movement of any kind in response. The same observation
was made with the spotted suckers. The fishes in both cases had been without food
for several days and were very hungry, but were obviously too much frightened to
respond to the food stimulus.

On another occasion the same conditions were prepared, except that a few dead
leaves were littered over the bottom of the tray. The fish when placed in the tray
immediately sought the shelter of the leaves, and, after a suitable interval to enable
them to become accustomed to the place, the feeding experiments were repeated.
Selecting a fish which was entirely concealed under a large leaf, save for a projecting
barblet, a Hit of meat on a slender wire was gently passed down into the water in
such a way as to touch the projecting barblet. it was instantly seized and swallowed.
This was repeated many times with several of the fishes.

In subsequent experiments the fish were not removed from their own tank, but
the water was drawn off so that it was only about six inches deep. Here they would
lie under the leaves and the experiment could be continued with a minimum of
disturbance to the fishes. The experiment of touching the barblet with meat was
repeated hundreds of times with an almost invariable result that the fish instantly
turned and snapped up the morsel. If the meat was merely held very close to the
barblet it usually produced no response. The reaction was obtained equally well, no
matter which barblet was touched.

In a later series of experiments I found that the fish would almost always turn
and seize the meat if he were touched at any point on the head or body. If the tail
of the fish projected out from under a leaf and the skin near the root of a tail fin
were touched with meat the fish would turn and seize the meat. This reaction was
not so uniformly made at first as that from the barblets, but after a dozen or so of
trials it followed with equal promptness and uniformity, the fish apparently requir-
ing a little practice to learn the movement perfectly.

The experiments last described were repeated the next day and by this time it
was found that the fishes had become so tame that they would take the meat if
offered to them in the open, without the shelter of the dead leaves, though not so
certainly as when under the cover of the leaves, often taking fright from the shadow
of the observer’s hand or from some other cause.

In none of these cases did the fishes appear to see the bait or to perceive it in any
way other than by actual contact with the skin at some point. If the bait were held

THE ORGAN AND SENSE OF TASTE IN FISHES.

253

a moment in front of them and then moved .slowly away they would not follow it.
If, however, it touched a barblet and then moved rapidly away before the fish had
time to seize it, then the fish would sometimes follow it a short distance.

At this point, the relations of vision and smell to these reactions should receive
some further consideration. These young- fishes, like their adults, spend much of
their time buried under the debris of the bottom, with perhaps a barblet or a por-
tion of the tail only projecting. Under these circumstances it is easy to apply the
stimulus to various parts of the skin with the assurance that the contact is wholly
invisible to the fish. Many such experiments show decisively that the reaction takes
place in the same way whether the fish is able to see the stimulus applied or not.
The visual factor being so conclusively ruled out, 1 have not thought it necessary to
blind the fish for further control.

This conclusion of course must be limited strictly to fish of the species and age
under investigation. It by no means follows that they may not subsequently learn
to use their eyes in finding food, as well as in escaping from their enemies. Indeed,
during the later experiments of this series, after the fishes had been fed for several
weeks almost daily with meat on the end of a wire, I saw some slight evidence that
they took note of the bait by the sense of sight, but the observations were in no case
conclusive. Whether the adult Ameiurus nebulosus ever uses the eyes in the capture
of food 1 have no definite information, though from the habit of spending much of
the time during the day completely buried in the mud and of feeding chiefly at night
it is very improbable that they do so. With the channel cat-fish, Ictalurns , the case
is certainly different.

Mr. I. A. Field tells me that while fishing for bass in the Black River, Ohio, he
has sometimes caught large specimens of Ictalurus with live minnows as bait. The
current was swift and the minnows were kepi off the bottom of the river and in
motion all the time. At the meeting of the American Association for the Advance-
ment of Science, at Pittsburg, July 1, 1902, in the course of a brief report upon these
experiments, I asked the question whether anyone ever caught a cat-fish on a spoon
hook. Dr. L. L. Dyclie stated that he has occasionally caught the channel cat (Icta-
lurus) on a spoon in a small lake, but only in bright sunlight. Dr. Eigenmann stated
that Ictalurus has much better eyes than Ameiurus. They are not only larger, but
the retinal pattern is more nearly like that of other fishes, while that of Ameiurus
is decidedly degenerate.

The part played by the sense of smell is much more difficult to determine. As
intimated above. I have evidence that the gustatory organs of the skin can function
only in contact with the sapid substance. The most highly flavored food can be held
within a millimeter or two of the barblet or lips without calling forth the character-
istic instantaneous reflex. I will narrate one experience which was many times
repeated in a variety of modifications. Three fishes were lying quietly under a small
water-soaked leaf. A bit of rather stale beefsteak, with a strong odor, was held on
the tip of a tine wire over the edge of the leaf under which they were lying and sepa-
rated by a centimeter or two from the nostrils of the fishes. The leaf was consider-
ably corroded by decay, and doubtless the odor could freely permeate it, though it
was nearly or quite opaque. After some ten seconds the fishes began to move rest-
lessly about in circles under the leaf, which was soon swept away by their movements.

254 BULLETIN OF THE UNITED STATES FISH COMMISSION.

As a rule the fishes swam in narrow circles close to the' bottom and for a lone- time
failed to find the meat, though they seemed to be aware of its general position for
they never circled far away. If the meat were very slowly moved across the aqua-
rium the fish could be drawn in this way after it for a considerable distance, though
the meat was never found unless in the course of their apparently aimless movements
one of the fishes came in contact with it, when it was instantly snapped up.

This aimless circling movement may be termed provisionally the seeking reac-
tion, since it is so different from the characteristic movement made when the stimulus
is in contact with the body — a sharp turn of the body and instantaneous seizing of
the bait — which I shall term the gustatory reaction. Unfortunately, I have not had
opportunity as yet to carry out extirpation experiments on Ameiurus to determine
decisively the part played by the olfactory organ in this reaction. (Compare the
experiments on the tomcod narrated below.)

The fishes upon which these experiments were performed have unfortunately
been lost. At the present time 1 have a fresh lot of Ameiurus fry under observation,
and have already verified many of the conclusions reached with the first lot. But i
this second collection of fishes has not, at the time when this report is submitted,
been in captivity long enough to become sufficiently accustomed to their new sur-
roundings to feed freely and fearlessly. After some months of further prelimi-
nary observation, I hope to carry on experiments which may shed some light on the
sense of smell in these fishes. But this must be reserved for a later report. A few
subsequent observations are noted on pages 270-271.

We must content ourselves at the present time, then, with the inference that
the sense of smell plays at least a small part in these reactions, for the animals
became slightly restless in the proximity of the stimulus, though they were not in
contact with it; this, however, appears never to provoke a definite reaction of seiz-
ing the food, but merely a vague reaction in search of food. On the other hand,
physical contact with the irritating substance causes a definite .and precise reaction
which is practically constant. This points either to touch or to taste.

To test the relative part played by stimulation of these two sets of sense organs,
the following series of experiments was performed. A half dozen fish in an aqua-
rium were tested a score of times with fresh meat on the tip of a wire, as in the
previous cases. The reaction was obtained uniformly, no matter what part of the
body or head was touched. Half an hour after the close of these experiments a bit
of cotton wool was wound around the tip of a wire and the fishes were tested with
this exactly as they had been with the meat. For the first six trials the barblets
only were touched. The fish in each case turned and seized the cotton as promptly
as the meat had been taken. The cotton would be immediately dropped. After a
few more trials the fishes would generally turn when touched, but would check their
movement before the cotton was actually taken into the mouth. Several specimens
were now tested on the trunk with the cotton. One or two turned completely around
and took the cotton, but generally there was a slight movement only toward the
cotton, which was checked before the cotton was reached. After a few further
tests, the fishes would usually pay no attention to a contact with the cotton on the
skin of the body and the reaction by the barblets became uncertain, until finally the
cotton could be freely rubbed over the barblets or lips of some of the individuals
without producing any response.

THE ORGAN AND SENSE OF TASTE IN FISHES.

255

These experiments were many times repeated, sometimes using' white cotton,
sometimes red cotton, and sometimes fresh meat. The reaction was uniformly obtained
with the meat. If at the close of a few experiments with the meat a minute pledget
of cotton was substituted for the meat, there was feeble or no response from rubbing
the body with the cotton, though upon touching the barblets the fish would usually
turn and often would seize the cotton and drop it again at once. After several repe-
titions, the fish became wholly indifferent to the cotton, no matter how it was applied,
or they would if touched upon a barblet turn toward it without biting it. They were
now again tested with bits of meat. This they took as eagerly and as precisely as
before, showing that they were still hungry.

After the interval of a day or two the fishes would still appear to remember the
cotton, and I rarely, after the first trials, got a prompt “gustatory” reflex with the
cotton. If they noticed it at all, they would turn slowly and touch it with the lips or
a barblet in a tentative or inquiring manner, only to turn away again without taking
it into the mouth. This deliberate movement may be designated, for reasons to appear
immediately, as the tactile reflex , as distinguished from the instant seizing of food, the
‘ ‘ gustatory reflex . ”

These experiments seem to show that in the reactions to the meat, both from the
barblet and from the skin of the body, the senses of taste and touch both participate.
This is in accord with the known innervation of the skin and barblets, for all parts
of the body surface receive general cutaneous (tactile) nerves, and all parts arc plenti-
fully provided with terminal buds (taste buds) which are innervated by communis
(gustatory) nerves. The experiments further suggest that these two sensory factors
can be experimentally isolated by training.

The fishes having become accustomed by brief training to make the simple reflex
of seizing the food under the stimulus applied to any part of the barblets or skin,
and doubtless utilizing both gustatory and tactile sensations, the gustatory factor is
eliminated by the substitution of cotton wool for the meat. The tactile sensation
alone proves to be sufficient to set off the reflex after the training previously given.
The stimulus is, however, never followed by satisfaction and is soon given up, the
fishes after further practice not reacting to the tactile stimulus alone. If, however,
the gustatory sensation is added, by the substitution of meat for the cotton, the
original reflex is given as promptly as before. This would seem to indicate that,
while the tactile sensation alone is not sufficient to maintain the reflex, the addition
of the gustatory element is sufficient, and therefore that the gustatory element is
the essential element in setting off the reflex. This hypothesis was tested by an
extensive series of experiments similar in plan to those last described.

In general there was no noticeable difference between the reaction to the white
cotton and that to the red, though in some cases, especially toward the end of the
series of experiments, after the fishes had learned to pay no attention to white cotton
when touched at any point by it, they would sometimes turn and touch the red
cotton with the lips or a barblet, immediately to turn away again without biting the
cotton as they did at first. The reaction is not the quick turn and instant seizing of
the bait, which I have termed the “gustatory reaction,” but a more deliberate move-
ment similar to what I termed above the “tactile reaction.” This occurred only
when the cotton was in plain view at the time of the contact and is probably in this

256

BULLETIN OF THE UNITED STATES FISH COMMISSION.

case partly a visual response, called forth by the similar appearance of the red cotton
and bits of beefsteak on which they were habitually fed. It was not by any means
constant, for, in general, after the first few days, contact with neither color of cotton
called forth any response whatever.

After this result was reached, I dipped the pledgets of white cotton in the
filtered juice of fresh beef and touched the body surfaces and bardlets with them in
the same way as before. In all cases I got a typical “gustatory” reaction exactly
the same as with the meat, and this reaction persisted after many trials with no
diminution. The cotton was taken instantly into the mouth and tugged vigorously.
No amount of training served to eradicate or to weaken this reflex.

I next prepared a small bulb syringe, w itl i the delivery tube drawn out to a
very fine point. This was filled with the water in which the fishes were and a fine
jet directed against their bodies. They either paid no attention or were disturbed
and swam away. I now substituted for the water in the syringe the juice of raw
beef pressed out and strained. When a jet of this fluid was directed against the side
of the body, the fish always instantly turned and tried to take the end of the syringe.
The reaction was identical with that produced when a corresponding part of the body
is touched with raw meat. 1 invariably got the reaction, both from the sides of the
body as far back as the root of the tail fin and from the skin of the head and barblets.

I also tested the fishes with bits of red brick held in forceps. The forceps
seemed to frighten the fishes. They either paid no attention to the contact with the
brick (when touched in such a way that they could not see the point of contact), or
else the harsh contact seemed to frighten them. I then touched them on various
parts of the body and the barblets with bits of brick which had been soaked in raw
meat juice. In most cases they would turn and touch the brick with the lips or take
it into the mouth, but often they seemed frightened and would swim away. I then
gave them a few bits of meat with the forceps and found that they took it eagerly,
being very hungry, but it had to be given more cautiously than with the wire, as
they were afraid of the forceps if they saw them clearly.

Next I dropped bits of brick which had been soaked in meat juice in front of the
fishes as they lay under leaves with the barblets projecting beyond the edges of the
leaves. In all such cases, upon touching the brick with a barblet, they seized the
brick and bit at it viciously. Often they would return to it a second or third time
and try to bite it. I dropped similar bits of brick which had not been soaked in
meat juice in front of them in the same way, but they paid no attention to them, or
in a few cases they would touch them with the barblets and then swim away again
(“tactile” reaction). They never attempted to bite them. Clearly they taste the
meat juice in the bricks when they are touched by a barblet, and the experiment
when the body was touched by a similar brick held in forceps shows that they taste
the juice by the body also.

On one occasion 1 tested the fishes with pieces of cooked meat that had been
long boiled so that nearly all of the extractives were drawn out. The experiments
were conducted just like those with the raw meat, but the fishes gave by no means
so clear reactions to it. Upon touching the sides of the body, the fishes usually paid
no attention to the stimulus, treating it just as they did cotton. I then touched the
barblets a few times, and to this they would generally react by turning and taking

THE ORGAN AND SENSE OF TASTE IN FISHES.

257

the meat, but not always nor so promptly as with fresh meat. Upon testing the
sides of the body again after this experience I got a reaction. The fishes would turn
and touch the meat with the barblet or lips before taking it, rarely giving the quick
reaction characteristic of fresh meat. Evidently the cooked meat has less taste to
the fishes than fresh meat and this interferes with the reaction. They eat the cooked
meat when they are sure that it is edible.

These experiments, all of which were many times repeated and controlled, I
think show conclusively that practically the whole cutaneous surface of Ameiurus is
sensitive to both tactile and gustatory stimuli, and that the latter call forth charac-
teristic reflexes which are of the greatest value to the fish in procuring food. The
fish normally reacts to contacts on the body by both types of stimuli — to the mere
tactile stimulus (if at all) by a tentative movement calculated to bring the doubtful
substance into contact with the more highly sensitive barblets or lips, but to the
tactile stimulus accompanied by the gustatory by an immediate, rapid, and precise
movement calculated to seize the food. This latter reflex is unvarying and is very
persistent under a great variety of forms of stimulation. The former (“ tactile”)
reflex is less stable, and may be readily eliminated by a simple course of training.
Clearly the gustatory element of the sensation complex resulting from a contact with
a sapid substance is more important than the tactile element.

It is clear that in order to call forth the characteristic “gustatory” reflex the
stimulus must be quite strong* and rather sharply localized. For when there is only
a small amount of meat juice diffused through the water, as by the presence of a
piece of fresh meat near the fish, he is not able to localize it accurately, but exhibits
only the “seeking reaction.” I have not as yet been able to convince myself whether
the fish could accurately localize a strong and sharply localized gustatory stimulus
with no tactile element. In all the experiments in which meat juice was directed
against the body with a pipette or syringe there was doubtless some tactile effect
produced by the impact of the jet. We know from the experiments that pure tactile
stimuli can be accurately localized on the skin, and there can be no doubt that under
normal conditions these assist in the localization of the food object. Compare the
further discussion in the Addendum, pages 270-271.

EXPERIMENTS ON GADOID FISHES.

The preceding experiments were all carried on in the zoological laboratory of
Denison University; the experiments on marine fishes which follow were made during
the summer of 1902 at the U. S. Fish Commission laboratory at Woods Hole'. The
feeding reactions of three types of gadoids were studied, viz, young pollock ( Polla -
chius virens ), about 10 cm. long; hake ( JJrophycis ten uis), about 20 cm. long, and
young adult tomeod {Microgadus tomcod).

As is well known, the hake and tomcod have a mental barblet which ds known to
be abundantly set with terminal buds and which receives both communis and general
cutaneous innervation. In all three types the lips are freely supplied with terminal
buds and there is a recurrent branch of the facial nerve, the ramus lateralis acces-
sorius, which carries communis fibers into the trunk to supply terminal buds found
on the fins, especially the free rays of the ventral or pelvic fins. These fins are far
forward under the throat. In the pollock they are but little modified; in the tomcod

F. C. B. 1902—17

258

BULLETIN OF THE UNITED STATES FISH COMMISSION.

two rays are about twice as long as the others and for about half their length they
project freely below the rest of the tin. In the hake all of the rays of this tin are
suppressed save these modified free rays, so that the tin is filliform, branched at the
end. Microscopic examination shows that the terminal buds are more abundant on
the more highly modified tins. The hake also has a free filament on the dorsal fin
produced by the extension of the third and fourth rays beyond the others. I have
not examined this free filament microscopically, but know that it receives communis
fibers from the r. lateralis accessorius, and have no doubt that it also has numerous
terminal buds, as the experiments show it to be very sensitive to gustatory stimuli.
The pollock have very large eyes and are excellent visualizers. When food is thrown
into the water, they dart for it and in general the}' take their food by the visual reflex.
So keen is the vision that it would be difficult to carry on any experiments, such as I
have done with the other two species, without first blinding the fish. Nor do they
habitually drag the bottom with the free ventral fin rays as the others do. I have,
therefore, not devoted much attention to this species, preferring to study more care-
fully those species in which the gustatory reflex plays the greater part in the life of
the fish.

The hake ( Vrophycis tenuis). — These fishes, like the tomcods, readily adapt
themselves to life in captivity, and are easily experimented upon in small tanks.
They are excellent visualizers, though not so much so as the pollock. When bits of
meat are thrown into the water they usually catch them before they fall to the
bottom, and their keen vision makes difficult such experiments as I carried on with
the cat-fishes. They do not seem to recognize by sight food lying on the bottom, but
only when it is in motion. But bits of meat, fish, or clam lying on the bottom are
usually found by the aid of the free ventral tins. These fishes spend much of their
time in slowly swimming in an apparently aimless manner close to the bottom of
their tank. During these movements the filamentous pelvic fins are so held that
their tips drag the bottom. These fin rays are quite long, and they are usually
directed obliquely forward, outward, and downward, with the two branches of each
fin widely divaricated, so that the four tips touch the ground in a line transverse to
the body axis at about the level of the mental barblet. In this way the bottom under
the fish and for a short distance on either side is thoroughly explored as the fish
swims over it. and all food particles with which the barblet or free fin rays come in
contact are taken by a quick and precise movement similar to that set off in the
siluroids by contact with their barblets. Bits of meat or clam on the end of a slender
wire could be laid on the bottom of the tank and then slowly moved up under or
behind the fish and the reflex from the ventral fins tested in this way. Such experi-
ments, however, had to be made with great caution and many times repeated to rule
out possible visual sensations which likewise call forth an immediate reflex.

Bateson ('90, a) records similar reactions with the rockling ( Motella ), a gadoid
fish with the same general structure and distribution of terminal buds as the hake,
but with better developed barblets. (On the structure of the pelvic fins of Motella
compare Bateson’s account on p. 211 with that on p. 234 of the same volume.)
Bateson, moreover, got the same reflex with fishes which had been blinded, and I have
not thought it necessary to repeat this experiment, for my fishes give sufficiently
clear evidence that this reflex from the fins is wholly independent of vision. We

THE ORGAN AND SENSE OF TASTE IN FISHES.

259

have, however, to investigate the parts played by tactile, gustatory, and olfactory
sensations.

Bateson’s remarks (’90, a , p. 214) in this connection on the rockling may be
quoted here. The three-bearded and the live-bearded rockling are nocturnal and lie
still all day.

Generally, both the animals take no notice of food until it has lain in the water some minutes,
when they start off in search of it. The rockling searches by setting its filamentous pelvic fins at
right angles to the body-, and then swimming about feeling with them. If the fins touch a piece of
fish of other soft body, the rockling turns its head round and snaps it up with great quickness. It
will even turn round and examine uneatable substances, as glass, etc., which come in contact with its
fins, and which presumably seem to it to require an explanation. The rocklings have great powers
of scent, and will set off in search of meat hidden in a bottle sunk in the water. Moreover, a blind
rockling will hunt for its food and find it as easily as an uninjured one.

The above, taken in connection with other passages, shows that this author con-
siders that the food is found largely by scent, and that the fin reaction is essentially
tactile, though he has seen the sense organs on the pelvic fins and recognized their
resemblance to taste buds.

Examination of stomach contents shows that the normal food of these hake is
largely crustaceans, particularly shrimps. I fitted up a tank with some seaweed and
put into it a large number of prawns ( Palxmonetes ), mostly living, but some dead.
Upon putting the hake into this tank, they immediately ate some of the dead prawns
from the bottom and afterwards caught the live ones, but very slowly and with many
failures. The response seems to be wholly visual. The fishes would repeatedly
pass directly over living prawns, touching them with the fins or being brushed by
their antennae, but so long as the crustaceans were quiet they seemed not to notice
them. If, however, a prawn was killed and crushed and thrown back into the water,
it was immediately found. Upon another occasion I put a live clam into the tank
with the hake, where it remained for several days, with siphons greatly extended.
The fishes repeatedly brushed over this siphon with their free fins, but never paid
any attention to it, though if a similar siphon were cut off from a live clam, so
as to allow some of the juices to escape, it would be immediately taken and eaten.
Evidently live food is not clearly located by the gustatory organs of the fins.

Besides observing as fully as possible the normal feeding habits of the hake, I
experimented upon the reactions to stimuli applied to both the pelvic and the
filamentous dorsal fins. As mentioned by Bateson, the pelvic fins are freely used to
explore all manner of substances which may attract the notice of the fish, whether
edible or not. After these fishes have become accustomed to being fed small bits
of meat or clam or mussel {Modiola) in their tank, they immediately swim toward any
small unfamiliar body with the pelvic fins thrust forward to touch it before the
mouth reaches it. Sometimes the tips of these fins close over it with a movement
strongly suggestive of grasping, though of course this they can not do.

Upon testing by contact with meat or other bait, the free dorsal filament is
found to be quite as sensitive to gustatory stimuli as the filamentous ventrals. The
reflex in this case is very characteristic and constant — the fish upon touching a
savory morsel checks its forward movement and immediately ‘‘ backs water” so as to
reverse the movement of the body until the object is directly above the mouth, when

260

BULLETIN OF THE UNITED STATES FISH COMMISSION.

it is taken at once. This reflex usually (though not so invariably) follows a contact
of meat upon any part of the dorsal fin, as well as the free filament. The reflex
rarely fails when any one of the filamentous fins is touched by freshly cut meat. j

After meat has been in the water for fifteen minutes or more it seems to lose its
savor and the fins may be repeatedly dragged over it without calling forth a
response, and the same is true of the barblet and lips.

I tested the filamentous fins with a wisp of cotton wool on a fine wire, as I did
the cat-fishes. It was rarely noticed at all by the pelvic fins, but at the first contact
with the filamentous dorsal the fish reacted just as he did to meat with which he had
been tested immediately before. Upon repetition, the response was soon discon-
tinued. For a few tests the fish would pause, and perhaps back up slowly so as to
smell the suspicious object or touch it with the barblet, but it was not taken into the
mouth. After from two to ten tests no further attention was paid to the cotton, or
the fish would pause a moment without backing up. This experiment was many times
repeated in the course of the first day of its trial and daily thereafter for some time.

If three or four hours intervened between two series of about twenty tests, the first
one or two tests of the second series might be followed by an incomplete reaction,
but after that usually no notice was taken of the cotton. The fishes apparently
remembered the preceding tests. But if more than twenty-four hours intervened
between tests, the process of training usually had to be gone over again.

The fact that the hake does not appear to remember the difference between the
pure tactile stimulus and the tactile plus the gustatory for so long a time as the cat-
fish does is probably to be explained b}r the fact that the number of taste buds on
the filamentous fins of the hake is much less than that on the bardlets of the cat-fish,
and therefore the gustatory element in the sensation complex is doubtless much less
in the hake. The whole course of the experiments indicates that the response is in
fact much more strongly tactile in the hake.

During the course of these experiments I often alternated bits of meat with the
cotton wool, and at other times substituted cotton that had been soaked in clam juice.

In these cases 1 always got the characteristic gustatory reaction by all of the filamen-
tous fins, no difference being observable between the reaction to meat of clams or fish
and that to cotton soaked in filtered clam juice.

I also tested the hake with gelatin which had been soaked up in cold water.
Shreds of the well-softened gelatin were fastened to the end of a wire and brought
into contact with the body surface. The reactions were identical with those obtained
with white cotton. The gelatin shreds are very nearly colorless and absolutely
tasteless to my tongue. But to the sense of touch they are almost exactly the same
as the bits of fresh clam meat with which most of these experiments have been con-
ducted. The hake at first would take the bait when the filamentous dorsal was
touched, but if the gelatin was taken into the mouth it woidd be immediately
rejected, and after a few trials the fish would no longer respond to the stimulus.

He acted in the same way when the pelvic fins were stimulated. Shreds of the
softened gelatin falling through the water were sometimes noticed, but rarely taken
into the mouth, and if so, were immediately rejected. Similar shreds lying on the
bottom were neglected, even though the barblet and filamentous fins dragged over
them repeatedly.

THE ORGAN AND SENSE OF TASTE IN FISHES.

261

I next took small clam shells that had been lying long in the tanks containing
the fish and were, thoroughly cleaned of fleshy matter and which the fishes had not
paid any attention to for days. These 1 dried and warmed and then filled
with melted gelatin which had been previously softened up in cold water. Upon
cooling there results a mass, colorless, tasteless, and odorless, which feels almost
exactly like the flesh of the clam, which has often been fed to the fishes in this way.
Upon dropping these shells into the water, the fishes eagerly snatch them up, feel of
them with the lips or barblet, and then bite into the gelatin. They immediately
reject the gelatin and they never repeat the process. Even if they draw the fins or
barblets repeatedly over the shells and the contained gelatin, they never again pay
any attention to them.

1 also repeated with the hake the experiments which 1 had previously carried
out upon the cat-fish, using a fine-pointed pipette and sapid solutions. The fishes
were in all cases first tested with sea water taken from the tank in which they were
swimming. • On one occasion (the first test made) a jet of water directed against the
filamentous dorsal was followed by the characteristic backward movement of the
fish, so that he finally received the jet in the face. He turned and tried to take the
point of the pipette in his mouth — a purely tactile reflex apparently. This response
1 never got again with this or any other fish, though occasionally the fish would
stop, hesitate a moment, and then swim on, paying no further attention to the
stimulus. If the jet of water is directed against the pelvic fin while it is extended
and searching the bottom for food, the fin is usually quickly withdrawn and pressed
against the side of the body.

The pipette was then filled with the freshly prepared and strained juice of the
mussel ( Modiola ), and this was directed against the fish in the same way. The fishes
responded instantly, just as when stimulated by meat, whether the jet was directed
against the filamentous dorsal, or the dorsal fin at any part, or the side of the body,
or the free pelvic fin. The reflex is immediate and unmistakable, more sharply
defined than I usually get by contact with the meat of the same mussel. The experi-
ment, was many times repeated, always with the result that the jet of water was
ignored or avoided, while the jet of mussel or clam or crab juice was eagerly sought,
the fish usually snapping- at the end of the pipette.

I have carried out no systematic chemical experiments to determine the gustatory
preferences of the fishes, having shaped my experiments so far as possible along the
lines of the normal feeding habits of the species studied. Nagel and some other
previous students of these problems have relied chiefly on reactions to unpleasant
stimuli, and the reader is referred to their works, though I consider this a less satis-
factory line of inquiry than the study of normal reactions to food substances. The
few fragmentary observations which 1 have made with chemical stimulants I shall,
however, record in their appropriate places.

Specimens of hake were tested with a 0.2 per cent solution of hydrochloric acid
made up in distilled water, the acid being directed against the body by means of a
fine pipette. The dorsal and ventral fins, the sides of the body, and the lips were
tested. When first tested on the fins one hake turned and tried to take the pipette,
much as he did with the clam juice. Afterwards this fish, as well as all the others
from the first, seemed rather to dislike the acid and would swim slowly away. There

262

BULLETIN OF THE UNITED STATES FISH COMMISSION.

is no constant reaction, however, and in fact the fishes act very much as they do
when a jet of simple sea water is directed against them. They do not appear to
dislike the acid intensely. Later I tested these fishes with a 1 per cent solution of
hydrochloric acid in sea water. This is decidedly unpleasant and is uniformly
avoided.

The experiments recorded seem to show clearly that the hake receives both tac-
tile and gustatory stimuli by means of the free fin rays and to some extent doubtless
by other parts of the outer body surface. What role may be played by the sense of
smell remains obscure. To test the powers of locating concealed food the following
experiments were tried:

In a tank containing two hake which were very hungry I placed a piece of fresh
clam meat concealed between two small, old, and thoroughly clean clam shells which
had been lying for some time in the bottom of the tank. The fishes did not seem to
smell the meat at a distance and so be attracted to the spot where the shells were, but
if in the course of their aimless movements along the bottom of the tank they passed
over the shells, they generally stopped a moment, smelled around, and then passed on,
first feeling over the whole area of the shell with their free fins. As time passed,
this reaction became less clear until after some fifteen minutes they generally passed
over the shells without paying any attention. They never found the meat. This
experiment was many times repeated with the same result. The sense of smell can
play no strong part in the locating of their food. It may play some small part,
though I incline to believe that the interest which the fishes show in the concealed
bait is excited by a vague stimulus to the terminal buds on the fins. Compare the
experiments made after extirpation of the olfactory organs in the tomcod described
below

The tomcod (Microgcidus tomcod). — These fishes are much less active than the
hake, spending most of the time lying quietly on the bottom of their tank. They
have not so keen sight as the hake and pollock, but still obtain much of their food by
this sense, catching food thrown in before it reaches the bottom. They do not catch
live prawns in captivity so well as the hake do, yet prawns and other active crusta-
ceans are found in the stomachs of specimens taken with the seine. The dorsal fin
lacks the free filamentous rays and is not especially sensitive to gustatory stimuli.
The ventral fins are, however, very efficient in locating sapid substances lying on the
bottom. They are shorter than those of the hake and are not thrust forward, but
incline slightly backward. Like the hake, the tomeods spend much time in slowly
exploring the bottom, though they assume a very different position, with the head
directed downward at an angle of some 30° to 45° with the bottom, so that the tips
of the barblet and ventral fins just drag the bottom. When food particles are located
they are snapped up by a quick lateral movement similar to that of the cat-fishes.
Sometimes, however, stimulus of the ventral fins is followed by a reversed swimming
movement, the fish backing up to take the bait. At other times the fish when explor-
ing the bottom swims slowly backward, so that no change of direction is necessary
when food is located.

I made a series of tests with cotton wool and cotton dipped in clam juice similar to
those described for the hake, and with the same results. I also repeated the tests
made with sea water and with strained clam juice by the aid of a pipette, with iden-

THE ORGAN AND SENSE OF TASTE IN FISHES.

263

tically the same results as with the hake. After a few tests the fishes ignore sea
water and plain cotton, but invariably respond to cotton soaked in clam juice and to
the juice itself as they do to meat. The tomcod reacts to bits of clear gelatin soaked
up in water essentially as the hake does.

I also tested the tomcod with hydrochloric acid, 0.2 per cent in distilled water
and 1 per cent in sea water. Both are obviously avoided. 1 filled a tine pipette
with a solution of quinine sulphate in sea water, about 0.1 per cent — a very bitter
solution. The tomcod swims away immediately if applied either to the lips or to the
pelvic fins, but appears not to notice it if applied to other parts of the body.

Within two old clam shells, which had been lying in the tank with the tomcods
for several days and had remained unnoticed, was placed a piece of fresh clam.
They were then closed together and laid on the bottom of the aquarium containing a
tomcod. Shortly the fish passed near it, appeared to perceive it, turned from his
course, and passed and repassed the spot until the shell was located, apparently by
smell, by a method of “trial and error.” Then he rooted at the shell vigorously
until the two halves were separated and he could get the meat. I repeated this with a
piece of squid within the shells with the same result. I tried two empty shells in the
same way. He saw me put them into the water, came up to investigate, smelled (?)
of the shells and went away without so much as touching them, and never came back
to them again.

These experiments were repeated in many forms many times. In most of these
cases the efficient organ in discovering the presence of the food was almost certainly
the pelvic fin. At least, this alone located it, for the fish swam about (possibly
feebly smelling something good), but did not make a definite movement toward the
bait until the fins were dragged over the crack between the two shells containing it,
from which the juices were doubtless being diffused out into the surrounding water.
Then he backed up in the typical way. If the bait was not found within a very few
minutes it was left unnoticed, even though subsequently uncovered.

These fishes almost invariably find a concealed bait, though the hake rarely does
so. The hake seems to perceive the odor or savor of the food, for he lingers about
the spot where it is concealed, but never makes a movement to uncover it. The
tomcod, on the other hand, actively pushes things about with his snout until the bait
is discovered. But, unlike the gadoid fishes which Bateson describes, these fishes do
not get the scent of the food at any considerable distance and then search for it.
They do not notice the bait until within a few centimeters of it, and there is no
evidence that the sense of smell assists at all in the localization.

To test this point the olfactory organ was extirpated in several tomcods which
had given the reaction last described clearly. Several ways of performing this
operation were tried. The most successful method was to etherize the fish sufficiently
to keep him quiet and then operate in a shallow tray with the mouth kept under
water, cutting off the olfactory nerves or crura with a sharp scalpel. The wound
suppurated badly, but appeared to give the fish no serious trouble, as they feed
normally from the second day onward. Without going into the details of the observa-
tions, I may say that after the third or fourth day the fishes took their food in all
respects like uninjured fishes, so far as could lie observed. They gave all of the
characteristic reflexes that have been mentioned above, including the discrimination

264

BULLETIN OF THE UNITED STATES FISH COMMISSION.

between cotton wool and cotton dipped in clam juice and between sea water and clam
juice applied with a pipette, etc. The operated fish would locate a concealed bait by-
means of the pelvic fins exactly as the normal fish does, and he would similarly7 root it
out and eat it. In short, the gustatory reflexes, so far as I have observed them, were
absolutely unmodified by the operation. That the olfactory apparatus was totally
destroyed was verified by autopsy dissections made after the close of the observations.

OTHER FISHES.

The sea-robin (Prionotus carolinus) . — The three finger-like rays of the pectoral
fins of the gurnards have long attracted the attention of zoologists, and the American
species of Prionotus have been made the subject of a careful research by Morrill
(’95). He finds that, as in the closely related European Triglct , the free rayrs are
totally devoid of terminal buds or other specialized sense organs and that the sensory
nerves with which these free rays are so abundantly supplied end free, like tactile
nerves in general.

lie also made some interesting physiological experiments. The normal food of
these species, so far as known, is small fish, young clams, shrimps, amphipods and
other small Crustacea, squid, lamellibranch mollusks, annelids, and seaweeds. (Linton,
1901, p. 470.) They are constantly feeling about the sand, turning over stones and
feeling under them, etc., with these free rays, and undoubtedly find their food largely
in this way, especially the annelids, mollusks, and Crustacea; but in captivity the
eyes are used chiefly in securing the food. Morrill writes further:

In order to test the use of the free rays independently of sight the crystalline lens and cornea
were removed from some fish, and in other cases the cornea was covered with varnish, balsam, or tar.
The repeated experiments were negative in their result, as the fish paid no attention to the food, even
when it was placed in contact with the free rays.

Morrill concludes “that the free rays have been modified for tactile purposes,
and that they are mainly, if not altogether, used in searching for food.”

Morrill’s dissections leave it uncertain whether the free rays of the pectoral fins
receive communis nerves, as they should do, of course, if these organs had given
evidence of gustatory powers. The only source of communis fibers for this fin
would be through the ramus lateralis accessorius (r. recurrens facialis). Stannius
(1849, p. 49) did not find this nerve in Trigla gumardus and T. hirundo. I dissected
a specimen of Prionotus carolinus and found the same to be true here, so that it can
be taken as assured that no communis nerves reach the pectoral fin in this species.

After an examination of the feeding habits of the adult sea- robin and of young
specimens about 10 cm. long 1 quite agree with Morrill that the reaction to food
particles by the free fin rays is tactile only, with no gustatory element. When
adults are fed with fresh clams or mussels, the shells split open to expose the meat,
they turn and bite out the meat as soon as a free ray touches the soft flesh. Young
fishes did not give this reaction so invariably, and evidently relied much more on
sight. Clean clam shells filled with melted gelatin were reacted to like the fresh
clams once or twice by each fish, but. usually were thereafter ignored.

The free rays constantly stir up the sand and gravel of the bottom. If soft
edible particles are touched the head may be turned to snap them up, especially with
old fishes.. With younger ones this usually does not happen unless the particle is seen

THE ORGAN AND SENSE OF TASTE IN FISHES.

265

while in motion. In fact, with these younger fishes the purpose of the activity of the
free rays seems to be in the main the agitation of particles on the bottom to bring
them into the range of vision. Almost any unfamiliar object, such as a bit of coal or
a brightly colored pebble or any soft particle, if seen while in motion, will he apt to
be taken into the mouth. The analysis is done here — not by the peripheral cutaneous
organs. All small objects thrown into the water are taken into the mouth as they
fall; bits of filter paper, gelatin, etc., will be taken and immediately rejected. The
same bit of paper or excrement may be taken and rejected a half dozen times in
rapid succession, the reflex following in a perfectly automatic way as soon as the
moving object is seen. Small worms when thrown into the water would be captured
before they had time to reach the bottom, but if placed on the bottom they would
seek shelter under pebbles and remain unnoticed until they were stirred up and sent
floating off, when they would be seen and taken at once. The free fin ray was
observed to touch the worm when concealed without evoking a response. A moment
later the worm was set in motion and taken at once.

1 got no evidence that the fishes smell or otherwise detect the presence of food
at a distance or concealed from sight and touch. Meat inclosed between clam shells,
which a tomcod would have secured within a minute or two, remained unnoticed,
though the outsides of the shells were repeatedly fingered over by the free rays and
similar bits of meat were taken at once if in motion near the fish.

The young sea-robins eat crab meat well. I made a strong- extract of crab meat
and filtered it. Now with a fine pipette a jet of clean sea water was directed against
the free pectoral-fin rays. There was no response, or if the jet was strong the fin
was folded against the body. The extract of crab applied in the same way with the
pipette gave the same result. Even when the jet is directed against the lips the fish
usually pays no attention or is disturbed and swims away. This would seem to
indicate that t he sense of taste is absent or very feeble on all of the exposed parts of
the body. Thus the absence of special gustatory sense organs, of communis nerves,
and of gustatory reactions from the free rays of the pectoral fins serve as mutual
controls.

The king -fish ( Mentieirrhus saxatilis). — These fishes have a short, thick mental
barblet, and they were studied to compare their reactions with those of the siluroid
and gadoid fishes. Most of the types of experiment made previously on the latter
fishes were repeated on the king-fish. Without going into details, the experiments
seemed to show in general that the king-fish is not a pure visualizer. though vision
is somewhat used in finding food. This seems to be in the main a tactile reaction,
as most of the food taken was by contact and non nutritious substances were
generally taken if they felt like food. For instance, colorless gelatin is taken at the
first contact and repeatedly thereafter for an indefinite number of times, though in
each case it is at once rejected as soon as it enters the mouth. The sense of taste
seems to be limited to the mouth, and I found no evidence of a gustatory reaction by
the barblet, though the experiments were not sufficiently numerous or varied to be
conclusive. They do not find a concealed bait.

The toad fish ( Opsanus tau). — These fishes were experimented upon at the same
time as the hake and tomcod, and by the same methods. The toad-fish never found a
concealed bait and never seemed to get food by any other reflex path than the visual

BULLETIN OF THE UNITED STATES FISH COMMISSION.

266

or tactile. The fleshy, cutaneous appendages of the skin were especially tested to
bring out possible gustatory reactions, but with negative results save for those bor-
dering on the lips, where it was impossible to exclude the participation of taste buds
on the lips. This agrees with the anatomical findings of Miss Clapp (1899), whose
careful study of the skin of this fish failed to reveal any terminal buds on these
appendages or elsewhere away from the buccal cavity . A jet of sea water directed
against these appendages or the body surface in general usually disturbs or frightens
the animal merely, if it is noticed at all. A jet of clam juice similarly applied calls
for the same reaction unless it is so directed as to reach the lips, in which case the fish
reacts to it just as the hake and tomcod do, attempting to take the tip of the pipette in
the mouth. The following solutions were applied in the same way by a tine pipette
to various parts of the body surface: 0.2 per cent hydrochloric and 1 per cent hydro-
chloric acid in sea water, and 0.1 per cent quinine sulphate in sea water. In all cases
the fishes paid no attention to the stimulus unless the substance was so applied as to
come into contact with the lips. The experiments lead me to conclude that the toad-
fish can taste only within the mouth and on the lips, and that if the cutaneous appen-
dages have any sensory function it is tactile only.

CONCLUSION.

The morphological and physiological significance of the terminal buds of fishes
is a problem which has exercised some of the ablest morphologists for over half a
century. The methods of the older anatomy have signally failed to yield concordant
results. Not until the innervation of the cutaneous sense organs was worked out
from the standpoint of nerve components was this confusion relieved. The older
morphologists (Schulze, Merkel, and others) discovered a morphological criterion,
the “hair cells,” by which the terminal buds could be distinguished from cutaneous
sense organs belonging to the lateral-line system. But this fact attained its signifi-
cance only when it was discovered that the organs of the lateral-line system, or neu-
romasts, which possess the “hair cells,” are always innervated by lateralis nerves
related centrally to the tuberculum acusticum, while terminal buds, which lack the
“ hair cells,” are always innervated by communis nerves which are related centrally
to the primary gustatory centers of the vagal and facial lobes.

Presumably, then, lateral-line organs and terminal buds have different functions;
and, further, the function is probably not tactile in either case, since all parts of the
skin receive general cutaneous nerves in addition to the special sensory components,
and these general cutaneous nerves are related proximallv to different centers from
either of the others. The lateral-line organs are known to lie used in the maintenance
of bodily equilibrium and the perception of mass motion of the water. (Compare
the recent works of Lee and Parker.) On the other hand, the terminal buds are
related in structure and innervation to undoubted taste buds of the mouth, and hence
the inference that their function is taste. This inference is abundantly confirmed by
the experiments here recorded, and the function and morphological rank of the
terminal buds are at last definitely fixed.

It may be regarded as established that fishes which possess terminal buds in the
outer skin taste by means of these organs and habitually find their food by their

THE ORGAN AND SENSE OF TASTE IN FISHES.

267

means, while fishes which lack these organs in the skin have the sense of taste
confined to the mouth. The delicacy of the sense of taste in the skin is directly
proportional to the number of terminal buds in the areas in question.

Numerous unrelated types of bony fishes from the siluroids to the gadoids which
possess terminal buds have developed specially modified organs to carry the buds and
increase their efficiency. These organs may take the form of barblets or of free
filiform fin rays. The free rays of the pelvic and dorsal fins of gadoid fishes are thus
explained, and indeed this is possibly the motive for the migration into the jugular
position of the pelvic fins of the gadoids.

In all cases where terminal buds are found on barblets or filiform fin rays gusta-
tory nerves belonging to the communis system are distributed to them. These
barblets and free fin rays likewise receive a very rich innervation of tactile or gen-
eral cutaneous nerves, so that they merit their popular designation — “feelers."
Both sets of end organs undoubtedly cooperate in the discrimination of food, and the
animal has the power of very accurate localization of the stimulus. Whether the
gustatory stimulus alone can be localized apart from its tactile accompaniment can
not at present be stated. A purely tactile stimulus with no gustatory element can
be localized precisely, and I have as yet no conclusive evidence that a pure gustatory
stimulus, even when strong, can be located by the fish. It is certain that feeble and
widely diffused gustatory stimuli can not be accurately located by the fishes which I
have experimented with, either by the terminal buds or by any other organs.

The fishes in which the cutaneous terminal buds are most highly developed are
in general bottom feeders of rather sluggish habit, and in some cases they are noc-
turnal feeders. The high development of this sense is compensated for in some
fishes by the reduction of others. The visual power of the fishes is especially apt to
suffer degradation. This degradation may be organic, a positive degeneration of the
visual apparatus, as in Ameiurus , or it may be merely functional. In the latter
case, though the organs of vision are not necessarily modified, these organs are not
actually used in procuring food, the fish being unable to effect visual reflexes toward
food substances or to correlate visual stimuli with the movements necessary to react
toward food substances. The fish may be perfectly able to effect other visual
reflexes, but is apparently unable to understand the significance of food when per-
ceived by the sense of sight only. This particular central reflex path has never been
developed, or has atrophied from disuse. Nature has here effected for the species
something similar to what is accomplished in individual men occasionally by disease,
in the production of certain aphasias.

The number of reflex activities habitual to an animal with a nervous system as
simply organized as the bony fish is probably far smaller than is commonly supposed,
and these activities are in general characterized by but little complexity of organiza-
tion. It is probably quite within the range of possibility to determine by observa-
tion and experiment for any given species of fish, to a high degree of accuracy, what
these habitual activities are and to work out by histological methods the reflex arc
within the nervous system for each of them; and since the human nervous system is
built up on the same general plan as the piscine nervous system it follows that such
a thorough and systematic correlation of function with structure would be profitable
from many points of view.

BULLETIN OF THE UNITED STATES FISH COMMISSION.

268

Terminal buds do not occur in the outer skin of all fishes; in fact, they are prob-
ably lacking here in the greater number of species. But whenever they do occur
they tend to be arranged according to one general plan. This is particularly true of
their nerve supply, for, though the details of the peripheral nerves of fishes are
exceedingly diverse, yet the main communis branches for terminal buds, when such
occur, are substantially similar from the Siluridae to the Gadidse. There are, how-
ever, striking resemblances in