meee) ban Hh bell ath Mbaone tt Bebe Hints bs Segoe as pe sn eOAt MACOMS eben imo he tnt heaee hie Sh et Se Be bois My rete the tay rere ey cies i) rh des patty es Nyt ivess sfeters D dee reyes ssh bebe ‘¥ + o Leet + shee pe vese CEbes erie tom ny. + Se-ear Tere h een ee ee lene wis Cera sry it y) 1 eH ote Pe uMy MPH)» hey Cai _e oe + Waele he prety Hoy a meg at ah ity Cote Peon hMM wy ghd lb etal eh Stroy nie edna mht be te Be te ie Pte ne neiiety eae be be ie rhe to Canin pleas pMdioe be pe y Gt, Peele OO eh mechs tote ,; Ve vane unin SA, Soarh be em ehh eon ‘) Mabey a seh ee Clit esc rre ea Vee shat a $e ue eee artery CBI Ryn ittietne etre as pe btbe eb hae 4 Pelicee aerate lededsetehalecalemeetatcit ar eae acl ae Saleh tel rye ¥ bs 00 he b> babe be Con bret iatet eh ny pene tty ahs yee * sis br Weibstigee sratrnasraps i Mien he 4 thigeh oe ee be bein aie ive ‘ EGO, JRF Phen \ bel henge py Taine eh rhea blab Ven eh ie . i Greate betedadatedote’ ) + ry be Sey Lon tt Gita 400) SMe hae ed. Sines! beh we mee weve bh, ft ern Beppe pulp tou =e ying’ ag eh eae hice Vhs oce Ceeuit pte aetane okt ey hw be 40) es M rebede ys ale: Fe (ie og he ins ba rh Meh Mee tb rd be eh ie peediad tah! Meal Attia a ean B eh Nie eb te wh bet te epee Meta iy es cas peerereeens MEV RE No isesess hae gly eee Nisieataath hh hbk “te re Seer ttt ry is yy) we he, asgeeh Aled ‘ ete be eee es hitlsta bt eh es eat) SL hee wet shew wal Li Haemeck iste ; Pretest: CEtairuler -* rere te eb ‘epee bbe he he tire pede patina err ‘1 atc! a i eae sate og he igh emits ri Dt be ety Me he bee fe be rw bet sede ete ne Me - Tem Waid Mb ‘ ty Mieke Peiche brisk HH debe bem be HIP WRN oe Cs easter. n> hae Me eho ehctity iets) veer Pati yee rhb Mid we ry b> ty wontly: abi atts Sry wht fe dp rae. PERK ree ot se 2 3 nae ot SP miten heheh Sg ee ENG > . Mis ee pee tout t “4 ey hes oe HP beds ee eh itis per iet! Ve ein bro ibe monte eb Pato we dope + be heh tela x is tf bt be * eneers aeareite abn t Rhye Op bpe ve) uote phabetes orp ae Cort itars ets hb a hee nee td Af Ht sabe betel 10 wore Hae be ele wr ae aera tnt ede evn be ot be ve Titet ett to sh er O94 jie we ve ee bbe pas - ee ar ee Intehar’ iy eben topes set att Mindebrarhein to ichy fa Bec wads Rede be Drie fe eh ey) por te by te be. ee bee hark oleh ets Whee hie ess t Tak altel he ekstaltbeietohotal i baa pit ot be inks $e ee py Fine «hee ce Hate hetelabel it eitieey ay al ‘ tea fey ish bes ehetetn late leteseheri’ sr estalgs bheloles * “ bbitetete tote ft segs ee Wincantyth! Jos bet beatae be beta boy perc ie eae eae: te alee =) mye Mbt beisy bat pe t) Veta halen oy set acaes Plath eetatae hMdehio {i tebe ies are ber eretegrrerseter fas he’ ' tet Nye peste, peerertents Hou a Nh no bh febeigtete ee beh mish ier beh Mactan: ie Jota Le Hebe beeas Parisien te te orn DrheM pa del shaun fab aby pele = ebee nde Make Levaoncn, i Debra be bebe Ghetian ete Ait eh ee ee ry be Soe 1s be Bebe ts #2 MS Yen Sheer oe hh bee Pt Vo is ete athe Ne CALL ME Mee be th ory est heset boty! MUD ie bbe Wie te ay apes bee pte bs Dhte reott ts ne bat iin tH Risen + be be be be behing let tuishs Birt anys Si hy | lei Ts te] nan i i} nie he Eeatteg Mp ai “ eet ieee to vienw nn dae isty Baar oats ene Mela lobuhaltheletah) aa ee ay. Sh» demesene be he tun 1 ge set etedibon, wee rey Ware bgebr ee ae deting es ve retried Inbsiors exdeintearst sabtin (iittala neh f° s Y SUPA PM RM eh Rhy ba ony FARTS Wen anne ary ne skeletal hum eh Se vets he yany on Lt be bs ie res ba bee Moria wate percep se agitate) lve eo): tebe AAAs HN Hog So wie be 1h cain ar eter ies bgt ea le eet Ret Cakghy bee init bape bebe eet erry ebb ape ayy est tie mre be pe +) Pe kiny sane Snr bity meh We wee i TO nae be Web itaeieoregin eh te be acre bebe os be bate hh his eboee ee ee o Pat be oe Cot iret RULE Cheers vibe » haere Actceatineta seater fi, Se meee Ma eet ote , +a ye hgeree teats Pane nee PMA nie Be ¥ ite fire vdhy viene Gots oi ‘3 oct tt et ree whee a hn bee “wee coined ¥ Bevel toe a mde Lhateteltelstaditabeeet stat AC eatelitens re Waihi e heel Pe eee ep ttatattyt tad ey CCpoC unt LE ane wpe eee prae hr Pg Pete d ohh phy PAA Mabel s thee) - dy yin ¥ ¥ Woe Pt * " ettedteteret Woe by oe CUE Ulall vi pele th pans os Chty Ahn ern Ceri Ttslleltal Peal ot ee ae Corre it Chi) Cus bs ¥. x > > * * > i 13%, Sita ery ne ” * : < ie ae fs x, = eats, bute rh ue YE res be un eb Ot ae oe Oe tie Nite ie, Sebteisi sesh io hseie pestis eistohee: OF etree Fete he eee missle tse hthehobdarhhowheustes waeney Ese mir ry os he ose vile a eS ty a » q Bota tlbeetter ate part tia yap as : mies Salleh . ie Inlepe hte Rrecsitieiits Batt ikke Ueki ee hey Seti eaten ae Halts his bebe re ieh with at waehbuteenteh Chee abrhra tr siete tptatri rte yt] Rehan, MalMdlistis ies a ce ball Nahe igi rin " ‘ to 9 te tN Rbk tdehet yy! feat ag Tyas t t ioe . Feith hh 1 eae, Mleleinhede ls Motels ts Seater 3 Si Minin be 4°4 bo ede $4 4 lege ym ae ahi item ets heals emt 4 6 babe oe dr oa bt ade Mott Web Wee ae eee He edo dehy} Peet Sup teh Perms tabete ia’ Sebi is he us Me be be tnt Mr adebees sf Seplig ga ide \ Mir be bene , its it shelly byte elpistilst ; sayetrs a phe be Ptatdst sett OV Wn a Oy 8 ee dhe vah badeie te yiyisgs, ane Sees SF, =. = = = = ort = i i Bets ns rg FN Sg) At ne age th! Woy S haw Tey eo ory ai ltedit i, thle Cra tn - Patt prneeel eb Ane Mm Aisi inks SAT nti BEG abia hin) Mie wh pice PR TS a Pei then hs AHIR Me HARV Ane Hota dctyttatyte EW sneeh shes tat Haaren ones, ” Wey aah. a4 nie ee alit ie ont? COLLERLLCE TOT Ae nee en avn Wien seat y sabi cbat a ene 7 staat alot stss el - ne Prneesee AAP am ere Heine } Sin itt att vem Pitittetattets Sterne we iy ts eae ier Dies fen Sette ¥ siete ” ott Ay ft eee PR Neessects e pybeete 1 or ee) trp it +sesay bes vob tiagurowuenyt say eet) Toe eite Nw lle www 4] Venere AN TTL gg Wwe 7’ yg" eens vee * tie iE iw! Vy pwerreet ve “qn? wd’ whi a ee By Aitass sys et Oa NS . 1] Tt} Sonal ae 'y : Spree He Eh wy ] wey wes 4 gue tf were AA WY ad a | es Js. qT ivwavens 1 we j \ . be ; ml ian iA eaeeane —" Pwr re POCL OAC Sa ag ~ RS Nea Vv “eave Wi niet Nw, wd~ “s ‘Sy J VW lw. d, YWORA) AAD ne ATT Hep el matt ee S| nVNnty Tati itnd Tih . R rig ; if Mas "Yee verre sh OPOEeS 7 a6 ' gE f $ ; NN & ge We atl oat 5 hd ‘were Al Lies yy MT | é re ‘ Laon \ a w lie fr fe ee aor. en s® ype nh UT. TI | Pa TUNeg wid Teel yee ® 4 came we eee eT iy es ceeaeee ~ pe ; vee vier | 4a ye “we arene. VIET yg ATPL qa 2 wa “Ale ‘we Ad | | ree Oe met PAL ive’ pene | NAY YP a etPE SS Sel aee Pad] anee eoee) ‘eve TTT Ly aM! Tyas bt WWoevge” i 3 as Be aaee . ~ eth AM. wre wyow oe ) MNO ey Wives oe’ Be Re Re Bie NYRR SSA AAAL, \} |) rd tis Tall gIA OO v rN se tas ty : vo? ow wyveryrrs . ee *] x hey. vee 7 wrw * WAI ogy verTNtT ALEa AS. eee ihe me F) nen y mp 4¢ eerie, AM awe’ ; se hd Teaee af “lay vw * ¥ UV rs 5 Bee he) NN we t Pe a ced | tig “eaten : ai weyenere eA agt y attest sf UT sau TEU . Ma Se. vv" || Sal wh Tests A} celine We SO NEN add 7 Nneevttn tl ay e ay HAL A ; veer A ve “rr ba sbi! LD Ween wnat ae fot eee 4 = pwede TTL TT A betel Seg yih.t Se ' + lel 8 . 4? w LN Lhd 4 ayn! Wey 1 P yy ST egy oh AA b Fey aah Nee err tt ati re yeet diovelttyty anne ? Tersaterenie, “Sera WA Sur wore} Wes Seer nen ¥ 4 lhe. re -47~ a ee UBUOSAGT Pe * se ws Se seme ot) 4. : . . Y uy re | Owes Pt ah Mihi PRE a ¥ Pe |) me | at oh, hdr WOO SUveM dy ps Wiol ob li . - ™ ' . } A Tha by ome ett) be iM LPL whit ee Hed] tere oe ytlvr, a, | ee TM ¥ dL ahane . ese 8 inc eee Te muster aan ayn NTO’ ' ha grat ows TPT “wd Uv e wry jdt SemeRT oe a hee ieee dad nv ya af PT] ve 3.8 | a ey (Mr meseet lt JLab oe ry rurvannl ub VO PY! Adel l-t | were SacManduitatliosul TV Sin we ‘ 4 _ . a ¢ wwy Fwe © ter verse ne SweersF | wu 1a" : ay a Ae Vue ty i pore ye we NI hated win w “ae 5 vee ah Nui! Hy wa" Se SVE NA Be a¥ty Pete Ww wot if OTT hd Vuwruryy NW TMS conenelt a: pw ena® Le ee ey, ™ . vow | Saal Hy] y bel) Ny yprytt wit eee “ ows 77 Yee re ie Pp), PRA i * ISG Res wViek = Avery iuee”, } eA Y V . aud ury ur! ere. He ‘ Corks ww 5 Od SST AYE A eee wee I rz Be é ’ Saws oL, olthhd MORPaALTTLL a eee e Ps a4 Ree Y lig 7 ier! Dr pial” rome eee Petty inabeet il Vig Wwe ww EB JT atatie) 1 I "| ; Wiens, bee Oak lietatadaal tt PW, i i) \ Wony’ Ih Wwir* Vf cM eR EZ tel for ety oe OY JB gel Wren UAL Wee ETT DROP AAD eed hs Seite = Sa Vy, Y u ; TEES NA RdPOT oyodt “wwe tepay tot ¢ qautes \ AM | Was tt MINS we Vane * - See Fy Se woweve oe Ce Sy =. Bias b “ALA i, er. ev ~@ v I wy ww 4 vN ene i Ve ‘x bed 4 s AR i A | a ttiverce J — WY ere Titel cgi LOS) ( ey Wi Rw vive PAL ESTEE Ay SS 1. Vvwe PW Ap Hitt Eye PINs cece ty wUne Sb bi | lil “ed | || | Tit L POW VEN ee, ee ee mace SARIN MARY -anke wo” na hwant™ He vs My edt de a a § i Rasa a a Ty TTTLETE UH eer eran ; 4 OA Ch Neareeyve | i} TL eal HU aan HU TTI ith, Ag Lees WN MT ETT > ~tieuytt tyMydy ae: THT Ou abs t ee nt ~ a aa we 4 = ee = od - a Se wate Malle LE re oat ye Wg ae PML OF ae valwweiwn © eer Uaye siaeunn HEE tts, ~ virevtvws 4 WWWwweeee eu. Nyt yen wT ertthy SN yh uw IANS ee | | ’ PALL) pad ‘ NL Nay w%/ Wayye swiertt Te a % AS Ty J - wer | At | iq rh il ane ei ~ sie Tenet) LEA aM Wey ea yy a Aa eeceattn atatiet! TT) Alt) Lact) Seeman ennai loud 4 : lee SS en So ArH Z icy MOLT Tt ues oe A SA wv re wre SO ae wees -~ wr ad id ‘ pes. i ~ ns if RR /™ a] fi q 5 ~~) ~ ve Ae ee ee ww ame ee wire” pyre v Wy y J Priiaae Ch MT LT] i VY were S Pars al S.— sy | | rer ‘i uy! 7 J . Y 4 , ri) ry : ‘ y A ; P OR Eh my, a Ait ry . VJ A 5 vir~ 7 % awe vy ‘ a i) ee - ‘i > ij \ i " . ‘ \ , a y “hia o s! ‘ iy pala hig ©, EGA nm V5 1 ty >| DEPARTMENT OF MARINE BIOLOGY OF CARNEGIE INSTITUTION OF WASHINGTON ALFRED G. MAYER, DIRECTOR PAPERS FROM THE TORTUGAS LABORATORY CARNEGIE INSTITUTION OF WASHINGTON L23 VOLUME II WASHINGTON, 1D ies Se PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON 1908 9 / ez bu a Los : ; fl (o5 op DEPARTMENT OF ee rocy OF 0 io , CARNEGIE INSTITUTION OF WASHINGTON Lenk Sayer ~ } ALFRED G. MAYER, DIRECTOR | PAPERS, FROM THE TORTUGAS LABORATORY OF THE CARNEGIE INSTITUTION OF WASHINGTON VOLUME II 2 So 5 =) > aT ee Mationsl Hosea =e WASHINGTON, D. C. PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON 1908 FFR2 1979 LIBRARIES CARNEGIE INSTITUTION OF WASHINGTON PUBLICATION NO. 103 PR THE NEW ERA PRINTING COMPANY |. HABITS, REACTIONS, AND ASSOCIATIONS IN OCYPODA ARENARIA. By R. P. COWLES. Il. HABITS, REACTIONS, AND MATING INSTINCTS OF THE WALKING- STICK, APLOPUS MAYERI. By CHARLES R. STOCKARD. II]. STUDIES OF TISSUE GROWTH: 1. AN EXPERIMENTAL STUDY OF THE RATE OF REGENERATION IN CASSIOPEA XAMACHANA. By CHARLES R. STOCKARD. IV. SOME INTERNAL FACTORS CONCERNED WITH THE REGENERA- TION OF THE CHELAZ OF THE GULF-WEED CRAB. By CHARLES ZELENY. V. A CONTRIBUTION TO THE LIFE-HISTORIES OF THE BOOBY AND MAN-O'-WAR BIRD. By F. M. CHAPMAN. VI. THE HABITS AND EARLY DEVELOPMENT OF LINERGES MERCURIUS. By E. G. CONKLIN. VII. TWO PECULIAR ACTINIAN LARVAE FROM TORTUGAS, FLORIDA. By E. G. CONKLIN. VIII. THE BEHAVIOR OF NODDY AND SOOTY TERNS. By J. B. WATSON. IX. AN EXPERIMENTAL FIELD-STUDY OF WARNING COLORATION IN CORAL-REEF FISHES. By JACOB REIGHARD. CONTENTS. I. Habits, Reactions, and Associations in Ocypoda arenaria. By R. P. (CON HG Saas Gu ho RD OE o CODOCIOC RIOR COMO TIOO CE Rae Ree Te II. Habits, Reactions, and Mating Instincts of the “ Walking Stick,” Aplo- USHMAN Crim Dyn Ghiatlesniwens tOGKkanG acne cemin|-tacacas cioceciice eras III. An Experimental Study of the Rate of Regeneration in Cassiopea xama- chong Eva Chanles: Rs Stockard seria siecissiers eles ecisie «eres aie IV, Some Internal Factors concerned with the Regeneration of the Chel of the Gulf-weed Crab (Portunus sayi). By Charles Zeleny.... V. Contribution to the Life-histories of the Booby (Sula leucogastra) and Man-o’-war Bird (Fregata aquila). By Frank M. Chapman.. VI. The Habits and Early Development of Linerges mercurius. By Edwin (Cry © orate emer ey eeyete ai nievesose icra vcsekel alerhoke usisre hate siciave osteoma ayers VII. Two Peculiar Actinian Larve from Tortugas, Florida. By Edwin G. On Riri pee eer errr eestor els, Sve aysearatehaa ice ails cis neue eae VIII. The Behavior of Noddy and Sooty Terns. By John B. Watson....... IX. An Experimental Field-Study of Warning Coloration in Coral-reef iKishes) (Bye Jacobrdkeishanduerermtastiiees see sae ortnrenn cess 61-102 103-138 139-151 153-170 Pie PLATE 1. ERO Lin oe lies erga a ee x 7 Hae C. Kellner. A HOEN 4S CO BAUPIMORE A. SMALL OCYPODA KEPT IN THE DIFFUSE LIGHT OF THE LABORATORY. B. SAME SPECIMEN AFTER EXPOSURE TO DIRECT LIGHT FOR 10 MINUTES. ~ C. ADULT SPECIMEN—NATURAL SIZE. le HABITS, REACTIONS, AND ASSOCIATIONS IN OCYPODA ARENARIA. BY ka ba COWLES; Instructor in Biology in Johns Hopkins University. 4 plates and ro text figures. HABITS, REACTIONS, AND ASSOCIATIONS IN OCYPODA ARENARIA.* By R. P. Cow tes. One of the most interesting crustaceans that inhabits the Atlantic Coast of the more southern States is the brachyuran Ocypoda arenaria, the so- called “ sand-crab,” and no one who sees this lively creature can help marvel- ing at the rapidity and dexterity of its movements as it scampers over the beach sands. While the adult has not been reported north of New Jersey, Verrill (1874) tells us young specimens have been seen by Mr. S. I. Smith at Fire Island, Long Island, by. himself at Block Island, Rhode Island, and that the megalops larva has been taken in abundance by Mr. Vinal Edwards in Vineyard Sound. Verrill has suggested that ocypodas found in the Northern States are carried there from the South by the Gulf Stream while in the larval condition and that each winter they are killed off by the cold weather, so that they never grow large enough to breed. This supposition seems to be very plausible, since none but half-grown specimens are found in those regions and since breeding occurs in the South at just about the right time to make it possible for larvz to be carried up in the early spring. The adult has been reported from Cobb Island, Virginia, and it is very common along the sea-beach in the region of North Carolina. It flourishes in the Bahamas, on the sandy keys of Florida, and is found as far south as the coast of Brazil. During the summers of 1905 and 1906 it was my good fortune to spend a few weeks in the Marine Biological Laboratory of the Carnegie Institution of Washington at Loggerhead Key, Florida. On this key specimens of Ocypoda arenaria were very abundant, and owing to the small size of the island it was an easy matter to study them at all times of the day while only a short distance from the laboratory. A preliminary report of my observa- tions during the summer of 1905 has been published in the Year Book of the Carnegie Institution of Washington, No. 4, 190s. My purpose in studying Ocypoda was to learn as much as possible con- cerning its behavior and to determine how far it was able to form associa- *I wish to express my thanks to the Carnegie Institution of Washington and to Dr. A. G. Mayer, Director of the Marine Laboratory at Tortugas, for many courtesies extended to me. = Ro) 4 Papers from the Marine Biological Laboratory at Tortugas. tions. I found a quantitative investigation of this latter point very difficult, owing to the inhibition of normal behavior under artificial conditions, but several experiments showed conclusively that they formed associations. METHODS. Much care was taken, while observing the behavior of ocypodas and while experimenting with them, to have the conditions as favorable as possible. For the observation of their normal behavior in nature a white tent was used for a time to conceal the observer, but it was found that this was unnecessary and that they could be studied equally as well from the window of a well-ventilated building built out partly over the water some distance from the laboratory proper. From this window, with the aid of a good pair of bird-glasses, many interesting points in the “behavior could be distinctly seen without in any way disturbing the crabs. The best time of day for watching the ocypodas seemed to be from sun-up until about 11 a. m. and from 5 to 6 p.m. During the middle of the day the adults were usually down in their burrows, but there were often a few of the younger ones out at this time. Individuals were also observed during the night, when one might approach quite closely to them without interrupting their feeding. A simple trap for collecting ocypodas to be used in experiments was made in the following manner: A deep hole was dug in the sand above the high-tide mark and in it was placed a large dry-goods box without a cover. It was so placed that the open end was flush with the surface of the sand and boards several inches wide were nailed horizontally around the upper edge so as to prevent the crabs from climbing out. On one occasion during the first night 3 individuals were caught; on the next night the number was increased to 6; on the next to 12, and on the next to 24. While in this special case food and water were kept in the trap, it was found at other times that the trap was equally as effective when nothing of this sort was present. Whether the ocypodas dropped into the trap by accident or whether they were attracted there I am unable to say. BURROWS. As is well known, Ocypoda lives in burrows which it digs in the sand along the seashore. At Loggerhead Key the large majority of the burrows are found on the sloping beach all around the island, extending from the shore-line to 20 or 30 feet from it, but quite often stray ones occur in the interior of the island many feet from the water’s edge. During my stay at Loggerhead Key I found several ocypodas living close to the steps of the laboratory, probably for the purpose of picking up stray pieces of refuse which might be thrown from the doors of the station. Around the kitchen, which is situated over the water’s edge, the habitations of the crabs were Habits, Reactions, and Associations in Ocypoda arenaria. 5 very numerous, and at certain times of the day large numbers could be seen moving around near it in search of food. Ocypoda usually makes its burrows on the sloping beach; in some cases, however, they occur on perfectly level ground, and in others, where the beach has been washed away, they are found dug into the vertical surface of the shore. There are three kinds of burrows, all very simple. One consists of a tunnel which is not perfectly straight, extending down into the sand at an angle of about 45° with the surface. This tunnel opens to the exterior at one end, terminates blindly at the other, and has a more or less vertical passage branching off from it and sometimes communicating with the ex- terior by an opening (see fig. 1). The opening of the vertical passage is Fic. 1.—Longitudinal vertical section through a burrow with two openings. a, entrance; b, opening for escape. always farther away from the water-line than the main tunnel. As a rule these burrows extend directly away from the shore-line, but sometimes they do not exactly. The depth varies considerably, averaging 1 or 2 feet. An- other kind of burrow, which occurs higher up on the beach, is like those just mentioned except that it is much deeper and does not have any branch- ing passage. One of these, when carefully followed to its end, was found to extend 4 feet 2.5 inches, ending in coarse, wet sand. It occurred to me that these deep burrows might be used by the females during the breeding season and also that they might be used for molting purposes, but I have no observations to support the truth of these statements. As a matter of fact I have never found females with eggs at the bottom of their burrows, nor have I found casts. It is also true, however, that I have never found the casts on the surface of the ground. Very young ocypodas make still another kind of burrow which extends vertically downward into the wet sand for only a few inches. The burrow first described is much more numerous than any other kind and is more interesting on account of the branching passageway connected with it. I have examined many of these burrows in the effort to find the use of the branch. As stated above, sometimes the passage opens to the 6 Papers from the Marine Biological Laboratory at Tortugas. exterior and sometimes it does not. In the first case I have found it used only rarely for entrance and exit. As a rule the opening of the branch passage is much smaller than that of the burrow proper and usually there are no tracks leading from the opening of the branch passage, while the sand around the entrance to the burrow is covered with them. These facts seem to indicate that the branch passage is not made for habitual entrance and exit. It occurred to me that this second opening might be a means of venti- lating the burrow, thus making it cooler. There are absolutely no facts to support this theory, especially since it is found that the deep burrows, which would need ventilation if any did, were without these branching passages. Some light was thrown on the question when an ocypoda was observed digging its burrow. After it had dug for a certain distance, bringing up the sand at regular intervals, it remained down in the burrow for some little time and then suddenly there occurred a disturbance of the surface of sand in about the region where the opening to the second passageway should be. The sand began to sink in a little, but no opening was formed, and it re- mained in this condition until I dug up the burrow, when it was found that the crab had made a branch passage, starting from the burrow and working up almost to the surface without quite breaking through. Such a procedure was witnessed on more than one occasion, and each time the branch passage did not quite break through to the surface. After observing this it occurred to me that the passage might be used in time of need to escape from enemies, and I soon witnessed a sight which con- vinced me that such was the case. One ocypoda was in possession of a burrow with a branching passage, when another appeared at the opening of the burrow. The owner immediately went down and the intruder followed for a short distance and then returned to the opening, followed by the owner. This was repeated several times, the intruder going down farther each time, when suddenly the owner climbed up into the branching passage, broke through the surface and scampered away, leaving the intruder in possession. The two crabs were evidently contending for the possession of the burrow, and it seems probable that when the owner was unable to keep the intruder out it escaped by the only means possible in order to prevent injury to itself when driven to the end of the burrow. This same procedure was witnessed several times, but occasionally it was varied by the owner going around to the opening of the burrow and driving the intruder up through the branching passage instead of running away after its escape. Frequently this chase in through the burrow opening and out through the branched passage would be continued for some time, making a very ludicrous sight for the observer. The behavior just mentioned, together with the facts that the opening of the branching passage is usually not made at the same time the passage is, and that there are but few, if any, tracks around the opening, leads me to Habits, Reactions, and Associations in Ocypoda arenaria. 7 believe that it is used for escape by the ocypodas when they are hard pressed. It seems most remarkable that these crabs should often dig the escape passage out almost to the surface without breaking through, thus leaving the place of exit concealed. While the above description of the burrows applies to the large majority, there is considerable individual variation to be seen both in the construction of the burrow and the manner of doing the work. Some have arched roofs with an almost level floor, and others are almost round in cross-section. While, as has been described, the burrow usually points directly away from the water, there are some that extend obliquely, and in one or two cases bur- rows have been found which lead toward the water. Some individuals, especially the older ones, finish these burrows carefully, while others do not; some work quickly and others slowly; some start many burrows, but finish none; and many drive individuals out of the burrows they have made and take possession. As would be expected, Ocypoda usually digs its burrow in a place where the sand is firm and damp below the surface. It scratches out a shallow hole with its walking appendages, making it deeper and deeper, until it is unable to simply scratch the sand away, when it begins to carry it out in the following manner (plate 2, fig. A): Using the first and second walking appendages and chelz of one side, it rakes up quite a good-sized load, hold- ing it firmly between the appendages and the body proper. It then crawls out of the hole and deposits the sand, sometimes throwing it by a quick movement of the chele and walking appendages for a distance of almost a foot, sometimes carrying it away 2 feet or more and depositing it quietly, and occasionally dropping it immediately in front of the mouth of the burrow, As is well known, Ocypoda has one chela decidedly larger than the other, and in nearly every case in which an individual was observed making its burrow it went down into the tunnel sideways with the side having the small chela foremost, using that side in the digging and carrying of the sand (fig. 2). Out of 17 cases in which ocypodas were observed going down into their holes, after the burrow was constructed, the side with the small claw led the way 16 times. The small-chela side is undoubtedly better adapted for digging, which may account for this phenomenon. Further- more, under these conditions the crab is better protected from any enemy which may come down after it, since its large fighting chela is presented to the intruder. On some occasions I have noticed individuals lying on their backs in the tunnels digging away at the roof, and this is probably the manner in which the vertical passage is started. Many ocypodas exhibit considerable care in the construction of their burrows, especially the entrance to them. Sometimes, when the approach to the opening is quite steep, the crab will pile the sand that it brings from the 8 Papers from the Marine Biological Laboratory at Tortugas. burrow in front of the opening and then very carefully tamp it down, using the broad surface of its large chela and the distal segments of the walking legs. Sometimes, if there is a small hollow in the surface, it will drag sand from another place into the hollow and press it down. Often an ocypoda will carefully tamp the sand that it has thrown away, for a radius of over a foot, and frequently it will smooth off the edges of the opening into the burrow. Almost invariably, unless startled, before an individual goes down into a burrow, either while constructing it or after, it will halt for some little time, apparently surveying the region around to see if there are any in- truders about. A most interesting performance that I have frequently observed is the closing of the burrow. This usually occurs after the crab has stocked it with food. Upon several occasions I have thrown a number of small fish to an individual, each one of which was carried down and deposited at the blind end of the tunnel. Then the crab would gather up a load of sand from out- 2 Fic. 2,—Diagrammatic longitudinal section through burrow which an Ocypoda is plug- Fic. eave of plug from above, showing ambulatory appendages tamping surface of plug; mouth reduced to small hole. side the burrow, draw it into the entrance, pressing it firmly on one side of the opening (fig. 2). After this it would come out of the half-closed entrance, take up another load of sand, return and enter, drawing the sand after it. This load would be placed alongside of the other and would practically close up the entrance, leaving, however, the distal joints of several of the walking legs outside. With these the ocypoda would carefully tamp down the plug now formed and finally quickly draw them in, leaving some- times only a very small hole and sometimes none at all (fig. 3). By this means the entrance to the burrow was very effectually concealed. Sometimes, when individuals are disturbed by other crabs or by man, they COWLES PLATE A. ypod: trying sand from burrow en i a ony Sax VV Ui a eee Habits, Reactions, and Associations in Ocypoda arenaria. 9 will run into their burrows for a few inches and push a plug of sand up from below, completely closing the entrance. It seems probable that not more than one ocypoda at a time makes a burrow its home, unless during the breeding season, although I have seen as many as three emerge from a single hole. Most of the burrows are not permanent living-places, since they are usually made below high tide. In- dividuals often drive one another out of their burrows, and the successful one usually appropriates the new home that it has won; on several occasions three in succession have been seen to gain possession of the same burrow. BREEDING SEASON. The breeding season of Ocypoda was practically at an end when I visited Tortugas in June, but my own observations and those of others lead me to believe that it occurs probably in the spring and early summer rather than throughout the summer, as suggested in my preliminary report. During the late summer of 1906 I examined many female specimens, but only in a few cases did I find any eggs, and these were always few in number, suggesting the last of a brood. Mr. Kellner, who was at Tortugas as early as March, states that at that time large ocypodas were quite scarce, but that very little ones were quite numerous. Later, however, according to his observations, the adults began to appear again, so that in June they were present in con- siderable numbers. It seems very probable that farther north along the Atlantic Coast the breeding season is somewhat later. There is scarcely any difference in appearance between the male and female ocypodas, except in the shape of the abdominal segments (plate 3) and the abdominal appendages. The females, however, are usually clean- looking and less scarred than the males. Both have the stridulating ridge, of which I shall have something to say later (page 28). During the time that I was at Loggerhead Key the males seemed to be much more in evidence than the females. In nearly every case when burrows were dug up and individuals were found inside they were males. The traps also always showed a larger number of males than females. FOOD AND FEEDING. Ocypoda is often found feeding during the daytime, but more especially at night (plate 2, fig. B). It is also a cannibal, and to test this characteristic a small trap was stocked with many small individuals measuring from 1 to 2 em. across the carapace. During the night two larger individuals, about 4 cm. across the carapace, dropped into the trap and in the morning most of the small ones were found torn to pieces, with the soft parts eaten out. It is interesting to note that when a trap is set only large individuals are found as a rule in the morning. In a trap which had caught 30 ocypodas during several nights no small ones were found, and it seems probable that they 10 Papers from the Marine Biological Laboratory at Tortugas. do not enter on account of the presence of the adults. If they drop in by chance they are almost always killed and eaten. Among the many different kinds of food that Ocypoda will eat are cocoanut, sea-weed, bread, turtle- meat, fish, coffee-grounds, potatoes, ham fat, and jelly-fish. Bebe, in his article on “ Birds of Cobb Island,” states that the sand-crab eats bird’s eggs, and Verrill (1874) has found that it feeds upon Talochestia longicornis and T. megalophthalma. When the dead body of a fish which has been washed ashore is too large to be carried away by an ocypoda, it is apt to build a burrow by the side of the fish and feed on it day after day. Dr. Mayer has observed that when Physalia are driven up on the beach the same thing occurs, and then there is usually only one burrow to each Physalia. Ocypodas are great scavengers and keep the beach almost free from decomposing animal and vegetable matter, but there are a few things that they will not eat ; one of these is the lime. Pieces of this acid fruit thrown to them are seized immediately, but one “taste”? seems to be sufficient. They immediately begin to rub the lime in the sand, apparently trying to get rid of a substance that irritates the sensory organs of the mouth. There is no doubt that this crab reacts more strongly to certain kinds of foods than to others. While feeding ocypodas usually congregate along the drift-line of the beach, but as far as I have observed they never enter the water for the pur- pose of seizing food. During the summer large schools of Atherina laticeps (sardine of the Gulf) are almost always found close to the shore of Loggerhead Key and they are an important food for Ocypoda (plate 2, fig. B). These small fish were frequently used in feeding experiments. Feeding was observed both while individuals were in captivity and while they were under natural con- ditions. The eyes do not seem to play an important part in the detection of food, although they undoubtedly serve to lead the crabs to objects which may be food. When atherinas are thrown near an ocypoda (the observer being concealed), the “sand-crab” usually jumps and runs toward the former as though it “knew ” the fish were food, but, as we shall see from other ex- periments, this is probably not the case. As soon as the fish is reached the Ocypoda touches it with the claws and then immediately grabs it. As a rule the Atherina is carried at once into a burrow, but often (especially in case of the first few specimens) it is pinched with the claws, which are then rubbed against the first pair of maxilla situated at the mouth opening. Usually this process is repeated several times, alternating one claw with the other, until finally the crab has succeeded in introducing some of the juice into the mouth. Then it may hold the fish in one claw, using the other to tear off pieces and to transfer them to the mouth-parts, or it may grab the fish with either end directed toward the mouth, gnawing on it with the mouth-parts Frequently the Ocypoda begins by picking out the eyes of the Atherina and eating them. Habits, Reactions, and Associations in Ocypoda arenaria. II Whether one calls it taste or chemical sense, there is no doubt that Ocypoda is able to distinguish between foods. When several different kinds of food are presented to a number of individuals, certain kinds will be eaten in preference to others. A rather interesting experiment was one where the feeding was observed in the case of an individual from which the small chela had been cut off close to the body. This specimen seemed to have great difficulty in adjusting a fish so that it could be eaten. An attempt was made to use the first ambu- latory appendage on the side from which the chela had been removed, but the crab was not very successful in this, and after many trials it hit upon the following method: The fish was grasped about in the middle with the large chela in such a manner that the end was directed toward the mouth, and in this position feeding was accomplished with some difficulty. It would push the end of the atherina up against the mouth-parts, where the mandibles would take hold and then pull the fish away, thus tearing off pieces which could be eaten. This was repeated many times on different occasions. Occasionally the “sand-crab”’ did not seem to be able to adjust the fish according to this method, and then holding the atherina with its one chela tightly against the sand, it tipped the body downward until the mouth-parts were in contact with the fish, when it proceeded to feed as usual. ‘“OLFACTORY ORGANS.” Nagel (1894) and Bethe (1895, 1897) have studied the crustacean Car- cinus m@nas in the endeavor to find out in how far this crab makes use of the senses of smell and taste, or, as these have been called together, the “chemical sense.” Both of these investigators have concluded that Car- cinus in its search for food is aided by the chemical sense, and Bethe has gone so far as to say that the chemical stimuli are the principal ones that lead the crab to food, the eyes aiding only slightly or none at all. The experiments devised and the observations made, as recorded below, were for the purpose of determining if Ocypoda would react to foods at a distance through other senses than that of sight. The so-called “olfactory organs”’ of Ocypoda are situated on the an- tenn, which are very much reduced in size (plate 4, fig. c). They consist of the typical “ olfactory hairs,’ which are open at the distal end and which have a nerve running part way up the axis. These hairs are not as numer- ous as the feathered tactile hairs which occur in a large bunch on the seg- ment next to the basal segment of the antenna. Dr. Mayer observed that ocypodas dug up decomposing fish which had been placed about the roots of young pineapple-plants, but this does not prove that the odor attracted them, for it is quite probable that the juices or small particles of the fish were left on the surface of the sand and that the ocypodas in their wanderings and search for food happened to pass over these places. 12 Papers from the Marine Biological Laboratory at Tortugas. Several experiments were devised to determine if Ocypoda would react to odors, and although the results were not conclusive they were of some interest. Experiment 1.—FPieces of coral and small fish (Atherina) were dropped alternately near a crab. They ran for each and often carried the coral as well as the fish away with them. Precaution was not taken, however, to have the pieces of coral free from the odor of fish. Experiment 2.—Atherinas soaked in carbon disulphide were thrown to the crabs and these were at once seized in all cases and carried into the bur- rows. As far as could be observed this ill-smelling liquid had no effect on the ocypodas. Experiment 3—Poured aqua ammonia down a burrow containing a crab. This burrow had two openings and a strong odor of ammonia came up through the second opening, showing that it penetrated through the entire burrow. This did not cause the crab to come out of its burrow until 10 minutes at least had passed, when it came out unaffected, as far as could be observed. Experiment 4.—An atherina was wrapped in several thicknesses of news- paper, the ends being left open and care taken not to get fish juice on the newspaper. (The paper extended an inch beyond the fish at both ends.) This was then wrapped in Swiss book, tied at the middle, and dropped near the opening of a burrow. The crab soon came out, pinched it, and attempted to bite it at the open end. It seemed to react as it would toward an uncon- cealed fish, and it is not probable that the juices of the fish came in contact with the mouth-parts. Finally the package was taken down into the burrow, which was then closed up. The same experiment was repeated, substituting a roll of newspaper for the fish and wrapping with Swiss book as in the former case. This package was then dropped and was almost at once grabbed and pinched, but soon left as though there were nothing attractive about it. Experiment 5.—In this experiment a bowl-like hole was dug and athe- rinas were put in the bottom of it. They were then covered over with a layer of sand about an inch thick, so that they could not be seen, but so that the odor could be detected by the experimenter. A screen made of %4-inch- mesh wire was then put over the excavation, so that the crabs could not reach the fish without digging under. Great care was taken not to allow the juices of the fish to get on the screen or the sand surrounding it. At the expiration of 10 hours the ocypodas had not made an attempt to dig down to the fish. There were many tracks on the sand, however, showing that the crabs had been examining the region. A crab was observed crawling over the screen, pinching the wires of the same and then rubbing its chela over the mouth-parts, which behavior is typical of the feeding reaction. This was repeated by the same crab, and seems to afford strong indications that Ocypoda may react to food at a distance through some other sense than those of sight or contact (figs. 4 and 5). Experiment 6.—A young ocypoda which had not been fed for a day was used in this experiment. The eyes were painted with a mixture of shellac and lampblack, so that they did not react to shadows or to the movements of the hand near it. A fresh atherina was held within about 2 or 3 mm. from the mouth-parts for about 3 minutes. So far as could be seen the crab did Habits, Reactions, and Associations in Ocypoda arenaria. 13 not react. There was no movement of the antennz, such as Bethe (1895, 1897) has observed when a piece of meat is held near the mouth-parts, but when the fish was touched against one of the chelz the ocypoda immediately jumped at it, took hold, pinched it, and put its chela up against its mouth- parts, after which it began to eat. There is nothing in the behavior of the crab in this experiment which proves that Ocypoda reacts to odors. How- ever, the rapidity with which it determined that the object was food after the chela was touched leads me to believe that it had been stimulated by the odor of the fish before the chela was touched and that it inhibited the reac- tion on account of the abnormal conditions to which it was subjected. It is true that the fleshy consistency of the fish may have been a food- determining factor. Figures illustrating experiment 5. Fig. 4, ver- tical section; fig. 5, seen from above. 5 In the following experiments an attempt was made to find out from how great a distance Ocypoda would react to food. A square 9.5 inches by 9.5 inches, divided up into 64 smaller and equal squares, was marked out on the sloping beach in a region where it could be seen by the experimenter while he was practically concealed about 10 feet above the sand in a house nearby. Small pieces of decomposing fish about 1.5 c.cm. in volume, fragments of coral, and pieces of black glass were used to attract the crabs. Experiment A.—A piece of fish was placed at the center of the square, as shown in fig. 6. An ocypoda came to point A, and began to make a burrow. After a few digs it went straight for the center of the square (A,) and grabbed the fish with no hesitation. The wind was blowing in the direc- tion shown’ and the meat was on the surface of the ground. An important condition in this experiment, as we shall see in another section, was that flies were moving about on the meat and frequently flying to and away from it. Experiment B.—The conditions in this experiment were the same as in Experiment A and the results were the same except that the crab stopped at B, for 2 or 3 seconds while on its way to the meat. While the crab in these two experiments may have reacted to the odor of the fish, the eyes undoubtedly received a stimulus that determined the move- ment toward the food. Experiment C.—A piece of fish meat was placed in the center of the square. A crab 2.5 feet from the center (C,) went directly to the meat *In experiments A to J the wind was blowing across the square from the upper left-hand corner to the lower right-hand corner. 14 Papers from the Marine Biological Laboratory at Tortugas. (C,), returned to C, and went down into its burrow. A brisk wind was blowing directly away from the crab and there is no doubt that the latter did not react to the odor of the fish. Experiment D.—In this experiment a piece of black glass 2 cm. wide and 4 cm. long was stuck in the SEE at pean ne condueatethe center or the square. A crab at D,, fac- ing away from the center of the square, moved slowly but straight for the black glass. To my surprise it went backward all the way, striking the glass squarely in the middle of the pos- terior side of the carapace. Later it was found that this was not an unusual way for a crab to approach an object. After digging around the piece of glass and pinching it several times it was abandoned. Here odor could not have been a factor in the beha- Fic. 6—Diagram of paths taken by vior of the crab ROE could ocypodas in experiments A to F. the movement of flies, for there were none present. Experiment E.—A piece of white coral about 2 by 4 cm. was placed at the center of the square. A crab whose burrow was at E, came out, but showed very little inclination to examine the coral. It was almost 5 minutes before it started for the coral and then it moved slowly. As in the case of experiment D, flies were not present. Experiment F.—Several pieces of white coral smeared with the juices of decomposing fish were placed at the center of the square. The subject of the experiment in experiment E did not move toward the coral at once, but as flies began to alight the crab approached and attempted to get something to eat off of it. Soon, however, it returned to its burrow. The flies which had been driven off by the attack of the crab then came back and with their return the crab quickly approached the center of the square. Instead of leisurely taking hold and pinching the coral it gave a little jump in the seizure, as it does in the case of a live fish. The flies were driven away, after which the ocypoda returned to its burrow. This performance was repeated fifteen or twenty times, approaching, jumping, seizing and return- ing, after which the coral was abandoned, the crab apparently paying no more attention (objectively speaking) to it nor to the flies collected upon it. The ocypoda then began to dig its burrow, and when other individuals came and attempted to eat from the coral it did not chase them away, but merely stood guard at the entrance to its home. The eyes in this experiment played the most important part in the behavior of the crab, and there is no doubt that Ocypoda associates the pres- ence of flies, or rather the difference in the intensity of light resulting from the movement of the flies, with the presence of food. Its behavior can Habits, Reactions, and Associations in Ocypoda arenaria. 15 hardly be explained in any other way, and, as we shall see below, the move- ment of objects has an important influence on the behavior of Ocypoda. Experiment H.—A piece of decomposing fish was placed in a hole at the center of the square, the hole was filled up and the surface then smoothed off. The observer then left, but after 30 minutes returned and found that a crab had dug the fish out and carried it away. Considerable precaution was taken to prevent the surface sand from touching the fish or its juices, so that the indications are that the ocypoda found the food through the stimulus of odor, the eyes not aiding in its detection. Experiment I—In this experiment a piece of decomposing fish was fastened to the end of a stick, which was then stuck in the sand, so that the meat was 60 cm. from the ground. A crab ran up to the stick, pinched it, and tried to reach higher than usual with its chelz, but was unsuccessful in what appeared to be an attempt to get the fish. After this the crab returned to its burrow and did not go back again. Experiment J—The conditions were the same as those of Experiment I, except that the stick was pushed down into the sand until the fish was 4 cm. from the surface of the ground. Flies were thickly clustered about the food when a crab approached. This individual did not reach up and take the fish, but made an attack by jumping quickly at the food, striking it with its claws and jumping away, after which it moved off quite a distance. Then the ocypoda repeated this same behavior several times, until the flies were scared away, when finally it reached up, standing on the last two pairs of walking appendages, and began to eat. The stick was then raised again until the fish was 50 cm. from the sand, after which the same crab returned, but did not reach up. The fish was soon abandoned. Several times during the afternoon a small crab attempted to climb up, 7. ¢., put its claws against the side of the stick and raised itself, but was never successful in reaching the food. While it can hardly be claimed that any of the above experiments fur- nish conclusive proof that Ocypoda is stimulated by odor, experiments 5 and H point strongly in that direction. The behavior in most of the rest of the experiments may be explained by any one of the following hy- potheses: (1) That Ocypoda reacts to odor. (2) That opaque objects are distinguished as a result of the difference between the intensity of light on the object and the region surrounding. (3) That Ocypoda actually has vision. (4) That any two or all of these factors determine the behavior. As we shall see later, the author was unable to obtain any evidence that Ocypoda has vision. It might be held by those who do not admit that crus- tacea see with their eyes or react to odors, that Ocypoda approaches food and other objects merely by chance, but those who have observed this interesting crab can not for one moment believe this to be the case. ISNOESE Anyone who, for the first time, sees an ocypoda running over the white sand is forcibly impressed by the prominent and relatively large stalked eyes which, being almost black in color, form a striking contrast to the rest of 16 Papers from the Marine Biological Laboratory at Tortugas. the body. Who has not been tempted to say, when he walks along the beach and sees these crabs in the far distance scampering in haste toward their burrows, that they have a keen sense of vision? Thus far, however, we have no proof that crustacea have vision—that definite pictures, such as we know are formed in the eyes of the higher vertebrates, occur in the eyes of these invertebrates. The stalked eyes of Ocypoda are capable of considerable movement. They may be dropped laterally into grooves under the anterior edge of the carapace, where they are quite well protected. Besides this lateral move- ment I have noticed individuals move the eye-stalks backwards and for- wards while sitting perfectly still on a level surface. It is of interest to note that individuals when starting down their burrows (they always go down sideways) drop the foremost eye-stalk into the groove under the carapace, leaving the hindmost one erect. The advantage of this procedure can easily be seen. When the eyes are erect the angle of possible movement toward the sagittal plane of the body is very small, so that in going down the burrow, if the eye-stalk were kept erect there would be danger of its being forcibly bent or broken off by striking against the walls of the burrow. The hind- most eye-stalk, however, does not run this danger. Furthermore, it is not probable that the crab receives any very definite light stimuli through the eye on the side presented to the dark end of the burrow, while the other eye undoubtedly serves as a means of detecting any lessening of the intensity of light at the opening of the burrow which might be caused by the approach of an enemy. In Ocypoda the part of the eye sensitive to light extends over quite a large surface and covers much of the distal end of the stalk. It is not evenly distributed; the largest surface exposed is on the anterior side; the next largest is on the outer sides; the next on the posterior side d and the smallest is on the dorsal surface (fig. a 7). This distribution of the sensitive surface corresponds with the attitudes commonly as- sumed when acrab is watching. In the majority @ . of cases it presents either the entire surface or the side of the body toward the object in ques- tion; less frequently the posterior surface. The © 2 dorsal surface of the eye which has the smallest sensitive surface is no doubt used the least, since no enemies approach from above and since Ba erg 2, eee the food is on the ground. As we shall see posterior view ; ¢, outer side; below, this crab seldom reacts to the movements ” immer Side; ¢, dorsal side. of objects directly above it. z K ypoa Oc Habits, Reactions, and Associations in Ocypoda arenaria. 17 The so-called compensatory movements of the eye-stalks, such as have been described for other Crustacea, occur in Ocypoda when it is tilted either from right to left or anteriorly and posteriorly. The angle in the latter case, however, is quite small. It was noticed that when an individual was tilted for- ward or backward until the limit of movement was reached the eye-stalks were immediately lowered into their sockets. Another interesting observation was the following: When an ocypoda was picked up, the normal position with reference to the ground being maintained, it almost invariably dropped the eyes into the grooves, but when the crab was lowered again the eyes were raised as soon as the legs were allowed to touch the ground. In these experiments I kept hold of the crab even after it had been placed on the floor. Occasionally the eyes were raised in mid-air, not while moving up or down, however: The eyes are usually drawn down into the sockets when they are touched by the experimenter, and in the case of some individuals it was found that the same behavior was brought about when the crab was scratched with a pointed instrument along the middle of the carapace. There is considerable variation among different individuals with respect to the reac- tion from scratching, but in the same individual the result is quite constant. Ocypoda lives along wind-swept shores and the wind seems to have no ill effect on the eyes. Even in a strong gale the eyes are seen standing erect, apparently unaffected. An attempt was made to cause individuals to lower the eye-stalks by blowing suddenly and sharply upon them, but it was not successful. In the case of the land-crab (Gecarcinus) the eyes were dropped immediately when the same experiment was tried upon them. It was found, however, that when an ocypoda was brought close up to the mouth of the experimenter and warm air was gently breathed on the eyes they were immediately drawn into the sockets. They remained erect when the crab was brought close up to the mouth without breathing on them. There is no doubt that Ocypoda can distinguish a large object, such as a person, many yards off (at least 50 yards), as any one can testify who has seen these crabs run away on the approach of man. By this statement it is not meant that this crab has vision, that it sees things as human beings do, although I would not deny the possibility of imperfect pictures being formed on their eyes, but that its eyes are sensitive to changes in intensity of light and that it is able to see the outlines of objects where the contrast in the amount of light reflected by the object and that reflected by the sur- roundings is great enough. While Ocypoda can see objects which are not moving, as the experi- ments D and E prove and as frequent observations of individuals ap- proaching objects indicate, its behavior shows that the eyes are much more sensitive to objects in motion. The following interesting performance, which was witnessed from the window of a laboratory built partly on the beach and partly out in the water, shows how well Ocypoda reacts to moving objects. One of the most 2 18 Papers from the Marine Biological Laboratory at Tortugas. common fish around the shores of Loggerhead Key is the gray snapper, Lutianus griseus, which patrols the waters close to the bank in search of food. Besides this rather large fish there are also schools of a very much smaller fish mentioned above, Atherina laticefs. These two species ordi- narily occupy rather definite positions with reference to the shore line, the atherinas close in and the gray snapper farther out. The latter, however, prey upon the former and frequently make excursions shoreward, driving the atherinas into the surf and finally out of the water upon the beach, where they flop about helplessly. During the chase the gray snappers often flounder in the surf, making considerable noise. When this occurs, almost immediately ocypodas, if there are any out on the beach, run down and cap- ture the little fish, returning with them to their burrows. They travel fast and in a straight line to their prey, leaving no doubt in the mind of the ob- server that the eyes are stimulated by the movements of the atherina. I have frequently brought about this same reaction by throwing small fish from the window of the laboratory upon the sand below, in which case the behavior would be the same as above, showing that the noise produced by the gray snapper in the surf is not the only stimulus, if it is any at all, that attracts the crabs. It might be claimed that the ocypodas reacted to the odor of the fish, but this is probably not the case, since the olfactory sense is not very well developed and since, when pieces of coral are substituted for fish, the be- havior is the same up to a certain point. They will run directly to the coral, often pick it up, but seldom take it to the burrow. In order to thoroughly test the influence of odor in the throwing ex- periments the hands were thoroughly washed and ten clean chips of wood were taken in the left hand and ten fish in the right hand. These were then thrown alternately to the crab, with the result that the chips were run after as if they were food. They were always examined, but not taken away. To prove that the ocypodas did not react through the tactile sense as a result of the jarring of the sand when the objects were thrown upon it, another experiment was tried. An atherina was tied to a long thread and was thrown from the window in such a way that it would come within about ro cm. of the sand, but would not touch it. Immediately an indi- vidual would run directly toward the fish, but finding nothing on the ground would attempt to capture the shadow cast by the fish. Failing in this the ocypoda would usually remain still for some little time, apparently watch- ing, and would then make another jump at the shadow. After several trials, during which it would move away some distance and then return, the crab would go back to its burrow or begin feeding along the drift line. If then the bait was drawn up and after an interval of one or two minutes, was thrown again, the same behavior would be repeated by the original individual. Finally, if the experiment were repeated many times the crab would no longer react by running toward the shadow. Habits, Reactions, and Associations in Ocypoda arenaria. 19 The results of these experiments and observations leave little doubt that in Experiment F the movement of flies around the coral was a factor in determining the behavior of the crab. The eyes of Ocypoda are stimulated much more strongly by moving objects and probably also by still objects on days when the sun is shining brightly than when it is cloudy; also much more during the middle of the day than early in the morning or late at night. It was found that the ap- proach of man on cloudy days did not cause the crabs to retreat to their burrows as quickly as on days when the sun was shining brightly; also that they did not run as promptly after fish thrown to them early in the morn- ing or late in the afternoon as they did after those thrown during the brighter part of the day. During the night-time it was found that Ocypoda would easily approach a man lying on the sand and even crawl up upon him. The writer has had them enter his pockets and on several occasions has received their rather severe pinches. They show but little fear at night, when the contrast in intensity of light on objects is small, but in the day-time such behavior does not occur or at least very seldom. Bethe (1895, 1897) has observed that when a light is directed against the eye of Carcinus menas, or when a dark object is placed in front of the eye, the crab reacts by a movement of the first pair of antenne. Such a movement does not take place in the case of Ocypoda, but it was found that when a dish containing several individuals was brought from the direct sunlight outside of the laboratory into the subdued light inside, they reacted by a sudden jump the moment the shadow of the door was reached and that when the dish was taken out they reacted again in the same manner as soon as the edge of the shadow was reached. After the change had been made rapidly and many times the crabs failed to react, but after allowing them to rest for a few minutes and then repeating the experiment the indi- viduals began again to react. While the anterior, posterior, outer lateral, and inner lateral surfaces are sensitive to the differences in intensity of light, the dorsal surface does not seem to be, or at least only slightly so (fig. 7). The writer was able to stand during the middle of the day at the open window of the laboratory mentioned in a previous section, without disturbing the crabs below in the least; not even the most exaggerated motions, such as the swinging of the arms or the waving of large objects, would cause them to run to their bur- rows or even move away. This was not a case of inhibition, for the same movements performed on the sand by the experimenter at the same distance, or a much greater distance, always caused a run for shelter. Many attempts were made to study the behavior of Ocypoda after the eyes were cut off, but they would not live long enough after the operation to recover sufficiently from the shock. Much care was used in amputating the eyes, but in no case was it possible to keep the individuals alive for more 20 Papers from the Marine Biological Laboratory at Tortugas. than three days. The base of the eye-stalks is undoubtedly one of the most vital spots, and it is probably true that in nature the eyes are never pinched off without causing death. I have seen many hundred ocypodas both at Loggerhead Key and Beau- fort, North Carolina, and while I have often found specimens with one of the chele or one or two of the ambulatory appendages missing, I have never seen an individual with even one eye gone. Those ocypodas that had both eyes cut off did not react in any way to light or shadow cast upon them, and there is scarcely any doubt that the eyes are the only organs that are sensitive. Notwithstanding the severe shock resulting from the amputation of the eyes, individuals were often found that would be feeding at the end of 24 hours. The effect of painting the eyes with a mixture of lampblack, shellac, and chloroform was also tried. When several coatings of this mixture were put on, these organs were not sensitive to light. It was found that imme- diately after the painting individuals were quiet, but that very soon they became more active and assumed the defensive attitude (plate 1, fig. c). As Bethe (1895, 1897) has observed in the case of Carcinus, they were well able to protect themselves from normal individuals. On land they were much more sensitive to tactile stimuli than ordinarily, and in the aquarium they were much more sensitive to the vibrations of the water produced by striking the glass than normal ocypodas. The angle of compensation when a crab was tilted was somewhat less than the normal, confirming the observations of Clark (1896) and Prentiss (1901) for the fiddler-crab and those of Lyon (1899) for the crayfish. The blackening of the eyes seemed to have no effect on the maintenance of equilibrium and did not prevent the ocypodas from running about in a nor- mal manner when stimulated. The most evident change in the behavior under these abnormal condi- tions was the lack of reaction to a sudden increase or decrease in the inten- sity of light, the absence of any reaction to moving objects and the failure to approach objects as they ordinarily do when the eyes are not painted. Several experiments were tried to determine if the eye of Ocypoda re- acted to one color more than to another. For this 18 atherinas were used; 6 were stained red, 6 were stained blue, and the rest were left unstained. These were thrown alternately to an ocypoda, but there was no special difference in the behavior toward the different ones. An attempt was made to see if they would form an association between colors and food made distasteful in some way. Many experiments using stained atherinas soaked in the acid juice of the lime were tried, but these did not give any results indicating an association of color with distasteful food. Quite a number of experiments with red and violet color-filters made of celloidin were tried, but these yielded no definite results. Habits, Reactions, and Associations in Ocypoda arenaria. 21 In conclusion it may be said that the eyes of Ocypoda are the most highly developed of crustacean eyes; that they are stimulated by differences in the intensity of light when these are large enough; that they are quite sensitive in this respect; that they do not react to different colors; that they aid much in the search for food, in the detection of enemies, and in the accuracy of locomotion. My observations and experiments afford no proof that Ocypoda has vision, such as exists in the human eye, but its behavior leads me to believe the eyes are so well developed that it almost amounts to the same thing. While they probably do not see the color of an object or the finer characters of its surface, they undoubtedly see its outlines and possibly some of the more evident irregularities of the surface, made evident by the differences in lighting. COLOR-CHANGES. It has long been well-known that some species of Crustacea change in color when placed under different conditions, but these observations seem to be almost entirely confined to those forms that live in the water. Cer- tain species of Hippolyte, Palemon, Crangon, Idothea, Nika, Gammarus, and others have been studied, but as far as I know observations on the color changes of only one of the Brachyura have been made. Color changes or pigment migrations have been investigated in the verte- brate skin, especially that of lizards and frogs, in the hypodermis of crus- taceans, in the retina of vertebrates, crustaceans, insects, and cephalopods. Various causes have been suggested which might produce these changes, such as light, heat, color environment, emotional states, and other nervous conditions. The carapace of Ocypoda is very lightly colored and shows practically no color-pattern. Any pigment which it does contain apparently undergoes no change when exposed to different intensities of light and heat, but the hypo- dermis underneath is rich in dark pigment-cells. As the carapace is almost colorless, and is translucent, the pigment-cells of the hypodermis, arranged in the form of a pattern, show through very plainly under certain conditions. The writer had been observing ocypodas and experimenting with them for a considerable period before it was seen that the color-pattern under the carapace changed from time to time, although it was often noticed that some individuals had a definite pattern, while others were almost free from it. This lack of color in some specimens was supposed to be due to the fact that they had recently molted, until finally, while testing the effect of differ- ent color-screens on the behavior of Ocypoda, new light was thrown on the subject. In this experiment 4 ocypodas were placed in a box which had colored windows made of gelatin. After being confined for about an hour they were taken out and by chance placed in a dish of cool sea-water where, much to my surprise, they began to turn dark, showing a very distinct color- 22 Papers from the Marine Biological Laboratory at Tortugas. pattern. Three of these were then put in a glass dish and placed in the direct sunlight on the white sand. Ten minutes later these had lost nearly all their color, showing almost no markings. The other one was kept in the diffuse light of the laboratory and did not change in color. It had the same dark pattern the next day. These observations left no doubt that the color- pattern under the carapace was subject to considerable change when placed under different conditions of light, heat, or moisture, and experiments were then undertaken to determine what factors brought this change about. Experiment I.—Three small ocypodas, quite light in color, were taken from the trap and placed in the box with colored windows at 3> 15™ p. m. The box was kept in the bright light near a window until 45 50™ p. m., when it was found that the crabs had not changed in color. So far as these ex- periments were concerned there was no indication that the colored windows had anything to do with the formation of a color-pattern. However, it is not claimed that careful experiments with color-screens might not show that certain parts of the spectrum may be more effective in bringing about color changes than others. After the specimens were removed from the box they were put into sea-water and kept in the laboratory, where they began to turn dark at once. They were then removed to a dry dish, where they continued to grow darker. Experiment II.—Two rather dark ocypodas were exposed to the direct sunlight for 15 minutes and became very light in color. They were then put into a dish of sea-water and placed in a rather shaded part of the labora- tory (88 12" a. m.). A dark plaid pattern soon made its appearance, after which the crabs were put in the direct sunlight on the white sand still im- mersed in sea-water (88 22™a.m.). At 8" 30™ a. m. they were much lighter in color and at 8 49™ a. m. they were almost colorless. The specimens were then put in a photographic dark-room (8 50™ a. m.), still immersed in sea-water. To my surprise (believing at the time that bright light was the cause of blanching) at 10% 45™ a. m. they were still almost colorless. However, upon being put in a dry dish and placed in a shaded part of the laboratory, they became dark after a few minutes. These experiments show that the appearance and disappearance of the color-pattern is not a simple process, but that it may be brought about by one or several factors. Experiment III —In this experiment the same individual was used that was kept in the laboratory in Experiment I. It was dark in color. It was immersed in sea-water and exposed to the direct sunlight (8 a. m.). At 82 23™ a. m. the specimen had not changed in color to any appreciable de- gree and at 84 30™ a. m. it was possibly a little lighter. The crab was then placed in a dry dish (8 31™ a. m.) and left in the sunlight. As would be expected, at 88 35™ a. m. the specimen was still light. At 8" 4o™ a. m. it was put in a dish covered with a black tray and placed in the photographic dark-room. When examined at 104 45™ a. m. it was found to be still devoid of a color-pattern, as in the case of Experiment II. When the individual was put in sea-water and placed in a shaded part of the laboratory it became quite dark after a few minutes. Experiment IV.—A rectangular glass dish, 35 cm. in length, 25 cm. in OWLES PLATE 4 Mase Upper figure. Ocypoda kept in diffuse light of the laboratory. Central figure. Same individual exposed to direct sunlight for 10 minutes. Lower figure. Otocysts of Ocypoda. ie aK r S ror a Habits, Reactions, and Associations in Ocypoda arenaria. 23 width and 7 cm. in depth, was divided into two equal parts by a partition in the middle. One of these parts was lined and also covered with black glass. The other part was left without lining except that the side of the partition was covered with white bristol-board. The cover was made of ordinary clear white glass. Sea-water was poured into the dish until it was 2.5 cm. deep and then two crabs with dark plaid patterns were put in—the darker of the two in the light side, the other in the dark side (8° 35™ a. m.). The dish was then placed in the direct sunlight until 97 47™ a. m. Both were equally light in color. (Temperature of water 45° C. at end of experiment.) The results indicated that heat was the factor which brought about the loss of color, so the following experiment was tried: Experiment V.—Two dark specimens of practically the same shade were used. One was placed in a dish of water kept at a temperature of 45° C. (temperature of the water in Experiment IV after exposure to the sun), the other in water whose temperature was 23° C., and both were put in a shaded part of the laboratory ; they were left for 15 minutes, after which the former became decidedly lighter, while the latter did not change. Experiment VI—The same experiment was repeated, except that the temperature of one was kept at 44° C. and the other at 23° C. The result was the same. The last three experiments undoubtedly show that a temperature as high as that on the sands of Loggerhead Key will cause a loss of color when the specimen is not exposed to bright sunlight. These results received further confirmation in the following experiment: Experiment VII—Two dark-colored ocypodas were used. At 2" 35™ p. m. one was put in a dry dish over a sand-bath, kept at 36° C., in a shaded part of the laboratory, and the other was placed under the same conditions, except that it was not heated. (Temperature of air 24° C.) At 3 p. m. the former had almost lost its color-pattern, while the latter had not changed. Experiment VIII—A light-colored ocypoda (1) which had its eyes painted with shellac and lampblack (it did not react to shadows) was put in a dish of sea-water and placed in a shaded part of the laboratory at 75 40" a.m. Another light-colored one (2), whose eyes were painted (it was not quite as light as the other), was put in the direct sunlight without water (78 30™ a. m.). At 7* 50™ a. m. (2) was brought into the laboratory, put in a dish of sea-water, and placed in the shaded part of the laboratory. Crab (1) was taken out of the sea-water, put in a dry dish, and then in the shaded part of the laboratory (74 56™ a. m.). Finally (2) was taken out of the water and left in the shaded part of the laboratory (8* 06™ a. m.). Throughout these changes neither (1) nor (2) changed in color, and at 8" 21™ a. m. they were in the same condition. This experiment affords quite conclusive proof that the eye of Ocypoda must receive light stimuli in order to bring about a distal migration of pigment under the carapace, i. e., in order for the color-pattern to appear. Experiment [X.—In this experiment two ocypodas whose eyes had been painted 24 hours before were used. (They did not react to light.) Both of the specimens were light in color, although kept in the shaded part of the laboratory (6 a.m. Temperature 23° C.). The paint was removed from the eyes of one of the individuals and in less than half an hour it became 24 Papers from the Marine Biological Laboratory at Tortugas. darker. At 75 45™ a. m. it was much darker, showing a distinct plaid pattern, while the other ocypoda was still as light in color as at the beginning of the experiment. These results I consider as evidence that the stimulus given to the eye by light is an important factor in bringing out the color- pattern, since the temperature was practically constant throughout the experi- ment. It might be claimed that disturbances in the nervous condition of the specimen affected the results, but in the light of other observations this does not seem probable. Experiment X.—A specimen with a very dark color-pattern was used and an attempt was made to bring about a color change by frightening it. The crab was stimulated several minutes by moving the hand in front of it. This caused the specimen to run about very vigorously, but there was no change in color. This method of stimulating the crab was continued until it showed signs of exhaustion, but still there was no change in color. When it was subjected to a strong electric shock two of the ambulatory appendages were broken off, but even this had no effect on the color-pattern. Finally the ocypoda was etherized for 2 minutes until it was perfectly quiet, and then allowed to revive, but this brought about no change in color pattern. The results obtained in Experiments IV, V, VI, VII show conclusively that the high temperature on the surface of the sand, which is a result of the exposure to the direct sunlight, brings about a proximal migration of the dark pigment under the carapace of Ocypoda (plate 4). Accordingly the following two experiments were tried, in order to determine if the direct sunlight would cause a blanching of the carapace when the temperature was kept comparatively low. Experiment XI.—An ocypoda which showed a dark color-pattern and which had been kept in the diffuse light of the laboratory (temperature of air 33° C.) was used. It was placed in a dish filled with sea-water and covered with a glass plate. The outside of the dish was then covered with a piece of heavy cloth saturated with alcohol (unfortunately no ice was at hand) and the whole thing was put in the direct sunlight on the sand at 2h 0o4™ p.m. The evaporation of the alcohol kept the temperature of the water at 35° C. At 22 35™ p. m. the specimen was still almost as dark as at first. The slight loss of color was probably due to the 2° increase in temperature. Experiment XII—Using the same individual, the above experiment was repeated after removing the cooling cloth (25 40™ p. m.). At 25 56™ p. m. the temperature of the water was 45° C. and the specimen had become very light. Experiment XIII.—The ocypoda used in Experiments XI and XII was brought into the laboratory and placed in a shaded place (25 56™ p. m.; temperature of air at 33° C.). At 3" 11™ p. m. it had regained its original dark color. Experiment XIV.—Experiment XI was repeated under the same con- ditions, except that the specimen was put in the darkness of the photographic dark-room. Although the temperature was kept at 35° C. by the above cooling device, the crab soon lost its color-pattern and dark color, a change undoubtedly due to the absence of a light stimulus on the eyes. Often during the above experiments individuals were observed to grow nan Habits, Reactions, and Associations in Ocypoda arenaria. 2 perceptibly darker after removal from water into a dry dish. This color change was undoubtedly due to a decrease in temperature resulting from the evaporation of water from the surface of the specimen. Many observations were made on ocypodas living under natural condi- tions. During the middle of the day, when the sun was shining brightly, specimens seen feeding on the sand were usually light in color, but occasion- ally dark individuals were found. These latter had probably just emerged from their cool, shady burrows (the entrance not closed). During the late hours of the afternoon and on cloudy days, however, most of the ocypodas feeding along the beach were dark in color, although occasionally a very light one would be seen. Probably the latter had been buried in a closed burrow which was dark. The following interesting observation of the effect of light on the pig- ment-cells of the eye was made: A small ocypoda was kept in a box, the bottom of which was filled with sand to a depth of 10 cm. There was a crack left between the bottom of the box and one of its sides, into which the direct sunlight shone. The specimen had burrowed to the bottom of the box in such a way as to have one eye exposed to the direct sunlight while the rest of the body, including the other eye, was in comparative darkness. When this individual was taken out the eye which had been exposed was almost devoid of black pigment, while the other one was black as usual. The color-pattern under the carapace, however, was neither very dark nor very light. After 15 minutes in the diffuse light of the laboratory both eyes were as black as usual. While no experiments were tried to determine the effects of light and heat on the pigment migration in the retina of the eye, it is probably true that the blanching in this case was due to the heat of the direct sunlight and not to the light itself. The medium dark coloration of the hypodermis under the carapace may be explained by the facts that the eye was stimulated by light and that the rest of the body was exposed to a medium temperature, two conditions which, according to my experiments, would result in a medium dark coloration. DISCUSSION OF RESULTS. Brooks and Herrick (1889) working on Alpheus and Palemonetes, P. Mayer (1879) on Jdothea, and Herdman (1894) on Virbius varians find that the stimulation of the eyes by light is an important factor in color change. Malard (1892) tells us that in the case of Hippolyte varians the intensity of the light has an effect on the color, and that in many crustacea the stimu- lation of the eye by light and the color environment are factors. In Jdotea tricuspidata Matzdorff (1882) finds that variations in temperature, light, and density of water do not cause color change, but claims that the latter is a “sympatische Wechselfarbung” which varies with the surroundings. Pouchet (1872), in his work on Palemon and Leander serrator, lays special 26 Papers from the Marine Biological Laboratory at Tortugas. stress on the stimulation of the eye by light, but also finds that toxic sub- stances and to some extent electric shocks may bring about color change. In a later paper (1876) he discusses in particular the influence of the back- ground on color change. According to Fritz Miller (1883) alarm causes the color of the male of a Brazilian species of Gelasimus to change, and this is the only observation I know of on any brachyuran. Temperature does not seem to be a factor in color change according to the above investigators, but Jourdain (1878) working on Nika edulis finds that heat affects the color. Not only has he observed that low temperature causes a blanching, but he also tells us that stimulation of the eye by light and internal stimuli are factors in color change. The only other observations on crustacea in which temperature is said to play a part in the color changes are those of Gamble and Keeble (1900). These investigators find that the color of Hippolyte varians is affected by ether or the recovery from its effect, by electrical shocks, by cold, and by light stimuli through the eyes. In Anolis, according to Carlton (1903), the green state is brought about by darkness, withdrawal of circulation, and possibly the cutting of nerves. Specimens in narcosis from ether, those treated with nicotine, and those which are dead, are green in color. Carlton believes that the green state represents the resting condition of the melanophores and the state to which they return when they no longer receive stimuli. This green state seems to correspond to the blanched condition of Ocypoda, being brought about by some of the same factors. The crustacean literature does not seem to afford any extensive investi- gations of color changes in land-crabs such as have been made in the case of the lizard, although Gamble and Keeble (1900, 1904) have published two papers which promise to be the beginning of a very thorough study. It is of considerable interest, I think, to find that the results of my work on Ocypoda—which lives under much the same conditions as Anolis caro- linensis, worked on by Parker and Starratt (1904), and especially Phryno- soma blainvillei, studied by Parker (1906)—agree in many respects with the results obtained by these two investigators. The experiments so far performed yield no results which would indi- cate that color environment or nervous condition are responsible for changes in color observed through the carapace of Ocypoda; nor does the immersion in water seem to be a factor, except in so far as the temperature of the water is concerned. Changes in the intensity of light and variations in tempera- ture are undoubtedly the main stimuli, if not the only ones, which bring about the proximal migration of pigment resulting in a dark color-pattern or the distal migration resulting in a disappearance of the color-pattern. In the absence of light when the temperature is anywhere between 22° C. and 45° C., and undoubtedly when it is lower or even higher, a light coloration occurs. Generally in diffuse light, and even direct sunlight, a dark coloration ap- pears, provided the temperature is not too high. Habits, Reactions, and Associations in Ocypoda arenaria. 247 Usually at low temperatures, not above 35° C., a dark coloration occurs, provided the eye is stimulated by light. At high temperatures, above 35° C., a light coloration is the rule, and it occurs independently of the intensity of light. The temperature limits given above are those obtained from a study during the hot summer months, but it may be that in winter, when the average temperature is much lower, the limits would be different. These results may be stated in another way. Dark coloration occurs at comparatively low temperature in diffuse or direct sunlight. Light coloration occurs at comparatively high temperatures when light is absent. Light coloration occurs at comparatively low temperatures and at medium temperatures when light is absent. Light coloration occurs at comparatively high temperatures in diffuse or direct sunlight. The blanching of individuals on the sands of Loggerhead Key is prob- ably due to the high temperature alone. (See plate 1.) The results of Experiment XI indicate that it is not a reversed light reaction. On the other hand, the blanching in the photographic dark-room may be due to the absence of light alone. The dark coloration of individuals occurs only when the eye is stimulated by light, and then only when the temperature is comparatively low. The indications are that it is not the result of the direct action of light on the pigment-cells or on the nervous system. EQUILIBRATING ORGANS—* AUDITORY ORGANS.” For many years an auditory function was ascribed to certain organs found in the basal joint of the antennules in decapods, but since the work of Bethe (1895, 1897), Beer (18098, 1899) and Prentiss (1901) we have strong evidence for the assumption that the so-called “auditory organs” do not have the auditory function, but that they are organs which are important in the maintenance of equilibrium. So far as I know we have not as yet any good evidence that the Crustacea hear, and in those cases which have been recorded the supposed reaction to sound was probably due to tactile stimuli. Although Prentiss (1901) stated in his paper on “ The Otocyst of Deca- pod Crustacea ” that Gelasimus pugilator, a brachyuran decapod, living much of the time on land, did not react to sound-waves, I was much interested to see if such an active and highly developed land-crab as Ocypoda was able to hear. The “auditory organs,” or otocysts, as I shall call them, are situated in the basal joint of the antennules, as in nearly all the decapods (plate 4), and they are partly protected by the rather broad but short rostrum. 28 Papers from the Marine Biological Laboratory at Tortugas. In order to determine if Ocypoda would react to sound-waves the follow- ing experiments were tried: A flute was played upon and large stones were hit together while the observer was hidden 12 feet away, but no movement could be seen that would indicate that the ocypodas were stimulated. The report of a pistol produced no apparent effect. During a heavy storm the peals of thunder were deafening, but several ocypodas which were along the beach did not return to their burrows nor cease their feeding. These ex- periments and observations indicate that Ocypoda does not hear, but they are by no means conclusive, and a series of such careful experiments as those performed by Fielde and Parker (1904) on ants and Yerkes (1905) on the green frog (Rana clamitans) would well bear repeating before it should be said that Ocypoda does not react to atmospheric sounds. Although it does not seem probable, it might then be found that notes causing a very small number or a very large number of vibrations of the air per second had an effect. Ocypoda while eating produces a grinding sound by the movements of its mouth-parts, and at times it also makes a noise resembling the “ peep” of young birds, but I have no proof that these sounds stimulate the otocysts of other crabs. Prentiss (1901), in his review of the crustacean literature (p. 228), has cited two well-known examples of noises, the one a case of stridulation in Palinurus vulgaris, described first by Mobius (1867) and later (more cor- rectly probably) by T. J. Parker (1878), and the other the pistol-like report of Alpheus described by Goode (1878). To these may be added the stridu- lation of Ocypoda, although I have never been fortunate enough to hear it. Often at night I have stayed on the beach among the feeding ocypodas in the hope of hearing the stridulation, and I have listened at the en- trance of the burrows which were occupied by individuals. I have good reason, how- ever, to believe that the stridulating noise is made, since I have often seen ocypodas go Fic. Se > boone ee during through the motions that would produce it : ; when they were attempting to gain possession of a burrow occupied by another specimen. Along the inner surface of the palm of the cheliped in the individuals there is a row of fine tubercles(stridulating ridge, plate 3), which when the cheliped is bent, as shown in text-fig. 8, comes in contact with a process on the basal joint. By moving the distal part of the appendage from side to side the row of tubercles is rubbed against the process and a sound is produced. I have seen this movement frequently in the case of individuals “Mentioned first by Leach in Malacostraca Podophthalmata Britanniz, 1815. Habits, Reactions, and Associations in Ocypoda arenaria. 29 that seemed to be excited, and it was performed with great rapidity. There is scarcely any doubt that the sound is produced, but I have no reason to believe that other ocypodas hear it. However, the vibrations produced by the rubbing might easily be transmitted to the sand and thence to the sensi- tive tactile hairs on the ambulatory appendages of another individual, thus producing a tactile stimulus. In Mier’s (1876) paper it is stated that in one species of Ocypoda the stridulating ridge is absent in the very young individuals. I have examined many specimens of O. arenaria, the smallest 1 cm. across the carapace, but have always found the ridge present. It is of interest to note that the land- crab, Gecarcinus, which is found in abundance on Loggerhead Key and which lives under very similar conditions, has no stridulating ridge. As has been pointed out by Prentiss (1901), it does not follow that an animal is able to hear because it makes a noise, and as yet we have no ob- servation or experiments proving that Ocypoda or any other crustacean has audition. Many attempts were made to repeat the experiments performed by Beer (1898) and Prentiss (1901), in which the behavior was observed after the removal of one or both of the otocysts. In every case in which both of these organs were removed the individuals did not live long enough to recover from the shock. As a general rule, even the removal of one otocyst caused such profuse bleeding that the specimen soon died. The behavior of oper- ated individuals, however, was very similar to that observed by Prentiss (1901). Those ocypodas which had their eyes painted, so that there was no reaction to sudden changes in the intensity of light, retained their equilibrium as well as normal individuals. The only difference in behavior was their tendency to remain quiet. While the eyes undoubtedly assist in maintaining the equilibrium, they are not the most important organs in this respect. The removal of one otocyst brought about some disturbance in equi- librium, part of which may have been due to the shock, but the removal of both otocysts caused very marked effects; individuals when placed on the sand usually turned one or more somersaults. While these disturbances may have been partly due to the operation, they were undoubtedly largely the effect of the loss of organs that regulate the equilibrium. TACTIEE SENSE: While Ocypoda probably does not hear, the tactile sense is well devel- oped and is a very important factor in regulating this animal’s activities. As would be expected, Ocypoda is quite sensitive to jars transmitted to the sand. I have often watched individuals which were part way down in their burrows, but not so far that the legs of one side might not be seen. Under these conditions, a movement of my foot in the sand or a step within 3 or 4 feet caused the specimen to react by a sudden movement, even when 30 Papers from the Marine Biological Laboratory at Tortugas. I was hidden from view. Often while observing ocypodas in the trap I have found that they did not move when a stick was waved in front of them, but as soon as it hit the side of the trap they would react either by a quick jump or a movement of the eye-stalks. However, there was con- siderable difference in the behavior of individuals, some being quite sensitive and others not so much so. Ocy'poda also reacts to vibrations of the water in aquaria, but it is not nearly so sensitive to these stimuli as to those caused by jarring the sand. LOCOMOTION. The locomotion of the Decapoda has been investigated by List (1897) and Bethe (1897), but both of these workers have confined their studies to those forms which live most of the time in water and which are not adapted to very active locomotion on land. The Ocypode, however, spend a large part of their lives on the beach sands and can travel for long dis- tances at a considerable rate of speed. In fact, their movements are so rapid that it is often impossible to determine the order in which the various ambulatory appendages are used. Bethe (1897) has observed that Carcinus menas ordinarily travels side- ways, but that to a very limited extent it can move forward and backward. According to this investigator the locomotion sideways is directly sideways, and not oblique as List (1897) has described for decapods in general. As in the case of Carcinus, ordinarily Ocypoda runs sideways and in a direction at right angles to the sagittal plane of the body. It seems to be much better adapted to locomotion in other directions, however, than any of the Brachyura. Unlike Carcinus, it will move obliquely sideways with considerable speed and will travel forward for long distances at a good rate. As mentioned in Experiment D, Ocypoda can move backward for quite a distance, but it does so slowly, and when a quick movement is re- quired in that direction it usually jumps. On the whole it seems to be better adapted to life on land, so far as locomotion is concerned, than any other decapod. Ocypoda usually runs toward an object with a considerable degree of accuracy. It often returns to its burrow almost directly in a straight line when disturbed by the approach of man, and I have often seen them ac- complish this from a distance of 15 or 20 feet. This same accuracy is shown when they approach objects which have been thrown on the sand, although it is true that sometimes they fail to go more than two-thirds of the distance. So far as I have observed, individuals very seldom run beyond the mark. Bethe (1897) on page 510 says: “ Mann sicht daraus klar dass von einem ‘ Sehen’ in unserem Sinne, von einer Perception der Lage und Ent- fernung des Gegenstandes, nicht die Rede sein kann.” Bethe’s statement Habits, Reactions, and Associations in Ocypoda arenaria. 31 that Carcinus does not see in the way we do is undoubtedly true also of Ocypoda, but I believe that the latter has, through the use of the eyes, a sense of position and distance; else how could it run so quickly and often so accurately after objects thrown on the sand several feet away? Not only is Ocypoda quite accurate in returning to its burrow after moving away from it in a straight line, but it also returns accurately after running around in different directions on the beach. However, individuals differ much in this respect. The following observation showed very nicely how an ocypoda could return directly to its burrow even after feeding along the beach and showed also that it could return when the dark opening of the burrow was hidden. This crab had made its home so that the open- ing was in a hollow formed by the heel of a shoe and on this account the entrance was concealed (fig. 9). Starting from A, the specimen made its way leisurely to B (12 feet), feeding on the way; from B, it slowly went to C (10 feet) continuing its feeding; then the approach of a man caused the crab to run for its burrow (about 13 feet away), and it did so directly and accurate- ly even though, as stated above, the en- trance was hidden. Similar exhibitions of Vase) = €@ accuracy, when, how- : < ever, the opening of the burrow was not hidden from _ view, Cc were frequently ob- B served. The young individuals are much eit one tee eae WATERLINE less accurate in this SS —S— respect, and while an observer is walking along the beach many of the small ones will be seen running about aimlessly as though lost, while the adults nearly always reach their burrows easily. While Ocypoda undoubtedly depends to a considerable extent on its eyes during movement toward a definite object or place, as is indicated by the inactivity of individuals with painted eyes, there are other factors which are almost equally important. The care which many of the older indi- viduals take in preparing the surface of the sand immediately in front of the entrance to their burrows leads one to believe that they are sensitive to the contour of surfaces. Such a sense of the contour of surfaces has already Fic. 9.—Illustrating accuracy of return to burrow when entrance was hidden from view. 32 Papers from the Marine Biological Laboratory at Tortugas. been ascribed to Pagurus by Bohn (1903). Ocypodas are often seen care- fully tamping down the mound in front of their burrows by means of the flat outer side of the large chela, making it round and smooth. Occasionally when there is a small hollow in the surface they will bring sand and fill it up. Sometimes I have seen individuals, which had apparently lost their bearings in the hasty return to their burrows, behave as though they were testing the surface of little mounds which were not situated in front of burrows. Several factors contribute in bringing about the accurate return of these crabs to their homes. They are undoubtedly guided by differences in the lighting of surfaces, by tactile stimuli, by differences in muscular effort, and by stimulation of the equilibrating organs resulting from a tilting of the body. BEHAVIOR TOWARD WATER. While Ocypoda lives most of the time on land, it is absolutely necessary for it to go into the water occasionally. I am unable to say how long it stays in the ocean during the breeding season, but at other times, so far as my observations go, it remains there only a short time unless forced to do so by some enemy that prevents its return. Adult individuals may spend hours in their burrows without going near the sea, but the young ones seem to be much more dependent on a fresh supply of water. This may account for the fact that their burrows are built closer to the water-line than those of the adults. When an ocypoda is startled or disturbed while far away from its bur- row it is very apt to run into the water, but even then it probably never goes out more than 4 or 5 feet from the shore. Ocypoda does not swim, but crawls along the bottom and is washed back and forth by the surf. When undisturbed, Ocypoda goes down to the ocean now and then in order to moisten its gills with fresh sea-water; but at these times the indi- viduals do not enter the water; they settle down about 6 or 8 inches from the water-line formed by medium-sized waves, with the ambulatory ap- pendages of one side presented to the ocean and those of the other side firmly embedded in the sand. In this position they wait until an extra high wave washes over them and then return to the higher parts of the beach. Sometimes, after remaining in a place for a considerable period without being wetted by a wave, the crab will change its position to one closer to the water. While Ocypoda does not live in the ocean much of the time, it is able to do so, as is shown by the following experiments: A large adult male was put in an aquarium filled with sea-water and left there for 6 hours. It was apparently in good condition when liberated at the end of this time. An- other specimen left in the aquarium for 24 hours was also active when released. Other specimens placed in fresh water lived only for 5 hours and made frantic attempts to “escape. Habits, Reactions, and Associations in Ocypoda arenaria. 53} Ocypoda is not able to live very long without water, although it may stay down in its damp, cool burrow for several hours. When in the direct sunlight without water it lives but a short time. Of 10 specimens placed in the sun in a wooden tub all but 3 had died after 4 hours’ exposure. Other specimens kept in a dry aquarium in the laboratory and not exposed to the direct sunlight lived almost 24 hours. Although Ocypoda can live probably not over 24 hours in a dry place, it remains alive much longer in damp sand. Several specimens placed in an aquarium and buried to a depth of 15 inches in damp sand were alive and active at the end of 48 hours. ENEMIES, DEFENSE, HIDING, SLEEPING, ETC. On Loggerhead Key, Ocypoda does not have many enemies. The gray snapper and man seem to be the only large animals that molest them, but they are undoubtedly troubled by parasites and sometimes they are killed by their own kind. The writer has often seen gray snappers darting about close to the beach when an ocypoda has been driven into the water. As a rule the latter keeps just out of reach, but occasionally a young one will be snapped up by the fish. The large adults defend themselves fairly well when dropped into deep water among a number of snappers. Upon the ap- proach of the fish the crab strikes out with its large chela in the same way that it does on land and this usually drives the gray snapper away. Eventu- ally, however, if it does not crawl into shallow water its ambulatory ap- pendages are bitten off one by one and the body is torn to pieces. Such observations can be easily made at Loggerhead Key, because the water is very clear. Bethe (1897) has described what he calls the ‘‘ Aufbaum Reflex” of Carcinus menas. This attitude, he tells us, is brought about by stroking along the back or head and also by moving an object in front of the crab. In the case of Ocypoda a similar attitude is assumed and it is undoubtedly one of defense. The crab rises up on the distal segments of its walking legs, the chelee are raised and spread for apart and the body proper is held well up above the surface of the sand. This defensive attitude is shown in plate 1, fig. c. As a rule, when Ocypoda is disturbed by man it either runs into its bur- row or, when hard-pressed, goes into the water; but this interesting crab has another way of concealing itself from man. Occasionally an individual is seen which, instead of trying to find a burrow or attempting to run into the water, will settle down in some little hollow, push its body backward into the sand, and then with its posterior ambulatory appendages throw sand over itself until most of the body is covered. Usually the eyes remain per- pendicular, but sometimes they are dropped into their sockets. Another condition in which Ocypoda is occasionally found is the resting or “sleeping” condition. I have seen a few individuals in this state during 3 34 Papers from the Marine Biological Laboratory at Tortugas. my walks along the beach. The body always rests on the sand, the legs are relaxed and stretched out, and the eyes aredown in their sockets. Unless the sand is jarred too much by the approach of the observer, they lie perfectly still and do not react to changes in light intensity caused by the movement of an object in front of them. They may be even picked up before they show signs of activity. Upon one occasion a piece of meat on the end of a reed was moved within about 5 cm. of the mouth-parts, but this brought forth no reaction nor did the crab react when the meat was held as close to the mouth-parts as possible without touching them. When, however, the ap- pendages of the mouth were touched by the food, they began to move slowly and then faster; after this the eyes came up, the ambulatory ap- pendages assumed the position necessary for locomotion, and finally the chelz seized the meat. While Ocypoda does not normally exhibit what has been called the death-feigning reaction, or, as Bethe (1897) has called it, in the case of Carcinus, the “ Starrkrampf reflex,’ it does so sometimes when placed upon its back and held in that position for a minute or so. MEMORY. Since the appearance of Bethe’s (1808) classical paper on the nervous system of Carcinus menas, in which he compares this crustacean to a reflex machine and denies it the ability to learn, several papers have appeared which furnish abundant proof that some crabs are able to profit by ex- perience and are even capable of forming habits. Yerkes (1902) investigat- ing Carcinus granulatus, Yerkes and Huggins (1903) studying the crayfish Cambarus affinis, and Spaulding (1904) working on the hermit-crab Eupa- gurus longicarpus, have found that these crustacea form associations. I hoped that Ocypoda would be a favorable subject for a quantitative study of habit formation, but as far as my experiments have been carried this does not seem to be the case. Ocypoda, we have seen, is a very active crustacean; its eyes are very sensitive when compared to those of other crabs; it is easily frightened, and when in this condition either runs away rapidly or remains perfectly still, failing to behave normally; when placed in a labyrinth with a solid bottom it usually scampers away from the hand that has released it and then often settles down in a corner without attempting to escape, or tries in vain to dig a burrow ; failing in this it tries to climb; when the bottom of the labyrinth is made of sand it usually digs a burrow, goes down in it, plugs up the entrance, and stays there for a considerable time. For the reason just mentioned it has been found very difficult to devise labyrinths which would be satisfactory, and the data accumulated on the length of time required for individuals to escape are probably not very trustworthy information concerning the rapidity with which Ocypoda forms a habit. Habits, Reactions, and Associations in Ocypoda arenaria. 35 Several different kinds of labyrinths were used, some of them similar to those employed by Yerkes (1902) and Yerkes and Huggins (1903), but these had to be reversed in the case of Ocypoda, because when the latter is disturbed it always runs away from the water. All the observations were made while the observer was practically hidden, and healthy active indi- viduals were chosen. It was found that labyrinths built of wood and closed up except at the exit were so dark that individuals would hide in them for hours at a time without attempting to escape, and that also labyrinths much more complex than those used by Yerkes and Huggins were of no use, on account of the length of time required for the crabs to find the exit. A labyrinth like that used by Yerkes (1902) in his experiments with Carcinus granulatus was made, but glass plates were substituted throughout for the wood, so as to allow plenty of light to enter. It also had a glass bottom and top and was placed on the beach near the water, properly pro- tected from any disturbance by the movements of people. The glass bot- tom was covered with a layer of sand about 1.5 cm. thick, and a screened pen was built around the exit so that the specimens could not escape entirely. It was found that individuals placed in this labyrinth usually only spent a short time in hunting for the exit; then they tried to dig a burrow, and fail- ing in this they attempted to escape by climbing. Finally they usually settled down and remained quiet, so that they furnished no data bearing on the object of the experiment. Finding that individuals nearly al- ways made some attempt to burrow, another kind of labyrinth was devised in which the burrowing instinct might be made use of in escape. While the ocypodas dug their way out in every experiment, they did not usually do so until night. Having noticed that nearly all individuals were inclined to climb, when confined, another labyrinth was made in which escape might be ac- complished by this method. A box Ir inches wide, 15 inches long, and 10.25 inches high was divided length- ways into two equal parts by a ver- tical wooden partition (a) which ex- tended from the top of the box to within 1.5 inches of the floor, thus leaving an opening connecting the two compartments (fig. 10). The Fic. 1o.—Labyrinth for memory experi- ments. 36 Papers from the Marine Biological Laboratory at Tortugas. ocypodas experimented with were introduced into the posterior compart- ment, shown in fig. 10, through the sliding door indicated (b). The front of the box (c), instead of being of wood, was made of ordinary fly- screen. At (d) and (e) on the floor of the box were placed two pieces of wood 2.4 inches long and 1.6 inches high, which divided the anterior com- partment into three passageways. In the top of the box at (f), (g) and (h) were set three windows 2.6 inches by 4 inches, made of ordinary window-glass. In the center of the middle window (g) an opening was left which was the exit from the labyrinth, and around this window was erected a barrier (7) which prevented the crabs from climbing out over the box. Similar ones (7) (k) were placed around the other windows so as to make the lighting the same in each case. In escaping a crab would have to pass from the posterior compartment into the anterior compartment, climb up the screen, pass through the opening in the middle window, and then climb down the outside of the screen to the surface of the sand in the pen. This pen was inclosed by panes of window-glass, 10 by 15 inches, set vertically. Six active ocypodas were used in the following tests, but one of these soon died and another was lost, so that in the table below only the time records of four individuals, A, B, C and D, are recorded. As a rule each crab was given two trials a day, and precautions were taken to guard the individuals from disturbing influences of moving objects. In general there seems to be a decrease in the time required for escape up to the eighth day, but after that the average time increases. This increase may have been due to the condition of the crabs. During the climbing after the eighth day they fell frequently, and this undoubtedly had a bad effect on them. Time required for escape from labyrinth, recorded in minutes. Date. A. B. | (ce | D. iverage. | Date. A. B. | C. | D. |Average. a _ = —| _ | — — lyon eOr wets les 127 38 | July 10 | 9 18 | 10 2 9.75 | aN na 1.75 | 65 20 22.68 | II | 0.25 | 34 sy ize 14.03 | WAZ) sa ALS ie ces 5 onl 16.75 || Ir] 4 8.5 |12.5 | 8.25 8.31 | 8 | ar mh lS | 19 mish |) 12 |14.5 |17 | 1.25 |23.5 | 14.06 8 | 13 OSS. eis |) 8.25 || 12) || 0 3 2.5 |16 5.62 | 8 | 17.5 | 15 a |) 3 12 | 13,| 0.5 | 75'| 225 | 10 nek | | hanes 25m NIG 4.03 | 13 | 10.75 | 41 6.5 | 49 26.81 9| 9 I 6 | 1.5 | 4.37 || 14] 5 36 5 | 4 11.37 10 | 10 8 | 25 | (os 6.18 || 15 |74 4.5 | 2.5 5 | 20.38 10 | 46 4 | ese | 14.87 | ES) ie3 18 | 32.5 | 13 | 16.62 During the first few experiments the behavior was as follows: Indi- viduals when introduced into the posterior compartment almost invariably ran to the left side; then if active they went into the left passageway and up the screen, usually following the corner of the box quite closely; on arriving at the top they would scratch the left glass window, apparently Habits, Reactions, and Associations in Ocypoda arenaria. 37 attempting to get out, but failing in this they would either crawl along the edge of the window toward the right until they reached the opening in the middle of the center window, where they would climb out and down into the pen, or they would descend again into the box and after a time repeat the climbing either in the left-hand or right-hand side and occasionally in the middle part of the screen. Although the time records do not indicate that the ocypodas learned the position of the exit, yet the behavior, after about 5 days, began to indicate that all the individuals had learned it more or less perfectly. They still continued to crawl up along the corners of the box, but they now often climbed across the screen in a diagonal direction toward the exit, from either the right or left passage-way. Toward the end of the series of tests, the ocypodas became quite inactive, often refusing for a long time to move, but when they did become active they frequently climbed almost directly toward the exit. The fact that in the above labyrinth experiments we do not obtain good quantitative evidence of memory does not show that Ocypoda is without memory. It must be admitted that the labyrinth used was a rather difficult one and that the tests did not extend over as long a period as they should. The injury experienced by frequent falls from the screen and the somewhat artificial conditions undoubtedly affected the results. Another experiment performed shows that Ocypoda profits by experi- ence. It had been noticed that individuals in the trap made use of the sea- water given to them in glass dishes in much the same way as individuals under natural conditions use the water in the sea. Accordingly a trap was thoroughly cleaned and stocked with several ocypodas caught in another part of the island. A study of the behavior was then made when a dish of sea-water was put in the trap. Care was taken not to allow any of the water to drop on the sand and the dish was buried until its rim was level with the surface of the sand. The ocypodas apparently did not pay any attention to the dish at first, but after a time, seemingly by chance, they crawled until their legs projected over the edge and into the water. Usually when this occurred to an individual it would settle down, digging the legs of the other side into the sand as it does when waiting for a wave. After a time it often climbed over into the water and moistened its gills. This ex- periment was repeated for several days, a dish of water being given to them twice a day, and many of the crabs soon went through the same behavior almost as soon as the water was introduced. Finally a dish without water was tried and it was found that some of the crabs behaved in the same manner as when water was in the dish, 7. ¢., they crawled up to its edge, settled as though waiting for a wave, and then finally climbed over into the dry dish. This behavior I interpret as a case of associative memory, in which the 38 Papers from the Marine Biological Laboratory at Tortugas. crabs formed an association between contact with the glass dish and the presence of water. In conclusion it may be said that Ocypoda, like the crustacea investi- gated by other workers, has memory, is able to profit by experience, and can form habits. SUMMARY. Adult ocypodas build two kinds of burrows. One consists of a single tunnel extending down in the sand for 3 to 4 feet. The other is similar, except that it is shorter and has a passage branching off from it, which is used for escape. Young ocypodas make short burrows, only a few inches long, which often extend vertically downward. Breeding in the region of Loggerhead Key probably occurs in the spring and early summer. Ocypoda is a scavenger and a cannibal. The eyes do not seem to play an important rdle in the detection of food, but they undoubtedly lead indi- viduals to objects which may be food. That Ocypoda is stimulated by odors was not conclusively shown, but certain experiments point strongly in that direction. The eyes are highly developed, so far as crustacean eyes are concerned ; they are quite sensitive to large differences in the intensity of light; they do not react to different colors; they aid much in the search for food, in the detection of enemies, and in the accuracy of locomotion. Ocy- podas probably do not have vision such as that of the human eye, nor do they see the color and finer characters of the surface of an object, but they undoubtedly see its outlines and possibly some of the more evident irregu- larities of the surface made evident by differences in lighting. The color-pattern seen through the carapace of Ocypoda changes in in- tensity under different conditions of temperaure and light. In the absence of light when the temperature is anywhere between 22° C. and 45° C., and undoubtedly when it is even lower or higher, a light coloration occurs. Generally in diffuse light and even direct sunlight a dark coloration ap- pears, provided the temperature is not too high. Usually at low temperatures, not above 35° C., a dark coloration occurs, provided the eye is stimulated by light. At high temperatures, above 35° C., a light coloration is the rule, and it occurs independently of the intensity of light. No indication of audition was observed in Ocypoda. The so-called “auditory organs’ are equilibrating organs. Ocypoda has a stridulating ridge on the palm of its large chela. Any sound which it may make is probably not heard by other individuals, but Habits, Reactions, and Associations in Ocypoda arenaria. 39 the vibrations of sound produced during the movement of this stridulating ridge against the basal joint of the chela are probably felt by other ocypodas. The tactile sense is well developed in Ocypoda. With the body orient- ated in a fixed position, the animal can move in practically any direction. It runs with a considerable degree of accuracy and undoubtedly has a sense of position and distance. In locomotion these crabs are guided by differ- ences in the lighting of surfaces, by tactile stimuli, by differences in mus- cular effort and by the stimulation of the equilibrating organs resulting from a tilting of the body. During most of the time Ocypoda lives on land, only going to the water occasionally for the purpose of moistening the gills. When exposed to the direct sunlight without water for more than 4 hours individuals usually die. They can live in their burrows without sea-water for at least 48 hours. The “ Aufbaum Reflex ” described by Bethe is an attitude of defense. Ocypoda often hides from man by simply settling down and throwing sand over its body. It is sometimes found in a resting or “ sleeping” condition when it does not react to many of the ordinary stimuli. The death-feigning reaction is exhibited under certain conditions. This brachyuran forms associations and habits. REFERENCES CITED. BeEeEr, T. 1898. Vergleichend-physiologische Studien ziir Statocysten Function. I. Arch. f. ges. Physiol., Bd. 73, 1808. 1899. Vergleichend-physiologische Studien ztir Statocysten Function. II. Arch. f. ges. Physiol., Bd. 74, 1899. BeETHE, A. 1895. Studien tiber das Centralnervensystem von Carcinus mznas, etc. Arch. f. Mikr. Anat., Bd. 44, 1895. 1897. Das Nervensystem von Carcinus menas. Arch. f. Mikr. Anat., Bd. 50, 1897. Bonn, G. 1893. De l’évolution des Connaissances chez les Animaux Marines Littoraux. Inst. Général Psychologique, Bul. No, 6, 1903. Brooks and Herrick. 1889. The embryology and metamorphosis of the Macroura. Memoirs of the National Academy of Sciences, No. 4, vol. v. Car.eTon, F. C. 1903. The color changes in the skin of the so-called Florida chameleon, Anolis carolinensis Cuy. Proc. Amer. Acad. Arts and Sci. vol. xxx1rx, No. Io, 1903. (Gufs, Ib ize 1896. On the relation of the otocysts to equilibrium phenomena in Gelasimus pugilator and Platyonichus ocellatus. Jour. of Physiol., vol. 19, 1806. DARWIN, CHARLES. 1875. Descent of man. 2d edition, London. John Murray, 1875, p. 271, 1875. FIELpDE and PARKER. 1904. The reactions of ants to material vibrations. Proc. Acad. Nat. Sci. Phila- delphia, Sept., 1904. GAMBLE and KEEBLE. 1900. Quart. Journ. Micro. Sci., vol. 43, 1900. 1904. Phil. Trans. Royal Society London. Series B, vol. 196, pp. 295-388, 1904. Goong, G. B. 1878. The voices of crustaceans. Proc. U. S. Nat. Mus., vol. 1, 1878. HeErRDMAN, W. A. 1894. Proc. and Trans. Liverpool Bio. Soc., vol. vi, session 1893-1804. Jorpatn, S. 1878. Sur les changements du couleur du Nika edulis. Comtes Rendus, t. 87, 1878. Krerpi, A. 1893. Sitzb. d. Kis. Akad. d. Wiss. in Wien, Bd. 102, 1893. Lis; 1897. Mitt. der Zool. Sta. zu Neapel, Bd. xm, 1897. TEVON, Bee: 1899. A contribution to the comparative physiology of compensatory movements. Am. Jour. Physiol., vol. 3, 1899. Matarp, A. E. 1892. Bul. Soc. Philomath. de Paris, 8éme Series, 1892. Matzporrr, C. 1882. Ueber die Farbung von Idotea tricuspidata. Jenaische Zeit., xvi, 1882. Mayer, P. 1879. Ueber Farbenwechsel bei Isopoden. Mitt. aus der Zool. Sta. zu Neapel, vol. 1, 1879. Miers, E. J. 1876. Report on the Brachyura collected by H. M. S. Challenger. From report on scientific results Exp. Voyage of H. M. S. Challenger, 1873-76. Zoology, vol. xvi. 40 Habits, Reactions, and Associations in Ocypoda arenaria. 41 Mostus, K. 1867. Ueber die Entstehung der Tone welche Palinurus vulgaris mit den aussern Fiihlern hervorbringt. Arch fiir Naturgeschichte. 1867. NacEL, W. A. 1894. Geruchs- und Geschmackssinne und ihre Organe, etc. Bibliotheca Zoologica, No. 18, 1894. Parker, T. 1878. Hate ‘on stridulating organ of Palinurus vulgaris. Proc. Zool. Soc. London, 1878. PARKER AND STARRATT. 1904. The effect of heat on the color changes in the skin of Anolis carolinensis Cuy. Proc. Amer, Acad. Arts and Sci. vol. xt, No. 10, 1904. Parker, G. H. 1906. Influence of light and heat on the movements of the Melanophore pigment, especially in lizards. Jour. Exp. Zool., vol. m1, No. 3, 1906. Poucuet, G. 1872. Sur les rapides changements de coloration provoqués expérimentalement chez les crustaces. Comptes Rendus, Lxxiv, 1872. 1872. Jour. de l’Anat. et de la Physiol., vol. vim1, 1872. 1876. Jour. de l’Anat. et de la Physiol., vol. xu, 1876. Prentiss, C. W. 1901. The otocyst of decapod crustacea: its structure, development and functions. Bul. Mus. Comp. Zool. Harvard College, vol. xxxv1, No. 7, 1901. SPAULDING, E. G. 1904. An establishment of associations in hermit-crabs, Eupagurus longicarpus. Jour. Comp. Neurol. and Psychol., vol. xiv, 1904. VERRILL AND SMITH. 1874. Report upon the invertebrate animals of Vineyard Sound and adjacent waters. From Report of S. F. Baird, Commissioner of Fish and Fish- eries, 1874. YERKES, R. 1902. Habit-formation in the green crab, Carnicus granulatus. Bio, Bul., vol. m1, No. 5, 1902. 1905. The sense of hearing in frogs. Journ. Comp. Neurology and Psychology, vol. xv, No. 4, 1905. YERKES and HvuccIns. 1903. Habit-formation in the crawfish. Harvard Psychological Studies, vol. 1, 1903. ZWAARDEMAKER, H. 1895. Die Physiologie der Geruchs. Leipzig, 1895. P. ~~ — aa ul m4 ex a” * STOCKARD—WALKING-STICK PLATE 1, C. Kellner. Fig. 1. A dark type Aplopus female with first pair of legs extending forward, inclosing the antenne. This position is often assumed and serves to give the anterior end of the insect the appearance of a Straight stick. Fig. 2. The pale-gray type female with first pair of legs apart showing curve in the femora, which fits closely to the head in Fig. 1. Fig. 3. Male Aplopus, indicating his smaller size and green legs. Figs. 1,2,and3 are two-thirds natural size. Fig. 45. Egg of Aplopus, natural size. 4». Seed of Suriana maritima, natural size. The two resemble each other closely in size and color. Me HABITS, REACTIONS, AND MATING IN- STINCTS OF THE “WALKING STICK,” APLOPUS MAYERI. Bye GiUNRIWIES, Re SWORD) Instructor in Comparative Morphology, Cornell University Medical College, New York City. 3 plates and 1 text figure. 43 % 7 ; a ATvew itive oeAry iA rar} ut Ril | Fiat. Linn | ine 4 v8 j g ) 7 ernie ; a ; +n | : 5 Pie oti ian Gath ar ie ¥ ’ li «itp wi * | me : th HABITS, REACTIONS AND MATING INSTINCTS OF THE ‘* WALKING- STICK,’” APLOPUS MAYERI. By CuHartes R. STOCKARD. An investigation of the behavior of a protectively adapted insect is im- portant to show definitely whether the actions of such an animal are co- ordinated with its protective structure. If an insect such as the “ walking- stick,” which forms the subject of the present discussion, was found to move about briskly on exposed portions of the plant on which it lives, and to show other habits which might attract the attention of enemies in spite of its apparent resemblance to the stems and leafy parts of the plant, then, notwithstanding this resemblance, it would scarcely be as well protected as other insects showing no such resemblance but remaining still and con- cealed among the leaves and branches of the shrub. It is clear to all ob- servers that the behavior of an animal is almost if not quite as important as its structure in determining whether or not the animal is truly protectively adapted to its surroundings. With the above as a premise, we may ask what would be the theoretical expectation for protective behavior of such an insect as Aplopus. Since movement attracts the attention of birds and other enemies almost or quite as readily as conspicuous appearance, we should first expect this insect to remain perfectly motionless during the day, while it may be seen, and to move at night; in other words, nocturnal habits would be ideal for its protection. Aplopus is unable to fly or leap; thus its most effective means of escape would be to drop bodily from a position when touched, and become lost by falling and alighting among the lower branches of the bush, where it might remain motionless and concealed. When moving from place to place it would attract less attention should it move slowly, and the extreme per- fection of movement would be to vibrate from side to side as it progressed, as a twig swings in a light breeze. The male is colored with more green than the female; therefore, he may move among the leaves to better advan- tage and she will be less conspicuous on the larger brown stems. Lastly, since these insects vary greatly in color, as do most protectively colored ani- mals, we might expect, for instance, that light-gray females would rest on the lighter-colored stems, while the dark-brown type would be found on darker stems. 45 46 Papers from the Marine Biological Laboratory at Tortugas. After studying the habits of the “ walking-stick ”’ one finds that the fore- going expectations for its behavior are minutely carried out, with the single exception of the last mentioned. In the following pages I shall discuss the habits of Aplopus and enume- rate the results of a number of experiments conducted to test its responses to various stimuli and changed conditions. Record will also be made of some experiments concerning their mating instincts, in which males were induced to copulate with the amputated abdomen of a female attached to a small stick supported by wire legs. The experiments were performed at the Laboratory of Marine Biology of the Carnegie Institution of Washington, Tortugas, Florida, and it is a pleasure to express my thanks to Dr. Alfred G. Mayer, the Director of the Laboratory, for many courtesies extended to me while there, and for the kindly interest he has shown in my work. BEHAVIOR OF APLOPUS IN NATURE. Aplopus mayert is a large insect, the females often measuring more than 8 inches from the tip of antenne to the tip of abdomen, while for the smaller male 6 inches is an adult length. The male’s antennz are longer than those of the female. The male is much more sensitive to stimuli and in nature is the more active of the two. Figures 1 and 2 of plate 1 illustrate females and figure 3 shows a male, all reduced to two-thirds natural size. A decided difference in color will be noted between the two females, one being very dark, while the other is a pale gray. A large variety of gradations exist between these two extremes. The males are more or less greenish, but they also vary considerably in color. In some males the abdomen is a rich dark brown, in others a pale drab. The legs in all are darker or lighter shades of green. Such variations are common within the family Phasmide, all members of which are more or less protectively colored and constructed. The young males are brown or grayish, resembling the females in color, and can only be distinguished from them by the absence of the oviscapt and the presence of a prominent organ of intromission on their ventral surfaces near the tip of the abdomen. At maturity, however, the males acquire the adult greenish color and may then be recognized at a glance. The wings of both sexes are rudimentary, but are capable of being raised when the animal is greatly excited, giving to it a lively and agitated appearance. Aplopus has spines and prominences on its body and legs suggesting the slight irregularities on the bark of twigs. These insects are found only on their food-plant, Suriana maritima, and on this shrub they are extremely difficult to detect. In color and shape the female resembles closely the stems of small branches. The greenish color of the male conceals it among the leaves, while one may find a close resemblance in size and color between the eggs of Aplopus and the seed of Habits, etc., of the “ Walking-Stick” (Aplopus mayert). 47 Suriana, both of which fall from the branches to the ground, where they are obscured among the débris (figs. 4a and 4b of plate 1). The “ walking-sticks ” are not easily collected during the day, though at night, when they become active, one may obtain them in large numbers by means of a lantern. The following instance may be cited as illustrating the extent to which these insects are concealed as they sit among the branches. A small group of bushes about 10 feet in diameter contained six large fe- males and two adult males. Three of the females were sitting near the edge of the bushes and within about 2 feet of one another. A person unaccus- tomed to searching for the insects, although familiar with their appear- ance, attempted to find the three, but failed to locate any of them. I then followed after him and succeeded in locating two, but failed to find the third until a second careful search was made. As a test of how readily birds might find aplopi while in their motionless attitude, five individuals were placed on the ground near a hen with several small chickens. The mother showed no evidence of recognizing the in- sects until within a short time the Aplopi became excited at finding them- selves on the ground and began to crawl toward a bush. When they began moving the hen immediately started after them. The observations and experiments that follow were based on the actions of 26 males and 81 females, some of which were kept in cages, while others were allowed to remain in their natural environment on isolated Swriana bushes. ATTITUDES WHEN AT REST. Aplopus assumes a decidedly protective attitude when at rest. The fore- legs, which are slightly grooved on their anterior surfaces, are directed for- ward, the grooves being approximated to form a tube inclosing the antenne, thus producing a resemblance between the anterior end of the animal and the straight end of a small dead stick (see fig. 1 of plate 1 and also figs. I, 2, and 3 of plate 2, photographs from life of the insects in such a position). With their forelegs straightened in this manner, the females usually rest in an obliquely vertical position on the leafless stems. The abdomen sometimes points upward, the tips of the forelimbs being directed down, grasping the branch on which they rest. Again, the abdomen may be turned downward and serve as a partial support, while the body, with the extended forelegs free, points obliquely upward. In either attitude the in- sect closely simulates a dead stick projecting from the supporting branch. The femora of the first pair of legs are curved near their proximal joints so as to fit closely around the insect’s head when they are extended forward (fig. 2 of plate 1). The females, as mentioned before, vary considerably in color from dark brown to light gray, as do also the stems of Suriana, though the insects do not at all times take advantage of this variation, the dark females being not uncommonly seen on light stems, and vice versa. 48 Papers from the Marine Biological Laboratory at Tortugas. The males do not so constantly assume the position with the first legs extended. This attitude is less essential for them than for the females, since they are found as a rule among the leafed branches, where their greenish legs are inconspicuous. Their slenderer proportions and smaller size also make them more difficult to see. At times, however, the males do straighten their first pair of legs forward and assume the position so common to the female. The more common occurrence of the greenish males among the foliage and of the brown females upon the brown stems of the plant suggest the case of Mantis religiosa. Di Cesnola’ records this mantis as occurring in a green and a brown form in Italy, the green form being always found upon green grass and the brown form upon grass burnt by the sun. He found that when 25 individuals of the green form were tied on brown grass all were killed in 11 days, while 20 tied on green grass were all alive after 17 days. The results were similar when the brown form was tied on green and burnt grass. MOVEMENTS OF APLOPUS. Aplopus is nocturnal in its habits. During the day it sits motionless among the branches, but as the sun’s rays weaken the males begin to move about first, and later the females. Their manner of walking is peculiarly interesting. They resemble sticks crawling about on legs. The legs are moved in a stiff manner and the insect progresses slowly as a rule, although they sometimes move at a rapid gait. While walking among the branches or on a flat surface they often show a lateral swinging movement, the foot- hold forming a fixed point, while by bending the tarsal and knee joints the body is swung sidewise, sometimes with an amplitude of more than half an inch. This movement suggests to the observer the swinging or waving mo- tion of branches and stems when shaken, or blown by the wind. The male has also a peculiar quivering movement that is sometimes performed at intervals while moving about the branches. This motion is more to be associated with the mating instincts, as is shown below. In the evening Aplopus shows a tendency to climb upward, being negatively geotropic. It will then often turn and go upward if the twig is inverted after it has reached the top. The males travel greater distances during the night than do the females. Three females under observation did not move more than 2 feet from their original positions within a period of two days, yet others were found to migrate to parts of the bushes 10 or 12 feet distant during a single night. The males usually travel many yards in a night and it is almost impossible to keep track of them for more than one day. Aplopus avoids or escapes its enemies in the following ways: On several *Di Cesnola, A. P. Preliminary note on the protective value of color in Mantis religiosa. Biometrika, III, p. 58, 1904. PLATE 2 WALKING STICK Oh sui | ii, 4 Habits, etc., of the “ Walking-Stick” (Aplopus mayer). 49 occasions females were touched lightly as they sat in their protective attitude. They remained motionless after being touched, as though they were inani- mate bodies. A male under observation about dusk was struck by a large insect which flew against it; the Aplopus jerked quickly back and remained motionless for more than a minute, after which it walked swiftly down the branch. The most effective means of escape for these animals is the “ dropping reaction.’ When one attempting to capture an Aplopus fails to seize it the first time, it often drops bodily from the limb on which it rested and catches on some lower branch that it may chance to strike. If seen and unsuccess- fully grabbed at for the second time, it will again drop and may sometimes fall entirely to the ground. One male in attempting to escape capture fell to the ground, striking on its back with its legs extended in the air. It feigned death perfectly and remained in this awkward position for more than 8 minutes; then turned itself over and moved away so quickly that it was lost sight of in the dim evening light. Since “ walking-sticks ” are unable either to fly or jump, this dropping reaction is a most important means of escape, and the dense growth of the Suriana bushes would apparently prevent birds from finding the insect a second time after it had so suddenly fallen out of reach. Their motionless attitudes during the day and close resemblance to the stems of the bushes no doubt serve to protect Aplopus to a marked degree from predaceous birds and other enemies. The food of Afplopus consists entirely of the leaves of Suriana maritima, the plant on which it lives. The only previously published statement regard- ing its habits is a brief paragraph in the catalogue from the supply depart- ment of the Marine Biological Laboratory at Woods Hole. This states that “the prey is seized by a quick movement of the forelegs.” Such an idea is, of course, erroneous, since all members of this family are known to be vegetable feeders. The statement is doubtless based on the opinion of some amateur collector. Aplopus usually feeds at night, although those resting on leafy branches are sometimes observed to feed during the day. In feeding they bite the leaf straight across the top and often eat it entirely away, or they may bite the leaf in an up-and-down fashion until it is con- sumed. They rarely make semicircular cuts in the leaves, as the locust often does. These characteristically bitten leaves serve to furnish a trustworthy index of the whereabouts of Aplopi, as they seem to be somewhat locally distributed on the island. MATING INSTINCTS. Mating occurs as a rule during the night, although several pairs were observed in copulo during the day. The active process is much the same as in kindred insects. The male takes a position on the back of the female, with his front feet resting on her metathorax, the second pair of feet 50 Papers from the Marine Biological Laboratory at Tortugas. grasping her abdomen about its middle, his third pair of legs usually hang- ing freely extended, the tip of his abdomen being firmly attached to a slight pit on the ventral surface of the seventh abdominal segment of the female. The intromissive organ of the male is then protruded and placed between the oviscapt and the last three abdominal segments of the female. In this position the male remains for from 30 minutes to several hours. His copulating organ is then withdrawn, although he may still re- main for a long time sitting upon the back of the female (fig. 4, plate 3). One male may copulate with several females during the same day. The male often gives periodical quivering movements while over the female, prob- ably for the purpose of exciting her to the sexual act. He sometimes shows a slow, swinging motion during copulation. The female is supplied with a long ovipositor, although it seems to be useless, as her eggs are allowed to fall carelessly to the ground as she sits motionless among the branches. The eggs resemble closely in size and color the seed of Suriana, as mentioned before, but differ from them considerably in shape. (Compare figs. 4a and 4b in plate 1.) All of the observations on the habits of Aplopus in nature would seem to indicate that the behavior of this insect is as truly protective as is its close simulation of the branches on which it lives. EXPERIMENTAL. EXPERIMENTS WITH LIGHT. Aplopus responds to light and darkness in a most interesting manner. The insects were observed to begin moving on the bushes by a much brighter light in the evening than that which served to stop their movement during the morning. They were seen feeding and crawling slowly about, at times, two hours before sunset; while they often came to rest more than half an hour before sunrise. The difference in intensities of the lights causing the two reactions is very great. It occurred to me that perhaps their response was periodic and not entirely due to the effect of light; that is, after being active for several hours during the night they become tired and cease to move for this reason, and not on account of any response to light, since the intensity of the morning light by which they come to rest is even less than that of the moonlight in which they are active. (The quality of the two lights is no doubt different.) The case is, however, made clear by the fol- lowing experiment : Sixty-five individuals in a wire cage had come to rest at 4" 50™ a. m. At this time daylight was scarcely perceptible. It was much darker than when they had begun movement during the evening, or even the moonlight of the earlier part of the night. It would seem, then, that a physiological periodicity had had some influence on their behavior. To test this the cage was placed in a dark-room at 5 a. m. In less than half an hour all were Habits, etc., of the “ Walking-Stick” (Aplopus mayert). 51 actively crawling, thus showing that they had responded to the faint light of approaching day and not to a tired condition. They continued to move actively for more than an hour in the dark-room, and were then put into the light, where they readily came to rest again. The fact that they begin movement by brighter light in the evening and stop by paler or weaker light in the morning may possibly be associated with a similar habit of some birds, which go to their roosts by brighter light than that by which they leave. A number of dark-room experiments were performed during the day to test the manner and time of response to light and darkness. The dark-room was one arranged for photographic work, having a red glass window that might be covered with a black oilcloth, so that no light was admitted. The door of the room opened into the closed side of the laboratory, thus per- mitting only weak diffused light to enter the room at any time. I was enabled to detect the first movement of the insects in the dark, as the gauze wire of the cage gave a perceptible clicking sound when their feet were moved upon it. We may first consider the reactions of normal individuals. Since the experiments gave closely similar results, one may be recorded for illustration. Three females and two males were placed in the dark-room at 10" 25™ a. m. After 15 minutes three were actively moving, while after 20 minutes all were in motion. The door was opened and light admitted at 11 a. m.; they came to rest in a little more than a minute, and remained so in spite of various loud noises until 118 11™ a. m., when the room was again darkened. The animals began again to move and all five were in active motion within 8 minutes. Bright daylight was thrown on them after all had been moving for 3 minutes; two were at rest in less than a minute, and all were quiet in about 2 minutes. When they had been at rest for 5 minutes the room was for the third time darkened, and after 11 minutes all were active. Light was then admitted and two ceased to move within 2 minutes, the others stopping after 5 minutes. The dark-room was again closed at 11" 45™ a. m. and left until 12 30™ p. m., at which time all of the insects were active. On admitting light they became motionless within 2 minutes. The dark-room was closed for the fifth time at 124 55™ p. m., after the five “ walking-sticks ’ had been quiet for more than 20 minutes; 10 minutes after they had been in the dark two were moving, and all were in motion after 15 minutes or at 18 10™ p. m. Light was admitted at 15 11™ p. m., and all came to rest within 30 seconds, a very quick response. They were then exposed to light for 10 minutes, then again put into the dark-room. The first one did not begin to move until 17 minutes had elapsed ; the others were moving after being in the dark 20 minutes. The first ones to move had disturbed others by striking against them, so that these probably moved earlier than they would have otherwise. 52 Papers from the Marine Biological Laboratory at Tortugas. They were allowed to remain active for 10 minutes and were then placed in bright light, where they again stopped all movement in less than 30 seconds. They were observed closely for 10 minutes while in the light, and not a leg or antenna was moved, though some had stopped in apparently awk- ward positions. The foregoing experiment was repeated several times on different in- dividuals and at various periods during the day, always giving similar results. Aplopus is thus seen to become active in the dark within from 10 to 20 minutes, and at times even more promptly. This activity is continued as long as it remains in the dark. When the insect is placed in bright light it promptly comes to rest within from less than 30 seconds in some cases to several minutes in others. Aplopus responds, therefore, more promptly to light than to darkness. The males appear to come to rest more readily than the females; they are also more active in the dark. These insects may readily be made to mate by placing a number of individuals of both sexes in a cage in the dark. The question next arises whether responses to light and darkness are due to the action of the stimulus on the optic organs or to the effects of light on the body-surface of the insect as a whole. In attempting an answer to this question several experiments were performed. First, a number of Aplopi were chosen and their eyes were well covered with a lampblack paste until they were apparently blind. These individuals were then subjected to dark- room experiments. When only the compound eyes were blackened, the simple eyes being uncovered, they still responded in the dark-room, though slower than the control. On one occasion a male and three females were used; the male moved slightly after 15 minutes, though almost an hour had elapsed before all four individuals had become active. When both the simple and compound eyes were blackened they responded still more slowly in the dark. Of four treated in this manner only one had moved after 30 minutes in the dark-room, and this one almost imme- diately came to rest again; so that after 50 minutes all were quiet, three of the four not having moved during this time. After 2 hours three were at rest and one was moving; one of the four had not moved at all during the two hours and the three that had moved did so only for a moment, not becoming really active, as they normally do in the dark-room. To test further the importance of vision in responding to changes from light to dark, I determined to blacken the compound and simple eyes of a number of Aplopi during the night to ascertain in what manner they would respond when the daylight appeared. Six females and two males were placed in a cage to themselves and their eyes were painted at 9 40™ p. m., while they were all very active. At 5" 30™ a. m. on the following morning three of the females and one of the males were actively moving. These animals were much more active than the control of about 50 individuals, all Habits, etc., of the “ Walking-Stick” (Aplopus mayert). 53 of which were now at rest, although they were caged nearby so as to expe- rience the same light conditions. At 6 o’clock four of the blind ones were still active, although the sun had been shining for half an hour. At 8 o’clock all were at rest, yet they were more than an hour later than the control in responding to the morning light. When these blind insects had been at rest for 2 hours they were placed in the dark-room, where all became active after about an hour. They were then brought into the light and assumed the attitude of rest within 12 minutes. In the evening the ones with painted eyes became active 30 to 45 minutes earlier than the normal ones did. At 6" 12™ p. m. not one of 50 normal insects had moved, while 5 of the 8 blind ones had been moving actively for 40 minutes. Aplopi probably appreciate light to some extent through their bodies, but more acutely by means of their eyes; thus night appears to come earlier and day later to the blind ones. The 2 blind males failed to pair with either of the 6 blind females, though normal males and females usually mated when they were caged together. Blind and normal females were observed in their natural environment on Suriana bushes. Here the blind individuals also became active earlier in the evening than the normal ones. When the strong light of a bull’s-eye lantern is thrown on a normal one of these insects at night, it turns its head from side to side and gives evi- dence of seeing the light. These experiments seem to show that Aplopus may respond to light and darkness through its general body surfaces, but that it does so much less readily, or slower, than by means of its optic organs. EXPERIMENTS WITH LIGHT RAYS OF DIFFERENT LENGTHS. It became desirable at this stage to know whether the insects responded to white light as a complex whole or to some of its constituent rays. Several experiments were conducted in the attempt to solve this problem. Light was passed through a vessel containing carbon bisulphide, which serves to eliminate the ultra-violet rays. This is the well-known experi- ment first performed by Sir John Lubbock on ants. A dark-jar was ar- ranged and two of the “ walking-sticks ” were placed in it. After they had been quiet for 10 minutes the jar was covered by a vessel containing sea- water in order to test the effect of the subdued light which was transmitted by the liquid. The animals under this condition remained motionless for 25 minutes. The vessel containing the sea-water was then removed, and the jar allowed to stand uncovered for 10 minutes. The insects still retained their daylight state of rest. A vessel containing carbon bisulphide was now placed over the jar, thus admitting daylight minus its ultra-violet rays. The Aplopi remained perfectly motionless in this light for 85 minutes and were then removed from the jar. It thus seems apparent that this insect is not 54 Papers from the Marine Biological Laboratory at Tortugas. brought to rest by the ultra-violet rays of sunlight, since they do not move in the absence of such rays. They probably respond, therefore, to the visible rays of the spectrum. Ten Aplopi were subjected to light transmitted through blue glass. (Spectroscopic analysis showed this glass to be impure, transmitting blue, green, and a little red and yellow.) The insects gave no definite response in this light, although during one experiment they became more active than usual, moving as if they were in the dark. Such a response was, however, not at all constant and I am inclined to think that the individuals of this experiment had become unduly excited from some other cause. The influence of red light, containing possibly a little orange, was tried on ten dplopi. This also failed to give any definite reaction. It seems likely, then, that these insects respond to sunlight as a complex light and not to a limited number of its rays—at any rate not to the few tested above. EXPERIMENTS WITH SOUND. Aplopus seems indifferent to loud noises; a loud voice or a strong rap upon a board is apparently unheard. A 32-caliber pistol was fired three times within 18 inches of three active individuals, one male and two females, yet they gave no indication of having heard the pistol. Before the pistol was fired the second time the “ walking-sticks” were made to assume awkward positions; still they remained motionless after the noise. MOVEMENT EXPERIMENTS. Aplopus may be made to assume almost any position, it matters not how apparently awkward, and it will often retain such a position for a long period of time. Such a response may be very useful in causing this animal to be passed unnoticed. If it be shaken or struck while resting on a limb it will not at first scamper off as most insects would, but remains perfectly still in almost any position it may chance to occupy after the shock. When such a disturbance is repeated for several times, Aplopus may become ex- cited and either drop or attempt to run away. One of either pair of legs may be lifted from its foothold and be straightened or twisted backward or forward and left with the foot free and unsupported. The leg may re- main motionless in such a position for long periods of time. Two or even three legs can be raised in such a fashion, and Aplopus will stand motionless on its remaining foothold. Not more than three of its feet can be raised at any one time, since it is unable to support its long body upon only two legs. The antennz may be directed in any direction the observer may wish and the insect will permit them to remain in such a position. One may actually lift an unexcited Aplopus by its long mesothorax and slowly place it back down upon a flat surface, where it will remain for 10 minutes or more with its legs pointing upward. A second individual may be placed in a similar manner over the first, and both will remain motionless for many minutes. As mentioned above, they show a death-feigning reaction. Habits, etc., of the “ Walking-Stick” (Aplopus mayert). 55 This insect suggests to one a papier-maché imitation with wire legs which may be bent or twisted in almost any manner and put in any position. It could scarcely be more stick-like. When walking, Aplopus often stops and waves its antenne about in a circle, apparently feeling for some object in front of it. If the antennz strike an object, a stick or a leaf, the first pair of legs reach forward and attempt to grasp it and pull the body of the insect up to it. When the insects have climbed to the top of a branch they usually wave their anten- ne, trying to find some object on which to continue their upward journey. The antennz of several individuals were cut away close to their proximal joints. The insects were slightly excited by the operation, but soon moved off, using their first pair of legs as feelers, stopping at intervals and waving either the right or left leg and at times circling both legs in front of the head, just as if the legs were efficient antennz. One of the first pair of legs was removed, and the remaining one then served the purpose of a feeler. The other first leg was then removed, leaving the animal without antenne or either first leg. The insect now progressed in a slower but surprisingly normal fashion upon only four legs. The point of especial interest is that first the one and then the other of the second pair of legs was raised and circled about as an antenna or feeler. Both of these legs could not be so used at any one time, since the insect is unable to stand on less than three legs. Normal insects were never observed to use either of the second pair of legs as feelers. The eyes of such a four-legged, antennaless animal were blackened so that it was unable to see. This confused the subject considerably and it turned several times in a circle before being able to progress straight for- ward. The progress was then slow and cautious. Such an Afplopus often turned its head from side to side, as if attempting to see; it also moved the stumps of its antenne and legs. It was able to climb among the branches and feed in a typical manner. On the following day it showed marked im- provement in its ability to progress. When at rest these crippled individuals directed the remaining proximal portions of their antennz and first legs forward, just as though they were assuming the attitude with legs and antenne pointing straight out in front of the head, which is so typical for normal individuals. A strong electric current causes Aplopus to move actively and may often cause its legs to kick violently for some seconds. MATING OF MALES WITH A PORTION OF A FEMALE ABDOMEN ATTACHED TO A STICK. Many experimenters have attempted in various ways to determine through what senses the male insect locates and mates with the female. Among the moths and butterflies the sexes are sometimes differently colored 56 Papers from the Marine Biological Laboratory at Tortugas. and observers have claimed that the adornment of the male or of the female was a factor in the selection of the other sex. This manner of viewing the case was seriously questioned by some interesting experiments performed by Mayer in 1900.1 A number of female moths were placed in an open- mouth glass jar covered by netting and five males when liberated 100 feet away flew to the jar. The experiment was then repeated with the jar in- verted, so as to close the opening. This time the males did not approach, although the females were visible through the glass. It thus appears that the male moth finds the female by the sense of smell rather than the sense of sight. Other females were inclosed in a box with an open chimney, and the males flew to the chimney, although the females were not visible. When abdomens of females were cut off, the males would fly to these rather than to the winged bodies. If the antennz of the male be removed he does not go to the female. Mayer also glued the wings of a male over the wings of a female, so that she appeared like a male; nevertheless she was found by a male and mated with normally. Males would pay no attention to other males with female wings, but would pair readily with a female both of whose wings had been cut away. In all of these experiments, however, the male and female were in healthy conditions during mating, so that they were capable of movement or actions by which the sexes might excite one another. Dr. Mayer informs me that during his experience a male would not mate with a fatally muti- lated or dying female. I wished to conduct an experiment that would eliminate the possibility of anything like a courtship or psychical action between the sexes. Since it seems to be the odor of the abdomen of the female that first attracts the male, I concluded to make papier-maché imitation females and smear the abdomen of these with juices from the abdomen of mature females; then, on caging a number of males with these imitation females, pairing might take place. The papier-maché imitations could not be obtained, however, so this experiment was abandoned, though it is probably well worth trying with a number of insects. It was then decided to construct an artificial female by fastening a por- tion of the abdomen of a mature female Aplopus on to a small stick. A Suriana stick was cut that approximated in thickness the female’s body and supported on six wire legs. One end of the stick was trimmed to a conical point and the abdomen of a female minus the first segment was pushed on over this conical end and made fast by winding thread about it. An ab- domen thus attached to a stick will remain alive and is capable of moving slightly, and indeed defecating after more than 24 hours. The head and *Mayer, A. G. On the mating instinct in moths. Annal. and Mag. Nat. His- g tory, Vv, 1900. WALKING STICK Fig. 4. A mating pair of Aplopi. The male has withdrawn his intromission organ but still grasps abdomen of female. Fig. 5. Male Aplopus immediately after he has withdrawn his intromission organ while copulating with the amputated abdomen of a female fixed to a wire-legged stick. Habits, etc., of the “ Walking-Stick” (Aplopus mayert). 57 thorax also continue to live and crawl about in the usual manner for several days after the abdomen has been removed. I induced two male aplopi to pair with such a “ stick-female” in a per- fectly normal manner (text-fig. 1 and plate 3, fig. 5). The experiment was performed as follows: It had been found, if a male was separated from a normal female while mating with her, that they would remate after a short time if placed in the dark. It had also been found by a previous unsuc- cessful experiment that the abdomen should be from a female that was mature, but that had not been mated with. This in mind, five males and ten females were put into a dark-room, where after 12 minutes one of the males had paired with a female. The pair was separated and the abdomen of the female cut off at the joint between the first and second segments and fixed ‘ it r Esl SS —— —_ Fic. 1.—Drawing of male Aplopus in full copulating attitude with the end of a female abdomen fixed to a wire-legged stick. to the stick with wire legs as described above. All of the females were now removed from the cage and the abdomen on the wire-legged stick was at- tached to the side of the cage in a vertical position and placed in the dark- room with four males. After 1.5 hours the “ stick-female” had not been disturbed by any of the males. It was now moved and placed in a horizontal position, as if hold- ing to the gauze-wire top of the cage by its legs, with its body suspended, the attitude of any insect while clinging to the under side of a horizontal sur- face (fig. 1). In such a position the abdomen was mated with in less than an hour. The male in this instance was not the same individual that had pre- viously mated with the entire female. He was in a perfectly normal copulat- ing attitude, his organ of intromission being inserted between the oviscapt and the raised end of the female abdomen, as is shown in figure 1. Figure 5, plate 3, is a photograph of the pair, although here the male has withdrawn his intromissive organ on account of the disturbance caused by shifting the cage into a favorable light for photographing. This male was finally, by the movements of the cage, made to leave the “ stick-female.” “ce mn (72) Papers from the Marine Biological Laboratory at Tortugas. A point of some interest is that Aplopus seems to prefer the female to occupy a horizontal position in mating. During the first experiment with the “ stick-female ” it was placed in a vertical position and caged with males in the dark for two days without a result. In the experiment above the “stick-female’’ was first placed vertically and remained so for 1.5 hours without attracting a male. When it was changed to a horizontal position a male paired with it in less than an hour. All of the normal pairs observed were in a more or less horizontal position. It may be that the attitudes of the male are difficult to assume unless the female stands horizontally. The cage with the “ stick-female” and four males was again returned to the dark-room and after 2.5 hours a second male was found standing on the decoy. He remained in this position for over 3 hours, which was a much longer time than I had known a male to stand over a normal female without copulating with her. After this time, however, he began the usual mating movements and copulated perfectly with the abdomen. This was, then, the second time that the amputated abdomen of the female had been paired with, and each time by a different male. No doubt, therefore, remains that the male Aplopus may pair normally with the female without any “ communica- tion,” “courtship,” or psychical processes having taken place between them. SUMMARY AND CONCLUSIONS. 1. The habits of Aplopus mayeri on its food-plant Suriana maritima are as truly protectively adapted as is its singular stick-like appearance. ‘The large females in their color and shape resemble the stem of this plant; the males are greenish and well concealed among the leafed twigs, while the eggs are peculiarly similar to the seeds of Swriana in size and color, although differing in shape. The insect is nocturnal and only occasionally moves in the daylight; then as a rule with a slow, waving motion suggesting the move- ment of a branch swinging in a light breeze. To escape enemies it may fall bodily from its position and become lost among the lower branches of the shrub, or at times it may fall entirely to the ground, where it will lie motion- less for several minutes as if feigning death. 2. Aplopus becomes active by a much brighter light in the evening than that by which it comes to rest in the morning. Both reactions are, how- ever, responses to light and not to a physiological periodicity, as may be shown with dark-room experiments. If these insects are blinded by paint- ing their eyes with lampblack paste, they still respond to light and darkness, although much slower than normally. 3. They gave no response to sunlight lacking the ultra-violet rays, and were equally indifferent to red and blue lights, acting in all as though they were in ordinary daylight. 4. Aplopus gives no indication of hearing sounds of various intensities. 5. These insects during the day, while inactive, may be made to assume Habits, etc., of the “ Walking-Stick” (Aplopus mayert). 59 an almost endless variety of positions, any of which they will maintain for a considerable period of time. They may actually be piled over one another, with their backs down and legs extended in the air, as if they were inani- mate sticks. Such stick-like indifference may often assist them to pass un- noticed by enemies that might otherwise be attracted by their movements. 6. While moving about, the antenne are often waved or circled in front of their heads, as if feeling the way. Should the antennz be removed, the forelegs are readily pressed into service as feelers, these being waved much as if they were true antenne. If now the first pair of legs are removed it is interesting to find that the legs of the second pair are alternately waved about and used as feelers, although normal insects were never seen to use either leg of the second pair in such a manner. Aplopus, considering the length of its body, progresses remarkably well with only four legs. If such a four-legged, antennaless individual has its eyes blinded, it becomes much confused and often turns in a circle, and twists its head from side to side as it walks. It improves in its movements with practice. 7. The abdomen of a mature female was cut off between the first and second segments and tied to a stick which was supported on wire legs. Males in a dark-room were found to copulate in a normal manner with this amputated female abdomen fastened on the stick. This is a unique case of a male insect’s pairing with a removed portion of the female. Such an experiment makes it evident that a courtship or psychical response is not essential between the two sexes in mating. ; hr) ment ay as > ann pc tax tg a Newey Ap Mastin! bg Tes deacon hele ee | ee ie a a y 4 ‘| one >On aes ~ © yy = & Ab@ Ii, SUUDIES OF TISSUE GROWTH. 1. ON EXPERIMENTAL STUDY OF THE RATE OF REGENERATION IN CASSIOPEA XAMACHANA (BIGELOW). By CHARLES R. STOCKARD, Instructor in Comparative Morphology, Pathological Laboratory, Cornell University Medical College, New York City. 29 text figures. 61 ia : ae. ATWO SUeAtT AG ST ete eS AWITAN. ANT TW FULTS JATAIMIAT INS Me AVATOAMAY AIMOWSAT AY HORA : : * WOM 4 _ tA C2, SSD ie - tabiuprred eullmeigen’) «| Grade 7 veal Poem jal & ——- _ Lan + am eS (pee | ew AN EXPERIMENTAL STUDY OF THE RATE OF REGENERATION IN CASSIOPEA XAMACHANA (BIGELOW). By CHArtEes R. STOCKARD, INTRODUCTION. The suggestion has been advanced by Zeleny (1903 and 1905) that the greater the degree of injury, up to a certain limit, the more rapid will be the rate of regeneration. Zeleny’s studies were based on the regeneration of the limbs in crustacea and the arms of the brittle-star, Ophioglypha. He was unable to offer any satisfactory explanation of why the regeneration rate increased with the amount of injury, but advanced several suggestive hypotheses which are subject to experimental test. First, he pointed out that the animal with the greater number of appendages removed might exercise the regenerating ones more vigorously than does the animal with the smaller number removed. In other words, activity should increase the rate of regeneration in animals. Child (1904) had also been led to think that some regulating influence was exerted over regenerating tissue by movement and nerve impulses in the flat-worm, Leptoplana. I have suc- ceeded in devising two different ways of testing the influence of rest and activity on the regenerating tissues of the medusa and find no increase in the rate of regeneration to result from activity. It was also suggested that the amount of available food might regulate the rate of regeneration. Those crayfish most injured have more food to draw from, since the other appendages are not present to take their share of it. Morgan (1906) has subjected this question to thorough investigation and finds that the amount or rate of differentiation of the regenerating organ is independent of the food supply, although the size of the organ is greater in well-fed individuals than in starved ones. “So long as there is enough food material in the blood or other fluids of the body to allow growth to take place at all it goes on at a rate determined by the peculiarities of each level, and largely independent of the food supply.” Here Morgan mentions one of the most interesting points connected with this subject—that is, the in- fluence of different levels of the body, or of an organ, on the rate of regen- eration. In the fish’s caudal fin it was found that new tissues regenerated faster the nearer the cut was to the base of the fin, and slower the nearer 63 64 Papers from the Marine Biological Laboratory at Tortugas. the cut to the free end of the fin. At first thought this statement seems only a different way of saying, “the greater the amount of injury the more rapid will be the rate of regeneration.” This is not true, however, as it was shown that the rate of regeneration varied with the shape of the cut in a manner not always correlated with the extent of the injury. Experiments will be recorded in the present paper which seem to contrast the two factors dis- tinctly, as well as to show the peculiar influence of the level at which the cut is made. Again, Zeleny offers the interesting conjecture that the uninjured chelz may be assumed to exert a retarding influence upon the growth or regenera- tion of all the others. When one chela is removed the number of uninjured limbs remaining is greater than when both chelz and the last two pairs of walking-legs are removed. The retarding influence with one chela gone, if the supposition be true, is greater than it is when more limbs are removed and correspondingly the rate of regeneration in the former is slower than in the latter case. Such an explanation when modified might be applied to regeneration in the salamander, the fish, and the medusa in the following way: When these animals are cut at various levels they regenerate faster the farther the cut surface, within certain limits of course, is from the ex- tremity or limits of the animal’s body. A fish’s tail-fin grows faster from a straight cut near the base than from a similar cut near the end of the fin. The medusa regenerates tissue faster the farther away from the periphery the cut is made, as though the more tissue removed the less uninjured body- surface remained to exert a retarding influence. The above considerations suggest the question of the limits of growth; as the body nears its adult or normal size the rate of growth becomes slower. It is also true that the regenerating tissue grows slower as it reaches the limits of the former body-surface. Morgan (1906) has expressed this idea as follows: If we can find the explanation for the cessation of growth at the proper terminus we can probably find also an explanation for the difference in the rate at different levels, for, as can be shown, the two things appear to be one and the same. In other words, as the new part grows larger its materials change, and this change is of such a kind that it leads to the cessation of growth. Hence starting under different con- ditions at different levels the same end result will be reached in all cases, and when the terminus is reached the growth should slowly decline, as we find in fact that it does, Emmel (1906 and 1907) has arrived at opposite conclusions after a study of regeneration and molting in the lobster from those cited above as ob- tained by Zeleny on the crayfish. Scott (1907), from a study of regenera- tion on the fins of Fundulus, reaches conclusions differing both from those of Zeleny and Emmel as well, since he finds that the degree of injury exerts no influence whatever over the rate of regeneration in the fish’s fin. Be this as it may, the fact remains that in the salamander, the fish, the earthworm, Rate of Regeneration in Cassiopea xamachana. 65 and the medusa the rate of regeneration does vary under various conditions of injury, but depends upon the body-level at which the cut is made. The crustacea seem rather unsatisfactory forms for the study of such problems as the rate of regeneration. They must molt before the regenerating portion can be observed and the time between molts is often greater than the time which would be expected as necessary for the given amount of regeneration to take place. There is likely a period of cessation of regen- erative growth preceding each molt. Animals which have a continuous growth of regenerating tissue seem much better adapted to these studies. The experiments here recorded were conducted in the Laboratory of Marine Biology of the Carnegie Institution of Washington, at Dry Tortugas, Florida, during the summer of 1907. I wish to express my thanks to the Director of the Laboratory, Dr. Alfred G. Mayer, for many kindnesses extended me while there. MATERIAL. The rhizostomous scyphomedusa Cassiopea .amachana is very hardy. It attains a large size, 15 or 20 cm. in diameter, and is particularly suited to regeneration studies, as several experiments or cuts may be performed on one and the same individual where the conditions are as near similar as would be possible to obtain. Further, since all portions of the disk seem capable of regeneration one may thus work on the animal’s body as well as on its tentacle-like appendages. Of exceptional importance is the fact that the circular disk will admit of variously patterned cuts which are impossible on animals with a differently shaped body. Finally, the disk pulsates rhyth- mically in a manner subject to the control of the experimenter, thus enabling him to test the influence of motion, or activity, on the regenerating tissue in a way not offered by any other animal yet experimented upon. These medusz are easily kept for long periods of time in small aquaria by merely changing the water every two or three days. They live for some time without taking food. One may collect them in abundance from the moat which surrounds the old Fort Jefferson at the Tortugas Islands. The water in this moat is about 4 to 6 feet deep, being rather stagnant at times. Here Cassiopea seems to thrive, and large numbers of them are to be seen lying upon the bottom with their mouth-arms turned upwards, resembling bunches of dark-colored moss. RATE OF REGENERATION FROM THE PERIPHERY OF THE DISKS WHEN CUT AT VARIOUS DISTANCES FROM THE MARGIN. It is well to consider first the less complex cases in which an attempt was made to determine the difference in regeneration rates from cut surfaces on the disk of Cassiopea at various distances from the margin. Medium-sized medusz were selected for the experiments, and the cut consisted in each case of the removal of a peripheral strip from the entire disk. Such an 5 66 Papers from the Marine Biological Laboratory at Tortugas. operation leaves the jelly-fish without marginal sense-organs and, therefore, its rhythmical contractions cease until a slight epithelial rim has regenerated, which serves to reéstablish the pulsation. This new tissue is itself unable to contract, yet it is the seat of the stimulus which causes the disk to pulsate. Two jelly-fish, each about 86 mm. in diameter, were cut around their entire periphery so as to remove a strip of tissue 10 mm. across (fig. 1). Two other medusz were cut, in a similar manner as near as possible, and in addition their mouth-arms were removed, so that they were entirely de- IS, Fic. 1—Diagram indicating method of cutting. : Fic. 2.—Stippled border shows newly regenerated tissue from cut periphery. New tissue widest where ends of mouth-arms (M4) press against it. prived of all means of obtaining food. The former are designated in table t as Nos. 1 and 1A, the latter as Nos. 2 and 2A. By referring to this table the rates of regeneration from the cut peripheries may be readily ascertained for the three medusze, Nos. 1, 1A, and 2; 2A died soon after the experiment had started. Nos. 1 and 1A were pulsating two days after the operation and in six days they had grown a rim of new tissue 3.3 mm. wide about their cut peripheries. The regenerating tissue then began to thicken and did not in- crease further in width until after the tenth day. On the fourteenth day the sense-organs were slightly indicated; from this time until the thirty-fifth day there was only a slight increase in the radial width of the regenerating rim until it reached about 5 mm. across, or was one-half as wide as the piece originally removed. During this period, however, the regenerating tissue was becoming thicker, until it had attained the normal thickness of the disk for the given level; further differentiation of the sense-organs was also taking place. At the same time it must be remembered that the animal as a whole was constantly becoming smaller for want of food, so that the disk of No. 1, which measured 66 mm. in diameter after the operation on June 13, measured only 40.6 mm. on July 18, or 35 days later. Thus the amount of regenerated tissue is to the diameter of the disk almost as much as the amount of tissue removed was to the original diameter after the operation was performed. Rate of Regeneration in Cassiopea xamachana. 67 Taste 1—Regeneration from the cut periphery of the medusa-disk after removal of circular strips of various width. Date. Remarks: Exp. Exp. Exp. Exp. Exp. Exp. Exp. € 1a 2. | 2a. an 3@- | 4. | June 13 | Disk diameter before operation. 86.7 | 86.7 | June 13 | Width of removed margin. ...... | 10 10 | June 19 | Width of new tissue rear | Bee) | June 21 | Width of new tissue ., See tS ere) | June 23 | Width of new tissue..... Yen REE | June 25 | Width of new tissue ..... 4.3 4.7 | June 27 | Width of new tissue..... 4.7 | 5 June 30 | Width of new tissue ..... 4.7 4.7 July 3 | Width of new tissue Send} le By 4.7 | uly Ge Disk, diameters..-.-.s2.58-sesceenes Bechy |) Zt | July 6 | Width of new tissue Adee | ess 4.5 | July 9 | Width of new tissue ..... SY fails | July 12 | Width of new tissue 5 5 July 18 | Disk diameter............ 40.6 33.3 July 18 | Width of new tissue Bal eedey, Date. Remarks. Exp. 4a. | Exp. 5. - 5a. Exp 50 | June 13 | Disk diameter before 93-3 | | operation. | June 13 | Width of removed | Disk center | All of disk | Small strip of | Same as 52. | margin. only re-| removed. disk tissue mained. | left. | June 19 | Width of new tissue..... 6.7 | INonel seerect ss None. | June 21 | Width of new tissue..... 6.7 | Regenerating | Regenerating from disk | from disk tissue. | tissue. June 23 | Width of new tissue.....| Gi7e |) Nonercec..ce Same as on Same. 2tst. June 25 | Width of new tissue..... Bia) |) Noneb....s:.. Film of tissue | Almost same over entire as 5a. | top. June 27 | Width of new tissue..... (thick) 4 | None.........., Same as on! Same. | 25th. June 30 | Width of new tissue..... ar | As on 25th. | As on 25th. | July 3 | Width of new tissue..... 3 As on 25th. Dead. July 6 | Disk diameter ........... 21 | ae See | ees ee uly 6 | Width of new tissue..... - fad Sn Wade lab meen Geen al es Thin film over top (aboral) | July 12 | Width of new tissue... 4. Seek: ek pear) July 18 | Disk diameter ............ | 20 poorer fig ome July 18 | Width of new tissue.....| 6.5 | PLCS Olds ssc. * Mouth-arms also removed. + Arched aborally, not measured. It was observed that those portions of the regenerating rim which were touched or pressed against by the mouth-arms of the meduse regenerated faster or grew out wider in a radial direction than the intermediate portions which were not so pressed by the arms (fig. 2). This condition may pos- sibly be due to the mechanical pressure of the mouth-arms against such places causing them to thin or flatten out, thus giving a more rapid radial growth, whereas the entire mass of tissue may be no greater here than from other parts of the regenerating surface. I made no observations, however, at the time of the experiments to ascertain whether the new tissue was thin- 68 Papers from the Marine Biological Laboratory at Tortugas. ner at these places where the mouth-arms pressed. Regeneration also pro- ceeded at a faster rate in the irregularities of the cut surface. This case will be fully considered in a following section. No. 2, which was cut in the same way as Nos. 1 and 1a and in addition had all of its mouth-arms removed, regenerated somewhat more slowly at first, although obviously the most injured of the three. Later, however, it showed almost as much regenerated tissue as either of the other two. During the observations this medusa showed a very peculiar condition; the periphery of the disk became arched aborally and the regenerating tissue was thus also directed aborally, being finally so folded over that the animal became cup-shaped (fig. 3). The regenerating tissue then grew toward the center, and by fusing the edges of its periphery changed the cup into a hol- low sphere. This condition was also observed in several other experiments and may be explained thus: The muscles being slightly out of the nor- Te @ Fic. 3.—A, top view of aborally arched disk; new tissue (R) regenerating from cut periphery grows toward center. B, cross-section of such specimen. C, new tissue completely fused over top, converting former disk into hollow sphere. mal condition of coérdination, those expanding the disk act more strongly than the oral contractors and the periphery is thus gradually directed more and more aborally. The new regenerating tissue has a tendency to fuse if two of its surfaces are brought together so that when its periphery is folded aborally and the edges come together they fuse and form the hollow sphere. A similar balloon-like condition has been recorded by Hargitt (1899) in Gonionemus. Wargitt was unable to produce such a condition artificially, although he tried in several ways to do so. Two other medusz, designated in the table as Nos. 3 and 3A, had a strip 16.5 mm. wide taken from the peripheries of their disks. These disks are, therefore, more injured than the first and they are also cut at a deeper level. After 6 days they had regenerated a rim of tissue slightly wider than that of Nos. t and 1A, and after 10 days the rims of the latter were only half as wide as those of Nos. 3 and 3A. From this time the periphery of No. 3 became abnormally arched and its regeneration was slightly modified, yet it continued ahead of Nos. 1 and 1a. The disk of 34 remained flat and the regenerated border here increased rapidly in width for 12 days and then commenced to thicken and ceased to grow in width; the sense-organs began Rate of Regeneration in Cassiopea xamachana. 69 to appear after 23 days. The disk had resumed its rythmical pulsation in 2 days after the operation. It will be found by a study of the table that after about 12 days the regenerated tissue began to decrease in width. This fact may be explained by the thickening which the new tissue commences to undergo at this time, or again it may result from the causes which tend to make the entire disk gradually decrease in diameter, until after 35 days it is little more than half as large as it was when the experiment began. Nos. 4 and 4A were cut so that only the center of the disk covering the bases of the mouth-arms remained. From No. 4 a strip almost one-third as wide as the entire diameter of the medusa was cut away. This disk was 90 mm. in diameter before the operation and only 34 mm. after the removal of the strip. It must also be kept in mind that the cut surface at, this level is very thick, since the disk is thickest at the center and becomes thinner as the margin is approached. No. 4 died soon after the experiment started, as is indicated in the table. No. 44 was healthy and within 6 days had regenerated a rim of tissue from its cut surface which was almost twice as wide as that observed in any of the above experiments. After 12 days, here again, the regenerated strip ceased to increase in width, but continued to become thicker. Finally, as is shown in table 1, the rim of new tissue actually began to decrease in width as it had in 3a. The deep-cut surfaces when regenerating first grow a wide, thin rim of tissue which finally begins to thicken at the expense of radial growth till the normal thickness of the disk at the given level is reéstablished. It will be seen that regenerating tissue from a cut surface near the disk margin widens slowly, but almost continuously, as at this level the disk substance is very thin and no subsequent thickening of the regenerated tissue is necessary. Three medusz were now cut so that in two individuals only a small bit of disk remained attached to the mouth-arms and in the third the entire disk was removed. The object of such operations was to ascertain whether the mouth-arms were able to regenerate a disk, or disk-tissue. It was found that the very small portions of the disks which remained would regenerate new tissue, but the mouth-arms were incapable of regenerating from their bases, although they healed the wounded surfaces and lived for 29 days after the entire disk had been removed. Other experiments on removing strips of various widths from the periph- ery were made and results closely similar to those above were obtained. One must then conclude that the disk of Cassiopea begins to regenerate its margin at a faster rate the nearer the cut is to the center of the disk. A small individual regenerates proportionately faster than a large one. These results are closely similar to those obtained by Morgan on the salamander, fish, and earthworm, and by comparison show that the rates of regeneration differ at different levels of the body, and further that (as in embryonic growth) the nearer the normal body-size or form is approached the slower 70 Papers from the Marine Biological Laboratory at Tortugas. will be the rate of the regenerating growth. Miss King (1898) finds in Asterias that the rate of regeneration is greatest from the disk and decreases directly towards the tip of an arm. It is also true that those medusa-disks cut nearer the center are the greatest injured and according to Zeleny would be expected to regenerate their removed tissue fastest, just as they really do. It so happens that the difference in level and the amount of in- jury are often closely associated. I shall, however, cite an experiment be- low which serves to contrast the two and shows the level of the cut to be the more important factor in regulating the rate of regeneration. RATE OF REGENERATION FROM DIFFERENT PARTS OF VARIOUSLY SHAPED CUT SURFACES. For the study of problems relating to the rates of growth from surfaces partially cut as compared with those entirely cut, and the rates of growth from different parts of the same cut surface, Cassiopea offers exceptional opportunity, since the disk-body itself may be cut in sundry patterns and the regeneration rate observed in the several cases. Morgan (1902 and 1906), from a study of regeneration in the fish’s fin, has contributed a number of valuable observations bearing on the question in point. The caudal fins of Fundulus and Carassius were trimmed in different ways, and it was found that partially cut surfaces regenerated slower than entire surfaces cut at the same level; also that new tissue grew out at a faster rate from certain parts of all cut surfaces than from other parts. Since Morgan’s experi- ments were confined to the manner of regeneration from fins or appendages, I determined to make similar cuts upon the disk or “ body” of the meduse to ascertain whether the same principles in regeneration would hold. The results show not only that the same manner of regeneration is adhered to in the body and in the appendages of the two animals, but further, that the forces controlling or determining the regeneration rate on various parts of the cut surfaces act similarly in animals as different as fish and medusz, almost at the opposite ends of the animal series. Straight cuts were made upon the disks of medusz in the following ways: First, a single piece was cut from the disk, as shown in figure 4. Second, two such pieces were cut off as indicated in figure 5, and lastly three pieces were removed as in figure 6. Five individuals were cut in each way and different-sized pieces were removed. The course of regeneration followed by each of the cuts in all of the 15 medusze was practically the same. The history of one set, consisting of one of each kind of individual, will answer for all. The specimen having one cut will be designated as A, the two cuts as B, and the three cuts as C. From A a portion of the disk was removed that measured 32 mm. wide at its broadest place and included 6 of the 16 marginal sense-organs. Four days later the regenerated tissue from the cut Rate of Regeneration in Cassiopea xamachana. 71 was shaped as indicated in figure 7. The middle part of the cut, which is the deepest part and nearest the disk center, regenerates faster than the sides. After a number of days the middle part goes a little slower and when the cut is 20 days old the regenerated tissues from different parts of the cut sur- face are about the same widths, although the middle portion is the thicker. Sense-organs commence to form from the new tissue at this time. The new tissue, being weaker than the other parts of the disk, is sometimes pulled aborally and somewhat folded or puckered, so that it is difficult to measure accurately, though during the first 25 days of the experiment the regenera- tion rate at different portions of the cut may be accurately measured. The manner of regeneration from the two cut surfaces of B is identical with that from the single cut of A. In both, then, the rate of regeneration Fics. 4, 5, 6—Diagrams indicating ways in which disks were cut to give one, two, and three straight cut surfaces. SO, sense-organs. Fics. 7, 8.—New tissue (stippled). Fic. 9.—Top and side views of disk cut as shown in fig. 6. During regeneration the intact corners became aborally arched, modifying the manner of growth and producing hollow spheres with opening at top. is retarded at the marginal corners of the cut, so that the mid-portion grows ahead of the lateral parts (see fig. 8). The three cut surfaces of C (fig. 6) follow the same course of regenera- tion as do those of A and B. Disks cut in this way, however, seem espe- cially inclined to turn their three intact corners aborally, and in so doing the cut surfaces, instead of remaining straight form angles. It will be shown more in detail later that regeneration proceeds much more rapidly in an angular cut than from a straight surface, since the two sides of the 72 Papers from the Marine Biological Laboratory at Tortugas. angle seem to reenforce one another in regeneration, a kind of summation of regeneration occurring. Through such a process the disk is converted into a ball-shaped body with a small triangular opening at the top, where the three uninjured corners are brought almost together (fig. 9). It will be recalled that the fins of the gold-fish and Fundulus, when cut straight across, begin to regenerate their new tissues faster in the middle of the cut and slower near the corners, a fashion identical with that fol- lowed by the disk of the jelly-fish. The meduse-disks were next cut in such patterns as to give what Mor- gan has termed “ partial cut surfaces” (figs. 10, 11, and 12). Such cuts were varied in the width of their different parts as well as in depth. Many individuals were prepared in the several ways. The deep part of the cut shown in figure 10 must be wide, since it shows a strong tendency to close its walls together after a week or two (fig. 14). The history of the regeneration from such a cut surface may be recorded in detail. The cut was made so that the bottom of the deep part was 23 mm. from the peripheral margin at its most distant point; this part was 26 mm. in width; the lateral shallow parts of the cut were each 19 mm. wide and 1o mm. below the margin at their middle point. Four days after the operation the regeneration was perceptibly greater from the deep-cut sur- face than from the lateral shallow surfaces, and within six days the middle part had almost overtaken the lateral surfaces (fig. 13). The regeneration in the deep cut really takes place from three surfaces, the bottom and the two sides of the cut, as here there is free opportunity for lateral regenera- tion, thus differing from the case of the fish’s fin, where the fin-rays seem to prevent lateral regeneration, since they are only capable of growing out from the stumps of the old rays. Ten days after the operation there was 13 mm. of regenerated tissue from the deep cut and only 10 mm. from the lateral shallow parts. After 14 days the deep cut had become so pulled together that there was only 5 mm. between its original walls. When 20 days old the regenerated tissue had rounded across its free margin and was now growing out as one piece. After 23 days the old sides of the deep cut were only 3.3 mm. apart; the regenerated tissue over it measured only 10 mm. and over the shallow parts 7mm. This loss in width may be either due to the thickening of the new tissue which is taking place, or may be on account of the general decrease in size which the medusa has undergone, measuring now 63 mm. in diameter, whereas it was 77 mm. across when the experiment began. The new tissue from the deep cut after the twenty-third day began again to increase slowly in width until when 35 days old it was 14.5 mm. wide and that from the shallow parts was 8 mm. The original walls of the deep cut were almost drawn together, being only 2.3 mm. apart. The entire cut had tended to contract, so as to take an angular form, as illustrated in figure 14. All Rate of Regeneration in Cassiopea xamachana. 73 medusz operated on in this fashion regenerated similarly to this one. Their manner of regeneration may then be briefly summarized as follows: The rate of proliferation of new tissue is faster from the deep partial- cut surface and slower from the lateral surfaces. The angles of the deep partial cut assist in the regeneration process and thereby help to make it proceed faster from this portion of the cut, whereas the corners of the lateral cuts seem to exert a retarding influence over the rate of regeneration Fics. 10, 11, 12.—Diagrams showing manner of cutting meduse disk to test regeneration rates from partial cut surfaces. Fics. 13, 14.—Stippled areas indicate course of regeneration from cut surface of pat- tern, fig. ro. Fics, 15, 16.—Course of regeneration from such a cut as shown in fig. 11. Fic. 17—Showing manner of regeneration from fig. 12. (see fig. 13). Such a conclusion is identical with that reached by Morgan in his study of the regeneration from similar cut surfaces on the fish’s fin. 74 Papers from the Marine Biological Laboratory at Tortugas. We may now consider regeneration from surfaces cut in practically the opposite manner from those just recorded. The lateral cuts are deep, with a high middle tongue-piece (fig. 11). Many medusz cut in this fashion regenerated tissue in a similar way. The exact history of one of the indi- viduals is as follows: The disk was cut so that the lateral surfaces were 26 and 41 mm. wide, respectively, and the high tongue-piece between them was 14.5 mm. wide and 10 mm. high, or above the level of the side cuts (fig. 11). Six days after the operation the newly proliferated tissue was widest on the two side portions and narrow from the middle piece. The corners of the high middle part seemed to exert a retarding influence on the regenerative processes, as did also the outer or marginal corner of the lateral cuts. The inner corners of the side cuts were, on the other hand, the places of greatest regeneration, as no doubt the lateral and basal surfaces both con- tributed to the process (fig. 15). Nine days after the operation the re- generated tissue from the lateral cuts was 5 mm. wide, while that from the middle piece was only half as much. On the twelfth day the conditions were about the same. The fifteenth day gave the side parts 7 mm. of new tis- sue, while the middle part had proliferated tissue only 2.3 mm. wide. At this time the old border of the middle piece is 8 mm. wide, while the lateral parts are 16 and 5 mm. respectively. When 21 days old the regenerated tissue had rounded its border (fig. 16) and measured 7 mm. deep over the side cuts and 3.5 mm. over the middle part. From this time until the twenty- seventh day the middle part continued to grow out new tissue, while the side portions seemed to have completed themselves. Regeneration from such a cut surface may be thus ginmanacd The lateral cut surfaces produce new tissue faster than the high middle piece. The outer corners of all the cut surfaces seem to exert a retarding influence on the rate of regeneration, while from the inner corners of the lateral cuts new tissue is formed at a very rapid rate, which is probably due to a sum- mation of regeneration. It will be again recalled that an identical condition exists in the regeneration from similar cuts on the fish’s tail. Medusz were also cut in such a way as to test the rate of growth at dif- ferent levels on one and the same individual. Here, obviously, the conditions of nutrition and vigor must be as nearly identical as possible. At one place on the rim of the disk a piece was cut out which was 10 mm. deep at its broadest part. Opposite this cut, or 180° away, a second piece of the disk, including an arc of the same extent, was cut away to a depth of 16 mm. from the highest point of the are (fig. 12). When one cut is narrower peripherally than the other, the rates of regeneration are not readily com- pared, since regeneration proceeds more rapidly from a narrow cut than from a wide one at the same level. After six days the regenerating tissue was broader from the deep than from the shallow cut, although here it has a thicker base of tissue to grow Rate of Regeneration in Cassiopea «amachana. 75 from; this is also true for the fish’s fin, where the deeper cut has a thicker base. Two measurements of the regenerating tissue were made, since the thick base was not exactly a perpendicular surface. The one was from the edge of the old tissue on the aboral surface to the edge of the new tissue, which measured 5 mm. over the deep-cut surface; the other measurement was from the oral border of the old tissue to the margin of the regenerated tissue, 3.5 mm. wide. The rate of regeneration was fastest at the corners in these cuts, being 7 mm. wide at this place in the deep cut. The shallow cut showed 1.5 mm. of new tissue from its middle and 5 mm. from its cor- ners (fig. 17). After 9 days the deep cut had regenerated tissue 7 mm. wide from its middle, while the shallow cut showed only 3.5 mm. of tissue. Both of the cuts were at this time 13 mm. in width peripherally. When 18 days old the cuts were 10 mm. across between the vertical edges of the old tissue, the deep cut had regenerated new tissue 7 mm. wide and the shal- low 3.5 mm., or half as much. Here again regeneration proceeds in one and the same individual at a faster rate from the cut surface at the level nearer the disk-center than from a similar more distal cut. I may now cite an experiment which was made to test whether meduse would regenerate their sense-organs faster when consecutive ones were re- moved or when alternate ones were cut away. The experiment threw no light on this question, but the result was curious and for this reason may be mentioned. Two healthy medusz, one with 16 and the other with 17 mar- ginal sense-organs, were treated as follows: Four adjacent sense-organs were removed from one part of the disk and three alternate ones from the region opposite these. After 23 days no definite trace of regenerating sense-organs could be detected, so all of the remaining old sense-organs were cut away to ascertain whether the new ones were sufficiently regenerated to maintain the pulsation of the disks. The disks became perfectly still after the last one of the original sense-organs was cut off, and only after a period of 6 days was one individual slowly pulsating. This is peculiar, as when the entire peripheral border with all sense-organs is removed the newly regen- erated tissue causes the disk to pulsate usualiy after two or three days. Further, a number of medusz with regenerated margins had produced sense-organs from their new tissue, while the two above had not regenerated them from their old bases. REGENERATION AFTER THE REMOVAL OF PIECES OF ORAL EPITHE- LIUM OF DIFFERENT SIZES AND AT DIFFERENT DISTANCES FROM THE DISK-CENTER— THE QUESTION OF ‘REGENERATIVE PRES- SURE.” These experiments were carried out with the hope of testing whether or not “regenerative pressure,” in the sense Morgan (1906) used the term, actually exists and exerts itself from the center radially to the periphery. In other words, is this force felt more towards the center and gradually 76 Papers from the Marine Biological Laboratory at Tortugas. less, as the limits or periphery of the body is reached? This pressure is responsible for the “ gradual slowing of regeneration as the normal form is approached, and it is apparent that this retardation will be the same, whether it occurs near the end of an old part or as a new part approaches completion.’ If this be true it ought also to follow that the pressure con- ditions of regenerative forces are greater in the center than at the periphery of the disk in Cassiopea. In the experiments, only the oral epithelium and thin superficial muscle- layers were removed. It may be that such experiments are not conclusive, since this pressure might exert itself outward from the face or cross-cut area only, and not so clearly on the surface. At any rate, as will be seen, the results do not lend particular strength to the idea of greater regenera- tive pressure near the center. Six medusz were operated upon as follows: From the oral surfaces of two individuals, Nos. 1 and 1a, two rectangular pieces of epithelium and underlying muscle were removed. The removed tissues had the same width in a radial direction and were equidistant from the periphery, while one piece was longer than the other in the direction parallel to the circumference of the disk (fig. 18). If the pressure exerts itself only in a radial direction, then the two cuts should regenerate at the same rate independent of their peripheral lengths, since they are equally wide. Two other medusz, Nos. 2 and 2A, had two equal-sized pieces removed from their oral surfaces, one piece being nearer the center than the other (fig. 19). The last two, Nos. 3 and 3A, had one piece running in a radial direction cut from each, as seen in figure 20. Mayer (1906) has shown that when the epithelium is scraped away on the oral surface of Cassiopea an electrical stimulus applied on one side of the abrasion is unable to pass over and stimulate the tissue on the other side. As soon, however, as a very delicate layer of tissue is regenerated over the cut place the stimulus will be transmitted across. This affords a delicate means of detecting the first trace of regeneration. Twenty-four hours after the above operations No. 3 was scratched on its oral surface so as to divide it into a series of concentric rings (fig. 21). The rings were then scratched across at a place opposite the removed radial strip. The rings of tissue were thus broken at one place, so that no impulse could pass from one of their halves to the other unless tissue had regen- erated over the radial injury sufficiently to conduct the stimulus. It would be expected that the inner ring should be the first to conduct. None of the injuries had regenerated sufficiently after 24 hours. Two days after the operation Nos. 1 and 1A did not transmit across their injuries. No. 2 transmitted the stimulus across the inner area only, although this was equal in extent to the more peripheral injury. The clear, transparent regenerating epithelium could now be seen, and it was noticed Rate of Regeneration in Cassiopea xamachana. ae that the growth was greatest at the two ends instead of in the radial direc- tion (fig. 194). This is probably due to the corners being nearer together at these ends, and regeneration takes place from both sides of the angles, such a summation causing it to proceed faster. Four days after the operation all of the scars had regenerated tissue sufficiently to cover them completely over. The smaller places had regener- ated sooner than the larger ones, yet a comparison of rates of regeneration is difficult to make, since the wounds tend to draw their walls together and IGA Fics. 18, 19, 20.—Positions and patterns of pieces of oral epithelium cut away to test force of regenera- tive pressure from disk-center radially outward. Fig. 194, New tissue growing over rectangular wounds. Fic. 21—Manner of scratching circles through oral epithelium to test rate of regeneration from dif- ferent parts of superficial radial wound. 2f thus close at the same time that the regeneration is in progress. On the whole this experiment is unsatisfactory. A somewhat similar experiment was arranged to test the rates of regen- eration of epithelial coverings over wounds of different sizes and others of the same size at different distances from the disk center. The sizes of the holes were regulated by means of a sharp cork-borer, which could be used to cut out small circles of exact diameters. Nos. 1 and 1A each had three cir- cular wounds 10 mm. in diameter at 10, 16, and 20 mm. from the margin. Nos. 2 and 2A had three circles scraped, each about 24 mm. from the disk margin and over radii leading to the sense-organs. The circles were 7, 8.5, and 10 mm. in diameter. On 3 and 3a four circles each, 8.5 mm. in diameter, were scraped, 20 mm. from the margin, two of the wounds being over radii leading to sense-organs and two midway between such radii. All four are, however, immediately below radiating canals, so that the difference in regeneration rate, should any be observed, might be attributable to their different nervous connections (fig. 23). 78 Papers from the Marine Biological Laboratory at Tortugas. Two days after the operation Nos. 1 and 1A showed their outermost circles with regenerated films about half over them; the inner circles were in the same condition, but the two circles occupying intermediate distances from the margins had regenerated films of epithelium which entirely covered the wounds. In No. 2, where the three circles were equidistant from the margin but of different diameters, the largest and smallest circles were still not completely covered over, while the one of intermediate size was entirely covered. In 2a all of the circles had regenerated coverings entirely over them. Nos. 3 and 3A showed all of the holes to be 5 mm. in diameter; thus they had become contracted to little more than half of their original size. Those on the sense-organ radii seem a little further covered than those on intermediate radii, though there is very little difference at all. Fic. 22.—Medusa disk with 3 equal sized circular wounds at different distances from margin. Fic. 23.—Disk with 2 circular wounds over radii leading to sense-organs, and 2 exactly similar wounds between sense-organ radii. SO, sense-organs. Fic. 24.—Regeneration from 3 circular oral wounds. With these circular cuts one eliminates the angular regeneration factor mentioned in the experiments above, and it was noted in all cases that the film was widest from that area of the circumference toward the disk center (fig. 24). This condition would be expected on the hypothesis of greater regenerative pressure near the disk center, though the deeper level of the cut at this part is a better explanation. Three days after the operation all of the circles were entirely covered over. These experiments are also difficult to draw conclusions from, since the wounds have a tendency to contract while they are healing and regen- erating new tissue. Those nearest the disk center contract most. Thus one might believe them to be more rapidly producing the new tissue. Rate of Regeneration in Cassiopea xamachana. 79 THE RATE OF REGENERATION FROM DIFFERENT AREAS ON TAPERING PERIPHERAL STRIPS AND REMAINING PART OF DISK — CONTRAST- ING THESEEVEL OR THE CUD ANDIEXTENT OF INJURY. The circular medusa disk offers exceptional material for certain opera- tions that could not be carried out on animals having a differently shaped body. It has been found difficult to perform an experiment which would clearly contrast the rate of regeneration from certain levels with the rate from parts more or less injured. According to Zeleny the rate of regenera- tion will be faster the greater the injury up to a reasonable limit, and ac- cording to Morgan’s pressure and growth idea the rate varies at different levels, being slower as the level is nearer the normal body limits. The con- ditions are usually open to either interpretation, since the least injured ani- mals are the ones with less body tissue removed and necessarily nearer the normal body limits than those with more tissue removed. Fic. 25.—Manner of cutting bias strip from periphery of medusa disk. Fic. 26—Such strip if straightened would form a long triangle with sense-organs, SO, along its base. Fic. 27,—Manner of regeneration from cut edge of the disk center. oc b 1523 Jb June 25 c b 7.9 b June25 | c b 1524 | 3 June 20 c b b b June20 > c b 1525 a July 5 c 6.4 | b b| July 5 c b 1520 3 June 27 c esp lll aie b June 27 c b #1527 23 June25 | c b b b June 25 c b 1528 June 20 c b b b>} “Janezo:|! b 1520 = July 3 c b b b July 3 ( b 1530 rot jaly 25) Bah 5.7 2.0) wiiily, sul arc 2.8 1531 ae July 2 c 3.6 6.0 b July 2 c b 1532 ae June 30 c 4.6 | c b June30 | c b eT 33 Oa ees July 4] c 4.0 8.1 38 July 4! c¢ 3:0 er5ga |) iste 24 || Ye oa b b July 2 c b 1535 rt June 30 c 4.6 7.4 4.4 June 30 c b 1530 = June 30 c b 6.0 b June 30 c b 1537 é Tuly 7a d 6.1 Ke 6.4 July 7 d | 6.4 1538 AS June 25 c b b b June 25 G b *1530 = July <5" || ¢ 4.6 8.2 4.4 July 5] ¢ 4.4 1540 a July 1 c b b b July 1 c b *I541 3 June 29 (s b b b June 29 c b 1542 = June 27 c b b b June 27 c b 1543 | of Jtilys 2) |) ic b b b July 2 c 2.8 1544 re July 5 c b b b July 5 c b *1545 3 July 3 c 3.2 5.8 b July 3 c b 1546 == \) Sjuness c b | b b June 25 G b *1547 ae July 1 c Ebi 5.6 b July 1. c b ThA ee July 2 | ¢ b 5.8 b July 2 C 28 1549 = July 9| d 4.1 HE 3.8 July 9| d 3.9 *15507 3 June 30] c 5.3 8.4 4.4 June 30 c 4.5 I551 3 jay 7) ‘ce b b b July 7 c b 1552 ie July 93) 1G 4.6 8.1 4.4 July 3 c 4.4 D553 hue June29 | c 3.0 ss b June 29 c b 1554 = July 1 G b b b July 1 c b 1555 3 July 1 | ¢ 4.5 7.9 b July 11 c b 1556 me July 4 Cc 3.1 5.6 b July 4 c b 1557 ais June28 | c Br b b June 28 c b 51426. June 25, left chela (probably regener- | ©1533. Plane of removal not at breaking joint. ated) is smaller than right; third | 71550. Not the same species as majority. right walking-leg missing. Regeneration of the Chele of the Gulf-weed Crab. 117 TasLe 1.—Table of data—Continued. Original. etl Sex. | Cephalo-thorax. Left Right chela. Date, Desc. == chela. —— Lg. Width. Date. Desc. Lg | 1558 ae July 1 c b | b b July 1 a b 1559 fof July 8 d 629)" surg 6.3 July 8 d 6.3. | *1560 =e 8July 9 | 8c 6 | rr4 6.1 July 9| c 6.2 1501 ey June 25 c (Sipe |) hie 6.3 June 25 c 6.2 1562 fof July 2 c 4.6 8.6 b July 2 c b *1563 3 July 1 c b 11.6 6.2 July 1 c 6.2 1564 = June 30 c 3.6 57 3.2 June 30 c 32 *15659 | July 4 c 5.1 b 4.8 July 4 c 4.9 1500 == June25| c 5.9 | 6.1 June 25 c 6.3 | 1507 3 June 25 c b | b b June 2 c b | 1568 =2 June 26 c b | 8.6 | 4.5 June 26 c b *1500 2s July 5 c bd bo. Ba ieeiialy 5) ac b *I570 fof June29 | c 4.0 8.4 | b | June 29 c b 1571 - July 2 c 3.0 6.2 | (| disiy Zoi) b *1572 oe ily 2a) c b 6.0 OY) eitly; 2 c b 1573 ae, June 28 c 3.4 5.4 b | June 28 c b 1574 = July 17 d b b b July 17 ad) b 1575 fof July 1 c 2.6 3.8 b July x) ¢ 7 | b 1576 ae July, 6) c 3.0 5.4 b| July 6 Cee 7, *1577 rol julys 2 2.8 | Bu) b July 2 c b 1578 SS \omiiive ar 4c: 3.2 6.0 Zone nly, x c 2.8 15790 -- | June25 CG b b b | June2s5 c b T5SOMN| con i) UM STs ce b b b| July 2 c b 1581 ae June 30 c 3.0 5.5 b | June 3o c b 1582 ae June 26 G a. 5.8 b June 26 c b *1583 -- | 10July 7| d 4.5 7.9 4:3 July 7| d 4.4 1584 | o July 7 c 4.4 b 4.1 July 7 c 4.2 1585 | o June 25 c b b b | June 25 c b 1560) |= June29 oc b 5.6 b = - June 29 c b 1587 | o& Niky Te oc b b el) Apeikee i c b *1588 | of June 30 c 4.6 8.2 b | June 30 c b | 1580 3 June27| c 3.2 5.8 b | June27| c b | 1590 3 June 24 CHa r3 5.9 28 June24 | ¢ 29 | *I5Q1 fof June24 | c 48 9.0 5.0 June24 | c¢ 5.0 — #159211, oi June24oc 3.3 5.4 31 June24 c Chit | 1503 3 June 24 c 4.7 8.5 b June24 | c 4.9 1504 ae June 24 c 28 3.8 b June 24 C b 1505 ae June 24 c 4.8 8.5 b = June 24 c b 1590 aa June 24 c b b b | June 24 c b 1507 3 June24 | c 4-7 8.7 4.7 June24 | c 47 *1508 | Jo June 24? oc b b 6b | June2q| c b | *I59912, June24 | c 2.4 b b June 24 | ¢ b 1600 = June 24 c b b b June 24 | c 4.9 | *j6or | oo June 23 c Bz) 5.6 b June 23 c b 1602 of June 23 Coal sak b b June 23 | G b 1603 | of July 12) ¢ | 409 8.6 4.0 July 12) 4.9 1604 | 3 June 30 c 5.0 8.0 4.9 June3zo |) c b 1605 | 6 June 27 c 4.6 8.2 4.3 | June27| c 4.4 | 1606 Jd July 2 (ee b b b July 2 c b 1607 Jd June 27 6 | b b b June 27 c b *160813 fi June 26 c b b b June 26 c b 81560. Died in molting. _ ar: ; 1599. Not the same species as the majority; ®1565. Not the same species as majority. fourth molt = July 31, k and m, 4.5, 107583. Died in molting. . 7.2, 4.0; July 31, 3d, 2.9. M1592. July 7, first regeneration, removal not 81608. Record not clear; time of original op- at breaking joint; fourth molt = eration not given. July 31, k, 6.8, 12.2, 6.8; July 30, 3d, 4.5. 118 Papers from the Marine Biological Laboratory at Tortugas. Taste 1.—Table of data—Continued. Original. | a Sex. Cephalo-thorax. ett Right chela. | Date. Desc. - Saal) ehelas — | | Lg. Width. Date. Desc. | Lg. | *1609 3 June 30 | c 4.5 7.6 | b Junez30 | c | b / 1610 a i) Sialy Whee 5.0 9.0 | G July 3 b | b IOI Se) ily ee b b b July 12 (RV || b | wqTo6r2 Cu Wh Maly, Gc b b b July, 7)" 6 |} b | *r6r3t4! og | Junezo)| c 3.0 5.4 bal wane:20))) ce" |) b | ¥1614 ==) i) Winlys 3 c 3.0 5.2 b July 3 fe | b | *I1615 ==) i) Seiune 26 c 3.0 5.4 b | June 26 c b | 1616 fof June25 | oc b b b jJune25 | c b *1617 roe June 27 c 4.6 8.1 4-4 June27 | c 4.4 | *1618 ==) |) wiinezo c b b b June 30 c b *161915,) _. | June 26 c 4.6 78 b June 26 G 4.5 *1620 3d | June 28 c 2.9 3.9 b June 28 c b *1621 fof June 28 c b b b June 28 G b *1622 Se) | Wiuly 8 c 2) 5.3 b July 8 c b 1623 a2) || sjune27 c 2.0 5.3 b July 8 c b | *1624 =e it Wine 2s: c ar 5.8 b June 24 c b | #1625 Pe Bulyete d 3.0 5.7 2.7, 1" a iialy; 1 d 207, | 1626 3 | June 29 (2 3.1 5.4 b | June 29 c b | © 1627 ao i July 6 c 4-7 8.5 44°) July; 6 c 4.6 *1628 o | July 6 c 4.5 8.3 b | July 6 c 4.4 #1629 3 June 27 c 4.5 b b | June 27 6 b 1630 = July 14 c 4.0 | 7.0 38 | July 14 c 3.8 1631 a5 July 4 G b | b 4.0 | uly a c b 1632 roe June 25 (5 b | b | b | June2s c b *1633. | -- | June24 c 3.2 | 5.9 | b | June24}. c b 1634 | 6 | June27 c b | b Bal) Siete 27.1 ic b *1635 -- | June 23 | ¢ b b by) Wine 237 Fc b *1636 | of | June 27 c KO) |) r0e7) | 5.8 June27 > oc 5.0 | *1637 3 | June 29 c b | b b June 29 c b *1638. ||. —= July 5 c 38 7.6 b faly, 5) |e 3.6 *763916 __ June 28 c 3.0 ty) b June 28 | c b | *164017) fi June24 | c b b b June 24 c b | 1641 x June 25 ( 4-7 8.5 b | June2s5 c b *1642 se June28 | c b b | b June 28 | c b *1731 aes July 1 c 4.5 8.7 b July, x) c b 1732 ==) Weadiuly: 23 c b b b July 3 c b 1733 2 July 11 c b b b July 11 (= b 141613. June 30, left chela (probably regener- ated) smaller than right. 183619. June 27, right chela (probably regener- ated) larger than left, 4th left walk- ing-leg absent; July 5s, both chele regenerated and of equal size; July 6, right chela (first regeneration) removed; July 21, right chela = small regenerating bud, left chela removed; July 27, both chele absent. | 1639. June 29. Right chela (probably regen- erated) smaller than left, 2d, 3d, and 4th right walking-legs missing; July 20, right chela smaller than left; July 27, both chele present. 1640. June 25, 2d and 3d left and 2d, 3d, 4th, and sth right walking-legs missing. Regeneration of the Chele of the Gulf-weed Crab. 119 Taste t.—Table of data—Continued. First molt. Catalog Sex Cephalo-thorax. | Tete Right chela. | i | Date Desc =| SESE, = —_———| | | | Lg Width. | Date Desc. | Lg. | 1304 oe July 4 | c | 3.9 6.2 b| June26| 1st} 28 | 1307 3 July 9| ¢ 4.3 7:7 4.1 July 9) ¢ 4.1 | *1308 ref July 5| c b b b June 25 | Ist | 182.9 1300 é uly; rr | .¢ 3.0 OM| 3.8 July 11 || oe) | *I400 3 July 6 | c 6.3 b | b July 6 c 6.5 *T4O1 _. | July 10 | ts 5.0 10.0 b June 29 a || BF *1402 = July 11 yenre b 0.8 5.4 June28} Ist | 4.6 | *T403 ref July 22 c 6.8 12, 6.7 July 8 191st | 5.4 | 1404 & | July 12 c 5.9 2 5.8 July 12 @ | ts 1405 oly 27: c 6.2 11.4 6.3 July 8 0 6.3 1400 ree July 8]| c 5.6 10.8 5.8 June 25 0 5.8 | *1407 ol) Muy. 78) | sc 7-3 13.9 7.9 June 25 0 7.4 *T 408 =|, Siunez6) sc b b 6.3 Junerg Ist *1400 Gu |) uly, oO c 7.9 15.0 8.1 July 9 c 8.4 | 1410 Go| wulyarst |! | b b b | June 30 0) 3.2 1554 | -- July 11 c b b b Julys 27a 0 4.4 1555 ri July 31 k 5.0 9.1 4.7 July 12 0 4.9 1550 ce July 19| d Bis) 6.8 B22 yalys 65 0 3.3 1557 ze July 8 c 3.5 | 7.0 gi July 8 c 3.2 nS || se July 12! c 5.5 | 10.4 5.6 July 2| o 5.6 1561 3 July to c 730 || sis yj June 26 0 78 1562 ret uly 150 6 0 || 10:3 5.6 July 3 0 5.7 #1563 3 July 17| c 69 | 134 6.0 July 2| o 7.1 TEG4N || ee July 10| c 42 | 7.0 4.4 July «| o 4.2 *1565 | of uly 0| eve 6.4 0.3 6.1 July 5 0 5.0 1566 a alys (Omar ||\eeeres 13.2 7.4 | June26 0 7.6 1567 3 July 8} c¢ | 60 IT.0 b June 26 0 6.1 1568 = July o] c 5-7 10.4 5.0 June 27 0 6.0 *1560 = July 13 c b 6.8 b July 6 0 3.6 *1570 fof 26July 18 C47 | aro) | 4.5 June 30 0 4.6 1571 =s July 5 d 4.2 75a 4.1 July 3 0 4.1 *1572 = July 13 c b b | b July 3 0 4.2 neyo Me july. Ol) <6 3.9 6.8 | 3.8 June 29 0 3.8 575) |) 1) July 10 CG b b | b July 2 0 2.0 1570) |) == July 31 k | 36 7.0 | 3.2 July 7 0 3.2 *1577 fof July 13 (3 aie) 5.2 b July 3 0 BOE 1578 at July 12 c Bu, 7.6 Bs July 2 0 3.5 1579 ee July 7 d 5.0 10.8 | 5.8 July 7 c 5.0 1580 rol July 12 c 3.4 5.5 | eras July 3 0 3.2 1581 = July I0| c 3.6 fg} Be July 1 0 3.4 1584 3 July 18| d 67) |) rest 6.7 July 8 0 5.0 1585 3 July 11 c 5:0) ||) our 5.9 July 11 c 5.9 1586 ee July 13 c 3.8 7.5 3.5 July 13 c 3.5 1587 fof July 10 eC 3.3 5.4 3.0 July 2 0 3.0 1588 3 July 12 c 5.4 10.0 | 5.5 July 1 0 5.6 1589 é July 8 c b | b | b June 28 0 3.8 1590 rei July 8 c b | Cy All 3.6 July 8 ( La *1s9127| of July to c (owe ee cae 2 6.5 June 25 0 6.7 #1592 a July 6| c¢ | b | b b June 25 | o 4.0 1503 3 July 13] c¢ | b b | b July 13 c b 1504 = July 9 CAEN yeh 5.7 | 3.2 June 25 0 Get 1595 ne July 11 ¢ || S:Onileezo!s | 5.9 June 25 0 6.0 % 1539. Died in molting. % 1547. Died in molting. % 1570, 27 Iso1. Date not certain. Curved. Regeneration of the Chele of the Gulf-weed Crab. 121 TABLE 1.—Table of data—Continued. First molt. See al"iSexe Cephalo-thorax. a | Ripnbchela: aly Date. Desc. | ———— -——- Gab. | == | Lg. Width. Date. Desc. Lg. 1590 = | July 7 c 4.3 LD Act June 25 0 4.2 1507 & | July i c 6.3 aisit | foe) July 11 c 6.4 *1508 3 | uly 5 c b b b June 26 0 4.4 *1500 fof July 4 c b b b June 25 0 2.5 1600 =e July 7 d 6.2 11.8 6.6 June 25 0 6.8 | *1601 | ot July 4 c 4.0 8.1 | 3.8 June 25 0 3.8 | | 1602 J July 11 c 4.1 8.2 3-7 July 11 c ay) | 1603 | o July 28 c 5.8 10.3 5.8 July 13 0 Say | 1604 | o July 19 d 5.7 10.6 5.8 July 1 0 5.3 1605 | o& July 13 c 5.7 10.7 5.8 June 28 0 5.8 | 1606 On) uly rac, 4.5 8.4 48 July 3] o 4.2 | 1607 & | Juner2 c 5.9 10.9 5.9 July 12 c 5.9 | *1608 oo | July 12 c 5.8 10.7 | b June 27 | Ist 5.0 *1600 3d | July 30 c b b July 1 whl es} | 1610 3 July 18 c 5.8 10.6 b| July 4 0 5.8 | 1611 = July 29| c 5.6 10.1 b July 13 0 5.5 r12 | @ | july 20 c b 5.0 July 8 0 5.0 *1613 & | July 16 d 3.5 6.9 27, June 30 0 3-3 *1614 -- | July 13 c b b b | 28July 4 0 2.0 *T615 a= ie elves d 3.5 6.8 | 3.1 | 29June 27 0 b 1616 o | July 8 c b Tr.2 | 6.0 June 26 0 6.1 *1617 3 July 12 c 5.2 9.2 | 5.0 June 28 0 5.30 | *1618 | = July 14 c 6.4 12.0 | 5.8 July 1 0 5.5 SIOIO) |) == July 5 c b b | 4.9 June 27 0 6.1 | *1620 | 6 July 8 c b b b June 29 0 2.9 | *162130, gi July 14 c 4.2 7.3 b July 14 c b | *1622 | ae July 28 d 3.4 6.5 | 3.1 | 31July o 0 i) TOSS eee oes He tye d 3.6 al Bs July 17 d 3-4 | *1624 | -. | July 10 c b b 2.9 | 32June 29 o | 30 | y626 | f || \Jalya2z| 3.6 7.0 b july i | ¢ | 33 | 1627 On ale phulys 17 c 5-4 | 10.0 b July 7 (i | SS | *1628 é& | July 18 c 5.1 0.6 5.2 July 7 0 5.2 | *1620 ee July to c 5.2) | 0.6 5.2 July 10 Ch RSS | 1630 Se ily d 4.9 8.0 4-3 | July 15 Ons a eatay || e633 |e July 15| d 5.2 88 hie | hele 9) a |) a) | | 1632 fof July 12 c b b b July 12 Can || b | | ¥*1633 ce July 3 c 4.0 8.0 | a7 June 25 Oo) “Ii 83:6 1634 Oe mlalyazon| mc 5.5 b b July 9/ c | 48 *1635 aes July 12 c 3.0 8.0 3.7 | June2s ope Sia | *1636 do July 12 c 6.9 13.7 Tas || Akela | re b pares. i id} July 11 c b b b | July 31 c b *1638 =e July 31 k 4.5 0.0 3336 | July 6 0 A | *1630 = (?) c b b | b | June29| o 2.6 | *1640 rot July 2 c b b b June 25 0 6.9 1641 = July 8 c 5.4 10.2 6.4 July 8 c 6.6 | *1642 ae July 13 c 7.2 1277, 7.5 June29 o CAS || *1731 = July 12 c b 66 | July 2}] o 6.6 | 1733 = July 14 c b 7.0 | 3.8 July 4] o 3.9 1733 =e July 24 ( 5.1 08 | 5.4 July 12 0 Bis || 1614. July 4, right chela (probably regener- | 813622, Date not certain. _ated) smaller than left. | 821624. Breaking plane ragged. ™ 1615. Right chela and third right walkingleg | 31638. July 6, both chele removed, left (prob- missing. ably regenerated) is smaller than 801621. Curved. right; July 27, both chelz absent. 122 Papers from the Marine Biological Laboratory at Tortugas. Taste 1.—Zable of data—Continued. | Second molt. | ee San a - | Cephalo-thorax. Left Right chela. ate. | esc. Ss. = " | | Taw lew | |__ Date. | Desc. | Lg. 1304 = July 6 d 4.0 7.0 39 | July 5 | 2d 3.0 | 1307 o | July 21 c 4.7 8.6 | b | July to o | b | *1398 3 | July at c 4.7 8.2 | 4.5 July 6) 2d 3-3 | | 1309 6 | July 19] d 4.5 7.6 4.1 July 12 0 4.1 *1400 3b | July 20 (es 7.3 14.0 | 7-7 | July 7 0 78 | *1401 Sees yuly: 2h enc 6.6 11.7 | 6.6 July 11 | ist _5.0 *140: z= July 31 | c 6.0 Il. | 5.9 July 12) 2d (85) *T403 Cll) wily sree 7.3 T37 7.4 July 23 362d 5-7 | 1404 Ge i Inet che Jae | Oy? 12.2 6.5 | July 13 0 6.7 | 1405 Coe wealyn si k 7.4 13.2 7.4 July 18 Ist 6.1 | *1400 6 | July 22 c 6.4 | 120 6.5 July 9 Ist 5.1 *140 roe July 24 | c¢ 86 | 15.8 0.1 July 9) Ist 77, 1408 ek July 11] c FE |) 13:5 7.2 June 27 37Ist 6.7 *1400 6 || viJiuly, 23) c 9.1 repiy. | 9.6 July Io 0 9.6 mr) |) oy || wWuly 27 c b b b July 14 | 2d 7a) I4II 3 July 30 c b b 9.6 July 13} Ist 7.0 | 1412 (of July 31 k 9.0 16.2 9.5 July 19 | Ist 6.0 | 1413 3 July 27 G b b |} July 10 | Ist 78 | I4I4 3 | July 20 c 9.2 16.9 9.2 July 10 | o 9.4 |} *I4I dé | July 20 c 8&8 15.9 9.6 July 4 381st 8.1 *1416 3 July 28 c b | b b July 12| o 10.0 *I41 ree 39July 2 c EO) | e2i-3) 12.1 July 11 0 T2335) *T418 & | July 28 G 0.7 19.4 10.8 July 9 41st 0.7 | 1419 So iW uly, 2. c 0.1 16.1 0.3 July 9) Ist 8.3 *1420 3 | July 29 c 12.7 24.0 14.3 July 12, 0 14.6 1421 Cul ulye sr k 11.6 20.0 ||| it-2 July 24 | Ist | 103 142. fof July 2 c 9.1 16.7 0.4 July 6] tst 8.2 1423 rot July 31 k | 11.0 20.6 11.4 July 21 | Ist 0.4 *1426 rch July 24 ¢ b 20.6 II.2 July 8 Ist 10.2 I5I 3 July 31 ke 1 s:6) ||| aos: 5.5 tae 28 |. Ist 3.9 Te | ae July 12| d 4-7 79 | 4.4 uly 5| o 4-5 1514 = July 2: c 4.6 8.5 4.4 July 11 Ist 3.6 | 1516 = July 28 d 4.2 | ke) 3.8 July 14 Ist 2.9 | aes 77 fof 41July 24 d 5.1 9.1 5.0 July 10 | Ist 3.8 *1518 fof July 25 or b 8.6 b July 12 421st 2.5 1520 fof July 20 d b b 7.5 July 20 | Ist 5.9 *1521 3 July 31 k 6.1 II.1 6.1 July 20 431st 3.8 *15 é July 2 c 6.5 11.6 b July to 441st 4.6 15 d July 21 Can aeG:2 IT.0 6.2 July 21 (c Ist 5.4 15 od July 22 Ce mo2 16.6 0.7 July 6] Ist 8.4 15 Jo July 31 EA) SHS 15.2 8.5 July 18 | ist 6.8 15 roe July 26| c 7.8 13.8 7.6 July 12 Ist 6.2 | 1528 | 6 July 2t/| c 9.6 17.7 10.1 July 7 0 10.1 | 1520 = July 28 c Gi 14.7 8.3 July 15 Ist 6.7 | 1530 3 July 31 | k 5.0 0.8 4.6 July 22, Ist 4.0 | 1531 Be! July 20| d 4.6 8.4 4.4 July 14 Ist 3.8 153 == July 2 (Ny eRe II. 6.3 July 14 Ist 5.2 | *1534 3 July 22 c 6.5 1.4 5.4 July 15 451st b | 1535 a July 31 | k 5.8 10.8 5.8 July 12} tst 4.4 | 1536 me July 26 d 4.9 8.6 4.7 July 11 | Ist 3.9 | 1540 2S July 26 d 5.0 9.2 Sar July 14 | Ist 3.8 | % r401. Died in molting. | 41418. First regeneration extends over two S85 r402. Deformed. | molts. 861403. Better classed as third regeneration. | 411517. Died in molting. 37 ;408. First regeneration extends over two 421518. Slightly deformed. molts. | 43 1521. Deformed. *8y415. Extends over two molts. | #1522. Deformed. 891417. Died in molting. | 41534. Lost in molting. Regeneration of the Chele of the Gulf-weed Crab. 123 Tas_e 1.—Table of data—Continued. Second molt. CHIE Nh Ga Cephalo-thorax. | a Right chela. . Date Desc: |——=—| =| Wehela. SSS = = Lg. Width Date Desc. Lg. | RETSA TS Wear 46July 25!) c | 6.4 11.6 6.3 July 13 Ist 5.5 i embgaa’ ad July 31 | k 59 | 12.5 5.8 July 19 | Ist 4.2 *I545 | of July 28 lim 16 4:2 | 8.4 4.0 July 16 |471Ist | b TSAOE ie Se July 13) d b | b 6.O\ql| yeaa ee th eee *1550 3 July 31 | Fk 6:8 | wey 6.0 July 18} Ist | 4.9 1551 Jb July 31 V2 || SGA || Wo) 8.1 July 26| Ist | 66 1554 ae July 25 G 5.1 | 9.1 4.9 July 12| Ist | 40 1557 Zs July 27 d 4.4 8.5 4.0 July 9 0 355) | 1558 == July 13 d b b b July 13 d b | 1501 roe July 23 (3 88 15.5 9.1 July 11] 1st | 83 | 1502 3 July 31 ee! 630 ree | 6.3 July 16 | Ist 5.2 | *1563 rH July 31 i> || Woy 16.8 9.2 July 18 |481st | b | | 1564 Pans July 30 d 47 8.0 4.5 July 11 Ist | 3.7 | *1565 fof July 31 k 7.03 ||) wlie.01. | 7.4 July 22] Ist| 67 | 1508 =5 July 24 d (xoy | ieee 6.8 July 10 | Ist | 5.6 | SrsGO, coo dh wily 2B ile C, | ws. | 78) woah. ebely 14,| 1St |iguais es72 Wee i 48uly sok cc |e a8 8.6 504. July 14| Ist] 38 1573 Se ils 2Sh len a Az 7.6 4.2 July 10) 1st} 32 | 1575 rofl July 23 c 3.9 6.6 b July 11 Ist 32 *1577 3 July 26 c 3.8 6.4 3.6 July 14 |511st 3.1 1578 SN aalysr | 2 | 45 8&8 4.0 July 13 | Ist 3.0 | 1580 gf | July 25) ¢ | 30 6.7 3.7 July 13] ist | 3.1 | T5e0) pose July 31} k | 4.3 8.7 3.0 Julyaras | site|) selon! MEC5) |) Oy | ulyee2r|) ves. )| b b b July 12 0 7.0 | TESOL == July 15 d b b b July 14 0 b | 1587 = July 24| ¢ b | b b July 11 | Ist 3.0 |} *1588 é 52July 26 6.2 11.4 6.4 July 13 | Ist 5.3 1589 3 July 31 k 4-4 | 09.0 4.1 July 9! Ist 2.7 | 1500 ey July 31| k 4.6 | 2 4.1 July 9 0 4.1 | *I501 3 July 23 | c¢ 723) \) 03:4. 7.5 July 11 Ist 6.5 #1592 3 july 17 | c 5.0 | 0.0 4.5 Niualys Fie rst 4.4 | 1503 3 July 290, c b b b July 14| o 7.6 | 1504 23 July 22 (2 3.9 6.8 b July to Ist 3.1 | 1505 a, July 31 k 6.4 11.7 6.5 uly 2 ist a7. 1500 za July 21 c 4.9 8&8 4.8 July 8 Ist 4.0 | 1507 3 July 24 c Wa angel 7.5 July 12| o 7.6 | #1508 3 July 21 c b b b July 6) Ist 4.5 *1509 Jo July 15 G 3.5 | 5.1 b July 5 Ist 2.6 *1601 Jb July 16 c 4.7 0.7 b July 5) 1st 3.6 1602 é July 31 k 4.7 0.9 4.4 July 12 0 4.5 | 1603 3 July 31 k 6.2 11.0 6.0 July 29 | Ist 4.9 1605 roe July 15 d 6.5 11.8 6.2 July 14 | Ist 4.6 | 1606 3 July 23 d 5.0 8.7 4.0 July 15 | Ist 3.9 | 1607 oh July 25 ¢ 6.8 13.0 7.0 July 13) 0 7.0 *1608 é July 31 k 6.6 12.1 6.6 July 13 | 2d 5.5 *1600 3 July 31 k 6.4 11.7 6.5 July 31 53k BG 1610 3 July 31 | 6.4 11.6 6.0 yeike aaoyl) ae || | 1612 3 | July 31 k 57 | To2 5.8 July 22 | Ist b *1614 -- | July 31 k 4.2 8.6 4.0 July 14 Ist 2.0 | 1616 roe July 21 c 7.2 12. 7.2 July 9} Ist 6.6 *1617 3 July 15 d 5.7 540.1 5.7 July 13 Ist 4.1 “61541. Died in molting. ®l 1577. Plane of removal not at breaking joint, “71545. Lost in molting. 827588. Left chela lost in molting. 48.1563. Lost in molting. 531609. July 20, right chela slightly smaller #91572. Died in molting. than left; July 27, both chelz present. 501572. Deformed. 51617. Approximate value. 124 Papers from the Marine Biological Laboratory at Tortugas. TABLE 1.—Table of data—Continued. Second molt. Cale: lasek Cephalo-thorax. Left Right chela. Date. DC eens MiChelat Lg. Width. Date Desc. Lg. *1618 es July 26 c 6.7 | b b July 15 551st 4.4 *I610 ae July 20 c 7.0 12.2 b July 6| Ist 6.0 *1620 3 July 23 c 3.8 6.3 3.4 | 56July 9 | Ist 2.8 *1621 3 July 31 k 4.9 8.4 4.4 July 15 | o b *1624 =s July 12 d 4.2 | 8.4 3-7 July 11 Ist 2180 | 1626 é July 18 d 4.0 8.0 3.5 July 13 0 3.8 1627 Jb July 20 c b b b July 18] 1st | 5.3 *1628 3 July 31 k 7.1 12.9 Fak July 19 571st b *1620 Jo July 24 c 6.2 he} 6.2 July 11 | 0 64 | 1632 | o& July 22 c b b b July 13 0 6.9 *1639) |i' S= July 15 c 47) | 9.6 4.4 July 4] tst 3.8 1634 3 July 22 c b | b b July to 0 6.1 *1635 -- | 58July 12 d 4.6 | 0.5 4.3 July 12 Ist 3.1 *1636 fof July 25 c 87 | 16.1 8.7 July 13 0 8&8 *1637 | o July 27 c be b b July 12 0 6.0 *1630 = July 28 d 4.0 — &.4 3.9 July 28 d 3.0 *1640 | oi July 19 c b | 09.0 uly? 30|) Sst 7.1 one || ae July 31 | & 63) || 702 6.1 July 9 o 6.0 | *1642 ae July 30 d 80 | 14.6 84 | 59July 14 Ist b | *1731 = July 24 c 6.9 | 1290 7.3 July 13 | 1st 6.1 | 17a) || oe July 31 k 7a || 8.3 4.6 July 15 | Ist 3.8 1733) Wele= July 31 k 5.8 | 10.0 5.8 July 25 | Ist 4.0 | Third molt. =, | ee — — = —s auilog Sex. | Cephalo-thorax. | 7 of Right chela Date. Desc. ana — — | Lg. | Width. Date Desc Lg. | | | 1307 é July 31 k 5.5 08 5.2 July 22 | Ist 3.3 *1308 3 July 31 k 5.2 8.7 4.6 July 22 | 3d 3.0 *1400 3 July 31 k 8.6 16.1 8&8 July 21 | Ist 605 | 1406 é July 31 k WBN L333 7.4 | July 23] 2d 5.4 | | *1407 3 July 31 | k 98 | 17.6 10.0 July 25 | 2d 61b | 1408 ae July 31 k 8.0 14.8 8.0 July 12| 2d 6.1 | *1400 é& | July 31 k 10.2 | 10.1 10.6 July 24 | Ist 8.6 1410 3 July 31 k 10.0 18.4 9.9 | July 28} 3d 80 | I4II 3 July 31 k 0.7 18.6 10.1 July 31 | 2d Fe 1413 a Madlyssr k 10.4 18.9 10.9 July 28 | 2d 88 | 1414 3 July 31 k 10.0 18.4 0.6 July 30 | 1st 7.2 *1415 a July 31 | & 9.9 18.0 | 10.0 July 21 | 2d b | *1416 3 July 31 k 11.3 Bint | Is July 29 |621st b 1418 | o July 31 k 11.8 2am |) er2is July 29 682d 10.0 | 1419 | of July 31 k 9.9 17-7) | 107) July 25 | 2d 8.0 *1420 3 July 31 k 14.5 27:7 || t6.9 July 30 | Ist | 144 1422 3 July 31 k 10.2 | 19.1 | 10.8 July 24 | 2d 8.5 | *1426 3 July 31 k 13.1 247) 73's July 25 | 2d 11.4 | |) Mistae Wa Seis ar |e gal) 662) “sa | “Sulyon | vad | ag *1518 é | uly ar k 5.0 10.6 4.8 July 26 2d 3.6 | *1522 Oo uly ste ke rhe || aia 7.2 July 25 642d 4.1 1523 é July 31 k 6.8 T2c0 6.9 July 22 | 2d 6.2 152 dé | July 3r | & 10.2 | 18.5 10.9 July 23 | 2d 8.0 8 1618. Injured. | ® 1642. Lost in molting. 581620. Not of same species as majority; July 8400. Right chela lost in molting. 9, plane of removal not at breaking 611407. Probably lost in molting. joint. ®21416. Lost in molting. 87 1628. Lost in molting. 68 y418. (?) 581635. Died in molting. 641522. Deformed. Regeneration of the Chele of the Gulf-weed Crab. 125 Tas_e 1.—Table of data—Continued. Third molt. ave Sexi ae = MT Cephalo-thorax. Left Right chela. | Se . | chela. | | | Lg. | Width. Date. Desc. | Lg. 1526 6 | July 31| & 8.90 16.5 9.0 | July 27| 2d 7.2 1528 Go|) uly sr ke Ror) “zor 11.8 | July 22 Ist 9.5 1520 =e July 31 | k 8.6 | 16.4 0.3 July 29 | 2d 7.3 153255) == July 31] & 7.0 12.4 7.0 July 25 | 2d 5.5 *1534 fof July 30| d b | b July 23 |651st b | *1545 3 July 31| &k 48 | 9.7 4.5 July 29 | 2d 3.0 | 1554 == July 31 | k 5.9 10.8 6.1 July 26 | 2d 4.4 1501 rol July 31 k 10.2 Wig) || a8) July 24 | 2d 8.2 *1569 | __ July 30! d | 46 Sole as July 29 662d | (66) 1575 3 July 31 | k 4.5 78 4.4 | July 24 | 2d 3.6 *1577 é July 28 d 4.2 7.3 4.3 | July 27 |672d 3.0 1580 3 July 31 k 4.4 8.1 4.4 July 26 | 2d 3.8 1585 3 July 31 k 8.0 14.8 8.1 July 23 | Ist 6.8 1587 3 July 31k AA 979 July 25 | 2d 3.3 *1588 | July 31 k 69 | 125 68h | July 27| 2d | 5.2 *I501 | of July 31 k 8.3 15.3 | 8.6 | July 24 \692d | 69 1502 | oo July 20 c b | b July 18 | 2d 4.0 1503 of July 31 | k 8.4 15.2 88 | July 30| Ist | 69 1504 ae July 31 k 4.4. | 78 4.1 ily 235) 2d \\ieesen 1500) | July 31) k 5.5 9.9 5.5 | July 22| 2d | 3.9 1507 | of July 31| & 83) 15.4 88 | July 25| 1st | 7.6 *1508 é July 31| & 58 | + 10:3 5.7. | July 22 |702d | b *1500 3 july sr |) vc 4.0 6.1 | 3.3 | July 16 | 2d 2.7 *TO01 3 July 31 | Fk 5.4 9.0 | Sit July 17 \712d a2 1607 Jb July 31 | k 8.3 TE.2 | 83 | July 26, Ist 7.3 1610 fof July 31 k 7.1 13.2 | Te Wey Suh | 2 da: 1616 3 July 31 | & 8.1 15.1 85 | July 22] 2d | 7.0 *1618 = July 20 d 8.3 b 8.7 July 27 \722d 5.7 *1619 cme July 31 k 8.3 14.5 8.3 July 20) 2d b *1620 o July 31 k | 44 7.6 4.0 July 24 | 2d 3.4 | 1627 3 July 31 | k 7.3 13.8 GY July 30 2d 6.4 *1620 ros July 31 k 7.0 12.8 7.0 July 25 731st 3.9 | 1632 3 July 31 | k 78 | 14.6 8.2 July 23 | Ist 6.6 *1633 Ze July 17 d ig} 10.8 5.2 | July 16 742d 3.8 1634 3 July 31 k 6.9 | 12.6 6.4 July 23 | Ist 5.2 *1636 3 July 31 k 10.0 18.9 10.5 | July 26 75rst 7.9 1637 rofl July 31 k 68 12.2 6.6 July 28 761st 4.2 #1640 3 July 31 k 0.7 17.7| 10.4 July 20 2d fe *1731 ae July 31 k 8.0 14.4 | 8.3 July 25 |?72d 5.0 ® 1534. Lost in molting. | 71618. Deformed. 883569. Deformed. | ™ 1629. Chela very stout. 871577. Deformed. | ™% 1633. Curved. 687588. Left chela lost in molting. 7 1636. Deformed. 8 ys91. Curved. | 7% 1637. Deformed. 70 1598. Lost in molting. 771731. Deformed. 711601. Deformed. ie 13.0 11.0 Cephalo-thoracic length (first regeneration period ) 10.0 ‘poriad Buryjow 126 = Papers from the Marine Biological Laboratory at Tortugas. * Superposition of two cases. d (first regeneration period). ing perio TauLe 2,—Correlation between cephalo-thoracie length and molt Regeneration of the Chele of the Gulf-weed Crab. TasLe 3.—Correlation between cephalo-thoracic length and molting period (first regeneration period). [Data arranged in order of length of molting period.) : aa odliGadtaicnll cs | ; Comoe | MOMNE | ‘horace || ONa"® | Sotode ength, 1509 J |e 1422 II 1575 9 39 ||_1524 | 17 | 1633 9 47 1516 12 1580 TON lnaiOn 1540 12 | 1581 10 4.3 1606 12 1504 10 4.7 1397 12 1601 10 4.7 1588 12 1530 10 4.9 1595 12 1554 10 5.1 1401 12 1405 10 7.4 1525 12 1632 TON io 1506 13 1578 II 4.3 1733 13 1580 II 4.4 1544 13 1514 11 4.6 1502 13 1531 It 4.6 1532 13 1573 Il 4.7 1541 13 1572 II 4.8 1568 13 1732 II 4.8 1034 13 1535 II 5.8 1616 13 1731 II 6.9 1507 13 1529 Ir a7, 16077 |73 1585 II 8.0 iat | Catalog No. Cephalo thoracic length. 9.1 1629 1020 1520 4.2 1410 5.0 1528 5.0 1504 5.5 1512 6.2 1017 6.4 1610 6.6 1501 8.5 1423 4.9 1603 5.8 1005 5.9 15901 6.1 1503 6.3 1412 6.4 1635 6.9 1421 6.9 1551 7.2 1420 8.3 1530 _ 83 1414 5.1 1523 Molting period Molting period, Cephalo-thoracic length (second regeneration period) Cephalo- thoracic length, 7.0 7.8 0.1 Taste 4.—Correlation between cephalo-thoracie length and molting period (second regen- eration period). * Superposition of two cases. 128 Papers from the Marine Biological Laboratory at Tortugas. TABLE 5.—Correlation between cephalo-thoracic length and molting period (second regeneration period). [Data arranged in order of length of molting period.] Catal Molti Cephalo- || Catalog | Molti Cephalo- |! Catal Molti Cephalo- ates |) peneds — Re) peace ie Nao | peeled shorecie 1304 9 4.0 || 501 | 13 83 1408 15 8.0 TSs20n |e 7.0 eee 10.2 || 1526 iss 8.9 | 1610 | 71 || xs87 | 14 | 44 || 14190 16 0.9 1627 a2) 7-3 || 1514 14 5.4 1524 17 10.2 1580 13 4.4 1506 14 5.5 I4II 18 0.7 1504 13 4.4 1554 14 5.9 1422 18 10.2 1575 13 4.5 1588 | 14 6.9 1413 18 10.4 1545 13 4.8 1406 14 7-3 1616 13 8.1 1520 | 14 8.6 Taste 6.—Correlation between molting period and amount of first regeneration. [Data arranged in order of length of molting period.] } Baht Right Right = . el i , i , yee | Monae | Siar | Cyr | Mee | Caer || Cue | me | Stat regenera- regenera- regenera- tion, tion, tion. | 1569 | 8 35 1585 II 68 1629 14 3.0 rss7i 9 | 30 1422 II 82 1526 14 6.2 1575 fe) 32 || 1524 | mt | 84 || 1410 14 8.3 163m eee OLN) |e es:Gnl e516 12 2.9 1528 | 14 9.5 | 1581 10 2.6 1307 12 3.3 1504 15 3.1 1580 10 3.1 1540 12 3.8 I512 15 3.9 bored 10 3.6 1006 12 3.9 1017 15 4.1 1504 ate) 3.7 1505 12 4.7 1610 15 4.5 1530 be) 3.9 1401 12 5.0 || I561 15 8.3 1554 10 4.0 1588 12 i 1423 15 9.4 1508 10 4.5 1525 12 68 1605 16 4.6 1405 10 61 | 1506 13 4.0 1603 16 4.9 1632 _ 5110) 3 |G. 1733 13 4.0 1412 160 6.0 i) | am [yy 1544 13 4.2 I501 16 6.5 1578 II 3.0 1532 13 5.2 1503 16 6.9 1573 II 2 1562 13 5.2 TALES | 316 76 1514 II 3.6 1634 13 5.2 1635 T7| ee 1531 II 3.8 1541 13 5.5 1421 wr 10.3 1572 II 3.8 1568 13 5.6 I55I_ 18 6.6 — 1732 II 3.8 1520 13 5.90 1420 18 14.4 | 1535 II 4.4 1616 13 6.6 1530 | 19 4.0 162 ul 5.3 || 1607 13 7.3 || 1414 20 7.2 1731 II 6.1 1507 13 |) 2 7.6) || esas 25 5.4 1520 II 6.7 1517 14 3.8 11.0 10.0 Amount of first regeneration. Se (o) 6.0 5.0 4.0 3.0 8 9 10 iH 12 13 14 15 16 lir/ 18 19 20 Molting period TasLe 7.—Correlation between molting period and amount of the first regeneration of the right chela. * Superposition of two cases; @ Superposition of three cases. 9 130 Papers from the Marine Biological Laboratory at Tortugas. Taste 8—Correlation between molting period and amount of second regeneration. [Data arranged in order of length of molting period.] | Hight Hight Right Catal Molting | chela, Catal Molti cee Catal Molti chela. eae srtecen ese | =e ou lllenericds = Me i petede ee tion. tion, tion, SS Ee || 13 7.0 | 1529 14 73 1532 II 5-5 1561 __ 13m 82 || 1408 15 6.1 1610 | Se | mp | a 3-3 || 1526 _ 15 72 1627 _ 12 6.4 1514 14 3.7 || 14190 16 — 89 : 1545 13 3.0 || 1506 14 3.9 || 1524 17 | 89 1504 13 3.1 1554 14 4.4 || 141r 18 77 1575 13 3.6 1588 14 5.2 || 1422 18 8.5 1580 13 3.8 1506 14 5.4 1413 18 88 15901 13 6.9 1410 TA 7.3 Amount of second regeneration 9 10 i 12 13 14 IS 16 17 18 Molting period TaBLe 9.—Correlation between molting period and amount of the second regeneration of the right chela. * Superposition of two cases. Regeneration of the Chele of the Gulf-weed Crab. 131 TasLe 10.—Correlation between the cephalo-thoracic length and the amount of the first and second regenerations of the right chela. [The specific amount of regeneration is the length of the regenerated chela divided by the cephalo-thoracic length. ] First regeneration Second regeneration. Cephalo: Wiehe ae Cephalo- Catalog Specific ae eRe | Che's tel Specific Catalog No. amount. ran epee eeu. sei amount. No. tion. tion, 1580 | 0.705 3.9 3-1 1504 795 3.9 3-1 | 1575 821 3.0 Be | 3.0 4.0 0.750 1304 1510 | .690 4.2 2.9 | 158r | 605 4.3 26 || | 1578 698 4.3 3.0 3.1 44 | -705 1504 | 15890 | 614 4.4 2.7 3.3 4.4 -750 1587 | 1569 | 778 4.5 3-5 || 3.8 4.4 864 1580 | 1035 | 674 4.6 By 3.6 4.5 .800 1575 1514 | 783 4.6 3.6 1531 826 4.6 3.8 1573 | 681 4.7 3.2 1601 -7606 4.7 3.6 1564 .787 4.7 3-7 1633 809 4-7 3.8 1572 792 4.8 3.8 3.0 4.8 625 1545 1732 -792 4.8 3.8 | 1536 796 4.9 3.9 1506 816 4.9 4.0 1540 -760 5.0 3.8 1606 | .780 5.0 3.9 1530 | .800 5.0 4.0 1517 | -745 5.1 3.8 |i 1554 784 5.1 4.0 3.7 5.4 685 1514 1307 600 5.5 3-3 | 3.9 555 | — -709 1596 1512 .606 5.6 3.90 || 1617 | -719 Be, 4.1 1733 690 5.8 4.0 1535 759 5.8 4.4 | 1544 712 5.9 4.2 4.4 5.9 746 1554 1562 | 852 6.1 5.2 1603 -790 6.2 4.9 } 1588 855 6.2 5.3 1523 871 6.2 5.4 1532 825 6.3 5.2 1610 -703 6.4 4.5 1505 -734 6.4 4-7 | 1541 850 6.4 5.5 1605 708 6.5 4.6 | 1401 758 6.6 5.0 1634 754 6.9 5.2 1568 812 6.0 5.6 5.2 6.9 | 754. 1588 1731 884 6.9 6.1 1629 557 7.0 3.9 Sas 7.0 | 786 1532 | | 1616 O17 7.2 6.6 5.5 | 7.1 775 1610 | 1591 800 7-3 6.5 5.4 | 7-3 -740 1406 1405 824 7.4 6.1 6.4 | 7.3 877 1627 1529 870 7. 6.7 | 1526 795 7.8 6.2 1632 846 78 6.6 1551 835 7.9 6.6 | 132 Papers from the Marine Biological Laboratory at Tortugas. Taste 10.—Correlation between the cephalo-thoracic length and the amount of the first and second regenerations of the right chela—Continued. Catalog No. 1585 1607 1507 1503 1525 1501 1412 1422 1419 1524 1414 1423 1528 1421 1420 First regeneration. Second regeneration. Right Right Specific Cephalo- chela, chela, Cephalo- Specific Catalog amount, Tae vepeaerie 2 senes- meee amount, No. tion. tion. 850 8.0 6.8 Ost 8.0 -762 1408 880 8.3 7g 70 | 8.1 864 1616 916 8.3 7.6 6.9 8.3 831 1591 821 8.4 6.9 800 8.5 6.8 7.3 8.6 849 1520 043 8.8 8.3 Ff | 8.0 809 1526 .607 9.0 6.0 -QOI 9.1 8.2 TG. 9.7 704 I4II .QI2 9.1 8.3 8.9 0.9 800 T419 O13 0.2 8.4 8.2 10.2 804 1501 -720 10.0 7.2 8.5 10.2 833 1422 855 II.O 0.4 8.9 10.2 873 1524 864 11.0 0.5 8&8 | 10.4 846 1413 888 11.6 10.3 A 14.5 14.4 12.0 11.0 © 3° Amount of first regeneration fo} ° me fo) 6.0 5.0 4.0 TABLE rra.—Correlation be line, line along w : Jae ai a ‘ — 4B al aes x - Seana | 3° Amount of first regeneration o © a hs x = ° 6. 2) ° 5. 4. °o 2 i -_ ieee! Le Bel 2 | 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10,0 11.0 12.0 13.0 i4.0 Cephalo-thoracic length at a tt TABLE Be conelation between cephalo-thoracic length and amount of first regeneration of right chela. * Superposition of two cases. Dotted ine, line along which all data would be arranged if specific amount of regeneration were the same for individuals of all sizes. y . ; ; [CHORE tea Hat Ai 5 =) é aU ¥ ae y hu + ‘7 q “ wi #| GS “% ie a — , Se! 14.0 13.0 12.0 11.0 10.0 Ist,0=2d.) 9.0 Amount of regeneration (X @ oO 6.0 5.0 4.0 TaBLE 118.—Comparison of generation; * superposi regeneration. 14.0 13.0 12.0 11.0 10.0 @ $0 fo) fo} Amount of regeneration (X=!st,0=2d.) x oO 6.0 5.0 4.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11,0 12.0 13.0 14.0 Cephalo-thoracic length during period in which regeneration takes place : . Bee : a i : irst re- TABLE 11B.—Comparison of first and second regenerations in all the individuals, arranged according to cephalo thoracic pa aae Pa Second generation; * superposition of two cases of first regeneration; © second regeneration; @ superposition of a a a regeneration. 14.0 13.0 12,0 11.0 S is) so So 90 o Length of left chela(unoperated individuals) xn o 6.0 5.0 4.0 3.0 4.0 TABLE r1¢.—Correlation be data would be arrange: @ =4 cases; arrows = Ss tS) so o 7% xx * Ve 2 o Length of left chela(unoperated individuals) 5.0 11.0 12.0 13.0 14.0 15.0 Cephalo-thoracic length (unoperated individuals) 3.0 4.0 5.0 6.0 7.0 8, TABLE 11¢.—Correlation between left chela-length and the cephalo-thoracic length in unoperated individuals. Dotted line, line along which all data would be arranged if specific length of left chela were the same for all sizes of individuals. —=1 case; *=2 cases; )—=3 cases; ® = 4 cases; arrows = cases which come beyond limits of paper. Ss 2 3 3 Es (syenpiapul Uoiyesauebaa 4S41J) 2/949 949) Jo yybu2 TABLE IID.— line, line sizes of 17.0 16.0 15.0 14.0 13.0 10.0 go oO & o Length of left chela(first regeneration individuals) 7.0 6.0 5.0 40 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 Cephalo-thoracic length (first regeneration individuals) TaBLE 11p.—Correlation between left chela-length and cephalo-thoracic length in first-regeneration cases. Dotted line, line along which all the data would be arranged if specific length of left chela were the same for all sizes of individuals. * Superposition of two cases. Regeneration of the Chele of the Gulf-weed Crab. 133 Amount of second regeneration 9.0 10.0 1.0 Cephalo-thoracic length TABLE 12.—Correlation between cephalo-thoracic length and amount of second regenera- tion of right chela. Dotted line, line along which all the data would be arranged if specific amount of regeneration were the same for individuals of all sizes. TABLE 13.—Comparisons of the first, second, and third regenerations in single individuals. [Data arranged in order of original length of the removed chela.] | Catalog | Original First Second Third Catalog | Original First Second i ‘Third No, length. regenera- regenera- regenera- No. length. regenera- regenera- regenera- tion. tion. tion. tion, tion. tion, 1500 2.5 2.6 2.7 2.9 | 1588 5.6 ie) 5.2 1575 2.9 a2 3.6 | 1610 5.8 4.5 5.5 1587 3.0 3.0 a 1616 6:1 6.6 7.0 1504 3.1 aul 3.1 _ 1410 6.5 (*) 7.3 8.9 1580 3.2 iat 3.8 1501 6.7 6.5 6.9 1514 3.8 3.6 3.7 | 1526 7.0 6.2 7.2 1596 42 4.0 3.9 | 1520 | 7.0 6.7 7-3 1523 4.3 5.4 6.2 1561 | 7.8 8.3 8.2 1554 4.4 4.0 4.4 Itt | 85 7.6 7-7 1627 5.5 5.3 6.4 1524 | 8.6 8.4 8.9 1532 5.6 { 5.2 5.5 * First regeneration record missing. 134 Papers from the Marine Biological Laboratory at Tortugas. 9.0 Amount of regeneration(X=!st,0 = 2nd,,@ =3rd.) 5.0 : . 8.0 9.0 Original length of right chela TABLE 14.—Comparisons of first, second, and third regenerations of right chela in single individuals and their correlation with original chela-lengths. Dotted line, line along which all the data would be arranged if regenerated chela-lengths were equal to origi- nal chela-lengths. > first regeneration; © second regeneration; © third regeneration. Regeneration of the Chele of the Gulf-weed Crab. 135 TABLE 15.—Correlation between the left and right chele and the cephalo-thoracic length in unoperated individuals, [Data arranged in order of cephalo-thoracic length. Asterisks (*) indicate individuals without catalog numbers, collected especially for these measurements. All of these starred cases, with one exception (*Q), were males. ] l ; i | . Catalog | Cephalo- | Left | Right ee ae Catalog | Cephalo- | Left | Right | iecken No. & thoracic chela chela to cephalo- a & thoracic chela | chela_ to cephalo- | length. length, length. thoracic length. length. length. | thoracic | length | | length. | 5 7 | ~|t — + anal | 1625 | 3.0 27 | 27 | 0.900 || (*) 4.5 44 | 45 | 978 1530 | 3.1 2.8 28 | <903 || (*) 4.5 45 | 4.5 1.000 | Too) |) 3:1 2.8 2.9 .903 | 1606 4.5 Ansa Ace | 1.067 Te7s, || 32 P20) || ARS) 900 1590 4.6 4.1 4.1 801 1504 | 3.2 32 | 31 | 3.000 | 1533 4.6 4.3 4.4 | 935 1587 | 3.3 30 | 3.0 909 || 1605 | 4.6 4.3 4.4 935 1502 3-3 Zp | Bi 939 || 1530 4.6 4.4 4.4 .057 (*) 3-4 3.0 3.1 882 || 1552 4.6 4.4 4.4 957 (*) 3-4 3.0 sbi 882 1Or7 || “4:0 | “4a 4.4 957 1580 34 3.1 3.2 gI2 1513 4.7 4.4 4.5 .930 1622 3.4 3.1 2.7 12 | 1602 4.7 4.4 4.5 .936 1613 3.5 Fy 33 771 1627 Hie NN) ho) .936 (SRS ees 3.0 3.1 857 || 1570 | 4.7 4.5 4.6 .957 (*) 3.5 But || 886 | 1507 | 4.7 4.7 4.7 | 1.000 1550 3.5 3.2 3.3 | O14 1510 48 4.4 45 | O17 1557 3.5 3.2 3.2 | 914 1402 48 4.7 4.8 979 1304 3.5 3.3 3-3 | 943 (5) 4.8 4.7 4.6 079 1543 3-5 3-3 3-4 043 || 1591 4.8 5.0 5.0 | 1.042 1548 3.6 3.2 3.4 | .889 || 1630 4.9 | 4.3 4.4 878 1564 3.6 3.2 3.2 889 || 1603 49 | 4.9 4.0 1.000 1570 3.6 2 3.2 889 | 1555 5.0 4.7 4.9 940 1581 3.6 3.2 3.4 889 1544 5.1 4.8 5.0 O41 1623 3.6 Be 3-4 | O17 1628 5.1 5.2 5.2 1.020 1545 3.6 3.5 35 || .972) ||| 2733 5.1 5-4 5:5 1.059 1578 aye || 6 3-5 946 || 1631 5.2 4.7 4.8 .904 1516 3.8 B65 3.5 | .o2r || 1552 5.2 5.0 5.1 .962 1586 38 35 3.5 21 1617 5.2 5.0 5.3 962 | 1635 3.9 3.7 38 | .040 1629 52 5.2 5.3 1.000 | 13090 3.9 3.8 3.9 974 | 1535 5-3 5.2 5-4 .982 1573 | 3.9 3.8 3.8 974 || 1541 54 | 5.2 | 5.5 963 1626 | 4.0 3.5 280 ie i875) i rs2n 5.4 Gai i] Gi) eesti 1633 | 4.0 3-7 316, | .925 1588 5.4 mas 5.6 | 1.010 | 1533 4.0 3.8 3.0 950 ~—|«1558 5.5 5.6 5.6 1.018 1601 4.0 3.8 3.8 950 (*) 5.6 5.5 5.6 | .982 1630 40, | 38)| 38 950 || 1542 5.6 5.6 | 57 | 1.000 1514 4:9 | (3:7) 338 902 ~+|-:1406 5.6 5.8 5.8 1.0360 1602 4.1 | 37 37 .902 1505 5.6 5.9 6.0 | 1.054 1540 4.1 3.8 3.0 027 1546 5.7 5-7 | §.8 | 1.000 1519 4.1 3.9 4.1 O5I 1522 5.7 5.8 6.0 | 1.018 1510 4.2 30 | 30 920 1005 5.7 58 | 5.8 | 1.018 1515 4.2 4.0 3.0 952 || 1604 5-7 58 | 5.3 | 1.018 1571 4.2 4.1 4.1 976 || 1568 5.7 Bo || 6.0 1.035 1504 4.2 4.4 4.2 | 1.048 || (*) 5.8 5.5 5.6 048 1306 AB | Ao | Zo 930 || 1603 5.8 58 | 5:4 | 1.000 1536 4.3 4.0 4.2 930 1404. 5.9 5.8 58 | 083 1307 4.3 4.1 4.1 953. || 1579 5.9 58 | 5.9 | 983 1506 4.3 4.1 4.2 953. || 1636 5.9 5.8 5.9 | .083 1508 4.3 4.2 4.4 077 || 1585 5.0 5.0 5.9 | 1.000 1557 4.4 4.0 35 | .909 || 1607 5.9 5.0 5.9 | 1.000 1584 4.4 41 | 42 | .o32 |} 1566 | Oy |) He || 6s) 1.034 1638 4.5 3.6 4.2 (*) | Go 605) 6x 1.000 1309 4.5 4.1 4.1 QIt 1422 6.0 6.1 6.1 1.017 1583 4.5 4.3 4.4 956 1560 6.1 6.1 6.2 T.000 136 Papers from the Marine Biological Laboratory at Tortugas. TABLE 15.—Correlation between the left and right chele and the cephalo-thoracic length in unoperated individuals—Continued. Cephal Lef Ane ees Cephal Left Righe | atest . = e) 1 . & | Cae: ae anaes cela ieeeoerd | Gale eecraeae chela- chee lee cesbalee : length, _ length. length. | thoracic | { length. length. length thoracic | =a Leet | length, | 1561 6.1 6.3 6.2 1.033. | 1561 ves | 7.7 78 1.055 1537 6.1 6.4 6.4 1.049 || 1407 7-3 7.9 7.4, 1.082 1538 6.2 6 | 63: 084 | T419 7.4 8.3 8.4 1.122 1629 6.2 6.2 | 6.4 | 1.000 || (*) 7.7 7.7 7.8 | 1.000 1405 6.2 (DEN) OR I.010 1413 78 8.2 8.3 1.051 1559 6.2 63 |) 163 1.0160 1409 7.9 8.1 8.4 1.025 1600 6.2 6.6 6.8 1.065 T4II 7.9 8.1 8.5 1.025 | (*) 6.3 6.1 | 6.3 968 1422 7.9 8.2 8.6 | 1.038 | 1507 6.3 63) |) 1640 | ooo; i) (5) 8.0 8.4 8.5 | 1.050 (*) 6.3 Geyel| 633 1.000 | (*) 8.0 8.5 8.7 1.062 I eS)) 6.3 6.4 6.5 1.016 1412 8.0 8.7 88 1.087 | 1618 6.4 5.8 BS) ls <00: 1416 8.2 8.2 8.5 1.000 | 1565 6.4 (hie |) SS 530 Hse) 8.6 8.0 Q.1 1.035 I501 6.4 6.5 | 6.7 1.016 || 1420 8.6 0.4 9:4 | 1.093 | (*) 6.5 OG OF 1.031 1636 8.7 8.7 88 | I.c00 | | 1551 6.5 Fes 6.8 1.123 1409 9.1 9.6 9.6 1.055 (*) 6.6 6.5 6.7 085 | 1414 9.2 9.2 9.4 1.000 1529 6.6 6.8 7.0 1.030 || (*) 9.5 9.7 8.9 1.021 1404 6.7 Oey || GH .970 | 1528 9.6 10.1 10.1 1.052 | 1584 6.7 o7 «| 5.0 1.000 | 1423 9.7 08 10.0 1.010 1520 6.7 6.9 6.3 1.030 1421 10.1 9.6 10.7 950 1607 6.8 7.0 7.0 1.02 (*) 10.1 10.0 10.7 990 | 1520 6.9 O81) 70 .986 || 1417 11.0 12.1 12.3 I.100 1563 6.9 (XO) |) Fan 1.000 | 1424 11.2 12.3 12.3 | 1 098 1507 WB 75 | 7.6 1.042 || (*) II.2 12.8 13.0 | 1.143 | 1642 7.2 rise SAS 1.042 || 1425 12.4 14.5 14.5 1.1690 | 1525 7.3 73) le aed. 1.000 || 1420 12.7 14.3 14.6 1.126 1566 73 7A 7.6 1.014 1428 14.4 17.0 1723) | one (*) 7.3 7.4 7.6 1.014 (*92) 15.0 16.1 16.5 | 1.073 | 1400 73 77 78 | 1.055 || Av. 5.77 5.77 | 5-84 | 982 Taste 16.—Correlation between the left and right chele and cephalo-thoracic length in first regeneration cases. [Data arranged in order of cephalo-thoracie length.] a ci Right pate of Goji ae Right | Ratio of 4 eft ele e o- se! rf | Cale iineie chela- ane Reese Catalog piowecic chela- poe sees: : length. | length. regenera- | thoracic ‘ length. length. | regenera- | thoracic tion. length. | tion. length. 1577 3.8 3.6 3-1 | 0.047 ee 4.7 ABN Sz, 057 1580 3.9 Ber 3.1 | .949 1732 48 | 46 | 3.8 958 1624 4.2 ay g8: | 88h Wisse, |) 40 | 47 | «36 .059 1510 4.2 3.8 2:9 | .005 | 1506 | 4.0 48 | 4.0 .980 1545 4.2 4.0 (2) | 952 1592 | 5.0 As ln ed .900 1581 4.3 8:0) |||) <2:6) ||) 007, 1530 | 5.0 46 | 40 -920 1578 4.3 4.0 3.0 | 930 1606 | ROM || kon ex) 980 15890 4.4 4.1 | 27 .932 || 1540 5.0 | 5.1 | 3.8 1.020 1569 4.5 4.3 35 956 | 1554 51 | 49 | 40 O61 1635 4.6 4.3 2:65 ||) O35 1517 Gu || EO | 3.8 .980 1514 4.6 4.4 3.6 LOpzae I a97n 5:5 5.2 3-3 045 1531 4.6 4.4 3.8 057 || 1512 | 5.6 Ii ae | 3.9 .982 1573 4.7 2 3.2 804 | TORZ | 5k) ||| | See | © ae aon 1633 4:7 4.4 3.8 .930 || 1612 ad Sh (a r.018 Regeneration of the Chele of the Gulf-weed Crab. 137 TABLE 16.—Correlation between the left and right chele and cephalo-thoracic length in first regeneration cases—Continued. Right Sey ence eee cl length, length. regenera- | tion, | 1535 5.8 5.8 4.4 1733 5.8 5.8 4.0 1544 5.9 5.8 4.2 1521 6.1 6.1 (*) 1562 6.1 6.3 5.2 1603 6.2 6.0 4.9 1523 6.2 6.2 5.4 1588 6.2 6.4 5.3 1532 6.3 6.3 5.2 1610 6.4 6.0 4.5 1541 6.4 6.3 5.5 1400 6.4 6.5 5.1 1505 6.4 6.5 4.7 1534 6.5 5.4 (?) 1605 6.5 6.2 4.6 T401 6.6 6.6 5.0 1637 6.8 6.6 4.2 1634 6.9 6.4 5.2 1508 6.9 68 so) | 1731 6.9 7.3 6.1 | 1629 7.0 7.0 3.9 1628 al 7.1 (?) 1408 Gali WP 6.7 1616 72 We 6.6 | 5901 7-3 7-5 6.5 | 1405 7.4 7.4 6.1 1410 7.4 7.6 ihe. || Ratio of Right di | Caen | Giiie | eh | eae thoracic || length. | length. regenera- length. | | tion, } 1.000 1520 7.7 8.3 6.7 1.000 || 1526 7.8 7.0 6.2 .083. || 1632 7.8 8.2 6.6 1.000 1551 7.9 8.1 6.6 1.033 || 1585 8.0 8.1 6.8 .968 || 1642 8.0 8.4 (2) 1.000 | 1607 8.3 8.3 733 1.032 1507 8.3 88 7.6 1.000 | 1503 8.4 8.8 6.9 937 || 1525 8.5 8.5 6.8 O84 1400 8.6 88 (?) 1.016 1501 88 Q.1 8.3 1.016 1412 9.0 0.5 6.0 831 || 1563 Q.1 9.2 (?) 954 | 1410 9.1 9.3 8.3 I.000 || 1422 9.1 0.4 8.2 971 || 1524 2 9.7 8.4 928 | 1414 10.0 9.6 2 086 || 1636 10.0 10.5 (*) 1.058 1409 10.2 10.6 8.6 1.000 || 1423 11.0 11.4 0.4 1.000 | 1528 11.0 11.8 9.5 T.o14 |! 1416 11.3 11.8 (?) 1.000 || 1421 11.6 11.2 10.3 1.027 1427 12.3 14.8 13.8 1.000 1420 | 14.5 16.9 14.4 1.027 || Av. 6.81 6.85 ue Ratio of left chela to cephalo- thoracic length, 1.078 -974 1.051 1.025 1.012 1.050 | 1.000 1.000 | 1.048 1.000 | 1.023 | 1.034 1.050 TABLE 17.—Correlation between the left and right chele and cephalo-thoracic length in second regeneration cases. [Data arranged in order of cephalo-thoracic length.] Catalog No. Z Right Ratio of | | Right Ratio of | Cephalo- Left chela, | left chela | Catalog | Cephalo- Left chela, | left chela thoracic chela- second to cephalo- | No & | thoracic | chela- second to cephalo-| length, length. | regenera-| thoracic | : length. | length. | regenera-| thoracic | tion, length. tion, length. 4.0 3.9 3.0 0.9075 1522 7.3 7.2 (*) 0.986 4.2 4.3 (*) 1.024 1406 7.3 7.4 5.4 1.014 4.4 4.1 3.1 032 1627 Hilo | eh 6.4 1.055 4.4 4.3 3.3 .077 1408 80 | 80 6.1 1.000 4.4 4.4 3.8 1.000 || 1731 80 | 8&3 5.0 1.037 4.5 4.4 3.6 078 | 1616 Sine al seeks 7.0 | 1.049 4.6 4.5 (*) .978 || 1618 8.3 8.6 (*) 1.048 4.7 4.5 3.3 957 || 1591 8.3 8.7 6.9* | 1.036 4.8 4.5 3.0 937. || 1520 8.6 0.3 Wea T.081 5.0 4.8 3.6 .Q60 1520 8.0 9.0 7.2 I.O1L 5.3 5.2 (*) 083 I4II 9.7 10.1 77, 1.041 5.4 5.1 (*) O44 1640 9.7 10.4 7.1 1.072 5.4 5.2 3:7 .963 1419 9.9 10.7 8.9 1.081 5.5 5.5 3-9 T.000 I415 90.9 10.9 (*) 1.101 5.8 5.7 (*) 983 1561 10.2 10.3 8.2 T.010 5.0 6.1 4.4 1.034 1422 10.2 | 10.8 8.5 1.059 6.0 5.0 (*) 983 1524 10.2 10.9 8.0 1.000 6.8 6.9 6.2 1.015 1413 10.4 10.9 88 1.048 7.0 7.0 5.5 1.000 Av. 6.99} 7-13 | ---- I.O1l 7.1 7.1 5.5 1.000 138 Papers from the Marine Biological Laboratory at Tortugas. 11.0 10.0 9.0 (oat oO on fo) Length of left chela(second regeneration individuals) 52 o) 3.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 Cephalo-thoracic length(second regeneration individuals) TABLE 18.—Correlation between left chela-length and cephalo-thoracic length in second regeneration cases. Dotted line, line along which all the data would be arranged if specific length of left chela were the same for all sizes of individuals. PLATE 1 CHAPMAN *QIUEJSIP 9Y} UT EI[esAUg JYORA UT Sa te "ISAM Suryoo, Auojop Aqoog IU} JO SSQUaLUIE} 9Y} Bue sNiy *Kuojog Aqoog ay Vv. A CONTRIBUTION TO THE LIFE-HISTORIES OF THE BOOBY (SULA LEUCOGASTRA) AND MAN-O’- WAR BIRD (FREGATA AQUILA). By FRANK M. CHAPMAN, Of the American Museum of Natural History. 6 plates. 139 A CONTRIBUTION TO THE LIFE-HISTORIES OF THE BOOBY (SULA LEUCOGASTRA) AND MAN-O’-WAR BIRD (FREGATA AQUILA). By Frank M. CHAPMAN. INTRODUCTORY. The expedition on which the observations herein recorded were made was undertaken primarily to secure specimens, accessories, photographs, and data to be used in the preparation of a “ habitat group” of the booby and man-o'war bird in the American Museum of Natural History. Dr. Alfred G. Mayer, in command, George Shiras, 3d, and the writer, sailed from Nassau, March 31, 1907, aboard the Physalia, for Cay Verde, about 230 miles to the southeast, where the birds desired were reported to nest. Unfa- vorable weather prolonged our voyage to the cay to nine days, and this delay, in connection with continued unpleasant weather and the absence of a harbor, made it undesirable to remain at the cay longer than was necessary for the accomplishment of our object. While, therefore, the trip was successful so far as collections for groups were concerned, the three days and nights (April 9 to 11) passed on the cay permitted us to make only the most casual study of the habits of the birds found nesting there. Nevertheless, with the codperation of Dr. Mayer and Mr. Shiras, information was gathered which appears to deserve record as a contribution to our knowledge of Bahaman bird-life, and particularly to the biographies of the man-o’-war bird and booby, about the nesting habits of which, in this region, little ap- pears to have been written. CAY VERDE. Cay Verde lies on the eastern edge of the Columbus Bank, 30 miles southeast of Little Ragged Island. It is about 0.5 mile long by 0.25 mile in greatest width, the longer axis lying approximately north and south, and, roughly estimated, contains some 40 acres. On the west and south, or shal- low, Bank sides there are steeply shelving beaches, where, under favorable conditions, a landing may be easily made; on the eastern side the deep-blue waters of the ocean break directly against the characteristic, water-worn limestone rock, of which Cay Verde, in common with other Bahama islands, is composed. At the northern end, where the islet terminates in a point, this rock is but little above sea-level. Southward it gradually increases in 141 142 Papers from the Marine Biological Laboratory at Tortugas. height and, with pronounced irregularities in coast line, reaches a bluff-like elevation of 75 feet at the southeastern extremity of the islet. About one-eighth of the surface of the island is covered with a dense growth, chiefly of sea-grape (Coccolobis uvifera), but with a liberal mix- ture, mainly about the borders, of a “ prickly-pear”’ cactus (Opuntia) and sea-lavender (Tournefortia gnaphalodes). Where sufficient soil has ac- cumulated, the remainder of the island supports a growth of coarse grasses, sparse on the higher and rockier portions, more luxuriant in the lower por- tions, particularly about the margins of a small salt pond, the size of which was dependent upon conditions of tide and wind. There is no fresh water on the cay. BIRD-LIFE. In the literature of ornithology, Cay Verde figures only in Bryant’s “List of Birds seen at the Bahamas from January 20 to May 14, 1859,” where it is casually mentioned’ as a breeding-place of the tropic-bird (Phaéthon flavirostris). This author writes at some length of the nesting habits of the booby and man-o’-war bird as observed on San Domingo Cay and the Ragged Islands, respectively, but does not refer to the colonies of these birds on Cay Verde. Possibly he did not himself visit Cay Verde, where doubtless both the species of birds named have nested for a pro- longed period; this cay, so we were informed, having, some ten years ago, been the site of a guano industry, which flourished until all the available deposit had been removed. The writer’s information in regard to the birds of Cay Verde was ob- tained from the late D. P. Ingraham, who, as a collecting naturalist, visited the cay about 1891. Mr. Ingraham’s information in regard to the presence of boobies and man-o’-war birds was fully verified. In May (he also wrote) great numbers of terns (doubtless Sterna fuliginosa, S. anethetus, and Anous stolidus) and a few tropic-birds come to the cay to nest. No land birds appear to be resident on Cay Verde, but it is evidently visited by numbers of migrants. During our stay the following species were noted: Common name. Scientific name. Audubon’s shearwater...........- Puffinus l’herminieri Sooty witernise sce asi scteeieoncres elec Sterna fuliginosa Great. bide! herotace ccc eceiseecee Ardea herodias Little blue heron... «-se+ = seein Florida cerulea Black=necked) ‘stilts < Serr. cles oer Himantopus mexicanus Greater yellow-leg............000+ Totanus melanolucus Littleyyellow=legs-.0/:, tesrietseraeeete Totanus flavipes east, sandpipers;..-i c i ie Io I2 CONKLIN—LINERGES MERCURIUS s 14 16 PLATE 3 ma CONKLIN—LINERGES MERCURIUS PLATE 4 Habits and Early Development of Linerges mercurius. 161 cavity is filled with the gelatinous or fluid substance which forms the ground-substance of the central area of the unsegmented egg (cf. figs. 30, 40). Whether this cavity is in any sense an artifact, and if so to what ex- tent, are questions which are difficult to answer, since the eggs are so opaque in life that their centers can not readily be seen. However, a cleavage cay- ity is a normal feature in the later stages of this and of most other eggs, and since the cavity present along the line of the first cleavage is directly continuous with this later cleavage cavity I conclude that it is normal and not an artifact. This cavity is found only in the region of the central area of the egg and it does not extend through the peripheral lavers to the sur- face; furthermore, it does not communicate with the cleavage head, if one may so judge from the fact that the fluid contents of the cavity do not escape. A large part of the ground-substance of the central area of the egg escapes into this cavity during the first cleavage (fig. 41, 43), and most of that which is left in the blastomeres escapes into the cleavage cavity during the second and third cleavages (figs. 44, 45). Owing to the escape of this fluid substance from the blastomeres the latter are left much more compact and with yolk spherules more closely crowded together than in the unsegmented ege (cf. figs. 37, 44). Another evidence that the escape of this central ground-substance into the cleavage cavity is a normal occurrence is found in the fact that although the cleavage cavity becomes quite large, the volume of the entire egg is scarcely greater in the 8-cell or 16-cell stage than in the t-cell stage (cf. figs. 37 and 46). The substance which escapes into the cleavage cavity probably represents a fluid yolk, which is gradually used up in the nourishment of the embryo. Second and later cleavages.—The subsequent cleavages are fairly regu- lar and in all of them, as far as I have observed, the nuclei divide by mitosis and in a manner similar to that described for the first cleavage. The divi- sions of the cell bodies are also similar to that of the first cleavage, though some of them merit a special description. The second cleavage begins along the line of the first, in the center of the egg, and cuts through to the periphery (figs. 15-18). This cleavage pro- gresses more rapidly on the side of the animal pole than on that of the vegetal pole, with the result that the connecting strand between the daughter cells is left at the periphery of the egg in the vegetal hemisphere (figs. 19, 20) ; later, perhaps by a slight rotation of these cells, this strand is carried still nearer to the vegetal pole (fig. 20). During this cleavage the animal hemi- sphere of the egg is highly arched, the vegetal hemisphere flat (fig. 19). The 4 blastomeres which result from this cleavage are approximately equal in size. The third cleavage is equatorial and divides the egg into 8 adequal cells (fig. 21). This cleavage also begins at the center of the egg adjoining the cleavage cavity (figs. 44, 45) and cuts through to the periphery, where a connecting strand is left (fig. 21). This is the latest stage in which I 12 162 Papers from the Marine Biological Laboratory at Tortugas. have observed the connecting strand of peripheral protoplasm between two daughter cells. This strand is evidently the result of the unilateral constric- tion of the cell body, and this in turn has been held to be due to the greater amount of peripheral protoplasm on one side than on another (Ziegler, 1903). If this be true of the third cleavage as well as of the first, the proto- plasmic layer should be thickest at this stage on the surface adjoining the cleavage cavity, but this is never the case. In my opinion, unilateral con- striction is due in large part to the presence of a more fluid central area, as I shall show in the last section of this paper, and, therefore, when the sub- stance of this central area disappears in the formation of the cleavage cavity and its contents, unilateral constriction during cell division also disappears. The escape of the central ground-substance into the cleavage cavity may often be seen to take place by the separation of small globules from the inner ends of the cells, in a manner similar to that in which the so-called “plasmic corpuscles” arise in the blastoccel of Phoronis (Ideka, 1901) and Terebratulina (Conklin, 1902). In the 8-cell stage shown in figure 44 the coagulated contents of the cleavage cavity are shown escaping at the vegetal pole, probably owing to shrinkage due to fixation. The 8-cell stage gives rise by meridional cleay- ages to the 16-cell stage (fig. 45). Figures 25 to 27 represent three suc- cessive cleavages of one and the same egg. Up to this stage the cleavages are normally quite regular; sometimes, however, they are more or less irregular, as shown by figures 22 to 24. These irregularities consist mainly in the temporary suppression of the division of the cell-body in one or more of the blastomeres. However, the nuclei in these blastomeres continue to divide by mitosis and the cell bodies subsequently divide (fig. 24) ; such eggs frequently give rise to normal blastule and gastrule. I have never observed such irregularities of cleavage in Linerges as have been seen by Hargitt (1904, 1906) in Pennaria, Endendrium, and Clava. In the transition from the 32-cell to the 64-cell stage every nucleus divides by mitosis, as is shown in figure 28.. In later stages the cell divisions have not been followed in detail, but I have nowhere seen any evidence of amitosis. The 64-cell stage (fig. 29) gives rise to a stage of about 128 cells (fig. 30), and the latter to a stage of double that number of cells. Blastula and gastrula—tIn a stage of about 500 cells (fig. 31) the cleavage cavity is somewhat eccentric toward the vegetal pole, and the cells at this pole are more rounded and less elongated than elsewhere; these rounded cells are endoderm. In many cases the polar bodies remain at- tached to the egg within the egg-membrane, and in such cases they usually lie at the pole opposite the endoderm pole, as is true of practically all ani- mals. The animal pole of the egg becomes, therefore, the ectodermal pole of the gastrula, the vegetal pole of the egg, the endodermal pole of the gastrula. After the 32-cell stage the entire embryo grows larger, apparently Habits and Early Development of Linerges mercurius. 163 through the increased size of the blastoccel (figs. 29 to 32, 47, 48). At the same time the contents of the blastoccel become more fluid and stain less deeply, as compared with the earlier stages. In a stage with about 1,000 cells (fig. 32), clear protoplasmic processes which resemble pseudopodia appear over the entire periphery of the embryo, save a small area at the vegetal pole. These processes are usually blunt- conical in shape, though some of them are very irregular. They lie under the egg-membrane, which they lift from the surface of the egg. These processes are the first steps in the formation of the cilia which ultimately clothe the entire outer surface of the larva. At first they move slowly and irregularly, but later, as they grow more slender, they vibrate in a typical manner. Gastrulation usually takes place by invagination (figs. 33, 34, 47 to 49). The small rounded cells at the vegetal pole are pushed into the blastoccel and become the walls of the enteron, a flask-shaped cavity which opens to the exterior through the blastopore at the vegetal pole. The blastopore soon closes, so that the enteron is shut off completely from the exterior (figs. 35, 30, 50). Sometimes gastrulation takes place by the immigration of a mass of endoderm cells at the vegetal pole, and in such cases there is at first no enteric cavity in this mass of entoderm cells. Later these cells separate and arrange themselves around an enteric cavity, and the end result is the same as in cases of typical invagination. The close relationship between unipolar immigration and invagination is thus clearly shown by the occurrence of both processes in different eggs of the same animal. In other genera of Scyphomeduse all forms of gastrulation (invagina- tion, immigration, delamination) occur. This fact indicates that the form of gastrulation is of no fundamental or general significance, but that it depends upon individual or environmental conditions. Planula—After the closure of the blastopore the embryo elongates and becomes a free-swimming planula (figs. 35, 36). The endoderm no longer forms a simple layer, but consists of a more or less irregular mass of cells, within which is the enteric cavity (figs. 35, 36, 50). In many cases (perhaps in all) several small ingrowths of ectoderm cells into the space between ectoderm and endoderm takes place (plate 8, fig. 50, text-figs. 1, 2). These ectodermal masses then become hollow. Owing to my failure to obtain material of the later stages of development I have been unable to determine their significance. The latest stage in the development of Linerges which I have seen, cor- responds to plate 8, figure 50 and to text-figures I and 2. At this stage there is no opening into the enteric cavity, though the ectodermal invagination shown at the narrower end of the larva in text-figure 2 may represent the formation of the mouth. With this stage—the free-swimming planula observations on the normal development come to an end. my 164 Papers from the Marine Biological Laboratory at Tortugas. e290 0 oo ° ° #0 °. ° ° 0 9 28222 Fics. 1 AND 2.—Longitudinal sections of the advanced planule, showing the columnar ectoderm, the more or less solid endoderm, containing irregular or radiating cavities, and the ectodermal invaginations which lie between the ectoderm and endoderm and are of doubtful significance. An ectodermal invagination at the narrower pole of fig. 2 may represent the stomodzal invagination. As compared with the development of other medusz, the entire embry- ology of Linerges is characterized by the regularity of the processes of cleavage and gastrulation; and although this regularity may suffer certain modifications, without preventing the formation of a normal planula, there is in this species none of that extreme irregularity which characterizes the development of Pennaria (Hargitt, 1904). EXPERIMENTS. Isolation of blastomeres—NMy observations on the development of parts ot the unsegmented egg and of isolated blastomeres are essentially similar to those of Zoja (1895) and Maas (1905). Parts of the unsegmented but fertilized egg may give rise to swimming larve; these are almost certainly the parts containing the egg and sperm nuclei. Isolated blastomeres, at least as late as the 4-cell stage, give rise to swimming larve, which are ap- parently normal; however the lack of clearly differentiated organs in the planula makes it difficult to determine in this stage whether the larve are wholly normal or not. When the egg fragments are small, or when the blastomeres are isolated at a late stage of the cleavage, the blastoccel is rela- tively small and the gastrulation is not normal. These results are essentially like those obtained by all investigators of the development of the Cnidaria. CONKLIN—LINERGES MERCURIUS BEATE Habits and Early Development of Linerges mercurius. 165 Centrifugalized eggs—It has been found by Lyon (1906), Lillie (1906), and Morgan and Lyon (1907), as well as by myself that the substances of the eggs of many animals may be separated into zones by means of strong centrifugal force. When the eggs of Linerges are centrifuged immediately after being laid, but little separation of the egg-substances is produced even if they be centrifuged at the relatively rapid rate of 12,000 revolutions per minute for 2 minutes. In later stages (just before and after the first cleav- age) the substances of the egg separate much more readily. Fertilized eggs in the stages just before and after the first cleavage when centrifuged for I minute at the rate of 10,000 revolutions per minute remain unaltered in structure and subsequently develop normally. If the same eggs are centrifuged at the same rate for 2 minutes the substances of the egg are separated into a clear, a blue, and a yellow zone, of which the first is the lightest and the last the heaviest. Many such eggs are evidently killed or so injured that they do not develop further, one such being shown in figure 38; others undergo irregular cleavages, which show abnor- mal distribution of these odplasmic substances. In some cases irregular planule are formed, with an abnormal distribution of the egg-substances, and these may live and swim about for at least 24 hours. For example, in one case the clear substances formed a prominence on one side of the larva, while the remainder of the planula was blue or bluish-yellow. Whether such a planula would give rise to a normal scyphistoma was not determined, since the larve were not reared to this stage. And this leads me to remark that in an organism in which there are so few differentiated parts as in a planula it is practically impossible to determine with certainty whether ex- periments on the egg have modified its potency; only the study of later and more complicated stages could yield conclusive evidence on this point. CONCLUSIONS. The organization of the egg of Linerges—The differentiations of the egg are limited to polarity and to the existence of concentric layers of differentiated o6plasm. Polarity is clearly marked in the egg before matura- tion, and as in practically all other animals the maturation pole becomes the aboral or ectodermal pole of the gastrula and planula, while the opposite pole of the egg becomes the oral or endodermal pole. The egg and embryo are radially symmetrical. The cleavages are ap- proximately equal and synchronous, and the earliest differentiation of the cleavage cells consists in the appearance of the rounded endoderm cells at the vegetal pole, while the ectoderm cells are long and narrow, form- ing a columnar epithelium (figs. 31-34). Of the different concentric layers of odplasm, the peripheral one is nearly free from yolk, and is slightly thicker at the animal pole than else- where. It becomes the peripheral layer of the blastula and gastrula and gives rise to the cilia which clothe the ectoderm. 166 Papers from the Marine Biological Laboratory at Tortugas. Beneath the peripheral layer is a layer of closely crowded yolk spherules in which the nuclei lie during the early and late cleavages: this yolk-rich layer constitutes the principal part of all the cells of the blastula and gastrula. The central area of the egg contains scattered yolk spherules within a semi-fluid yolk or matrix. During the cleavage this matrix is poured into the cleavage cavity, where it seems to serve as a kind of fluid yolk for the nourishment of the surrounding cells. The central area of the egg is thus the precursor of the cleavage cavity and its contents. The view expressed in my preliminary note on the development of Linerges (1906), that the 3 layers of the egg give rise to the ectoderm, the endoderm, and the mesoglcea is not confirmed by further study. Portions of each of these 3 layers are found in all the cells of the blastula and gas- trula, and consequently in both ectoderm and endoderm; therefore these substances are not organ-forming with respect to the germinal layers. Never- theless, each of these substances, under normal conditions, produces a specific part of the embryo. The fact that parts of the unsegmented egg or isolated blastomeres may give rise to entire planulz may be due to the fact that such parts or blastomeres contain the odplasmic substances of all of these 3 layers, owing to the concentric arrangement of these substances. It is not easy to isolate these substances and observe the development of each, for although they may be more or less completely separated by strong centrifugal force, they do not, in most cases, undergo further development ; and the more com- pletely these substances are separated the less capable they are of develop- ment. Nevertheless, as far as my experiments go they indicate that these odplasmic substances are not individually totipotent. Mechanics of cell-division—The peculiar form of cell-division found among ccelenterates has attracted much attention. Ziegler (1898, 1903), Rhumbler (1899), and Fischel (1898, 1906) have dealt with this problem in a comprehensive manner. In brief, Ziegler holds that cell-division is brought about by the activity of the outer protoplasmic layer, the unilateral constriction being due to a heaping up of this layer to form a “cleavage head” at the animal pole; this heaping up he regards as the result of the action at a distance (“‘ Fernwirkung’”’) of the centrosome. Rhumbler also finds the cause of unilateral constriction in the heaping up of the peripheral layer in the cleavage head, due to the astral rays, and in the increased membrane formation in the plane of cleavage, due to the escaped nuclear substances which lie in this plane. Fischel believes that the cleavage is explicable on the assumption that the astral rays are contractile threads, and that the unilateral constriction is due chiefly to the peripheral position of the nucleus and centrosome. In the case of Linerges a glance at figure 6, and the photomicrographs 38 and 42 shows that while the nuclei and centrosomes lie near the animal pole of the egg they do not lie in the peripheral layer of protoplasm. Furthermore, Habits and Early Development of Linerges mercurius. 167 these nuclei are no nearer the periphery than in the case of many other eggs (mollusks, annelids, etc.) in which unilateral constriction does not occur. Again, the heaping up of the protoplasm at the animal pole is by no means as great as in many other animals, such as the annelids and mollusks, in which unilateral constriction does not take place. Therefore, the cause of the peculiar form of cell-division found in the early cleavage of the egg of Linerges and other ccelenterates is not satisfactorily explained by any of these hypotheses. In most animals the cleavage of the egg begins at the animal pole, where the protoplasm is most abundant, and the cleavage-furrow then extends around the egg and gradually cuts in from all sides toward the center. In such cases the center of the egg is quite as firm as is the peri- phery. On the other hand, the center of the egg is less firm in some coelenterates (perhaps in all) than is the periphery, and it seems probable that the unilateral constriction in the division of these eggs is partly due to this condition ; for after the furrow has cut through the peripheral layers to the more fluid central area it would then progress rapidly toward the vegetal pole. This type of cleavage prevails during the first, second, and third cleavages, in fact as long as the more fluid central area is present, but with the disappearance of this area in the formation of the cleavage cavity and its contents, this type of cleavage disappears, and the ordinary type thereafter occurs. Taken in conjunction with the heaping up of the peri- pheral layer at the animal pole, this is, I believe, an explanation of the uni- lateral type of constriction of the eggs of ccelenterates. In conclusion, I desire to express my thanks to the Carnegie Institution of Washington and to Dr. Alfred G. Mayer, Director of the Tortugas Labor- atory, for the generous assistance given me while a guest of the Laboratory. REFERENCES. Cuitp, C. M. 1907. Amitosis as a factor in normal and regulatory growth. Anat. Anz., Bd. 30. Conk1in, E. G. 1902. The embryology of a brachiopod. Proc. Am, Philosophical Society, vol. 41. 1906. Preliminary report upon the structure of the egg of Linerges. Carnegie Institution of Washington Year Book, No. 4, 1905, p. 115. FIscHEL, A. 1898. Experimentelle Untersuchungen am Ctenophorenei. Arch, Ent. Mech., Se 1906. Zur Entwicklungsgeschichte der Echinodermen. Idem, Bd. 22. HAECKER, V 1902. Ueber das Schicksal der elterlich und grosselterlich Kernanteile. Jena. Zeit. f. Naturwiss., Bd. 37. Hareitt, C. W. 1904. The early development of Pennaria tiarella. Arch. Ent. Mech., Bd. 18 1904. The early development of Eudendrium. Zool. Jahrb., Bd. 20. 1906. The organization and early development of the egg of Clava leptostyla. Biol. Bull., vol. to. 1901. Observations on the development, structure, and metamorphosis of Actino- trocha. Jour. Col. Sci., Tokio, vol. 13 Lrg, F. R. 1906. Observations and experiments concerning elementary phenomena of devel- opment of Chetopterus. Journ. Exp. Zool., vol. 3. Eeyvon, 7b: 1906. Some results of centrifugalizing the eggs of Arbacea. Am. Jour. Physiol., vol. 15. Maas, O. 1905. Experimentelle Beitrage zur Entwicklungsgeschichte der Medusen. Zeit. wiss. Zool., Bd. 82. Morean, T. H., and Lyon, E. P. 1907. The relation of the substances of the egg, separated by a strong centrifugal force, to the location of the embryo. Arch. Ent. Mech., Bd. 24. RuuMB ter, L. 1899. Allgemeine Zellmechanik. Ergebnisse der Anat. und Entwickl., Bd. 8. ZieGier, H. E. 1898, 1903. Experimentelle Studien ueber die Zelltheilung. Arch. Ent. Mech. Bd. 8, 16. Zoya, R. 1895. Sullo sviluppe dei blastomeri isolati di alcune Meduse. Arch. Ent. Mech., Bd: a: 168 DESCRIPTION OF PLATES. All the figures of plates 1 to 6 represent entire eggs of Linerges mercurius, either drawn in the living condition or stained and mounted entire. They were drawn with the aid of the camera lucida under Zeiss apochromatic obj. 8 mm., oc. 6, at table level; in the process of reproduction they were reduced about one-third, so that as they now appear they represent a magnification of about 230 diameters. Plates 7 and 8 are photomicrograplis of sections, magnified about 175 diameters. PLATE I. Fics. 1, 2. Eggs before maturation, showing the peripheral position of the large germinal vesicle. Fic. 3. First maturation division; the maturation spindle is extremely small and apparently without centrosomes or astral rays; at the vegetal pole is a lobe of clear protoplasm which may represent the “ yolk lobe” of other animals. The peripheral layer of clear protoplasm is marked off from the deeper-lying substances of the egg by the crenated line and the spherules, which represent yolk. Fic. 4. Second maturation division; the spindle is here smaller than in the first divi- sion; within a clear area at the vegetal pole a chromatic body is found which probably represents the sperm-head. Egg showing the two germ nuclei side by side. Anaphase of the first cleavage-spindle, showing the asters, chromosomal vesicles, and connecting fibers. Fic. Fic. au PLATE 2. Fic. 7. Appearance of cleavage furrow and “cleavage-head” at the animal pole; each of the daughter nuclei consists of two parts, the gonomeres. Fics. 8, 9. Further stages in the formation of the first cleavage furrow by unilateral constriction. Fics. 10, 11. Eggs in which the cleavage head turns to one side after reaching the vegetal pole. Fic. 12. Egg viewed from the vegetal pole, showing the connecting strand between the two blastomeres. PLATE 3. Fic. 13. Egg showing the cleavage head turned still farther to one side. Fic. 14. 2-cell stage from the animal pole, showing the blastomeres flattened against each other. Fics, 15-18. Successive stages in the formation of the second cleavage-furrow, show- ing the cleavage-head advancing from the center of the egg toward the periphery. PLATE 4. Fic. 19. Side view of fig. 18, in the line of the first cleavage, showing the arched condition of the egg at the animal pole and its flat appearance at the vegetal pole; the outlines of the four blastomeres are shown at the animal pole, but the cleavage-furrow has not yet reached the periphery, and hence the daughter cells are still held together here by the con- necting strand. Fic. 20. Side view of an egg at the close of the second cleavage, seen in the line of that cleavage; the connecting strands between the daughter cells are shown near the vegetal pole. Fic. 21. 8-cell stage; side view of an egg at the close of the third cleavage, showing at the periphery the connecting strands between daughter cells. Fics. 22-24. Eggs showing irregular cleavage, due chiefly to the suppression of the division of the cell bodies; fig. 24 shows the nuclei and the appear- ance of the delayed cleavages in some of the cells. 169 170 Papers from the Marine Biological Laboratory at Tortugas. PLATE 5. Fics. 25-27. Successive stages in the cleavage of the same egg; fig. 25, 8 cells; fig. 26, 16 cells; fig. 27, 32 cells; the daughter cells are connected by arrows. Fic. 28. Stage transitional between 32 and 64 cells; every cell shows a nucleus in some stage of mitotic division. Fic. 29. 64-cell stage, showing (in dotted outline) the enlarging cleavage cavity. Fic. 30. 128-cell stage, showing the cleavage cavity much enlarged. PLATE 6. Fic, 31. Stage of about 500 cells, showing the rounded endoderm cells at the vegetal pole; the cleavage cavity is eccentric toward this pole; several of the cells are dividing by mitosis. Tic. 32. Stage of about 1,000 cells, showing the appearance of pseudopod-like proc- esses from the ectoplasm, which later become cilia. Fics. 33, 34. Optical sections of gastrule, showing gastrula invagination, rounded endo- derm, and mesoglcea cells, and the high columnar ectoderm cells. Fic. 35. Optical section of a stage after the closure of the blastopore. Fic. 36. Optical section of an elongated planula. PLATE 7. Fic. 37. Photomicrograph of egg after the formation of the polar bodies, showing the thin peripheral layer of protoplasm, the dense layer of yolk, and the central more fluid area. Fic. 38. Egg centrifugalized for 1 minute at 12,000 revolutions per minute, and fixed in Flemming’s fluid 6 hours later. The clear protoplasm is aggregated chiefly at the animal pole; the germ nuclei are large clear vesicles which have not undergone division; the central area is still less dense than the peripheral yolk layer. Fic. 39. Normal egg, showing the germ nuclei in contact near the animal pole, and the substances of the egg arranged as in fig. 37. Fics, 40-42. Successive stages in the first cleavage of the egg, showing the furrow cutting into the central area and the formation of the cleavage cavity; one or both of the nuclei of the two cells show in all the figures. Fic. 43. Oblique section through an egg at the close of the first cleavage, showing the cleavage cavity filled with the central matrix of the egg. Fic. 44. Vertical section of an 8-cell stage, showing the contents of the cleavage cavity escaping at the vegetal pole. Priate 8. Fic. 45. Horizontal section of an 8-cell stage, showing the mitotic spindles for the fourth cleavage; the blastomeres are partially torn apart. Fic. 46. Section through a 32-cell stage, showing mitotic figures in some of the cells; these figures are surrounded by small, deeply-stained spherules. Fics. 47, 48. Vertical sections through blastule at the beginning of invagination; nuclei are shown in many of the cells; the cleavage cavity is greatly enlarged and its contents stain less deeply than in previous stages. Fic. 49. Invaginate gastrula, showing the thin-walled endoderm and the thick-walled ectoderm, together with blastoccel, gastroccel, and blastopore; in the ectoderm the cells are faintly indicated by the vertical rows of yolk spherules, Fic. 50. Longitudinal section through a planula, showing ectoderm, endoderm, and three invaginations of ectoderm into the blastoccel; the cells of these ectodermal invaginations are quite unlike the endoderm cells, while they closely resemble those of the ectoderm. PLATE 7 CONKLIN, LINERGES MERCURIUS PLATE 8 CONKLIN, LINERGES MERCURIUS VII. TWO PECULIAR ACTINIAN LARVAT FROM TORTUGAS, FLORIDA. By EDWIN G. CONKLIN, Professor of Zoology, University of Pennsylvania. 4 plates, 5 text figures. 171 wet (RV AL MAI TORA Mah Leb Ow POTISVa TS? PRY uty: HipsAsr ur. ci OA A + (eS Wee OU Perr) %& Sipe “ae -¢ TWO PECULIAR ACTINIAN LARVA? FROM TORTUGAS, FLORIDA. By Epwin G. Conkiin. During a residence of about three weeks at the Marine Laboratory of the Carnegie Institution of Washington at Tortugas, Florida, in April and May, 1905, some peculiar actinian larve belonging to two different types were several times taken in the tow. These larve were usually taken toward the middle of the day, rarely in the morning or evening, which fact suggests that they come to the surface during the brightest part of the day and again sink to greater depths when the light becomes faint. Several of these larvae were kept during the whole of my stay at Tortu- gas, and Dr. Mayer, director of the station, kindly reared them and col- lected other specimens for me after my departure, for a further period of six weeks, but during this time no one of them became sedentary or transformed into an adult form. Again, in the summer of 1906, Dr. Mayer collected many of these larve and kept them in aquaria for sev- eral weeks, but no one of them underwent metamorphosis. Finally, I found these larvee in considerable abundance at Nassau, Bahamas, in April, 1907. Although some of the individuals were quite large, no one of them had passed the larval stage. They can not, therefore, be definitely referred to any known species of Actinozoan, and although in their structure they show many adult features, they must still be regarded as larval, or at least immature forms. When first taken these larve were wholly unknown to the writer, and indeed, while they were still living it was not certain to which phylum of the animal kingdom they might belong. One type bore a superficial resemblance to an annelid larva, while the other was apparently unique, but when they were killed and prepared for microscopical study it was easy to see that they belonged to the Actinozoa, and that they were immature or larval forms. On my return from Tortugas a consultation of the litera- ture on this group showed that similar forms had been found in various tropical or subtropical seas and that they probably belong to the family of the Zoanthide of the order Hexactinia. The most striking peculiarity of these larvee is a band of locomotor cilia, which is beautifully iridescent, like the swimming-plates of ctenophores. These cilia are long and extremely numerous, and in living specimens they 173 174 Papers from the Marine Biological Laboratory at Tortugas. seem to adhere into a plate or membrane, as in the case of ctenophores, but after being fixed and sectioned they are frequently found separate. In one type of larva (fig. 1) this band is longitudinal, extending from the mouth-opening along one side of the body through about two-thirds of its length; the body itself, in this form, is pear-shaped, the mouth being at the narrower end. In the other type (fig. 2) the ciliated band is circular and surrounds the body about the level of the inner end of the pharynx; at this place the body is deeply constricted, the cilia arising from the bottom of this constriction. The first type was originally described by Semper (1867) and it has since been generally known as “ Semper’s larva.” He found it near the Cape of Good Hope in the Mozambique Channel and on the coast of Java. The second type was also observed by Semper, but was insufficiently described by him and was supposed to be only an earlier stage in the development of the first type. Subsequently E. Van Beneden (1890, 1898) found two specimens of the first type of these larve and one of the second in the material brought back by the Plankton expedition of Hensen; the former came from a region just south of the Cape Verde Islands, the latter from the Guinea Current. Van Beneden pointed out the resemblance of these larve to the Zoanthee (microtype of Erdmann, 1885) in that, in common with this group, these larvee possess a pair of perfect ventral directives, a pair of imperfect dorsal directives, and between these on each side two pairs of mesenteries, the dorsal member of each pair being perfect and the ventral member im- perfect. Van Beneden made a careful study of the morphology and his- tology of these larve, reference to which will be made later. He proposed for these zoanthid larve, the adults of which are unknown, the follow- ing provisional names: For type I with the longitudinal band of cilia, the generic name Zoanthella; for type II with the circular band, the generic name Zoanthina. MecMurrich (1891) has also described a larva of this second type, 5 specimens of which were collected by the aid of the surface-net at Beau- fort, North Carolina. He was unable to determine the adult form to which this larva belongs, but he agrees with Van Beneden that it is the larval stage of a zoanthid. Still more recently a somewhat similar form, though showing cer- tain notable differences, has been found and studied at Beaufort by Cary (1904), who reared the larve until they transformed into the adult form, which, however, he was unable to identify, although he suggests that it may be some species of the genus Amophyllactis. Finally Heath (1906) has described a larva of type I which was taken near the Galapagos Islands, and which is specifically distinct from the forms described by Semper and Van Beneden. The fact that I have had Two Peculiar Actuuan Larve from Tortugas, Florida. 175 for study more specimens than any of the investigators named and that my material was excellently fixed for histological examination has in- duced me to give a rather detailed description of these peculiar larvz. CLASSIFICATION. As already noted, Van Beneden proposed for these two types of larve the generic names Zoanthella and Zoanthina, with the express statement that these names are to be regarded as provisional and that they are to be dropped as soon as it is possible to refer these larval forms to any known adult actin- ians. The following are the characters of these genera, according to Van Beneden: Zoanthella: Pelagic larve, attaining a length of 13 mm. Body elongated, pro- vided with a flagellar plate of distinct cilia, or with a vibratile fringe, extending parallel to the axes of the body along the anterior median line. Twelve septa, of which six are macrosepta and six microsepta, disposed as in the Zoanthariz (micro- type of Erdmann). Aboral pore present or absent. Under this genus Van Beneden recognized two species, viz: Z. semperi: Body cylindrical, with spiral torsion of median plane and vibratile fringe, the latter running from one pole to the other; body not incurved on ventral side at oral end; with an aboral orifice. Mozambique Current. Z. henseni: Body elongate pyriform; large end aboral, actinosome terminal at small end; body incurved on ventral side at oral end; no aboral orifice; vibratile fringe in the upper (oral) two-thirds of body, lacking in aboral third. Guinea Current. To these Heath (1906) has added a third species with these char- acteristics : Z. galapagoensis: Body spindle-shaped; ciliated fringe slightly spiral, exposed cilia-bearing portion of fringe of much smaller extent than side in contact with mesoglcea; body not incurved; cilia of fringe not fused into membrane; no aborai pore. Near Galapagos Islands. In almost every particular the specimens which I obtained at Tor- tugas and Nassau resemble Z. henseni; the only difference of note is that in Van Beneden’s specimens the cilia of the vibratile fringe are not closely adherent into a continuous membrane, whereas in my specimens the cilia are intimately connected together in life, though they may be more or less separated after fixation. Inasmuch as Van Beneden had only two pre- served specimens for study it seems probable that the separation of cilia may have resulted from the fixation, and that in life the cilia adhere closely together, as in my specimens. The Tortugas and Bahama specimens are, therefore, in all probability, examples of Z. Jenseni, and the fact that they are found in the Gulf Stream as well as in the Guinea Current indicates their very wide distribution. The characters of the genus Zoanthina are indicated by Van Beneden as follows: 176 Papers from the Marine Biological Laboratory at Tortugas. Pelagic larvae, body ovoid in early stages; a circular constriction, bearing flagellz, divides the body into two unequal parts, an upper smaller part containing the mouth and the pharynx, and a lower larger part; 12 septa are present, 6 macrosepts and 6 microsepts, as in the microtype of Erdmann. Z. nationalis: Larve longer than wide (one with 12 septa measuring 2.2 mm. in length, 2 mm. in width) ; ciliated furrow very deep; oral part of body attached to aboral by a sort of peduncle; no ectodermal papillae near the mouth; no canals in mesoglcea; an orifice at aboral pole. Guinea Current. Z. americana: Larve less elongated in the chief axis (one with 12 septa measures 1.4 mm. long, 1.5 mm. wide); ciliated furrow less deep; ectodermal papille present near mouth; canals in mesoglcea; no aboral orifice. Beaufort, North Carolina. It is not possible to determine with certainty whether the specimens from Tortugas belong to either of these species; in fact it is not certain that the specimens obtained by Van Beneden and MecMurrich belong to dif- ferent species. In most regards the Tortugas specimens closely resemble both of these species, though in some respects they are different. From Z. nationalis the Tortugas forms differ in having no trace of an aboral pore, and if Van Beneden had not made a careful histological study of this form it might be doubted whether such a pore actually exists. As it is, it is at least possible that such a pore represents an artifact rather than a normal structure. Another difference is found in the length of the actino- pharynx; in Van Beneden’s specimen it does not extend as far inward as the level of the constriction; in the Tortugas specimens it does. The Tortu- gas form differs from MeMurrich’s larva only in the lack of oral papilla and of large canals in the mesoglcea at the bases of the septa. These differences may be due to the fact that McMurrich’s oldest larva was more advanced in development than any I have examined. On the whole, then, there does not seem to be sufficient ground for considering the Tortugas form as the type of a new species. Numerous specimens of Zoanthina were taken at Nassau, Bahamas, in April, 1907. Of these there were two different kinds; one, a small yellowish form, identical with the Tortugas species, the other frequently much larger and of a violet color. Many individuals of this fatter form were as much as 5 mm. long and 3 mm. in diameter, but no one of them showed any trace of tentacles. The larva described by Cary (1904) is evidently very distinct from the forms just named, and it is doubtful whether it can be included in the same genus; in this form the circular band consists of bristles instead of cilia, and it forms a ring open at one side instead of a closed one. This summary shows that these larve have been found chiefly on the high seas in the Atlantic, Pacific, and Indian oceans. The specimens taken at Tortugas were virtually from the Gulf Stream, and it is not improbable that those obtained by McMurrich at Beaufort and by myself at Nassau were also from this same ocean current. What the habitat of the adults may be is purely a matter of conjecture. CONKLIN, ACTINIAN LARVA PLATE 1 Apu ~ Se see Ry ¥ Vie Si A la 2a Fig. 1. Zoanthella henseni (?), from Tortugas, Fla. Drawn from life by Dr. A. G. Mayer. x10. Fig. la. Photomicrograph of a longitudinal section of this species. The specimen is strongly flexed toward the ventral side, and associated with this is the wrinkling of the body wall at the aboral end of the ciliated band; the body wall is not wrinkled in the region of the band. x10. Fig. 2. Zoanthina americana (?), from Tortugas, Fla. Drawn from life. x14. Fig. 2a. Photomicrograph of a longitudinal section of Zoanthina; although the section passes through the middle of thepharynx and cuts the ciliated groove on both sides of the mouth, the ciliated epithelium is shown on one side only. This is the only case observed in which the ciliated ring was incomplete. x75. ~ ual 4 \ i. ihe" a a, Te» Y ems. | % 0 6 a SAAN Ba aonea « atin) Po, Cae eR ian ati ee Ei rat * + CONKLIN, ACTINIAN LARV PLATE 2 Transverse sections of Zoanthella 2 mm. long and 0.5 mm. wide. Photomicrographs by H. G. Kribs, taken with monochromatic blue light, of 448 « wave-length. x125. Figs. 3-5. Sections through the region of the pharynx. Figs. 6-9. Sections in region of ciliated band below the pharynx. Figs. 10-11. Sections in the aboral region, below ciliated band Two Peculiar Actinian Larve from Tortugas, Florida. 1747 NATURAL HISTORY. Apparently no one save Semper has studied any of these forms in a living condition. My observations on the living larve, although by no means complete, may therefore be of some interest. These larvee were taken in the surface-net, near the middle of the day, and usually in regions where the water was deep. They are quite hardy and will live indefinitely in small aquaria, though they do not grow in size or undergo metamorphosis, even though they be kept for several weeks. Zoanthella orients itself so that the blunt aboral pole is directed up- ward, the pointed oral pole downward, and in this position it swims about near the surface of the aquarium, rotating on its long axis in a clockwise direction, when viewed from the oral pole. Its position and movements in the water are the results of the activity of its ciliated band, the undula- tions of which may be plainly seen; floating is evidently an active process, for when the cilia cease to beat the animal falls to the bottom. The rotation of the larva on its long axis is evidently due to the fact that the ciliated band runs in a slightly spiral course toward the right (clockwise) when viewed from the oral pole. Although the undulations which run from one end of the ciliated band to the other are slow enough to be easily seen, the beating of the cilia at any one level may be so rapid that the individual cilia can not be seen. Under these circumstances the ciliated band appears hazy and broader at its free edge than at its attached border. Zoanthina larve usually lie on the bottom when brought into small aquaria and are motionless, except for the occasional contraction of the ciliated band. These contractions, due to the simultaneous beating of all the cilia toward the oral pole, are not unlike the pulsations of the bell of a jelly-fish, and are so slow and feeble that they do not serve to move the animal. Occasionally, however, many individuals may be found swim- ming rapidly near the surface of the water. This happens especially when stale water, in which they have remained for some time, is replaced by pure water. At such times these larvz assume a conical form, the aboral pole be- ing pointed and the oral pole truncated. The deep groove from which the ciliated band arises lies at the border of this flattened oral area and the band itself beats rapidly and violently, the stroke being toward the oral pole. Undulations or irregular contractions of the ciliated band also occur, running around the band in an anti-clockwise direction, when viewed from the oral pole. The resulting movement is quite rapid, the pointed pole being directed forward and the animal rotating in a clockwise direction when seen from the oral pole. When strongly stimulated Zoanthina becomes spherical in shape and the ciliated band disappears within the groove, which closes up. Although some of the larve which I have studied are large and well- developed, no food has ever been found in the ccelenteron; even diatoms and microscopic algz are lacking. Occasionally fine threads of a coagulum 178 Papers from the Marine Biological Laboratory at Tortugas. are found in the ccelenteron, but judging by the considerable number of specimens which I have sectioned I am inclined to believe that solid food is rarely ingested. One of these specimens (figs. 12-18), which was 3.5 mm. long and 1.2 mm. in diameter, after fixation and preservation for a year in alcohol, had no mouth opening, so that it could have taken no solid food, but its considerable size would suggest that it must have grown a good deal beyond the size of the egg. To a considerable extent this growth may be due to the formation of the hollow ccelenteron and to the absorp- tion of water, as Davenport showed to be the case in the early growth of the tadpole, but it is possible that in this case nutriment may be received from another source. In all of these larve there are considerable numbers of large round cells with dense nuclei and with vellowish-green granules in the cytoplasm. These are symbiotic alge, resembling Zoé-anthella, and it is quite possible that they play an important part in the nutrition of the larve. These alge occur in both Zoanthella and Zoanthina and in both the ecto- derm and endoderm, though they are more numerous in the latter layer; they also occur in the youngest larvee, with imperforate mouth, as well as in the oldest ones. The fact that these larve are found at the surface at a period of the day when most pelagic larve have settled to deeper and darker levels may be associated with the metabolism of this symbiotic alga, and it may be that if these larve had been kept in aquaria which were exposed to bright sunlight the later stages in their development might have been secured. MORPHOLOGY. Zoanthella——The general shape of this larva has been sufficiently de- scribed already and may be seen in plate 1, figs. 1, 1a. The color is a greenish or brownish yellow, mottled with darker spots, and it seems probable that this general color is due, in part, to the large number of Zod.xranthelle present in the body-walls. The size of the larve varies within wide limits, as Van Beneden has remarked, and the size is not in itself a measure of the degree of development. My smallest specimen is 2 mm. long and 0.5 mm. wide (plate 2), but it is much further developed than another specimen 3.5 mm. long and 1.2 mm. wide (plate 3). The largest specimen which I have measured is 8 mm. long and 2 mm. wide after having been fixed and cut into longitudinal sections. Many specimens were sectioned without having been measured, and accurate measurements of these can not now be given, but their relative sizes may be determined by the number of sections in each series, since the sections are of uniform thickness; these vary from 250 to 600 in number. In none of the larvee which I have seen is there any trace of tentacles or oral papilla. In Zoanthella the mouth and actinopharynx is formed at a relatively late stage. In the larva, 3.5 mm. long, shown in plate 3, the are pharynx has just begun to invaginate and the mouth is still imperforate, Two Peculiar Actinian Larve from Tortugas, Florida. 17¢ g / although all of the macrosepta are well-developed and the microsepta are already present. In the oldest of these larvee, 8 mm. in length, the pharynx is only 1.1 mm. long. In cross-section the pharynx of Zoanthella is quad- rilateral in shape, the four sides of the quadrilateral being incurved toward the center (plate 2, fig. 3). The longitudinal plications of the walls of the pharynx correspond in position to the primary septa (macrosepta) and below the pharynx these plications are directly continuous with the mesenterial fila- ments of these septa (plate 3, figs. 4-0). The septa consist of three pairs of macrosepta and three of microsepta, and, as Van Beneden has shown, their arrangement and sequence corresponds to that of the Zoantheze (microtype of Erdmann), 7. ¢., the dorsal directives are imperfect, the ventral are perfect, and of the two remaining pairs on each side the dorsal member of each pair is perfect and the ventral one imper- fect. Van Beneden holds that the order of appearance of the septa, as judged by their size, is for the macrosepta: 1, lateral; 2, dorsal; 3, ventral; and for the microsepta: 4, dorso-lateral; 5, ventro-lateral; 6, dorsal direct- ives. In the youngest Zoanthella which I have sectioned the dorsal macro- septa are smaller than either of the other pairs (plates 2 and 3), and judged by the standard hitherto used they are the last-formed of the macrosepta. Three pairs of microsepta are present in these youngest larve and they are all about equal in size. The macrosepta are triangular in cross-section, as shown in plates 3 and 4 and text-fig. 4, the base of the triangle lying at the central border of the septum. To this thickened central border the cylindrical mesenterial fila- ments are attached. A similar condition was observed by Heath in Zoan- thella galapagoensis, but Van Beneden does not figure or describe it in the other species of this genus. In the interspaces between macrosepts and microsepts are endodermal thickenings, which are sometimes as prominent as the septa themselves. These thickenings consist of vacuolated endoderm cells and in the younger larve (plate 3) they fill up a large part of the cavity of the ccelenteron, especially at the aboral end; in the older larve (text-fig.) they are much less voluminous, probably representing the remains of the primitive endo- derm which, in earlier stages, filled the entire enteron. Zoanthina.—The shape of this larva varies greatly with different stages of contraction; it may be elongated as shown in plate 1, figure 2, or con- tracted so that the oral-aboral axis is no longer than the transverse one (plate 1, fig. 2a, text-fig. 1). In some specimens the oral portion of the body, in front of the ciliated band, is relatively much larger than in the specimen represented in plate 1, figure 2. In general this larva is shorter and stouter than Zoanthella. In color it resembles the latter, being of a light brownish- yellow, mottled with darker spots; some of the Nassau specimens are of a violet tint. My smallest specimen of this genus is about 1 mm. long and 1.2 180 Papers from the Marine Biological Laboratory at Tortugas. min. wide, while the largest is 4.3 mm. long and 2 mm. wide. These measure- ments were made on material which had been preserved in alcohol for a year. In all of my specimens of Zoanthina the mouth and pharynx are fully formed and the size of the septa and absence of the endodermal thickenings between the septa indicate that these larvae are more advanced in develop- i) \\ wi a Fic. 1.—Entire specimen of Zoanthina, viewed from oral pole; specimen stained and mounted in balsam. Ectoderm shaded by radial lines, mesoglea by oblique lines, and endoderm by light stippling; ciliated groove deeply stippled. ment than those of Zoanthella which I have studied. The mouth-opening is round in outline, but the pharynx is compressed laterally, as shown in text- figure 1. In this genus the pharynx extends inward at least as far as the circular constriction, about one-third the length of the entire body (plate 1, fig. 2a). The walls of the pharynx are thrown into longitudinal plications, which become continuous with the mesenterial filaments of the macrosepta (plate 4), as in Zoanthella. The septa are relatively larger than in Zoanthella, and this is especially true of the microsepta, but Van Beneden holds that their arrangement and order of appearance are the same in the two genera, 7. e., the dorsal direct- Two Peculiar Actiman Larve from Tortugas, Florida. 181 ives are imperfect, the ventral are perfect, and of the two remaining pairs on each side the dorsal member of each pair is perfect and the ventral one imperfect, while the order of appearance is, for the macrosepta, 1, lateral; 2, dorsal; 3, ventral; and for the microsepta, 4, dorso-lateral; 5, ventro-lateral ; 6, dorsal directives. McMurrich agrees with Van Beneden that the arrangement of the septa of Zoanthina is characteristic of the Zoanthee, but he suggests that the sequence observed by Van Beneden is really due to the retardation of the development of the dorsal directives, which should, according to his view, stand fourth in the order of development. In all my sections of Zoanthina the dorsal directives are smaller than the other imicrosepta (text-fig. 1 and figs. 24, 25), and if the order of appear- ance is to be judged by the relative sizes of the septa, the dorsal directives are the last to appear of these 6 pairs. Therefore, my observations as to the sequence of the septa in Zoanthina agree with the conclusions of Van Beneden. The individual septa are not triangular in cross-section, as in Zoan- thella; on the contrary they are nearly as wide at the base as at the free border, and they all show a greater thickening on one side of the mesoglea than on the other (figs. 24, 25 and text-fig. 1). In the ventral and dorsal directives these thickenings face the median plane, in the laterals and ventro- laterals they face the ventral side, in the dorso-laterals and dorsals they face the dorsal side. Although the microsepta are well-developed, the macrosepta only are united to the pharynx (fig. 21). Van Beneden found that the septa in the oral and aboral portions of the body were not continuous through the region of the constriction. In the Tortugas specimens the septa are small in this region, but they are not interrupted (figs. 22 and 23). Sections through the outer fold of the ciliated groove show the presence of 12 pockets, formed by the 12 septa (fig. 23). HISTOLOGY. In most respects the histological character of these two types of larve is similar, though there are certain minor differences. Ectoderm—The ectoderm consists of greatly elongated cells, among which are numerous cell-spaces. At the free border of the epithelium no cell-spaces are visible and the cells are here more darkly stained than at deeper levels. The inner ends of the cells are narrow and apparently few of them run through the entire epithelium from the free border to the meso- glcea. Among the epithelial cells are nematocysts of two types, smaller ones which lie near the surface, in fact with one end of the nematocyst at the free border of the epithelium, and much larger ones which lie at a deeper level. In text-figure 2, which is a section through the ectoderm and meso- gloea of Zoanthella, both kinds of nematocysts are shown and at the base of 182 Papers from the Marine Biological Laboratory at Tortugas. the epithelium near the left side of the figure is a nematoblast from which the larger type of nematocysts is formed. Two types of gland-cells are also to be found in the ectoderm; one of these consists of elongated fusiform cells containing many small granules Fic. 2.—Section through body-wall of Zoanthella, showing char- acter of ectoderm and mesogloea. x 333. and lying with the outer end near the free border of the epithelium; the other type consists of rounded cells containing larger granules lying at the base of the epithelium; these cell granules stain intensely with plasma stains (text-fig. 2). In Zoanthina there are gland-cells which stain intensely with nuclear stains and which constitute a third type (text-fig. 5). Other cellular elements of the ectoderm are the ZoOxanthellz, which are especially abundant in Zoanthella; these are small round cells with dense nuclei and with yellowish or greenish chromatophores. They are found most abundantly in the outer portion of the ectoderm. At the base of the epithelium and adjoining the mesogloea is a layer of fine fibrils which run in all directions and which consequently appear in sec- tions as fine dots or short fibers. These are probably the fibrillar bases of the epithelial cells. The description just given applies to the general ectoderm of the larve. In the region of the ciliated band this epithelium is remarkably altered. The nematocysts and gland-cells are here lacking, while the ordinary ectoderm cells are replaced by exceedingly slender elongated cells (text-fig. 3). The nuclei of these cells lie in the deeper part of the epithelium, while the cell- bodies consist of slender filaments which are continued from the nuclei to the periphery of the epithelium and then into long flagelle, which constitute the vibratile band. The free border of the epithelium is marked by a faint line. It seems probable that this line marks a plane along which the various cell filaments fuse together. The filaments are apparently as numerous within the epithelium as without it and they are more numerous and more powerfully developed than in any other epithelium I have ever seen. At the base of this ciliated epithelium are a few rounded cells, some of which Two Peculiar Actinian Larve from Tortugas, Florida. 183 are plainly zoOxanthelle, and between the nuclei of the ciliated cells and the mesogloea is a finely granular or fibrillar layer, which resembles the fibrillar layer at the base of the ordinary epithelium. The boundary be- tween the ciliated plate and the ordinary ectoderm is sharp and distinct and there are apparently no transitional cells between the two. The ordinary epithelium does not overgrow the ciliated band at its margins as in the case of Zoanthella galapagoensis. In the latter species the ciliated cells are much longer, as compared with the ordinary ectoderm, than in the Tortugas species. © | | 4 Dy ys Fic. 3—Section through ciliated band and adjoining ectoderm of Zoanthella. Flagellated cells extremely long and slender, appearing like masses of spermatozoa; among them a few larger cells, and at base of epithelium is a granular zone. Mesoglea is very thin beneath ciliated plate, but at edges very thick. XX 333. In Zoanthina the ciliated band is circular and of nearly uniform width; in Zoanthella it is longitudinal and varies much in width, being widest in the middle and narrower at either end. In some specimens of Zoanthella the band is apparently double, being divided along its middle by a line of clear, non-ciliated cells (figs. 13-16). I was at first inclined to the opinion that the forms with the divided band were specifically distinct from those in which it is not divided, but further study makes it probable that this is only an individual variation. A similar splitting of the ciliated plate in its aboral portion was observed by Heath in the Galapagos specimen. The epithelium of this ciliated band has essentially the same structure in the two types of larvee. It is characterized by extremely small and com- pact nuclei, which resemble the heads of spermatozoa, and by very long, slender cell-bodies, every one of which starts from a nucleus and runs as a fibril to the periphery of the epithelium and is then continued into the free 184 Papers from the Marine Biological Laboratory at Tortugas. flagellum. The iridescence of the ciliated band is due to the great number of these parallel flagella, which diffract light like the lines of a fine grating. I have not been able to distinguish muscle or nerve cells in the ecto- derm, though the basal fibrillar layer may represent processes from one or both of these kinds of cells. The epithelium of the pharynx is more compact than that of the general ectoderm and it contains a larger number of gland-cells; consequently it stains more deeply than does the general ectoderm. Endoderm.—With the exception of the mesenterial filaments the endo- derm stains less deeply than does the ectoderm and the cell outlines are less distinct. Indeed, the general endoderm consists of a spongy layer in which TL; Uieazs AGL Lp Zz HG jj GHG Gb, GEE: GE Zz Fic. 4.—Cross-section through body-wall of Zoanthella, showing ectoderm, mesoglea, endoderm, and two septa with mesen- terial filaments; the gland-cells of the latter are especially evident. X 333. are numerous nuclei near the free border, and zoOxanthelle in the deeper portion of the layer, but in which cell boundaries are not distinct, except close to the free border. In Zoanthina the cell boundaries and nuclei are more distinct than in Zoanthella, and along the free border they form a definite epithelium, but in the deeper portions of the layer there are few nuclei and no cell outlines save those of the symbiotic alge, and of a few wandering or amoeboid cells (text-figs. 4 and 5). The mesenterial filaments are rich in gland-cells filled with a granular secretion, while adjoining epithelial cells contain no granules, text-fig. 4. CONKLIN, ACTINIAN LARV PLATE 3 ctions of Zoanthella 3.5 mm. long and 1.2 mm. wide. Photomicrographs by H. G. Kribs, taken with monochromatic blue light of 448 & wave-length. x75. Fig. 12. Section through the oral region, showing endoderm filling the coelenteron; at this stage the mouth is imperforate and the pharynx has not formed. Figs. 13-17. Sections through the region of the open coelenteron; the ciliated band is plainly divided down the middle. Fig. 18. Section through the aboral region, showing the spongy endoderm filling the entire coelenteron. At various places cells are seen in the mesogloea. PLATE 4 CONKLIN, ACTINIAN LARV A= Transverse sections of Zoanthina americana (?), 1.8 mm. long and 1.2 mm. wide. Photomicrographs by H. G. Kribs, taken with monochromatic biue light of 448 &@ wave-length. x75. Figs. 19-21. Sections through portion of body on oral side of ciliated groove. Fig. 22. Section through ciliated groove, showing a portion of the outer wall of the groove. Fig. 23. Section through deepest part of ciliated groove, showing twelve endodermal pockets in outer wall of groove. Figs. 24-25. Sections through the aboral part of the body. The endoderm cells contain many small round bodies, the Zooxanthellz. Two Peculiar Actinian Larve from Tortugas, Florida. 1§ Lal MESOGL(CA. The mesoglcea is generally thicker in Zoanthella than in Zoanthina and varies considerably in thickness in different parts of the same larva. In both types of larvee it is thickest in the aboral portion of the body and thinnest beneath the ciliated band. The extreme thinness of the supporting lamella in the region of the ciliated band as contrasted with its thickness else- where is well shown in text-figure 3 and in plates 2 and 3. The same figures show that in Zoanthella this layer is thinner opposite the point of origin of septa than in the region between septa, while in Zoanthina the re- verse is true. Cells and canals are found within the mesoglcea in both larve, though they are more abundant in Zoanthella. There is considerable evi- Fic. 5.—Cross-section through body-wall of Zoanthina, showing mesoglea and endoderm. XX 333- dence in favor of the view of Van Beneden that the mesoglcea is in life quite soft and that the cells found in it are wandering cells and the canals merely the tracks of these cells. In my oldest specimen of Zoanthina the mesogloea in each of the macro- septa is drawn out into many fine branches on the thicker side of the sep- tum, thus giving support to the longitudinal muscles of the septa. In Zoan- thella the flexure of the body toward the ventral side gives evidence of the presence of strong longitudinal muscles along that side, though I have not been able to distinguish them in my sections. Van Beneden has seen and described these muscles in the specimens which he studied. In none of the specimens which I have examined were there any em- bryos within the ccelenteron, such as Van Beneden discovered in Zoanthina nationalis. These embryos he shrewdly concludes were merely ingested by an older embryo, while all were contained within a ccelenteron of the vivi- parous parent. If this is the true explanation of their presence in Van Beneden’s specimen, it may be concluded that a more extensive study of the 186 Papers from the Marine Biological Laboratory at Tortugas. zoanthinas found at Tortugas will show that some of these contain embryos also. In the oldest of the larva which I have examined there is not a trace of germ-cells and they can not therefore be capable of sexual reproduction. Although these peculiar larvae have been taken only a few times and then in widely separated parts of the earth; they are certainly not ex- tremely rare, as their abundance at Tortugas and at Nassau shows, and I do not doubt that with the excellent facilities now afforded at the Tortugas Laboratory, some investigator more fortunate than myself will be able to observe their transformation into the adult form, and thus to determine be- yond any doubt their systematic position. REFERENCES. Semper, C. 1867. Ueber einige tropische Larvenformen. Zeit. wiss. Zool., Bd. xvii. Van BENEDEN, E 1890. Les Anthozaires pélagiques récuillis par le professeur Hensen, dans son expédition du Plankton. I. Une larve voisine de la larva de Semper. Arch. de Biol., Lt. X. 1898. Les Anthozaires de la “ Plankton-Expedition,’ Res. de la Plankton-Exp. d. Humboldt-Stiftung. Vol. II. McMourricu, J. PLAyFarr. 1891. Contributions on the morphology of the actinozoa. III. The Phylogeny of the Actinozoa. Jour. Morph., v. GarnvaaR: 1904. Notes on a peculiar actinozoan larva. Biol. Bull., vit. HeatH, Haron. 1906. A new species of Semper’s larva from the Galapagos Islands. Zool. Anz., Bd. xxx. VIIl. THE BEHAVIOR OF NODDY AND SOOTY TERNS. By JOHN B. WATSON, Professor of Experimental and Comparative Psychology, The Johns Hopkins University. 11 plates, 2 text figures. 187 »- . : ; : Fr am SUNT OTTO MR A) Oe a eee ~ i} PST ei rere al > tink i aay eames? | bar as ey] lager bee THE BEHAVIOR OF NODDY AND SOOTY TERNS. By Joun B. Watson. INTRODUCTION. During the spring of 1907, on the invitation of Prof. Alfred G. Mayer, Director of the Marine Biological Laboratory of the Carnegie Institution of Washington, I spent three months upon Bird Key, a small island belonging to the Dry Tortugas group. The specific object of my stay was to observe as far as possible the details of the lives of the noddy terns (Anous stolidus) and the sooty terns (Sterna fuliginosa) during their nesting season on that island. My thanks are due first of all to Professor Mayer for his unfailing kindness in liberally supplying my needs upon the island. Bird Key is unin- habited and is some distance from the key upon which the Biological Station is located. Since I lived upon the island, it was necessary to bring all sup- plies and apparatus from the laboratory. During this period I was supplied with a motor-boat and a servant. My thanks are likewise due to the Audu- bon Society, which not only rendered me financial assistance for acting as warden of the island, but also allowed me complete control of the birds. On an island as small as Bird Key it would have been difficult to have car- ried on the work if my authority had been divided with that of a regular warden. Mr. Carl Kellner spent many trying hours with me in attempting to photograph the birds in action. Owing to the peculiarities of the climate and to the extreme rapidity of the characteristic movements of the birds, we had scant success in our efforts. Our interest was not so much in obtaining good photographs of the birds, but rather centered around the portrayal of their activities. Many hundreds of photographs were taken, but many of the most interesting exposures, which were often made after hours of waiting, failed to develop properly for one reason or another. As will be seen by reading the present report, the nature of all the work has been preliminary. Indeed, if I were convinced either that I myself or some one else could immediately take up the work, I should cheerfully delay publication until a fuller account of the activities of the birds could be given. But since the immediate continuation of the work is not assured, and since *In this connection I wish to thank Prof. James R. ‘Angell for arranging for me a three months’ leave of absence from the University of Chicago, with which institu- tion I was connected at the time the present research was made. 189 190 Papers from the Marine Biological Laboratory at Tortugas. ” work of this kind is more or less “ impressionistic,” the attempt is here made, while the material is still fresh in my mind, to enumerate some of the more important problems to be found in the study of these birds and to set forth my tentative efforts to solve them. OBSERVATIONS UPON THE INSTINCTS AND HABITS OF TERNS DURING THE NESTING SEASON. A GENERAL DESCRIPTION OF THE TWO SPECIES OF TERNS. THE NODDY TERN (ANOUS STOLIDUS). The noddy tern is described by Saunders,! as follows: Adult Male in Breeding Plumage—Forehead nearly white at the base of the bill, passing on the crown into lavender-gray, which deepens on the neck into lead color; lores and orbital region black with a faint whitish superciliary streak; upper parts chiefly dark-brown, the primaries, tail feathers, and their shafts nearly black; under- parts dark brown on the abdomen and breast, passing into deep lead color on the throat; bill blackish; tarsi and toes reddish-brown, fully webbed, webs ochraceous. Total length about 16 inches, culmen 2.1, wing 10.25 to 11, tail 6 to 7, the fourth feather from the outside the longest, tarsus 1, middle toe with claw 1.55. Adult Female—Very similar, but as a rule somewhat browner on the shoulders and with less lead color on the throat, slightly smaller and with a weaker bill.” Hab. Tropical and juxta-tropical America; chiefly on the Atlantic side, but also on the Pacific in Mexico and in the central region; Atlantic down to Tristan da Cunha (breeding) ; intertropical African and Asian Seas, up to Yeddo; Australasia down to about 35° S.; Islands of the Pacific up to Laysan, etc., and as far as Sala y Gomez, 105° W.; also Chatham I, Galapagos (fide Ridgway), but not on the coasts of Peru or Chile. Breeding, as a rule, where found. Once obtained off the south coast of Ireland. THE SOOTY TERN (STERNA FULIGINOSA). Rothschild’s* description of the sooty is as follows: Adult—Forehead, sides of head, and entire lower parts, including lower wing coverts, white with a very delicate bluish tinge on the abdomen under wing-coverts and under tail-coverts when the birds are alive or quite fresh. Lores and upper parts, including the hind neck [which is whitish in H. anoestheta (Scop)], uniform sooty brown. Primaries black, but the shaft and outer web of the first primary white below, except on the outermost tip. Tail-feathers sooty-black; all the shafts white below, and the shafts of the outer pair, as well as their outer web and basal part of inner web, white. Total length about 17 to 17.5 inches, wing 11.6 to 12, outer rec- trices (if not abraded) 7.5 to 8, central pair 3, culmen 1.7 to 1.8, tarsus 0.85. (Speci- mens from America and Kermadec Islands are exactly similar.) Iris dark-brown, bill and feet black. Hab. Tropical and juxta-tropical seas, breeding wherever suitable islands and reefs exist; occasionally wandering to Maine in North America, and to Europe, even as far as England. Almost unknown on the South American side of the Pacific; otherwise very generally distributed. 1 Catalogue of the Birds of the British “Museum, vol. XXv, pp. 136-140. *I could not find any difference in appearance between the males and females. The two sexes in life are indistinguishable. This applies to both noddies and sooties. * Avifauna of Laysan, etc., p. 30. The Behavior of Noddy and Sooty Terns. 191 On account of the wide distribution of these two species of birds, of their great numbers, and of their habit of assembling on islands during their nesting season, frequent incidental reference to them is to be found in the writings of naturalists. Saunders gives a fairly complete reference to this literature. Extended statements concerning the instincts and habits of these birds are not extant. Dr. Thompson!’ gives the most comprehensive statement con- cerning their habits. His observations, like my own, were made upon Bird Key. The statements in his paper are apparently made upon the basis of intermittent visits to the island and are not always trustworthy. This is not to be wondered at when we consider the complexity of the life of the birds and the limited time which was at Dr. Thompson’s disposal. Descriptive and none too exact statements of the nesting behavior of the birds may be found in Henshaw? and in Rothschild.’ In the latter, most of the observations were made by the naturalist Palmer. The work of Walter K. Fisher* is especially worthy of mention. The above references bear only in a general way upon my own studies. In many cases, the observations to be found in them do not agree with my own. Iam not able to account for these discrepancies. The account of the instincts and habits of these birds given here is made largely upon the basis of my own observations.® Nearly all of the statements concerning the habits of these birds, like my own, refer to the nesting season. So far as I know to the contrary, almost nothing is known of their life outside of this period. Many of the reactions during the nesting season could be understood more easily if we knew the complete history of their life-cycle. GEOGRAPHICAL SITUATION AND HISTORY OF THE PRESENT COLONY OF TERNS. Bird Key is a small coral island about 300 yards wide (east and west) by 400 yards long (north and south). It is 65.8 statute miles due west from Key West. The island is partially sheltered on the east and on the northeast by a coral reef (fig. 1). Northeast of the island, about 1.125 statute miles distant, stands Fort Jefferson, now practically deserted. Still farther to the northeast other low coral islands are to be found. Loggerhead Key lies * Bird Lore, vol. v, 1903, p. 77 ff. * Birds of the Hawaiian Islands, etc., 1902. 2 Op. cit. *Birds of Laysan and the Leeward Islands, Hawaiian Group. Bull. U. S. Fish Commission, vol. xxX1t, 1903, pp. 767-807. °As examples of differences in observations I cite the following: One writer speaks of seeing these terns “swimming in the water.’ During my three months’ stay I never saw one of these birds in the water, except by accident (see fig. 26, plate 10), and then the bird, if the tide is against it, can never reach the shore, so poorly does it swim. Another statement is to the effect that these birds are often seen flying at night at great distances from the shore. My own observation is to the effect that the birds return to the island at night and leave it at daylight. 192 Papers from the Marine Biological Laboratory at Tortugas. about 4 statute miles to the west of Bird Key. Immediately outside of these islands is to be found the water of the Gulf of Mexico. The situation of the island shows that it is adequately protected from all but the severest southwest storms. The Tortugas as a whole are rarely subject to heavy storms during the nesting period of the birds. During the past season ’ ° s 05 83 55 8250 45, 82 40 35 QR 24 40 Q Sand Key e sé€ast Key Ft. * Middle Key > 4 Wg sk oa Bird Key go> as wo eS Loggerhead f }Jefferson 3 us| ae eee EE Rebecca Shoal Light 24 20 Fic, 1.—Showing the relation of Bird Key to the other islands in the group. (The dotted lines show a launch trip made to determine the feeding habits of the birds.) (1907) only one severe storm visited the island, and this was not very destructive to the life of the birds. Owing to its juxta-tropical location, its slight elevation, and the condi- tion of its surface (largely coral sand), the actual surface-temperature of this island is very high, ranging at times during the hottest days from 124° to 143° F. With the exception of the bay-cedar bushes, which are very abundant upon the central and western parts of this island, little vegetation exists. Ona certain limited portion of its surface (southeastern) a dense growth of cactus is to be found. Both cactus and bay-cedars are utilized by the noddies for nesting-places. No accurate data exist concerning the number of years these two species have migrated to this island for the purpose of rearing their young. The oldest inhabitants of the neighborhood say that as long as they can remember WATSON PLATE 1 . | < * = Yay M ~— ting attitudes of Sooties. iv IN ih The Behavior of Noddy and Sooty Terns. 193 the birds have been going there year after year. The terns arrive at approxi- mately the same time each year (during the last week in April), live there until toward the first of September, and then begin their southern migration. FOOD AND FEEDING HABITS OF THE TERNS. In a locality where marine forms are so abundant as in this favored Gulf region, the terns collect their food with little difficulty. They feed upon small fish of different kinds, which are present in great abundance.’ I have searched in the literature for statements concerning the methods utilized by these terns in catching their fish, but I was nowhere able to find any state- ments of value concerning this. I made a careful study of the water habits of these birds. To my great surprise, I found that the birds never swim nor dive. As a matter of fact, they never touch the water except when drink- ing or bathing. The bird drinks the sea-water as it skims the surface of the water with open beak. Bathing they perform in much the same way, never coming to a stop in the water nor completely immersing the body; usually the breast and head are the only parts dipped into the water. The birds fish by following schools of minnows which are being attacked by larger fish. The minnow, in its efforts to escape, jumps out of the water and skims the surface for a short distance. The terns pick off these minnows as they hop up above and over the surface of the water. The rapid- ity and accuracy of visual-motor adjustment in this reaction is wonderful. The birds feed singly or in groups, usually in groups. The group may be composed of both noddies and sooties and may contain sometimes as many as 50 to 100 individuals. All during the day groups of noddies and sooties may be seen at work. As the minnows cease to jump above the surface of the water, the group disbands and scatters in every direction. An instant later, as an attack is made upon the minnows in some other locality, the birds immediately rush there and renew their feeding. Whether there is a true following instinct at hand in this reaction can not be stated, but when one considers that such an instinct is probably present in the sunning reac- tion (which will be described presently) one feels justified in assuming that the act of feeding in groups is likewise a sign of gregariousness. In view of my experiments upon the function of distant orientation in these birds, it became very necessary to ascertain when the birds leave the island, how far they go, and when they return. Taking up these questions in order, I suggest the following probable answers: Both species of birds leave the island at early daybreak. In order to observe this more accurately, I rowed out from the island at 3 o'clock in the morning to a distance of 1.5 knots. No birds were seen on their way to feed until daylight began to appear. From that time they appeared in ever-increasing numbers—noddies * Examination of the stomach contents of both young noddies and sooties showed the presence of representatives of the two families of fish Casangide and Clupeide. 14 194 Papers from the Marine Biological Laboratory at Tortugas. and sooties leaving the island singly and in small groups. They apparently continue their flight until the jumping schools of small fish meet their eye. This may happen soon or late; consequently they may feed near the island or far away. Apparently, at the end of 2 hours the noddy has supplied its needs, for at this time it returns to the island and relieves its mate at the nest. The latter then comes out upon the water and takes, roughly, a two-hour turn at fishing, then likewise returns to the nest. This routine of spending 2 hours at the nest and 2 hours on the water is engaged in by all of the noddies dur- ing the seasons of brooding and of rearing the young (see p. 206). Before the egg is laid, however, the male does the fishing for both himself and the female at intervals which I could not determine. The male sooty, on the other hand, as will be shown later, during the laying and brooding season probably stays out upon the water all day long, returning at nightfall during the former season to feed the female, and during the latter season to take his turn at brooding the egg (see p. 209). I was especially anxious to determine the distance to which the terns go for their food. I have the following observations to report bearing upon this subject: So far as I could learn from questioning the residents at Key West, 65.8 statute miles to the east, these birds never venture as far as that for food. The lighthouse-keeper at Rebecca Shoal Light,’ 17 knots distant from Bird Key (east), tells me that he has never seen either of these species of birds fishing in that locality. He reports that on sunny days pelicans, frigate-birds, cormorants, boobies, etc., appear there in large num- bers. I am inclined to think that his statements with respect to the terns are correct. Both on approaching Bird Key from the east and upon leaving it to the west, I was not able to observe these birds feeding farther out that 4 to 10 knots. As a final test in the matter I made a trip, with the assistance of Mr. Kellner and Mr. Hooker, in the laboratory launch. The chart is shown by the dotted lines in figure 1. We left Bird Key at 8" 30™ in the morning and journeyed in a straight line almost to Rebecca Shoal Light. The birds were seen fishing in numbers until a distance of 9 or 10 knots had been traversed, after which fewer and fewer birds were sighted. From 13 knots on to 15 no birds were sighted. At this point we stopped and waited for an hour. No birds appeared. In returning to the island, we partially retraced our path and then turned toward the south and steered in a circle, keeping Bird Key within about 10 knots. On the return trip, 20 birds were counted before we began to circle the island. Just as the turn was made to the south, three large groups were seen feeding, and from that point on the birds were numerous. From my own observations I conclude that these birds rarely leave the *See fig. I, p. 192. The Behavior of Noddy and Sooty Terns. 195 island for distances greater than 15 knots. Further observations are sadly needed at this point, specially with reference as to whether the birds may not possibly journey further in other directions than in the easterly one. Also as to whether the birds have to go further during stormy weather in order to obtain their food. At all of the above distances, the lighthouse on Logger- head Key, 160 feet in height, could be “sighted” by the birds at the elevation at which they fly. From other observations, too numerous to mention separately, I con- clude that all birds return to the island at night. Many times just at sun- down I have come from Loggerhead Key to Bird Key. The terns are com- ing in by hundreds and thousands, flying low over the water. By the time twilight has faded the water is entirely deserted. Several trips made to Fort Jefferson late at night showed that these birds do not leave the island at night. The moment the island is reached, however, no matter at what hour of the night, one finds the sooties busily flying from one place to another on it. While I made these observations as carefully as I could, I realize that conclusive deductions can not be made until a more extended study has been undertaken. THE MATING OF THE NODDIES AND OF THE SOOTIES. According to Thompson,’ both species mate before reaching the island. Whether or not this statement is true can not be answered from my own observation. I did not begin my work upon the birds until May 4. Since they reached the island upon April 29, I could not observe their early be- havior. But as I arrived only 5 days later than the birds? and found both species actively engaged in nest-building and some beginning to lay,’ it seems quite clear that mating is either a very simple process, requiring little time, or else it had been accomplished before or during migration. A com- plete account of mating can scarcely be given until we know more of the life of these birds before their northern migration begins: Are they gregari- ous previous to it; if so, to what extent? Are the partnerships formed dur- ing the mating season kept after the young are reared? What are their habits in feeding and in roosting, etc.? Do their habits change at the ap- proach of the time for migration? What is the length of time for migra- tion? What are their habits during migration (7. e., do they stop on land at intermediate points), etc. ? My notes contain a rather full account of a striking series of reactions between two noddies, which I took to be a case of mating and choice of * Op. cit., p. 78. : *I am quite sure that I reached the island before many of the birds. Apparently the birds, the sooties at least, arrive in groups stretching over a period of about two weeks. *Two noddy eggs were found May 4, while the sooty eggs were first found on May-7. 196 Papers from the Marine Biological Laboratory at Tortugas. nest-site, but since it occurred late in the season and did not lead to a com- pleted nest I advance it tentatively: ‘ One day I observed several noddies “sunning” upon the wire covering of one of my large experimental cages. Suddenly, one of the birds (male) began nodding? and bowing to a bird standing near (female). The female gave immediate attention and began efforts to extract fish from the throat of the male. The male would first make efforts to disgorge, then put the tip of the beak almost to the ground and incline it to the angle most suitable to admit her beak. She would then thrust her beak into his (the ordinary feeding reaction). The feeding reaction was alternated with the nodding. After this series of acts had been repeated 20 times, the male flew off and brought a stick. He deposited this near the female and then again offered to feed her. She again tried to feed, then the male attempted sexual rela- tions. She immediately flew away, but almost immediately returned and alighted at a slightly different place. The male again brought the stick and again bowed and offered to feed her. She accepted the food, but again flew away when the male attempted to mount her. At this juncture the island was disturbed and my observations could not continue. If the above is a genuine case of mating, the process is very simple. It consists in the female’s accepting food from the male and engaging in sexual relations with him at a given nest locality. Such a process might well take place en masse during the first few hours after the birds alight on the island. 1 was not fortunate enough to obtain a corresponding set of obser- vations of even this unsatisfactory kind upon the mating of the sooties. THE CHOICE OF THE NEST-SITE, AND THE MATERIAL USED IN CONSTRUCTION OF THE NEST. THE NEST OF THE NODDY. The noddy constructs its nest from (1) loose dead branches of the bay- cedar bushes ; (2) of seaweed; (3) of a combination of these; (4) of a com- bination of either or both of these with various kinds of sea-shells and coral. When the shells and coral are employed, they are often placed as an inner lining to the nest and the egg is deposited directly upon them. The nest itself is a quite variable structure, and usually loosely put together. It is very shallow, and this is rather singular, since the wind often blows the egg or the young to the ground (see figs. 4, 5, 6, and 7, plate 2). The nests remaining from year to year are utilized by the birds at suc- cessive nesting periods; whether or not by the same pair can not with cer- *This nodding reaction is one of the most interesting and ludicrous acts of the noddy tern. It is quite elaborate. Two birds will face each other, one will then bow the head almost to the ground, raise it quickly almost to a vertical position, and then quickly lower it. He will repeat this over and over again with great rapidity. The other bird goes through a similar pantomime. If a stranger bird alights near a group, he salutes those nearest, and is in turn saluted by them. During the pantomime a sound is rarely made. The Behavior of Noddy and Sooty Terns. 197 tainty be answered at present.1. On account of this utilization of the old nest from year to year, some of the oldest nests have grown to enormous size, due to the addition of new materials at each successive season. The photograph of the group of nesting noddies (fig. 22, plate 8) shows the tallest of the bay-cedars (about 12 feet) and what I judge to be some of the oldest nests. Figure 5, plate 2, shows a newly constructed nest, which is much smaller. The statement has been made that the noddy sometimes lays its eggs directly upon the ground, but this is not quite true for noddies on Bird Key.? Very often the nest has the appearance of being constructed directly upon the ground, but a closer examination usually shows that it has been built upon a tuft of grass or upon the stem of a bush, the branches of which have been broken off close to the ground. Figure 22, plate 8, shows the characteristic groupings of the nests. It is typical of many localities on the island. Attention is called to the fact that the height of the nests above the surface is quite variable. The noddies apparently do not seek to nest in the thickest parts of the bushes. Although isolated nests are present even where the shrubs are most dense, by far the majority of them are to be found in bushes which border upon open spaces. When we consider the size and delicacy of their wings this fact has biological value in that nests in such situations are easy of access. Apparently there is no instinctive tendency to secrete the nest. The cactus growth contains about 20 per cent of the total number of nests. The nests there do not differ in construction from those found in the bay-cedar bushes. By means of a mechanical counting device it was found possible actually to count the total number of (active) noddy nests. The count gave 603 nests. In some places, where the bay-cedar bushes are very dense and the area has to be covered “‘ dog fashion’? (or at times even still more primi- tively), and in others where the cactus growth is very luxuriant, error in counting was easily possible. On account of these possibilities of error, I be- lieve that 700 nests is a more representative number. Since 2 birds occupy one nest, we have a total of 1,400 adult noddies on the island. THE NEST OF THE SOOTY. The nest of a sooty, when a nest is made, consists of a shallow oval de- pression in the sand. This depression varies greatly in depth, depending *A test designed to answer this question was made. Before leaving the island, I caught three birds on three separate nests. I marked the nests and placed a plati- num band around the legs of the birds. During the nesting season, just past (1908), Dr. Mayer wrote me that he had captured the birds occupying these marked nests and that none of the birds so captured was marked with the platinum band. It is recognized that the number of birds tested in this way is too small to afford a basis for generalization. “The nearest approach I found to the laying of the egg upon the bare ground was in the case of two nests built on a bare horizontal board lying among the cactus growth. In each of these cases the egg was laid directly upon the board, but some dozen or two small sticks retained the egg in position. 198 Papers from the Marine Biological Laboratory at Tortugas. upon the nature of the surface. It is rarely over 5 cm. in depth, even in loose sand. The northern and northeastern sections of the island are free from bushes, but are covered by a shallow growth of Bermuda grass. These areas con- tain by far the largest number of nests. The group photographs of the sooties appended (see especially figure 17, plate 6) are taken at one or the other of these sections. The eggs in these areas are laid literally on the grass and bare earth in no kind of nest structure. The eggs are often de- posited in open sandy places, but nest depressions are not always made, even where the nature of the surface easily permits it. A reference to figure 18, plate 6, will show quite clearly the absence of any complex nest-structure. A rather interesting variation in nest-structure appears among certain nests which are built under the bay-cedar bushes. The leaves from the bushes sometimes form a carpet over the sand. The nesting sooties often gather up these leaves and place them around the rim of the depression. Under no circumstances are the leaves collected from a distance further than the birds can reach with their beaks while remaining in a sitting posture in the nest. The nests of the sooties are assembled into groups. Roughly speaking, there is a southeastern, a central, a northern, and a northeastern group. An approximate count of the total number of the sooty nests was made in the following way: Those parts of the surface of the island containing nests were subdivided into ten separate areas. The number of square feet in each area was next determined. The average number of nests (spots where eggs were deposited) per square foot was then determined separately for each area. By means of these data, the total (approximate) number of nests on the island was found to be 9,429. Multiplying by two, as in the previous case, we have 18,858 as the total number of adult sooties. It may be said that the above determination was made late in the brooding season, after all the eggs had been laid. It may also be of interest to note that in localities where the nests are very numerous they often are not more than 10 to 12 inches apart. On account of this close grouping of the nests, and of the quarrelsome nature of the brooding birds, exact localization of nest and recognition of nest and mate easily became the most important features in the lives of the sooty terns. This situation affords a convenient starting- point for a psychological study of the behavior of these birds. My tentative beginnings in this field are described on pages 221 ff. REACTIONS OF THE NODDIES OBSERVED IN NEST-BUILDING. My notes, written during the observations, contain a large amount of material relating to the way in which the noddy and sooty nests are con- structed. The greatest difficulty in obtaining accurate notes lies in the fact that in neither of these two species are the differences in visual appearance The Behavior of Noddy and Sooty Terns. 199 between the male and the female marked enough to afford a basis for deter- mining sex. But with close observation, the differences in behavior are so marked, at least during the nest-building and egg-laying stages of the nest- ing period, that the following statements, which refer to the division of labor between the two sexes, are fairly accurate. Later on in the work, I found that observation was greatly aided by marking one of the birds with oil paint, and then from the behavior of the marked bird record whether it was the male or the female.' On May 11, the following notes were obtained from a pair of noddies at work upon a nest which had been started a few days previous: Both birds work, bringing sticks, sea-weed, shells, and coral. Both birds shape the nest clumsily by pecking and pulling at the sticks. They never weave the sticks so as to form a compact and durable nest. The stick is dropped on the rim, then drawn into position. Frequently, first one bird, then the other sits in the nest and shapes it. In order to do this the bird rises on its feet and depresses its breast and turns round and round. ‘The material is obtained both far and near. Floating sticks and seaweed are gathered from the water. They frequently alight under the nests of other birds and gather up the fallen branches. They even take the material from other nests which are left momentarily unguarded. Frequent fights ensue. The birds work neither steadily nor rapidly; 10, 15, 20 minutes may elapse before either makes a trip. On one of the trips the male grasped a large, dead branch which was fast. Another bird came up and also grasped the stick. A fight ensued in which the intruder was worsted. The male next picked up a large stick and attempted to walk out into an open space in order to rise. An obstruction barred the way. The bird, standing on the outside of the barrier, tugged and tugged at the stick, but unavailingly. Finally he stooped under the obstruction, grasped the stick and backed out. This observation is of in- terest in that it shows a rather wide range of instinctive adaptations. Both birds are busily engaged for the next hour. Then the male leaves, while the female remains sitting on a limb near the nest. Two hours later the male returns and feeds the female.” On another occasion in watching the transfer of material from one nest- site to another it was found that one bird did all the work; the other re- mained sitting on a nearby branch. Still another observation was made in which it appeared that the male did most of the work. The observation began just after the male had fed the female. The male flies away (4" 16™ p. m.), but returns with a fine straw at 4" 18™. Leaves and returns with straw at 4" 19™. Leaves at 4" 20™; returns at 4" 23™, bringing no material; sits on nest for a moment and preens feathers. Leaves at 4" 26™ and at 4” 27™ returns with another straw. Several straws are then brought at intervals of about one minute. On one trip female is sitting in nest. The male returning with straw forces her out, deposits the straw, and shapes the nest. He leaves and returns ?See p. 202 for details. 200 Papers from the Marine Biological Laboratory at Tortugas. white shell and deposits it in the nest. This she repeats two or three times. Both birds are active from now on until 5" 05™, at which time the male leaves for food. At 7" 05™ he returns and feeds the female. The remarkable thing in all of these acts was the accuracy of orientation. Many nests intervened between their own and the open spaces and many other nests were in process of construction in the same bush. The nest was localized with great exactness by each of the two birds at every trip. Finally, it may be said that the nest-building instinct is not so transitory as certain others. All during my stay on the island the noddies were carry- ing sticks; even those caring for young do not resist the impulse to gather up sticks. All my efforts to get control of their reactions by supplying them with food and water were unavailing, but I could easily induce activity in them by collecting a bundle of twigs and tossing them up in the air. Hardly would the sticks fall before the noddies were after them. A noddy in mid- air carrying a stick and dodging a dozen other birds in order to maintain possession of the prize is one of the most common sights on the island. So active and alert are they on the wing that if a stick is by chance dropped while the bird is in fight, it is often caught before it strikes the earth or the water. REACTIONS OF THE SOOTIES OBSERVED IN NEST-BUILDING. The building of the sooty nest is quickly accomplished. The obtaining of a nest-site is the difficult part of the reaction. As has been said, the sooties build their nests very near one another. For this reason it is ex- tremely difficult to make complete observations. My observations began late one afternoon, before any eggs had been laid. Hundreds of the birds were grouped together, incessantly fighting and screaming. It quickly became apparent that most of them had chosen a nest-site and were defending it against all late-comers. Both male and female were present. Each pair in this particular locality defended a circular territory, roughly 14 inches to 2 feet in diameter. Other birds in wandering around would stumble into this sacred territory and a fight would ensue. The fights would often lead to encroachments upon the territory of still other birds. The number of those fighting would thus be constantly increased. I have seen as many as 14 sooties thus engaging in a fight. Birds 1o and 15 feet away would rush into the fight and the noise and confusion beggared description. Some- times as many as 10 or 15 such fighting groups could be observed in the area Of 1,000 square feet. Quiet would momentarily ensue and then be broken by another series of fights. During the choice of the nesting-site the fights continue day and night, with only intermittent periods of quiet. Within this charmed circle the two mated birds remain relatively quiet. At this time sexual activity is at its height. It frequently happened in the sexual process that the two birds would step outside of their own territory WATSON PLATE 2 The Behavior of Noddy and Sooty Terns. 201 and a general fight would ensue. When the sexual reaction is in progress it fs a signal for the surrounding males to encroach. Coition is thus com- pleted only after much fighting. I have seen the male attempt to mount the female 12 to 15 times and at each attempt be interfered with by neigh- boring males. The actual construction of the nest, when a nest-structure is formed, begins after an undefended area has been found. The process of nest- building is somewhat as follows: The bird puts the breast to the ground, thereby supporting the body and leaving the legs comparatively free. The feet are used as a combined scraper and shovel. A few backward strokes of the feet are made, which serve both to loosen the sand and to remove it from beneath the body. The bird then turns slightly and repeats the process. When it has turned 360° (or less) it begins to use the breast as a shaper. By continuing this process, the depression is soon made to assume the required diameter and depth. My notes show that the bay-cedar leaves are often gathered up and placed around the rim of the nest as the hole is being dug. I can not say which sex does the work, but I believe that both male and female engage in it. As soon as the depression is made, both birds begin to defend it. Naturally, where no nest is made, the nest-site alone is chosen and defended as described above. If we compare the behavior of the noddy during this period with that of the sooty as described above, we find that the former is quietly building its nest and engaging systematically in a fixed routine of instinctive activi- ties. The contrast in the behavior of the two species is always marked, but never more so than at this period of the nesting season. THE DAILY RHYTHM OF ACTIVITIES. ACTIVITIES OF THE NODDY BEFORE THE EGG IS LAID. While observing the noddies at work upon the nest, it soon became apparent that the daily routine of the female was different from that of the male. From many hundreds of observations it was also evident that the male feeds the female at more or less regular intervals. Incidental mention has already been made of this difference in the activities of the male and female in the previous sections, but a more detailed statement is in order if we are thoroughly to understand the economic conditions obtaining on the island. After both birds have worked upon the nest for some time, the male leaves and is gone for varying lengths of time, depending upon the ease with which food is obtained. While he is away the female rarely leaves the nest. She sits usually upon a nearby limb and rarely shows signs of activity. It is not unusual to see four or five females thus sitting motion- less and stupid for hours at a stretch. Observation at such a time becomes exceedingly trying. Quiet is occasionally broken by the males from other 202. + Papers from the Marine Biological Laboratory at Tortugas. nests attempting to establish sexual relations with these temporarily un- guarded females. This is in all cases unsuccessful. Disturbance to some extent is also caused by other birds attempting to poach straws. If the day is hot, the female may make frequent trips to the water. The nest is thus left for a time unwatched. The bird quickly returns, however, and resumes watch over it. The male returns with a full-laden crop. He alights directly upon the nest or near the female. The female at once shows signs of life, and as they approach each other they begin nodding. Then the male invites the female to feed by putting his beak down to a position convenient to her. She gets the food by taking it directly from the mouth of the male, the male disgorging it by successive muscular contractions of the throat and abdo- men. The impression one gets from this ludicrous performance is that the bird is choking to death. During the whole of the process of feeding, a soft, nasal, rattling purr is emitted, presumably by the female. This purr- ing sound is an invariable indication that feeding is taking place. It is to be heard on no other occasion. At times the male upon his return is not so ready to feed the female. The female then strikes the bill of the male sharply with her own. I have seen the female thus strike the male 18 to 24 times before eliciting the proper response from him. On other occasions, the female is reluctant to feed the proper length of time, whereupon the male gently taps the female and puts his beak near her own again and again. The controlling stimuli throughout this reaction seem to be organic and visual. We might schema- tize the principal features of the reactions, as a whole, as follows: The male fishes until intra-organic pressure of food in the crop reaches a certain intensity. This acts as a stimulus to return (proximate and dis- tant orientation discussed on pages 224 and 227 respectively). The visual stimulus of mate (and nest and nest locality) coupled with the intra-organic stimuli just mentioned, condition the feeding reaction. On the part of the female we have the intra-organic (hunger) stimulus and the visual stimulus induced by the movements of the male. The male disgorges until there is a cessation of the excessive intra-organic pressure, at which time his feeding movements cease and the female may strike his beak in vain. The female in her turn feeds until there is both a cessation of hunger and a normal intra- organic pressure established. If this takes place before the male is ready, he in turn attempts to further stimulate the female by a slight change in behavior (7. e., “coaxing” by tapping the female and putting his beak down near her). The feeding reaction completed, the birds often sit near each other, nest- ling and nodding vigorously. This is the time usually chosen for coition, which takes place frequently up to the time the egg is laid. Feeding may occur at any time of the day, but the best time to observe it is at sunset, when the males are returning in numbers. The Behavior of Noddy and Sooty Terns. 203 At night the two birds usually remain in branches near the nest, but if disturbed, both fly away for a short distance and circle back almost imme- diately to the nest. In flying at night both the noddy and the sooty break their graceful flight into short, ungraceful, and ill-directed choppy swoops, very similar to the way the night-hawk breaks its flight when flying after dusk. During the egg-laying period, which is at the same time the nest-build- ing period, we may summarize the chief points in the lives of the noddies by saying (1) that there is common activity in the building of the nest; (2) that the female guards the nest while (3) the male procures food for both. Both birds are quite wild during this period. If the nest is approached they fly away and make no effort to defend it. For this reason it is difficult to capture and to mark the birds. They fly away at night at the first approach of a lantern or torch. Their behavior at this time, in this respect, is quite different from that observed later on during the brooding period. ACTIVITIES OF THE SOOTY BEFORE THE EGG IS LAID. I can say little concerning the separate daily activities of the male and female sooties during the corresponding period. My time was centralized around the nests of the noddies for the first two weeks of my stay, and consequently I lost my best opportunity to study this period of the life of the sooties. The birds are so numerous and the confusion so great at this time that detailed and sustained observations of habits are well-nigh impossible. That feeding of the female occurs | am sure, but I am not sure that the female never fishes for herself. I spent several continuous hours, at dif- ferent times, observing the nests of the sooties, and the only feeding reaction I saw took place late in the afternoon. The details of the feeding process are very similar to those described above for the noddies, except that when feeding between two sooties begins it is the signal for the approach of dozens of other birds, which precipitates many fights. The noise of the colony as a whole is so deafening at all times that it is impossible to say whether a special sound or series of sounds is made while the female is feeding. During the period before the egg is laid the sooty, like the noddy, will leave the nest if one approaches, and unless one is quite a distance away the bird will not approach the nest locality. It will circle in the air again and again, giving out the shrill nasal alarm cry of “ éah, éah, éah.”” It is the most restless and noisy bird I know, and almost as much so at night as dur- ing the day. Sleep apparently is taken during both day and night by dozing momentarily at intervals. How the bird maintains its vigor with no more continuous rest than it takes is a mystery. This peculiarity of the sooty has led to the popular nickname of ‘ wideawake tern.” ACTIVITIES OF THE NODDY AFTER THE EGG IS LAID. The noddy lays one egg. It may be laid almost as soon as enough straws are placed together to support the egg or it may not be laid until after sev- 204. Papers from the Marine Biological Laboratory at Tortugas. eral weeks of nest-building. It has been mentioned that the noddy never fully completes its nest; but after the egg is laid the gathering of additional nest-material is a sporadic activity. The first eggs were noticed on May 4. On May 6 I marked 16 nests and visited each twice daily. The first egg was found in these nests on the 11th; not until the 25th did each nest contain its egg. My chart shows that the majority of the eggs were laid from the 11th to the 16th. Fisher? has the following to say concerning the appearance of the egg: The rather acute ovate egg is a creamy white, sparsely spotted with light gray, burnt umber, and walnut brown. Most of the brown spots are on the larger half, and are sometimes small and at other times quite large (4 to 8 mm. across). One egg has no dark marks, but is scantily spotted and streaked with light Mars-brown. Specimens vary from 58 by 48 to 51 by 35 mm. After the egg is laid, a marked change appears in the behavior of both the male and the female. The birds will now attack even a human intruder, and their defense of the nest against their own kind becomes even more strict than before.? Oftentimes the birds will sit on the egg and allow themselves to be caught, striking viciously all the while with their long, keen, pointed beaks. Individuals vary greatly in this respect. On my daily rounds, as I approached the vicinity of a group of nests, several noddies would usually advance to meet me, striking viciously at my head. Their attacks would continue until I withdrew. Many times I have had my hat knocked off and the blood brought from my scalp by their vicious attacks. This change in the behavior of the noddies, which is so marked and which begins so abruptly, will be spoken of again further on in the paper. It may be said here that the stimulus to the change is to be sought for in the tactual and visual impulses aroused by the egg. Still another marked change occurs in the habits of the birds: The male no longer feeds the female. Each bird takes equal turns at brooding the egg. My attention was first called to this while I was watching the habits of the birds before the egg was laid. Several nests in the vicinity of the place of observation already contained eggs. At these nests I was never able to observe the feeding of the female by the male. At this period the two birds become practically automata. Their life is taken up in alternately brooding the egg and in feeding. ‘The birds spend little or no time together except at night. The one comes to the nest, the other flies away to feed. When the returning bird (which I designate R) arrives at the nest, the bird on the nest (which I shall designate O) may or may not imme- diately react to it. To take the most rapid case of immediate “relief” at the nest first, we find R returning and alighting on the rim of the nest. R nods to O. O nods in return. This continues for a moment, when O takes flight. FR immediately covers the egg. * Op. cit., p. 783. E *See Thompson, op. cit., p. 81. WATSON Fig, 8. Fig. 9. Fig. 10. Eight Days. Fig. 9. One Day. Fig. 10. Three D Fig. 11. Thirty Days. The Behavior of Noddy and Sooty Terns. 205 In other cases both O and R will stay on the nest together for as long as 20 minutes. All the time R is “crowding” O more and more; suddenly, as R receives contact stimulation from the egg, a more pronounced movement is made, thus forcing O aside. O is thus freed from contact stimulation. O gets up, nods to R, and flies away. FR (now OQ) turns the egg with the beak and settles gradually down into the nest. A few examples are here quoted from my notes: ‘ May 12, 1907. One “relief ’’ noted. O had been on egg for a long time. J returned and gradually pushed O aside. Nodding reaction very much abridged. O circled around nest, then flew away. May 13, 1907. O had been on nest 3.5 hours. At 12 o’clock R appeared. R alighted upon the rim of nest and gave one or two upward flirts with the beak. O was then gradually shoved aside. After O had been pushed from the egg, both birds occupied the nest together for 15 minutes. O then disappeared. In a nearby nest the R perched motionless for 25 minutes on the rim of the nest before attempting to push the O aside. The egg is generally covered day and night. Occasional trips are made by O to the water for drinking and for wetting the breast feathers. This latter reaction has its value possibly in keeping the egg at the proper tem- perature. The sun is so hot that if the egg were left uncovered for any great length of time it probably would not incubate. Occasionally, however, O will perch for 10 to 15 minutes on a nearby limb, leaving the egg exposed. After I had observed some 12 to 15 cases of the above interchange of activities at the nest, and found that the average time required for a shift to take place was 2 hours, it occurred to me that the birds might have some mechanism which might function as a time-sense (organic stimulation of some kind, probably). In order to carry out the work carefully, one bird (sex not determined) at each of 3 nests was marked with oil paint (aniline dyes do not resist the action of the salt water). The nests were visited every half hour, beginning at 6 a. m. and continuing until 8 p.m. The observa- tions were continued 2.5 days. In table 1, WM refers to the presence of the marked bird at the nest, U of the unmarked bird. The letter p shows the presence of both birds at the nest. From the table it appears at a glance that the period of occupancy of the nest is a variable one, ranging from 30 minutes to 5 hours. In some cases, however, the sequence is somewhat regular (see May 22, 1907, nest No. 3, from 8 a.m.to 8 p.m). It must be remembered that FR is the bird which controls the length of time O occupies the nest. O remains until R returns.1 It would be interesting to determine the stimulus which leads R to re- for several days at a time show that O will remain on the nest without going for food for 24 to 48 hours, but not longer. After this period elapses, O will go for food, leaving the nest and egg unguarded. 206 Papers from the Marine Biological Laboratory at Tortugas. TABLE 1.—Various “ shifts” made at the nests of the noddies, May 21, 22, and 23, 1907. ] Nest I. | Nest Il. | Nest III. Time. — | | May 21. | May 22. | May 23. | May 2x. | May 22. | May 23. | May 21. | May 22. | May 23 | = i Le giEeh |e AEST hemes 220 | eeeeeay | 6 0o™ a, m.| M M M M Wie) we et aie U M 6 30 | M M U M M M U Mp M 7 00 M M M M U Cee Up | U 7 30 M M M M U U Utveeihadhie yO 8 oo M Up M M U U Up ipa ae 8 30 M U M M Tint a M We i wv | 9 00 M | M | M i | AE A ea Ui en | 9 30 M U M M | M | M | M py sine Io 00 U U M Up Mp | M M Wise fl SU | Io 30 Pw) U U M | M M M | M ie 11 00 Up | U a U | M Se Ni M | M II 30 M M M |; M eee. || M | | 12 oom. | M M M U See alll M | I2 30p.m.| M M M U | M M I 00 \) SUD ey | Mp U | M U I 30 U U Mp Pn Tie hh 6 | 2 00 U U Up i || U U 2 30 M U U Mp U U | 3 00 M U U U | aM 3 30 U U M U M M 4 00 U Up | M M M M 4 30 U M | M M M M | 5 00 U M | | M M | M M 15 30 U M | | M M Ui ew | | 6 00 U Ww | | M M WS Wier | 6 30 U U Li a et fl) |), He 7 00 Ce ee M M eo yh | AR | 7 30 U U M M We | Us 8 00 U U M U M | M is reached. J often joins a group of other noddies sunning upon the beach or house-top (preferably the beach) and goes from this place to relieve O. We shall here describe the “ sunning reaction ” a little more in detail. In the brooding period, this “ sunning reaction’ (mentioned by Thomp- son, Palmer, et al.) is engaged in largely (1) by K’s before they return to the nest; (2) by O’s just leaving the nest, and (3) by birds which, I be- lieve, are not attached to any nest. Figures 27 and 28, plate 11, show the noddies collected near the beach. The house-top, the dock and an old wreck, the top of my experimental cages, etc., were all utilized by the birds for this purpose. Although the reaction is at bottom gregarious (similar to the feeding reaction) the birds are stolidly indifferent to one another’s presence. They sit silent, head to the wind, elaborately preening their feathers, pecking first at one toe, then at another. Occasionally when another noddy joins the group a mutual nodding is engaged in which at times for no observable reason ends ina fight. The birds here as elsewhere are silent. It is interest- ing to note that a definite distance is maintained between birds engaged in this activity. The distance is determined, I believe, by the long diameter of the body of the bird—they must have a free space in which to turn. I have seen 10 to 12 birds upon the comb of the roof of the house separated The Behavior of Noddy and Sooty Terns. 207 from one another by distances so regular that the unaided eye can with difficulty distinguish inequalities in the spacing. The above reaction was discussed in this connection because of its bear- ing upon the question of the presence of an “ organic time-sense’” in the noddy. The birds (R’s) often leave these chosen spots to go to the nest and relieve the mate. The probable conclusion to be drawn is that organic impulses furnish the stimulus leading to the return to the nest. If this be the case, we should expect the functioning of such a time-sense to be as inaccurate as our chart actually shows the case to be. Summing up this long section on the behavior of the brooding noddies, ‘we find: (1) that the presence of the egg brings about a change in the dis- tribution of labor between the sexes; (2) the male no longer feeds the female, but each sex separately obtains its food; (3) the egg is brooded constantly day and night by both sexes, the male and female relieving each other at intervals varying from 30 minutes to 5 hours, the average interval being in the neighborhood of 2 hours;! (4) the most significant general reaction caused by the presence of the egg is the change in the disposition of the birds. Before the egg appears, the birds are shy and leave the nest at the slightest disturbance; after the egg is laid, the birds will defend the nest against even human invaders. It is a little hard for the student absorbed in mammalian behavior alone to understand the fixed character of all these responses—the relative lack of any large store of latent adaptability and plasticity. These reactions preserve the birds and have preserved them for ages in this favored environment. There is no need for a larger repertoire of reactions. There is good material here for the study of the acquisition of habits and the permanency of such habits in an organism which is by nature already so largely adapted to its environment. ACTIVITIES BY THE SOOTY AFTER THE EGG IS LAID. The sooty, like the noddy, as a rule lays one egg, although I counted some 25 cases where a nest contained 2 eggs and, in one observed case, 2 birds were actually hatched and reared. The first egg was laid on May 7. By May 15 thousands of eggs were present. Fisher? has the following to say concerning the markings of the sooty egg: The ground-color is white or occasionally a cream buff. One type of marking consists of deep burnt sienna and grayish vinaceous spots, with occasional nearly black scrawls scattered rather evenly over the whole surface. These spots are 1, 2, and 3 mm. in diameter, with occasional larger and smaller ones. Another less prev- alent variation consists of heavy, very deep burnt sienna blotches (5 to 15 mm. in *It would be interesting to determine whether the birds relieve each other at night, or brood the egg on alternate nights, or finally, whether one sex always re- tains possession of the nest at night. “Ope cit. ps7 208 Papers from the Marine Biological Laboratory at Tortugas. extent) congregated in a zone near the blunt end, and lesser pale grayish vinaceous and deep burnt sienna spots sparsely scattered over the rest of the egg. The general disposition of the sooty, like that of the noddy, changes after the egg is laid and in the same way. Some of them become far bolder than the noddies in a corresponding situation. It was possible for me to lie down within a few inches of a brooding sooty and have it remain on the nest indefinitely. If the hand is extended toward the sooty it will attack vigorously, but I have never had a group of flying sooties attack me as I approached the vicinity of their nests, as was sometimes the case when I ven- tured too near the nests of the noddies. The birds are very variable in this respect. When one approaches a neighborhood containing many nests, the majority of the birds will fly up into the air, circling round and round, screaming all the while. If one remains quiet, the birds will gradually re- turn and cover the eggs. Gradually the nests nearest one’s position will be cautiously approached and then occupied. A certain small percentage of the birds will remain on the nest, no matter how violent the disturbance. My study of the instinctive reactions of the brooding sooties was again beset with difficulties because of the large number of birds present. My method of studying their behavior was similar to the one just described for the noddies. I would choose a favorable spot where several nests containing marked birds could easily be observed at once, then record what went on at each separate nest. I speedily found that the nesting reactions of the brood- ing sooties are quite different from those of the noddy. I would sometimes spend 4 to 6 hours at the nests without seeing a single bird leave its nest except to make short excursions for the purpose of fighting or to obtain water. The following day would sometimes find the same bird on the nest, sometimes its mate. In order more accurately to obtain data upon the ques- tion of the division of labor between the two sexes, I made a table (table 2) similar to the one presented for the noddies. At first the observations were taken every half hour, but as soon as it became apparent that so many ob- servations were needless the period was lengthened. It appears conclusively from table 2 that the shift at the nest is roughly a diurnal one, but that at times it may not occur except once in 48 hours. All of the factors in this reaction do not appear in the table. Apparently most of the shifts are made at night. I attempted on many occasions to determine the hour of shifting by leaving a lantern near the nest and mak- ing observations during the night, but the light could not be arranged so as not to frighten the birds, and their reactions consequently were not natural. The birds would refuse to cover their eggs if the light were made intense enough to be of value to me. Then again, in certain cases, the shift is not made within the 24-hour period. Where is the absent bird? All my efforts to discover its presence near the nest failed. Does it go such a great distance for food that it must WATSON PLATE 4 PIEK, ae - Bd P et ia ‘ » ra Fig. 14. Young Noddies. Fig. 12. One Day. Fig. 13. Three Days. Fig. 14. Three Days (born white). ee an " My if 7 7 Hib i wie Ms e's ae i as v Pina vii Fi Ay tS akan The Behavior of Noddy and Sooty Terns. TasLe 2.—Various “shifts” made at the nests of the sooties. | Date. Time. Nest IV. Nest V. Nest VI, May 27, I907— 7 00 a.m. Nest I. Nest II. Nest III. May 21, 1907— 6" 00" a. m, U M M U U 6 30 U M M U U 7 00 U M M U U 7 30 U M M U U 8 00 U | M M U U 8 30 U M M U ) 9 00 U M M U 17 a 9 30 U M M U U 10 00 U M M U U 10 30 U M M U U | II 00 U M M U Hyp ed II 30 U M M U Hy AE | 12 oom, U M M U | ew I2 30 p.m. U M M U a) I 00 U M M U U I 30 U M M U U 2 00 U M M U U 2 30 U M M U U 3 00 U M M U U 3 30 U M M M ) iw 4 00 U M M M 1 Wy) 4 30 U M M Mie cu 5 00 U M M Mp U 5 30 U M M M U 6 00 U M M M U 6 30 U | M Mp M U 7 00 U | Mp Mp M U 7 30 U | Mp Mp M U May 22, 1907— 6 oo a.m, Nw) U M M 7 00 M U U M M 8 00 M U U M M 9 00 M U U M M IO 00 M U U M M II 00 M U U M M I2 oom, ih a) U M M I oo p.m, M U U M M 2 00 M U U M M 3 00 M U U M M 4 00 M U U M M 5 00 M | U M M 6 00 eeu U M M 7 00 M U U M M 8 00 M U U M M May 23, 1907— 6 oo a.m U M M M U 7 00 U M M M U 8 00 U M M M U g 00 Ui ee M M U IO 00 U | M M M it II 00 U M M M U 3 00 p.m, ie || ane M M U 5 30 Us | M M M U i fps U M U M U May 24, 1907— 7 304.m. || AG) U U M 9 00 Ni WU} U U M IO 00 Nig tus U U M 3 00 p.m, WE HU) U U M 5 00 M U U U M 7 00 M | U U U M May 25, 1907— 8 ooa.m. U | M U U me a0 I2 oom, U M U U U | 8 oo p.m, 00 a ith U U U May 26, 1907— 8 ooa.m. U M U M M 7 oo p.m, U M M U M My) U U U U SSSSSCGCCCCCCEGC! iiiiiiiipiii:ipi::iiiiiiii: M 209 210° Papers from the Marine Biological Laboratory at Tortugas. remain away over night? The nearest land (and the birds never swim nor rest on the water) is 45 miles away (Marquesas, a very small key), and there is no evidence, so far as I know, that these birds visit there (none remains over night on the small islands belonging to the Tortugas group). On page 194 it is shown that in all probability these birds do not leave the island for distances greater than 15 knots in their search for food, and that they return at nightfall (or shortly thereafter) and leave at daybreak. My explanation of this failure to “ shift the watch ” is found in the peculiar restless nature of the birds. In actually observing their behavior during the shifts which occurred toward nightfall, I found that R would come and push O aside. O would leave for a few moments, then would return and scrouge back upon the nest. This would sometimes be repeated three or four times, one bird leaving the other temporarily in possession of the nest. If this procedure were gone through with in the early morning hours, it is easy to see how the wrong bird might easily join a group of other sooties leaving the island for food. O would thus be left in possession of the nest for two days in succession. The settlement of this point is one ex- tremely to be desired. If later controlled experimentation is ever to be undertaken, it is necessary for us thoroughly to understand their feeding and nesting habits.* Occasionally my records show the presence of both birds at the nest for several hours at a time. The general behavior of the brooding birds can best be understood by citations from my notes, which were made at the time and place of observation. By the following citations I hope to give a general impression of the behavior of the sooty colony at a time when its life is most complex, about the middle of May. Some of the birds have laid and are brooding their eggs, while others have not yet laid, but are present defending the nest-sites. In order to facilitate description, I shall use the letter S to designate the bird brooding the egg (the “sitter’’) and G to designate the mate of this bird, which is usually standing guard (when present). ‘The term “layers ”’ is used to designate those birds which have chosen a nest-site but have not yet laid the (one) egg. Reference to the habits of the latter should more properly have been made on page 203, but since the behavior of the “layers” influences the behavior of the brooding birds to some ex- tent they enter into the observation. May 13, 1907. Position taken near large group of “layers” and “ sit- ters.” The S’s droop their wings and waddle around the egg, finally sitting down upon it in a stiff and clumsy fashion. They are very much more rest- less than the noddies. They are constantly engaging in fights, leaving the eggs in order to do so. They turn round and round frequently, orienting the head differently at each adjustment. In some cases the G is present. In one *It is suggested that the nesting behavior at night could be studied by means of flashlight photographs if the birds and nests were prominently enough marked. Fig. 21, plate 7, shows poorly the possibilities in this method. The Behavior of Noddy and Sooty Terns. 211 case it stands about 2 feet from the nest, fighting all birds which approach. Once when this G drove a strange bird too near S, S left the egg to assist in the fight. Often in sitting down on the egg S rocks it gently, turns it with the beak, and gradually brings her weight upon it. Birds on alighting, if orientation has not been exact, have to run the gauntlet of hundreds of sharp beaks before they finally reach their own nests. Since this is constantly occurring, it greatly adds to the general confusion. Where the nests are under the bay-cedar bushes the birds on re- turning usually alight near the edge of open spaces and then run on foot to their nests, while if the nests are in the open spaces they circle around until the nest is located, and then alight. During this observation I saw three cases of feeding at nest-site where no egg had been laid. Feeding was accomplished with great difficulty. Dozens of other birds attempted to interfere, especially the G’s of nearby nests.” One case of shift at the nest was observed. S had been on the nest dur- ing the whole previous time of observation. Suddenly G went to the nest and gradually pushed S aside. G sat on egg for a few moments, then got up. S sat down on egg immediately. G then returned, pushed S aside, adjusted carefully, and remained on the egg. During this observation and in many others my attention was called to a peculiar reaction in the male. When approaching his mate and at times other females, he would arch his neck, droop his wings, lower his head slightly, turn the head to one side, and strut around and around the female, at times raising his head and then lowering it. The reaction is very similar to the one exhibited by the cock of our common barnyard fowls when he has called the females to feed. Figure 19, plate 7, shows 2 males in this attitude. My notes show that in some cases during the brooding period the birds sit quietly on the egg the whole day long, even in the midst of the turmoil which is ever present. In summarizing this unsatisfactory section we may say : (1) The presence of the egg brings about a change in the disposition of the sooty which is very similar to the one already noted in the noddy. (2) One important difference between the two species of birds is to be noticed—whereas the shift at the nest in the case of the noddy occurs once in 2 hours, the shift in the case of the sooty occurs once in 24 hours. This leads us to conclude that there must be an enormous difference in the way in which organic data function in controlling the reactions of the two species. The general impression one gets from close observation of the sooty is that auditory and visual stimuli play a more important r6le in its life than in the life of the noddy. 7 see fig. 2, plate 1. “This same reaction was often noticed when the parents attempted to feed the young. 212 Papers from the Marine Biological Laboratory at Tortugas. ACTIVITIES BY THE NODDY AFTER THE EGG IS HATCHED. The period of incubation varies for the noddy from 32 to 35 days. This fact was determined on the basis of 16 observations. The young began to appear on the island about May 9 (1907). The development of the young of both species is discussed separately further on in this paper. Under the present heading I shall mention only the change that the presence of the young brings in the behavior of the parents. At this period there is an increased tendency to defend the nest. They will now attack with vigor other noddies which approach too near the nest, the sooties and the frigate birds (Fregata aquila). From the writings of others I had drawn the con- clusion (1) that the frigate-bird attacks the terns and forces them to disgorge, and (2) that it feeds upon their young. I spent many weary hours in attempting to discover the relation of the frigate-bird to the terns, especially its relation to the noddies. Since the noddies build their nests in the bushes where the frigate-birds roost, it was presumed that there if anywhere the devouring tendency of the frigate-bird ought to appear.’ I found that the cause of the disturbance between noddy and frigate-bird lies chiefly in the fact that the latter, in attempting to find a bush in which to rest, sun, or roost, will oftentimes alight upon or very near to a noddy nest, whereupon the noddy most immediately concerned and those nearby will attack the frigate-bird and at times even rout him. It is a common occur- *When I first went to the island, about 50 frigate-birds were present. The number gradually increased during my stay, until at the end 400 to 500 were present. I caught and forced a number of these birds to disgorge (which they readily do), for the purpose of finding whether or not the remains of young terns or of very small minnows, the food of the terns, could be discovered in the stomach contents. I found that the staple article of diet with them is the flying-fish, which they can easily catch. At times I found fairly large herring and mullet. I also killed 2 of these birds, but found no traces of young terns. In one case the stomach cavity contained no food. This latter observation was interesting in view of the fact that for days during unpleasant weather these birds apparently do not leave the island. This fact made me think at first that possibly the young terns were preyed upon by the frigate-birds in bad weather, but I am convinced that the latter can go with- out food for long periods of time. On the other hand, it must be confessed that after the young terns appear the number of frigate-birds increases and that all during the day, when the sooties nesting in the open places are feeding their young, the frigate-birds line the bushes bordering upon the nests. At times a frigate-bird will swoop down and fly near the ground. This causes all the sooties to fly up and raise a terrible commotion. This soon subsides, only to be repeated an instant later when another frigate-bird flies to or near the ground. I built an observatory on top of the house and a tent in the midst of the nesting-places, and kept the whole island under close inspection with a field-glass (in this connection I wish to thank the firm of Bausch and Lomb for the use during my entire stay of an excel- lent stereo-binocular field-glass) in order to determine whether the frigate-bird when it made these movements actually picked up the young or whether it was attempting to pick up food which the parent bird had dropped while feeding the young. My efforts were unsuccessful, so far as showing that the frigate-bird really does vio- lence to the young tern or filches its food. The sooty, in contrast to the noddy, never attacks the frigate-bird. All that I have read about the ability of these birds to control their movements during flight is true. I have observed them many times in mid-air forcing a victim of their own species to disgorge a fish, which one of the pursuers would catch long before it struck the water. WATSON PLATE 5 15 E D Fig. . Thirty Days. The Behavior of Noddy and Sooty Terns. 213 rence, especially late in the afternoon, when the frigate-birds are returning, to see hundreds of such fights. The noddy is always careful to attack the frigate-birds by sudden thrusts (usually made from below), dodging quickly to avoid their fearful and powerful beaks. The curlew is also an occasional visitant on the island. The appearance of this uncouth bird brings forth fierce attacks from the noddies at this period of their stay. Figure 25, plate 10, shows an attack upon one of these birds. The parent birds alternately feed the young. The length of the intervals between feedings varies from 2 to 4 hours. The appearance of the young does not alter the feeding relations of the parents, which were found to hold in the table shown above. The interesting additional factor is that now when the birds go out to feed they must bring back a supply for the young bird. The question immediately arises: Does the parent actually catch more fish after the young is hatched, or does it catch the normal supply and share that with the young? If it catches an additional supply, what is the stimulus which leads to a more abundant filling of the crop? If an additional supply of fish is not caught, we should expect (1) either that the birds would feed oftener or (2) that they would become very much emaciated while they are caring for the young. Since my observations show that the birds do be- come emaciated during the rearing of the young, and that they apparently do not go out to feed any oftener, I am inclined to think that no additional supply of fish is (instinctively) provided for the young. The parents simply disgorge when the appropriate stimulus is at hand, 1. e., the sight and contact of the young. Since this process would tend to make the parent birds very hungry (or rather enormously to decrease intra-organic pressure) they prob- ably actually catch more fish on any one trip than they would otherwise.’ But this larger catch may be due to an emptier stomach and not to any instinctive tendency to provide food for the young. The apparatus for disgorgement is easily thrown into activity in these birds—very much more so than is the case with the sooties. My appearance at a group of nests in which young were present would call forth many such responses. As has already been mentioned, the noddy stays close by to defend its young. The bird usually sits on the nest and scolds in its peculiar, rasping way. Apparently this vocal effort can be better accomplished after the crop has been relieved of its contents; the bird can also probably fight better. Outside of these two possible reasons, I can offer no additional suggestion as to the biological value of this function when it is exercised in this connection. At this point, as well as at all others, one finds extreme monotony, fixed- ness, and lack of variability in the responses. nary conditions be attributed to them. Increase or decrease in the intra-organic pressure could function in the same way. 214 Papers from the Marine Biological Laboratory at Tortugas. The young are cared for in the nest until they become strong enough to leave it and live upon the ground. The young birds born in low nests, even at a very early age (20 days and even earlier), clamber from them with alacrity and hide in nearby bushes when danger is imminent. In many cases these young birds can not get back into the nest. Under these circum- stances they remain near the nest locality, and the parents on returning first alight on or near the nest and later hop to the ground and feed the young bird. It is interesting to speculate upon the method of recognition between parent and young. There can be no doubt at least of an accurate functional recognition. Since the noddy is always silent when contented, the evidence is good that recognition occurs wholly in terms of vision. Whether recogni- tion of young (or of mate by mate) would take place outside of the nest locality is a problem which ought to be solved. On account of my early departure from the island (July 18) I can say nothing of the methods by which the parents induce the young to fly, to leave the island and to feed. This period is probably one of great inter- est. At the time of my departure many of the young noddies were on the ground and were attempting to fly. If I may be allowed to advance state- ments which lie beyond my actual observation, I should say that the parents are not active at all in “training ” the young to fly and to feed. As the young birds advance in age the parents more and more often engage in the “ sunning reaction.” My opinion is that the young birds first instinctively collect upon the beach, as soon as their wings will support them. In a short time the instinct to follow the adults leaving from this point to go to food will lead them out over the water. The sight of the jumping fish is a stimulus leading to the movements used in catching the fish. This activity engaged in a few times so perfects the hereditary mechanism that the bird soon becomes independent of the parent in an environment where food is abundant. In summarizing this section we may say: (1) that the parents alter- nately feed the young at intervals varying from 1 to 4 hours; (2) that their general conduct is not greatly changed at the time of the appearance of the young, the changes actually observable being an increased tendency to pro- tect the nest and to disgorge on being disturbed. My own observations naturally contain nothing concerning the behavior of these birds when preparing for southern migration. Thompson? writes as follows (concerning both noddies and sooties) : Towards the end of September the birds begin to leave. They leave in great flocks and at night. The entire exodus consumes, apparently, but two or three days; and some morning the observer will find the island absolutely deserted, save for a few crippled birds that have been injured and are unable to follow their comrades. *As the young advance in age (20 days and at all later ages) the parent will readily leave the nest when disturbed. The tendency in this respect is to revert to the behavior exhibited during the egg-laying season. *@p: citps o2: The Behavior of Noddy and Sooty Terns. 215 From the reports of most of the residents, I found that the birds generally leave from toward the middle of August to the first of September. That they leave at night is most improbable, unless “at night’ means early morn- ing, at or shortly after daybreak. ACTIVIRIES BYa DEE, SOODY APGER THE EGG IS WATCHED: The period of incubation of the sooty egg is 26 days. This observation is based upon 16 marked nests. The development of the young sooty is deferred to a later place in this paper. The appearance of the young produces a profound change in the in- stinctive reactions of the sooty. A general change in the disposition of the bird is also noticeable. During the first three days after the appearance of the young, the sooty is reluctant to leave the young and nest on disturb- ance. Later, the adults fly away at the slightest disturbance, much as they do during the “laying” season. It is interesting to observe at every disturbance of a nesting-place how quickly the ground will be deserted by both young and old, after the young have reached the age of 3 days. As they leave, the alarm cry is sounded and the commotion spreads to all the nearby nests. When quiet is restored the birds again alight near the nest and gradually approach it. The young birds meantime have run to the bushes, where they remain motionless after sticking their heads into the crotch of some bush or depressing the body against any convenient solid object. The protective coloring of the young sooties is marked. When motionless, as above suggested, they are difficult to find. When the adult returns to the nest, the young birds gradually come from their hiding-places at the peculiar, clucking call of the parent. The parents (after the first few days) recognize their own offspring with ease and accuracy, often going to meet them as they emerge from the bushes. If by chance the wrong young bird is met, it is struck with great force. Naturally this is productive of fights between the adults. If disturbance occurs before the young birds are 3 days old, they will “sham death” in the nest or advance a few feet from the nest and sham death on the open sand. Lying outstretched upon the ventral surface of the body, with head flat upon the ground, it is with difficulty that one believes that life is present. The slightest contact stimulation will cause the young bird to attempt to get on its feet and struggle away. From I to 6 days is a critical period in the life of the young sooty. Hun- dreds of them are killed by the adults, and were it not for the “death-sham- ming instinct’ thousands would perish. This tendency on the part of the adult birds, especially of gulls, to kill the young has often been commented upon by nature students, but the phenomenon has not been very well un- derstood. The situation as it exists at the Tortugas colony is, I believe, not difficult of understanding. In the first place, we must consider the 216 Papers from the Marine Biological Laboratory at Tortugas. appearance of the young as being strange to the adult sooty. The advent of young is sudden; their movements, marking, coloration, etc., are striking visual stimuli to the adult—especially to those which have no young of their own and which are still brooding the egg. As previously stated, after dis- turbance the adult returns to the nest first, then the young begin to struggle back from the bushes; and those shamming death in the open become active and crawl back into the nest (these latter are especially liable to be de- stroyed). In doing so, the young must often pass by areas defended by brooding adults. Each adult begins to attack the small bird for exactly the same reasons that it would attack a persistent sand-crab, or a noddy or another sooty. The young bird by its peculiar cries and movements con- tinues to offer stimulation to the adults, and not until the former reaches its own nest or takes refuge in a motionless attitude and in a partially pro- tected place do the attacks of the adults cease. As the young birds gain in agility (and in experience) they rapidly learn to avoid defended areas and to dart quickly by and under attacking adults. On many occasions I have seen from 4 to 8 adults thus attacking a young tern, chasing it for 10 to 15 feet. At first the parent does not individually recognize its young, but reacts to it by reason of its presence at the nest.1_ Very early, however, there is mutual recognition between parents and young. The parent will advance toward its own young, even when several feet from the nest, and feed it, often forcing it back to the nest by a cumbersome rolling process. The young approaches the parent and nestles under its body. The parent ad- vances toward the nest, often upsetting the young bird, but at the same time advancing it. Both visual and auditory data are used in the mutual recog- nition of parent and young. ‘That auditory stimuli are functional can not be doubted. Often I have observed a returning bird give a call a long way off ; both the adult and the young answer the call and show changes in their reactions. ‘The parent at the nest will get up from the young and the young will stand up, flap its wings, and leap several times a few inches in the air. A moment later the returning bird will alight and feed the young. In order to observe the influences of auditory data on nest localization, I watched the areas containing nests many times at night. One evening I ‘This was determined by exchanging young birds. The exchange is not noticed when the birds are very young, but is noticed after the young is a few days old. My notes are very inaccurate at this point. In one case I took the young at birth from one nest (marked) to rear by hand. Two days later I put a new-born sooty in this nest. The reactions of the adults were very curious. The male first came up to the stranger young and gently struck it; the female came up and the young bird attempted to nestle under her and to feed. Other birds from neighboring nests came up, but were driven off. After this the young was treated in the usual way. There is room for interesting and careful work upon this subject. (Some of the noddy young are born with black juvenile plumage, some with plumage that is almost pure white (see fig. 14, plate 4). Now, if a white bird is exchanged for a black one, or vice versa, no disturbance is noticed on the part of the parent, even if there is a difference of 3 or 4 days in the ages of the respective young birds.) WATSON 17. Fig. sroups of r The Behavior of Noddy and Sooty Terns. 2147 lay quiet and hidden for 3 hours near a large group of nests which was situated under some very dense bushes. The adult would circle over the area and give a call; it would be answered and random movements would give place to direct. The bird would steer immediately for the source of the call. By peculiar chuckling sounds, which are emitted at this period when mates return, one can be sure that the proper nest has been located. I ob- served this many times during one evening. After the young were 20 to 30 days old I have heard the young birds answer the call of the parent back and forth a dozen times before the latter actually alighted. The parents alternately feed the young, but instead of a diurnal period of feeding, such as the parents have before the appearance of the young, the intervals vary anywhere from 4 to 7 hours. My observations are few on this point. Though the parents feed the young at any hour of the day, feeding can be most easily observed at dusk. It has already been mentioned that the sooties hurry home at nightfall in great numbers. From 4 until 8 p. m. this feeding process keeps the island in commotion. The feeding of the young birds has many interested spectators. While I have never seen the terns from the neighboring nests, which may be observing the process, attempt to rob the young bird, I judge from the actions of the feeding parent that such is occasionally the case. If the parent happens to disgorge more than the young tern can take into its beak and the food is allowed to fall to the ground, it is ludicrous to watch the rapidity with which the parent picks up the food and reswallows it. Oftentimes the mate of the feeding parent is near; its role is a purely passive one, except when the “ spectators” attempt to approach too near. Its part is then to assist in warding them off. Neither young nor old is quiet during this period of the nesting season. On the contrary, the noise is practically doubled. In addition to the ordi- nary sounds made by the adults and the new cries which are added at this time, there is present the high-pitched, insistent “ peep-peep”’ of the young terns. Momentarily the sounds of the adults will cease and the cries of the little ones remind one very strongly of a poultry-yard on a tremendous scale. A reaction very similar to the “sunning reaction ” of the noddies, while present to some extent before the appearance of the young, now shows itself in completed form. As may be judged from the feeding habits of brood- ing birds, practically only a half of the total number of birds is present on the island during the day, and that half is busied in brooding the eggs. Conse- quently there is little leisure at this time in the sooty colony. After the appearance of the young, the number of birds present on the island at any one time is much larger. When FR returns to feed the young, S usually leaves the nest, but as in the case of the noddies does not always leave immediately to feed. They collect upon the beach and sun themselves, preen- ing their feathers and standing idly about in a way which is quite similar to ’ ” 218 Papers from the Marine Biological Laboratory at Tortugas. the noddy. Figures 23 and 24, plate 9, show the birds engaged in sunning. The peculiarity in the reaction is that the birds always choose this one spot (the one shown in the cut) in which to assemble. Occasionally the noddies are to be found in the same spot. Figure 23, plate 9, shows the noddies leaving the beach in advance of the sooties. They “sun” nearer to shore than the sooties. Until we know more of the history of the life of the sooty, it will be difficult to understand the meaning of this reaction. There is one rather interesting difference between the habits of the noddy and those of the sooty which may be mentioned here: Every stake, buoy, or possible resting-place upon the water is utilized by the noddy. It will sit almost motionless upon any object projecting from the water for long periods of time. This habit of theirs is like that found in the cormorants, boobies, and pelicans which are present in the neighborhood. I have never observed a similar reaction in the sooty. I think the sooty always leaves the island and returns to it without at any time having ceased its flight. This seems rather remarkable when we take into account the fact that the sooty leaves the island in the early morning and oftentimes does not return until toward nightfall. The sooties often soar round and round, getting higher and higher until lost to sight. They usually join the frigate-birds in this reaction. I am in- clined to think that the sooty when sufficiently fed spends a large part of its time in such maneuvers. I have never observed the noddies engage in this reaction. In a note in the previous section, p. 212, we have already discussed the possible disturbance which the frigate-birds produce in the life of the sooties. When I left the island, the oldest of the young sooties were about 30 days of age. They were still in their juvenile plumage. From 20 days on, these young birds could be found quite a distance from the nest, but the nest locality still exerted its influence to a marked degree, both upon the parents and upon the young, as shown by the fact that when one of the parents re- turned from its feeding expedition it always alighted at or near the nest. At the sight or the call from the parent, the young bird would hasten toward the nest to receive food. This ability on the part of the birds to approach within a few feet of a spot which to our eyes had no distinguishing marks, but which had served as a nesting-place, is little short of wonderful. The care of the young, especially from 20 days on, must be an exhausting process for the parents. They become emaciated and somewhat bedraggled in appearance. This is not to be wondered at when we consider that a healthy young sooty can eat anywhere from 20 to 40 minnows of no insig- nificant size in a day. It may be of general interest to note that after the first few days the parent always recognizes and feeds its own young and no other, and furthermore, the young tern recognizes its own parents and attempts to feed only from them. Never but once out of many thou- The Behavior of Noddy and Sooty Terns. 219 sands of observations did I see a young tern begging food from a stranger. The statement has often been made by certain observers that young gulls feed indiscriminately from any adult which happens to be near. Such is certainly not the case among the sooty terns. In summarizing, we may say: (1) The presence of the young changes the general disposition of the parents; for the first two or three days after the appearance of the young the parents are more ferocious, but as the young bird gathers strength and can get away from danger, the parents become more wary and leave the nest upon the slightest disturbance. In this latter respect, as is similarly the case with the noddies, they, as it were, tend to revert to the habits formerly ex- hibited during the laying period. (2) A tremendous change appears in their feeding habits. During the brooding period the birds apparently are away all day; at the appearance of the young conditions are so changed that the birds relieve each other at the nest at intervals varying from 4 to 7 hours. (3) The birds have more leisure; they utilize this leisure in collecting upon the beach for sunning. (4) The birds become exhausted in caring for the young. PRELIMINARY ATTEMPTS AT CONTROLLING THE REACTIONS OF ADULT BIRDS. The experimental part of my work centered mainly around the young terns, because I could rear them by hand and could control their reactions through hunger. An experimental study of the adult birds is beset with many difficulties. The previous part of my work shows that until the egg is laid the terns can not be approached closely for experimental purposes. Only a short month was open to me after the egg was laid before the young appeared. After the young appeared, practically all of my time was devoted to them. From the middle of May until the middle of June, however, I had time to try the experiments upon the adult birds which are reported below at some length. I found that a month was too short a time in which to capture and tame adult birds for experimental purposes. In captivity, the birds are wild and restless, but unless they are in captivity they refuse to eat even live fish, even when the fish are placed near the nest. As was shown earlier, in the discussion of the feeding habits of these birds, the natural stimulus to the feeding reaction is the sight of schools of minnows jumping over the surface of the water. I may say in passing, however, that two sooties with broken wings learned to appear at my experimental cages whenever I fed the young terns. These two birds learned to eat both live and motionless fish either from the hand or from a dish. I think it not improbable that, had time permitted, I could have captured the birds and taught them to feed from the hand. Had this been done, hunger could then 220 Papers from the Marine Biological Laboratory at Tortugas. have been used as a stimulus for controlling their reactions. In future work it would be worth while to give this method a thorough test and to compare it with the method described below. A little experimentation soon convinced me that the method of using the nest locality in place of food was desirable under the circumstances, since the stimulus of the nest locality is almost as potent in its effects upon the reactions of these birds as the stimulus of food on the reactions of other animals. While this method of forcing the bird to overcome difficulties in order to reach the nest may not be so accurate nor so uniform in its stimu- lating effect as is the case when food is similarly used as a stimulus, it certainly has present in it all the elements of naturalness which even a Wesley Mills could demand. Before giving the results obtained from the use of this method, certain experiments bearing upon the general problem of recognition will be discussed. TESTS WITH NODDIES AS TO RECOGNITION BETWEEN MATES. My observations at the nests show that there is present in the noddy a very accurate functional recognition of both nest and mate. It will be re- membered that, to our eyes at least, the male and female noddy are indis- tinguishable. The question then becomes a pertinent one as regards the way in which the female recognizes the male and vice versa. Is there recog- nition between the two birds or is the nest alone recognized and reacted to? From much observation, unsupported however by experimental work, it was forced upon me that the male and female could recognize each other, at least within the nest locality (actually observed recognition of this func- tional kind took place within a radius of 10 to 12 feet). A few preliminary unsatisfactory tests of the following kind were made upon this subject: Two birds were caught from marked nests and treated in the following way: The black neck feathers of noddy No. 1 were dyed red with Higgins’ ink. The white head and eye-spaces of noddy No. 2 were dyed red in a similar manner. Noddy No. 1.—Released at 4" 28™ p. m., 30 feet from nest. Flew to water at once and “ dived” six times into the water. Circled back to nest, alighted, and covered egg at once (mate being absent). A few moments later the mate returned, would not alight, circled around the nest several times, then sat down near the nest for a long time; then hopped upon the nest and pecked at the marked bird. Sat behind the marked bird for a long time, then both birds began nodding to each other in a very ludicrous fashion. After 40 minutes the unmarked bird crawled upon the egg, the marked bird taking up a position on the rim and nodding to the unmarked bird. One minute later it again flew to the water and began diving. Two minutes later it returned and alighted near the nest. No further family dissension seemed to be caused by the dye. Noddy No. 2—The behavior here was in marked contrast to the above. No. 2 was likewise released 30 feet from the nest. Flew immediately to w — The Behavior of Noddy and Sooty Terns. 2 nest and covered egg (mate being absent). Both the noddies and the sooties in the vicinity of this nest were badly frightened by this bird. The mate on returning flew round and round the nest, but would not alight near it. The mate would circle out over the water and then back to the nest. Finally the unmarked bird alighted near the nest and fought off an intruder. The unmarked bird sat near the nest, but made no attempt for a long time to take its turn upon the egg. The marked bird became restless under these circumstances. It would hop up and expose the egg, turning round and round and make signs by nodding the head to the passive mate sitting nearby. The mate made tentative efforts to approach closer to the nest and seemed almost persuaded to take its turn, but the relief was not finally effected until 3 hours had passed. These two tests as they stand prove nothing definitely. They only sug- gest that any change in the visual appearance of the birds breaks down the customary and habitual responses which take place between birds attached to a single nest.1_ It occurred to me afterwards that since the noddy has be- come accustomed to the appearance of the sooty, it would have been a more reliable test had I painted the noddy in the guise of the sooty. The vivid red, if the birds sense color-tone (or even the changed brightness) obviously might have been exciting and might have aroused fear. It is just possible that if I had modified the appearance of the two noddies by painting them with different shades of gray that the change in the visual appearance might not have been noticed by the mates on their return. TESTS WITH SOOTIES AS TO RECOGNITION BETWEEN MATES. A similar test was made upon the sooty. Two birds were captured at marked nests, the night before the day upon which the tests were to be made. The white throat, breast, and spaces between the eyes of sooty No. I were completely covered with burnt sienna (in oil). Sooty No. 2 was painted as No. 1, but with permanent blue instead of burnt sienna. Sooty No. 1.—Mate on egg. Marked bird tried to alight. All the sooties nesting in the vicinity flew up, raising a tremendous uproar. ‘The bird alighted within a foot of its nest. Its mate and all the other sooties nearby drove the bird away from the nest vicinity, striking it violently with their beaks. It flew away, but returned in 8 minutes. While in the air it called to the mate (still upon the nest), mate got on feet, answered the call, showing evident signs that the call was recognized. Marked bird again alighted within a foot of nest. Violent commotion was again raised in the colony. Mate struck the marked bird violently with the beak and drove it away. Marked bird tried several other times to alight near the nest, but this was not permitted. Wherever in the whole island this bird tried to alight, commotion was aroused. It then disappeared from the island, after vainly trying for 3 hours to approach the nest. It did not appear again that day in the vicinity of its nest. The following morning, however, the marked bird was calmly brooding the egg. *There is a possibility, too, that the disturbance might have been due partially at least to the olfactory characteristics of the dyes. 222 Papers from the Marine Biological Laboratory at Tortugas. Sooty No. 2.—This bird was carried to a position within 20 feet of its nest and released there. The mate was not present. The marked bird, instead of going directly to the nest and covering the egg, alighted some 10 or 15 feet from it and attempted to walk past other nests in the vicinity. A commotion was immediately raised. The other birds brooding their eggs left them to attack this strange object. This bird ran rapidly out of my sight, being fought at every step, and disappeared for 10 minutes. At the end of that time I saw it approaching its nest. Three feet away from its nest it was halted by a row of hostile beaks. It left again and at the end of half an hour put in an appearance. Again it could not reach its nest. (A rather amusing incident occurred at this point. A noddy in search of a stick alighted near this marked sooty. The noddy is usually utterly ob- livious to the presence of the sooty when in search of his precious sticks. Such was not the case in the present instance. He dropped his stick, peered at the sooty, extending his head toward it and craning his neck in a most peculiar and unwonted fashion.) After some 2 hours had passed this bird finally reached its nest and covered the egg. The mate on returning* showed evident signs of restlessness and disturbance, but finally took up a standing position near the nest. At my approach some 2 hours later the unmarked bird crawled from the nest and the marked bird crawled on. Evidently the family difference had been satisfactorily arranged. These tests upon the marked sooties, in so far as they show anything of a satisfactory nature, lead us to believe that a change in the visual appear- ance of the bird is immediately noticed both by its mate and by the other sooties on the island. The disturbance caused by the appearance of these strangely marked birds was very much more pronounced than was the case with the noddies. Such experiments naturally leave the problem of recogni- tion of mate by mate almost untouched. It is a problem, however, which I believe can be attacked experimentally. TESTS ON RECOGNITION OF THE EGG. Since the nest is also accurately localized by both the noddy and the sooty, I desired to test whether the egg, the nest, and the nest locality were all of importance in this reaction. I made tentative efforts to arrive at some conclusion as regards this question. As a preliminary step I colored the eggs of both the noddy and the sooty and then watched their reactions under the changed conditions. The noddy eggs were colored with vermilion, blue-green, and violet aniline dyes. The birds immediately covered these eggs without the slightest change in their behavior being apparent. Eggs similarly colored with Higgins’ black ink produced no disturbance. Hen eggs and sooty eggs and eggs made of magnesium sulphate were likewise accepted without question. In this connection, an interesting incidental observation was made. A noddy before it lays its egg has habits different from those which char- “Since the bird was released in the morning, I was forced to sit near this nest until the mate returned—about 8 hours later! The Behavior of Noddy and Sooty Terns. 228 acterize it after the egg is laid (see p. 201). I found that by putting an egg in the nest of such a “ laying” noddy, I could change its habits from those of a “layer” to those of a “sitter.” One can observe under such circum- stances an almost immediate change in the general disposition of the birds. Before the egg is put down in the nest, the bird, which may be sitting on a nearby limb, will fly away at the slightest disturbance. When the egg is put into the nest, the bird on returning will alight near the nest and sit stolidly on a limb as before. Suddenly it is visually stimulated by the egg. It peers down at it, extending the head and withdrawing it, turning the head slightly to one side. It then alights on the nest. Contact with the egg seems immediately to change the disposition of the bird. The bird will now remain upon the nest, “rattling” in its gruff, hoarse way, and attempting to strike if one approaches too near. The reverse of the above behavior can be noticed if the egg is removed from the nest of some bird, even if it has been sitting on the egg for several days and consequently has had exercise in all the instinctive activities present during the brooding-season. The sooty reacts quite differently to colored eggs. They were dyed with the same dyes as were the noddy eggs. These tests were made upon three separate parts of the island. The sooties on returning and finding the col- ored eggs exhibited signs of great uneasiness. They walked round and round them, poked them with their beaks, rolled them out of the nest and then rolled them back. In one set of tests the eggs colored with vermilion dye were absolutely rejected. In another set of tests, made with different birds, vermilion was accepted; green was not accepted in one case, but in two other tests it was accepted. The black egg was rejected in one case, but was accepted in two other cases. On the whole, each color and black was accepted, but in every set of tests at least one of the dyed eggs was rejected. In one case the bird whose nest contained the vermilion egg dug a new nest alongside of the old one and made nearly Ioo trips between the old nest and the new. They, however, accepted each other’s eggs and noddy eggs without question. One sooty sat down immediately upon its nest after a hen egg had been put there in place of its own. Several days later, on again examining this nest, I found that the hen egg had been pushed aside and a new sooty egg deposited. From these experiments on the dyed eggs, it becomes apparent: (1) Under the conditions of the above test the noddy is not at all affected by changing the hue, brightness, and markings of its egg. (2) The sooty is affected by changing the visual appearance of its egg, but whether in the latter case the disturbance was due to the change in bright- ness or the change in hue or in marking, is not determined by the above experiments. (3) Neither the noddy nor the sooty recognizes its own egg. 224 Papers from the Marine Biological Laboratory at Tortugas. TESTS WITH NODDIES ON RECOGNITION OF THE NEST AND NEST LOCALITY. The nest of the noddy is not individually recognized. I found that I could exchange a large nest for a small one or vice versa; that I could tear out the old nest and construct a rough one of bay-cedar limbs, etc., without the noddy’s reactions being in the least affected. I then made tests of the following kind upon the nest locality, one of which I shall cite in detail: An isolated noddy nest, placed in the crotch of a limb near the stem of the bush, was moved 3 feet farther out on the limb, but was still left in plain sight. The noddy, on returning, flew to the old position of the nest. After some delay it alighted on the nest. It then flew back to the old posi- tion. It then flew to a limb a few inches above the old position of the nest and waited there for a time. It then made 9 trips between the old position of the nest and the new. After half an hour it settled down in the nest in the new position and remained quietly brooding the egg for a few minutes. It then became uneasy, got up, and made several more trips to the old position and back to the new. Finally it settled down upon the nest in the new position and made no further attempt to return to the old position of the nest. I next made it a nest out of bay-cedar limbs and put it in the old position, putting therein a sooty egg which I found at hand. On re- turning, the bird naturally had the choice of going to its old nest and its own egg in the new position, or going to a makeshift nest containing a sooty ege in the old position. It went immediately to the old position and set- tled down on the sooty egg with apparent satisfaction. These tests were re- peated on other noddies with similar results. I conclude from these possibly insufficient data that the nest locality exerts the stimulus for nest orientation and that the nest and egg as such are not important factors in this situation. Whether or not this return to the old position is accomplished in terms of visual data, my experiments do not show. Provided the one limb is left which supports the nest, the rest of the bush and the surrounding bushes may be cut away and the whole visual en- vironment greatly altered without the birds’ reactions being changed in the least. TESTS WITH SOOTIES ON RECOGNITION OF THE NEST AND NEST LOCALITY. The study of nest and nest locality recognition are the most interesting problems in the study of the life of the sooty. If one recalls the conditions under which they lay their eggs, namely, in open spaces and at distances apart sometimes not greater than 10 to 14 inches, one can not but admire the exactness and ease with which the sooty approaches its own nest. I have made numerous experiments upon the distance to which the sooty nest can be moved without disturbing the habitual adjustment of the birds to it. These tests, while not satisfactory as regards the determining of the sensory factors entering into this function, at least will serve to show the nicety with which the sooty makes its adjustments to the nest locality. WATSON STE Fig. 19. The strutting movements of the male Sooty. Fig. 20. The two Sooties on the left of the cut are beginning to fight. ight of Nesting Sooties To show possibility of studying behavior at night. ‘Cs 1 vi A at ae eo Ly, hie ct ay aes 1 : an ANS 2 Ay t ‘aoe at (Pan WATS ue 7 7 ae eo) ty * - y V ey iat ee Ly ‘ Judie + a7 ae ar Rs a Maaee : a ck ria b oe te mY / ahh er af PAT a ey , bf ‘ ; alls ey oe ats 4 A, ey X 4 ee 7 : Oey : We ie i es iH mi bas AT 7 i] yrs nay trae Sif e* » 7 i 4 ’ i P t , ' ‘ ; " “ U . A =a i" i; me as Pay a) Bn i , Ler Vee Lee 7 t ir y™ ” Das xlb | a Ae is ’ ; fia" ih 2 28 S S i 7 ’ é The Behavior of Noddy and Sooty Terns. 22 al EXPERIMENT I, (a) Anisolated nest under a bush was chosen for purposes of experiment- ation. Nest was left intact, but the egg was removed and placed in a new nest dug 18 cm. due north. Bird returned to the old nest and stood im- passive. Attempted to settle down on the nest and to poke egg, then looked up and saw egg, crawled over to it and sat down upon it. Bird got up and adjusted the egg and nest elaborately. Time for adjustment to situ- ation, 30 minutes. (b) I then allowed the bird 30 minutes at the new nest. At the end of this time I scared it away and placed its egg back in the old position, put- ting another sooty egg in the nest I had made. Result: Bird went back to old nest and covered egg without so much as looking in the direction of the new nest. (c) Experiment (a) was repeated with similar results. (d) Twenty minutes allowed for repose after above experiment. New nest No. 2 was then made 18 cm. due west of new nest No. 1. Result: Bird went back to original nest, shaped the nest, and scratched around in it and then walked over to new nest No. 2 and sat down on egg after adjusting it. (e) Twenty minutes again allowed bird in this position. I then fright- ened it away and noted the pathway of return. The bird walked by old nest, inclined toward it slightly, then walked on directly to the egg in new nest No. 2. EXPERIMENT II. A nest in the sand in an open space was chosen. A large tuft of grass was situated near. This made a very prominent visual characteristic. I pulled up the tuft of grass, obliterated the old nest (marking it with a pebble), making a new nest 88 cm. due north of the old, inserting the tuft of grass as nearly as I could in the same relations to the new nest as it stood to the old. Bird on returning stood for 8 minutes at old nest, then put down beak and attempted to arrange the egg just as though it were present. Later, bird walked over to the new nest, partially sat down, then got up and went back to the old nest, turned round and round in the exact position of the old nest, walked over to the new nest, arranged the egg, went back to the old nest, remained there at a loss for several minutes, turning head round and settling down with body exactly as though egg were present, then went over to the new nest and sat on the egg for a moment or two, then back again to the old nest. Finally adjusted the new nest elaborately and remained there in peace. I then frightened it away. On returning, it re- peated the above reactions, but with fewer trial movements. A second time I seared it away ; again it returned to old nest and tried to get contact with egg. This second time it stood at the old nest for a long time, fighting all the other birds away. Finally, at the end of ro minutes again walked over to the new nest and sat down. I scared it away again after allowing it to sit on the new nest for 12 minutes. Bird gone for 30 seconds. On return- ing, alighted from the air exactly upon the center of the old nest and again tried to adjust to nest in the old way. At end of 1 minute and 35 seconds waddled over and sat down on egg in new position. Was driven away again after 10 minutes of repose. Again alighted upon old position. Repeated it again; again alighted on old nest, but this time left it for the new in 12 seconds. Again alighted near old nest, ran to it and stopped there for an instant, passing rapidly on to new nest. Again alighted on old nest, but 16 226 Papers from the Marine Biological Laboratory at Tortugas. found new in 10 seconds. This routine was repeated 16 times, practically an afternoon’s work, without the bird adjusting itself perfectly at any time to the new situation. I repeated the above tests on 3 other birds which were nesting out in the open spaces, with absolutely identical results. In other cases I found that the nests could not be moved in any lateral direction for more than a few centimeters without the birds being badly disturbed. As a control test to the above I obliterated several nests and then redug them in the old positions. In no case were the birds disturbed by this. EXPERIMENT III. A nest was chosen in an open space, but very close to some bushes. [ obliterated the nest as the bird had constructed it, inserted a black pan, filled this with sand, and constructed a nest inside of it. This gave me an opportunity to move the nest wpward as well as laterally. On return- ing the bird alighted on the nest without showing any signs of disturb- ance. An hour later I came back and pulled the pan out of the sand and put a few sticks under it. The bird returned, but was not disturbed by this slight change. I then drove in four stakes 10 cm. high and mounted the pan thereon. This served to raise the nest upward without disturbing the other relations of the nest. The bird on returning alighted immediately on nest. The other birds gathered around, craning their necks and peering upward. The bird then stood up and came to the edge of the pan and peered down. This seemed to disturb it and it flew to the ground, but hopped up again immediately, covered the egg and sat there in comfort the rest of the day. Raising the nest 10 cm. in the air requires almost no adjustment on the part of the bird. On account of a storm on the island, which lasted for 2 days, no further experiments were made at this time on this nest. I next raised this nest 100 cm.; bird alighted immediately squarely on top of the nest; did not make a false movement. On craning neck over the edge of the pan a little later, however, became disturbed and alighted on the ground, and remained there for 45 minutes without attempting again to get on nest. I torced bird to fly up. Again alighted on the nest and began to brood the egg in comfort. On my return several hours later it was still sitting quietly on nest. On the second day after this (when this same bird was at the nest again) I lowered the nest back to 10 cm., its first vertical position. On re- turning the bird alighted squarely on the nest, making perfect adjustment. I scared the bird away. On its return the bird again adjusted accurately. 1 next moved the nest back to the height of 100 cm. Bird returned and alighted on egg and adjusted to it before I could get back to my position in the bushes. Adjustment in the vertical plane is made with exceeding rapidity and ease. I then moved the nest 100 cm. to the east, leaving it 100 cm. above the ground. Behavior of bird very interesting. Would not alight on nest. Alighted at the former ground position. After a long time flew from the old position and up to new position of nest. Immediately hopped down and began a most peculiar performance. Bird would hover in space, attempting to adjust to the nest in the air at its former position and height. It would then fly away again and come back to the old position and try to alight in space. This was done 20 times. At the end of 20 minutes the bird The Behavior of Noddy and Sooty Terns. 227 alighted upon the pan in its new position and sat down on egg. I then scared the bird away, 5 successive times, to see if it would alight imme- diately upon the pan. Each time on returning the bird alighted at the old ground position and proceeded from this point to the new position of the nest. I then put the pan back in its old position. Bird returned and alighted on pan immediately. In this position I then raised the pan to a height of 200 cm. This raised the nest well up above any of the surrounding bushes. This did not cause the bird the slightest disturbance. I forced it to make three or four adjustments to the nest in immediate succession. It made them all with equal precision. I continued my experiments on this bird in a similar way for several days and by repeatedly moving the nest now to the east, now to the west, etc., I succeeded in getting the bird to the point where it would immediately adjust to the nest regardless of its position." As was the case with the noddy, I found that the nest environment could be markedly altered without the bird’s being disturbed in the slightest so long as the position of the nest was not disturbed. SUMMARY. In the case of both the noddy and the sooty, the nest locality is the im- portant factor, the nest itself being reacted to by virtue of its location within this locality. Since environment can be greatly changed without disturbing the bird’s accurate adjustment to the nest, it is evident that if the adjust- ment is made in terms of visual data the visual environment which serves as the stimulus must be complex and have a wide extension. I am not pre- pared to admit from the above experiments that adjustment takes place in terms of vision alone. SOME EXPERIMENTS ON DISTANT ORIENTATION. In the present connection, I shall not take up in detail the various theories concerning the factors entering into distant orientation. Anyone familiar with the literature on the subject knows that the facts, as well as the theories, are in a chaotic state. I wish in the present instance to present a few facts bearing upon the subject. It is generally supposed that the homing pigeon possesses the function of orienting itself from a distance in a higher degree than any other animal. It is also supposed that even in the case of the pigeon training is necessary in order to get the bird to return to its home from a distance. The method? usually adopted is, first, to allow the bird to get thoroughly habituated to its cote; then at successive trials the bird is allowed to return to its nest from distances beginning at 0.125 mile, then ‘These tests are all confined to a radius of some 4 to 5 meters. I had intended carrying the test further to see if I could force the establishment of so strong an association that I could move nest from one part of island to another, but a storm which continued for several days made it impossible to continue the work. * Hodge: Method of Homing Pigeons; Popular Science Monthly, April, 1804, pp. 758-776. 228 Papers from the Marine Biological Laboratory at Tortugas. from 0.5 mile and so on. In a short time the bird, on account of the in- creasing distances to which it is carried, combined with its keenness of vision, establishes visual landmarks throughout an enormous territory. A well- trained carrier pigeon could thus hardly be taken into a neighborhood which would be entirely new to it. This presupposes on the part of the bird the ability to establish visual associations at an enormous rate. All the labor- atory tests which animal psychologists have made upon pigeons so far seem to show that the pigeon has no extraordinary ability to establish such asso- ciations.* It occurred to me that any migrating bird ought to possess the func- tion of distant orientation. Asa test I made the following experiments: EXPERIMENT I, Six noddies were captured one evening and marked characteristically and individually with oil paints. These birds were put on board the labor- atory launch, which happened to be making a trip to Key West on the following morning. The nests of these birds were all close together and were tagged with a large card in order to facilitate observation. Two of the birds were released at Rebecca Shoal Light, 31.38 km. (19.5 statute miles) from Bird Key; two at Marquesas, 72.75 km. (44.75 statute miles) ; and two at Key West, 106.02 km. (65.8 statute miles). [ kept their nests under constant observation the whole day long. Natur- ally, since the birds had been without food for some time, and since I had no guarantee that they would immediately seek the nest after reaching the island, I expected the return to the nest to be irregular. The results were as tollows: The two Rebecca birds, released at 9° 30™ a. m., returned about 12 m. The two Marquesas birds, released at 2" 15™ p. m., returned together at 4 p. m. The two Key West birds were released at 6" 30™ p. m. One returned at 7" 30™ a. m. the next day, the other at 5" 05™ p.m. These two birds, bearing out my statement that these terns do not fly at night, probably slept in the neighborhood of Key West and left early the next morning. A heavy gale and rainstorm set in very shortly after these birds were released and I doubted very seriously whether they would ever return. Apparently one of the birds was not affected by the storm, while the other was probably blown from its course. The respective mates of these birds remained on the eggs the entire time, going neither for food nor water (7). EXPERIMENT IT. Three noddies and two sooties (one of the sooties was known to be a male) were captured and marked as above. Their nests were likewise prominently marked. On the early morning of Thursday, June 13, these birds were put into a large insect cage and given in charge of Dr. H. E. Jordan, who was returning to New York. He carried these birds via the See Rouse: The Mental Life of the Domestic Pigeon, Harvard Psychological Studies, 11, pp. 581-613, and Porter: Further Study of the English Sparrow and Other Birds, Amer. Jour. of Psy., vol. xvu, pp. 248-271. The Behavior of Noddy and Sooty Terns. 229 government tug to Key West. There food was purchased for them (min- nows). At 3 a. m., Friday the 14th, Dr. Jordan boarded the Mallory boat Denver, which left at that time for New York. On board the boat the birds were both watered and fed. On Sunday, the 16th, at 9" 20™ a. m., the birds were released at lat. 35° 8’, long. 75° 10’ (12 miles east of Cape Hat- teras approximately). The wind was fair and fresh for several days after the birds were released. I kept their nests under constant observation, but had almost given up hope of their returning when, to my surprise, on June 21, at 8" 30™ a. m., I found both marked sooties on their respective nests. None of the marked noddies was ever found at its old nest, but several days after the sooties had been observed at their nests, by chance I ob- served one of my marked noddies attempting to alight on its nest. On account of the mate having formed new “ affiliations” this was not per- mitted, and I immediately lost track of the bird.t I have little doubt that the other noddies also returned to the island, but likewise were not permitted to return to their nests. The distance from Hatteras to Bird Key in a straight line is approxi- mately 1367.9 km. (850 statute miles). The alongshore route, which is the one in all probability chosen by the birds on their return, since they were gone several nights, is approximately 1739.6 km. (1,081 statute miles).? EXPERIMENT III. On Monday, July 8, two noddies and two sooties were captured and marked and given into the charge of Dr. Hartmeyer, who was returning to Germany by way of Havana. The birds were in such poor condition, owing to the enormous strain of several days’ feeding of their then quite large young, that we decided to release them at Havana instead of taking them farther out. On the oth the birds were carried by Dr. Hartmeyer on board the Government tug and taken to Key West, where they spent the night and part of the following day, the 1oth. They were carried in Dr. Hartmeyer’s stateroom to Havana on the night of the 1oth. Early in the morning of the rith the birds were released in Havana Harbor. All re- turned to Bird Key on the 12th. Since they had had to spend three days without food or water, they were in poor physical condition. They prob- ably spent one day and night around the shores of Cuba, leaving there early *At one of the sooty nests the egg had hatched. The egg at the other nest had hatched before the bird was captured. Apparently the 2 sooties which were left at the nests cared for the young birds without aid from the outside, the young being simply left in the nest while the parent sought the food. The behavior of the noddies left at home presents an interesting contrast to the “faithfulness” of the sooties. After 3 days had passed, one of the noddies took a new mate. At the other 2 nests one of the most peculiar incidents of my stay was happening. These two nests were in the same bush, one about 6 inches above the other. Both the birds remained stolidly on their nests for 48 hours without going for food; they then began leaving the nest regularly for food and water, brooding the egg and feeding at intervals closely approximating the normal. Finally the bird in the upper nest began bringing food to the bird in the nest below. Each time on bringing the food, the bird from the upper nest would nod and bill and coo to the bird below—reactions wholly similar to those engaged in by newly mated pairs. The eggs in both nests were neglected, no effort being made to keep them constantly covered. Sometimes the bird from the upper nest would spend a half hour or more in the lower nest. Sometimes the bird in the lower nest would spend its time in the upper nest. Again at times both birds would be away from the nests simultaneously. *It might be well to mention that the birds were transported in the hold of the Denver. 230 ~=©6© Papers from the Marine Biological Laboratory at Tortugas. in the morning of the following day. The noddies were observed on their nests at 7 a.m., while the sooties were noted for the first time at 6" 30™ p. m., of the same day. The distance in a straight line from Havana to Bird Key is approxi- mately 173.8 km. (108 statute miles). I think that these tests are significant. The return from Cape Hatteras is really startling. Cape Hatteras is hundreds of miles outside the range of distribution of the noddy and sooty terns. If my statement that the birds rarely leave the island for distances greater than 15 knots for purposes of feeding corresponds with the facts, it becomes extremely improbable that they could have formed visual associations throughout such a vast territory as that described in these experiments. While these experiments are not in any way crucial, the facts obtained from them are extremely difficult for current theories of distant orientation to explain. SOME PRELIMINARY EXPERIMENTS WITH SOOTIES UPON THE LEARNING OF PROBLEM BOXES. In all the following experiments, small wire boxes were put down over the egg and the bird was forced to overcome certain difficulties before it could reach the egg. The experiments will show that this method could have been used with profit if time had permitted. On account of the difficulty of making suitable problem boxes, our experiments are of a very rough and ready kind. EXPERIMENT I. A simple labyrinth was placed over a sooty nest. This labyrinth offered only one blind alley, but the bird was forced to change its direction three times and traverse a distance of about 3 feet before it ‘reached its egg. The movements of the sooty when this situation confronts it are char- acteristic. It first takes up a position which offers the plainest view of the egg and then attempts to push its head straight through the meshes of the wire. Leaving this position, it walks around and around the labyrinth as a whole and makes no attempt to enter the open door, although the latter is made very prominent by virtue of its being marked with two large up- right sticks. The first four or five trials consumed about an hour each, the birds working persistently most of the time. So little improvement was manifested and the test bade fair to consume so much time that I abandoned taking continuous notes upon it. I left the simple labyrinth in place, however. By the end of 3 days, both male and female had adjusted themselves perfectly to it and oan go in ae out with Se joe when LW hen I left the eee I carried a noddy to a diseanee of 4o miles. in pies to observe its behavior during the trip. The bird was turning in a circle, twisting, and poking at the mesh of the wire cage incessantly. The turns and movements of the vessel did not influence the movements of the bird in the least, so far as I could observe. When released the bird flew down near the surface of the water and started in the direction of home. Dr. Mayer. informed me that this bird returned to its nest. Reynaud’s law of “contre-pied”’ has in my opinion not the slightest basis in fact. In order to duplicate in reverse order on its return all the mov ements made on the out- going trip the bird would have had to fly back, revolving mainly in a circle! The Behavior of Noddy and Sooty Terns. 231 I attempted to get accurate tests of their time the birds became excited and tried to go through the meshes as before. EXPERIMENT II. The following series of experiments was tried upon a marked male sooty. A cubical wire box, 35 by 30 by 25 cm., was inverted over the nest. A simple opening, 9 by 12 cm., gave access to the egg. This opening was placed due west. The male on returning trotted round and round the box; found the door in 3 minutes; was uneasy; attempted to get out but could not find opening. Stuck his beak in and out between the meshes. Became excited and tore at the wire with beak for several minutes. A slight disturbance in the neighborhood at which the other birds flew up caused him to re- double his efforts to escape. Got out at the end of 14 minutes. Walked away for a foot or two, trotted in again immediately. Was uneasy and came out again, this time without useless movement. On trying to reenter a moment later, he missed the opening and went halfway round the box and returned before entering. Out again immediately. In returning again missed his way in and went three-fourths around the box, turned and en- tered. This time he turned the egg with his beak, but would not sit down upon it. Out again immediately. Entered after a few useless move- ments. Tried to come out again immediately, became confused and fought the wire. Stopped to adjust the egg, but again would not sit down. Came out again immediately, then entered and sat upon the egg, this time in appar- ent comfort. The box with its opening produced no further disturbance. Time for adjustment to new situation under above conditions, 30 minutes. I leit the box in position the rest of the day. On the following morn- ing, since the marked bird was still on the nest, I carried this test further by piling up loose sand around the entrance. The bird on returning was not in the least frightened by the change. Went immediately to the door, but finding the sand, walked round and round the cage trying to force his way through the meshes of the wire. Tried to get into door again and again, but would not scratch at the sand. Divided his time pretty well be- tween the east side of the box, where the egg could easily be seen, and the west side, where he had formerly gained admission. Bird was very persis- tent, but at times would walk away for a few feet and then run hastily back to the box and continue his useless movements. I then scraped away the sand so as to expose an inch of the opening. The bird alighted and passed by the door again and again. Apparently no perception of the situ- ation as a whole. I then exposed two inches of the opening so that only two inches of sand remained in front of the opening. Under these circum- stances, the bird mounted the sand pile again and again and attempted to peck his way through the meshes of the wire above the opening. Finally by accident he poked his head through the opening and squeezed through, making no effort to enlarge the opening. Time of whole experiment, 1 hour. After 5 minutes I drove the bird away (lifted up the wire box and allowed him to fly) and piled the sand up to the height at which he had previously been successful. Time for adjustment: 1.16 minutes. I next piled the sand up so as to completely cover the opening again. After 20 minutes of random movement and no success, I scraped away the sand so as to expose one inch of the opening. The bird came up and 232 Papers from the Marine Biological Laboratory at Tortugas. with great effort squeezed through the hole in 0.25 minute. I then re- moved all the sand and left the box in position as before. On the follow- ing day, since the unmarked bird (female) was on the nest, no further tests were made upon the male. On the second day from the above I continued my tests. As in the last test, 2 days before, I first piled up the sand within an inch of the top of the opening. Time for entrance: 2 minutes. I next piled up the sand so as to completely cover the opening. Finally, after 38 minutes of random movement, while attempting to poke his head through the wire mesh above the opening, he accidentally poked bill and head through the sand pile. Withdrew his head, walked round and round the cage as before, then came back to the hole and poked his head into it six or eight times. This was done apparently simply to get nearer the egg. Success at end of 40 minutes from the beginnine of the test. He went to the nest and sat down. Apparently the nest did not suit him as regards depth, for he immediately began to hollow it out by scratching. Apparently the scratching impulse arises only when the bird is at the nest. I think it exceedingly curious that this reaction was not utilized in such situations as the above. No further tests were made on him that day. Two days later I again took up work with him. I first tested him with the sand left one inch from the top of the opening. Time: 0.20 minute. I next covered entrance over completely. Result: Again many useless move- ments similar to the ones already described. Time: 7 minutes. I repeated this. Almost no useless movements. Bird dived at once for the opening, made a small hole, pulled his head out, ran half way round the box and returning squeezed through the opening. Time: 0.66 minute. Exactly the same procedure was followed with reference to the female with results identical, except that the female uniformly required more time to make the adjustments and was not so active and eager in her move- ments. From these experiments I concluded (1) that the egg and nest locality may be used in the place of food as a stimulus to the formation of new habits; (2) that the sooty tern can form associations by the trial and error method. EXPERIMENT III. After the above tests had been completed I tried the effect of placing the opening of the box in the other cardinal positions (no obstructions being placed to the opening). It must be remembered that all the adjustments of the bird had been made with the opening facing west. The opening was now turned due north. Results: The bird (male) on returning went immediately to the west, stood for a second and then walked to the south; went back to west, then went to the east, then back again to the west, then went north. The moment he saw the opening he went into it and covered the egg. Time: 0.96 minute. Second trial: Alighted on west, but ran immediately to north. Time: 0.08 minute. Third trial: Bird alighted on south, swerved slightly toward the west, but ran immediately to the north and entered. Time: 0.10 minute. The opening was then made to face east. Bird alighted at south, went to north by way of west, paused at north, then seeing door to the east dashed into it. Time, 0.10 minute. Second trial: Alighting south, ran to WATSON Su The Behavior of Noddy and Sooty Terns. Bae west and fought the wire for 3 seconds, hesitated at the north, then dashed round to the east and into the opening. Time, 0.13 minute. Opening was then turned south (the bird usually alights at the south). Result: First trial: Ran into opening before I could move away from the nest; not the slightest useless movement. Second trial: Ran into box be- fore I could time him. I repeated the above tests on the female with similar results. If the behavior of the birds under the conditions of the above test is contrasted with the behavior of the rat under conditions of a similar test, one is struck by the very great rapidity with which the birds make these adjustments. Turning the box as was done above and presenting it to the rat would have caused him the greatest difficulty. I conclude from this that the bird makes these adjustments largely on the basis of visual data, whereas in all probability the rat makes the same adjustment by means of kinzesthetic data.* EXPERIMENT IV. After these animals had been accustomed to having their nest coy- ered with this box for some 2 or 3 weeks, it occurred to me that if the recognition of the nest is accomplished largely by means of visual data, they ought to react to the box, even though it were moved from place to place without being disturbed very markedly by the change so long as it was not carried outside of their range of vision from the point where they alight. The box was first moved 88 cm. due north (entrance west as usual), while the egg and nest were left in the old position. Result: The bird (male) sat on the egg immediately without so much as looking at the box, which was in plain sight 88 cm. to the north. After sitting on the egg for a moment, the bird became uneasy, left the egg and went over to the box and entered it, and then returned to the egg. He repeated this procedure three or four times. On the last two trips, after entering the box and not finding the egg, went round and round the box. While he was doing this the female, which was standing on guard nearby (up to this time inactive, however,) rushed over to the egg and covered it. The male came back and drove her away and remained in comfort on the egg. I then went to the box, scooped out a nest in the ground inclosed by it, and put in an egg froma nearby nest. Result: Returned and sat on his own egg, but craned his neck and peered at the box, showing evident signs of “interest” in the box. I next removed his own nest, leaving the box as in the above test. Re- sult: Bird ran first to box, then back to old position of nest. He turned round and round in this spot, attempting to find egg. He then dashed for the box, entered and sat on egg, but was not quite comfortable and kept peering out the door and into the corners of the box. He then got up, walked around the cage once or twice, came back to the egg and adjusted it, and sat down in apparent comfort. Time for this readjustment: 0.66 minute. *Watson, John B. Kinzsthetic and Organic Sensations, ete. Monograph Suppl. Psy. Rev., No. 33, p. 85. 234 Papers from the Marine Biological Laboratory at Tortugas. lf the behavior of this bird be contrasted with that of others whose nests had been disturbed (p. 225), it will be found that his readjustment to the nest in a new situation was very much more rapid than in the case of the former. Apparently, a partial but incomplete visual association had been established between the nest and the box. It is clear in this test that if rec- ognition of the nest is not accomplished entirely by means of visual data, such data can nevertheless play the fundamental rdle under certain conditions. EXPERIMENT V. A heavy focusing cloth of rubber was placed over the box, leaving only the entrance, which was west, uncovered. The egg could be seen dimly, but the rest of the box was extremely dark. In order to see the egg the bird would now have to approach the nest and go very close to the door. Result: Bird apparently frightened at the dark object as a whole. Hov- ered over the box, but would not alight. Finally alighted and went up to within one foot of the entrance and peered in. His reactions were very curious. Went over to east corner of cage, peered at the covering, then backed off. Approached the entrance and peered in. Started in, but his courage deserted him. He did not go round and round as formerly. This probably was because the visual stimulus of the egg was cut off. He finally poked his bill into the door, pulled his head back, then began a curious pro- cedure of flying 10 feet away and flying back to the nest, repeating this in rapid succession. In conscious terms, this behavior suggested that the origi- nal orientation was in some way recognized by the bird as being wrong, and that by leaving and again approaching the bird sought to secure better orien- tation. This lends support to my statements above that if the stimulus to the nest depends upon visual factors the latter must be complex. The bird then came to the door, stood near it, poked head in and withdrew it, repeating this 15 times without moving from his position. At the end of half an hour he entered and stayed for a moment, but rushed out again. Entered two or three times again, but would not remain. At the end of 45 minutes, after many such timid entrances, the bird walked in and sat down in com- fort. I scared him away. He returned immediately and brooded the egg. He had become entirely adapted to the changed conditions. These experiments, inconclusive and unsatisfactory as they are, never- theless show that a definite question has been raised concerning the modus operandi of nest recognition and nest orientation, which is entirely open to experimental treatment. They likewise show that the nest locality, what- ever may be the stimulating factors present, may be used at will as a stimu- lus in controlling the reactions of the birds. SOME PRELIMINARY EXPERIMENTS WITH NODDIES UPON THE LEARNING OF PROBLEM BOXES. A few experiments similar to the above were tried on the noddies. A cubical box, 45 cm. to the side, made of wire mesh (12 mm.), was inverted over the nests of noddies built near the ground. A 9 by 12 cm. opening was inserted in one side of the box on a level with the top of the nest. The The Behavior of Noddy and Sooty Terns. 235 noddies were allowed to accustom themselves to entering and leaving the nest by way of the opening. When they had become thoroughly accus- tomed to the box it was possible to interpose obstacles in the way of their adjustments to the nest, by putting certain simple obstructions at the open- ing. Only enough tests were made to prove the applicability of this method for controlling the reactions of the noddies. The following extracts from my notes show the behavior of the birds in adjusting to the box when no obstructions were interposed at the door: Both birds present (one marked). They alighted on top of the box and attempted to poke heads through the meshes of the wire. After 15 minutes of random effort, the marked bird found the entrance and imme- diately covered the egg. I scared this bird out and before it returned the unmarked bird found the entrance and covered the egg. Both birds adapted themselves to the situation very quickly. After allowing the birds several days in which to get entirely accustomed to the box, a white bristol board sector was inclined at an angle to the cage entirely covering the entrance when approached from the front. The end of the cardboard, which rested on the ground, was pointed so that when the cardboard was struck or pushed it would fall over, thus exposing the en- trance. First trial: Bird alighted on a limb very near to the entrance and stood for a moment, then flew away. Returned and alighted as before, opened its mouth and gazed at the egg in a stupid, inert way. After standing for some time the bird again flew away and again returned. This was repeated four times. It then began flying away and returning to the top of the cage. This behavior of flying out over the water and returning to some new position on the cage was repeated many, many times. At last it alighted on the perch and “stuck its beak in at the crack made by the cardboard’s being inclined at an angle to the opening. By degrees the cardboard was pushed aside and the bird entered. Time: 41 minutes. Second trial: (10 minutes later). Flew against the cardboard, attempt- ing to insert beak at the same time. Repeated several times. This was a clumsy but ingenious method. Time: 15 minutes. Third trial: (One day later). Alighted on top of the box, flew away, re- turned, then flew at the door as in the last test, pecked in a slow and foolish fashion, finally pecked it open. Time: 1 hour. Fourth trial: (10 minutes later). Many trial movements as in the above case. Time: I hour and 15 minutes. Fifth trial: (One day later). (Egg had hatched during night.) Bird flew at door, mate sat stupidly on top of box. Flew against the door again, finally opened it by this means. Time: 13 minutes. Sixth trial: (10 minutes later). Many useless movements as before. The bird still attempted to remove the obstruction by flying against it. If this failed it fought at the wires on top of the cage. Mate stood stupidly on top of the box the whole time, but made no effort to imitate, follow or assist the marked bird. The cardboard was finally knocked down with the wings. Time: 16 minutes. Seventh trial: (One day later). After flying against it several times the cardboard fell. Time: 6 minutes. 236 = Papers from the Marine Biological Laboratory at Tortugas. Eighth trial: (Immediately afterwards). Very definitely done. Time: 3 minutes. Ninth trial: (Immediately afterwards). Shoved cardboard aside with beak. Time: 0.16 minute. Tenth trial: (Immediately afterwards). Opened with beak. Time: 0.16 minute. The above experiment was repeated on the mate (unmarked) of this bird. Its time was very much better than that of the marked bird. First trial: Flew against the cardboard and knocked it over immediately. Time: 5 minutes. Second trial: Flew against it vigorously. The wind assisted it. Time: 7 minutes. Third trial: This bird worked ten times harder than mate. Flew against the cardboard very definitely. Time: 3 minutes. Fourth trial: Not a useless movement. Time: 0.10 minute. Fifth trial: Interrupted work twice to drink. Time: 3.50 minutes. The above trials were all given on the first day and in immediate succes- sion. This change in method was necessitated by the increasing shortness of my stay. Sixth trial: (One day later). Flew against cardboard, the wind aiding it. Time: 1.50 minutes. On account of the wind, no further tests were made that day. The heavy winds continued for six days and the experi- ments had to be discontinued. Seventh trial: (Six days later). Opened by flying against it. Time: 6.50 minutes. Eighth trial: Opened by flying against it. Time: 9.68 minutes. Ninth trial: (One day later). Time: 7 minutes. Tenth trial: Time: 1 minute. Eleventh trial: Time: 0.75 minute. No further tests were made on this bird. Useless movements were rapidly eliminated during the first five trials given on the first day. Appar- ently these five tests and the one given on the following day were not suffi- cient to fix the association definitely enough for it to be carried over the six days in which no trials were given. The test is suggestive of a very low order of retentiveness. Other problems were submitted to the noddies in a similar way, but the results were not sufficiently definite to report. In conclusion, however, we may say that the above method ought to be one of great importance in ob- serving the reactions of the noddies under conditions of control. By it I am sure that in the end we may gain a knowledge of the variety and the com- plexity of the problems which the noddy can learn. The method would work equally well with respect to the length of time which such associations can be retained. The Behavior of Noddy and Sooty Terns. 237 THE DEVELOPMENT OF THE YOUNG NODDY IN CAPTIVITY. The following study of the development of the young terns in captivity is based both upon field observations and upon the observation of young birds reared in captivity. I was enabled to rear 3 young noddies and 8 young sooties from birth to 30 days of age.t At the end of 30 days they were still in good health, but I was forced to leave the island on the 18th of July, consequently my observations are concerned only with these first 30 days of the life of the young terns.’ On account of the great difficulty of securing suitable food, the rearing of the young terns entailed enormous labor both on my own part and on that of my servant. Each morning it was necessary to take a seine and a power-boat in order to catch a supply of minnows for the day. During the first few days of their lives, the young terns can swallow only very small, perfectly smooth fish. In order to keep the fish in good condition it was necessary to tow a live-fish car with us, into which the minnows could be emptied as soon as caught. During the first week, the young birds were fed about five times a day. This consumed an enormous amount of time. Each bird had to be fed individually and only one fish at a time could be given. When one considers that a young, live, healthy tern can eat from 12 to 40 minnows a day, depending upon the age of the bird and the size of the minnows, one can form some notion of the labor entailed in rearing the birds. The main difficulty we experienced was in catching the fish. On stormy days it was almost impossible to seine successfully, and even on some fair days, for some inexplicable reason, the fish were not to be found near the shore on any of the islands. As the season advanced, the minnows became scarcer and scarcer. The fact of the increasing scarcity of fish during July and August will be a serious handicap in the future for anyone who may desire to rear these birds and observe their growth for a longer period of time than the above. First day: The young noddies began to appear on the island about June 9 (1907). The first few hours after birth they are extremely helpless. During the first day of their life they exhibit few signs of fear, making little effort to shrink away from the hand. However, their eyes are very mobile. The 8 young birds which I observed on the first day were able at 5 hours of age to maintain the Age and body fairly veh ina noe position. At the 1y Benain ith 8 “pedaies and 12 sooties, fue owing to my attempting to feed item upon « salted fish” during a scarcity of live fish, 5 noddies and 4 sooties died. * The photographs of the young terns serve roughly to show some of the stages in their development during the period in which they were under observation. Sooties 1 day, 3 days, 8 days, and 30 days of age are shown on plate 3. Noddies 1, 3, 18, and 30 days of age are shown on plates 4 and 5. * About 65 per cent of the young are born with black plumage, the other 35 per cent are born almost pure white. Thinking this might mean a sexual difference, I asked Dr. Charles R. Stockard to make an anatomical investigation. It appeared that the color at birth is not correlated with sex. 238 Papers from the Marine Biological Laboratory at Tortugas. end of the first day the birds were able to stand fairly erect and to move their heads with some freedom, following my pen-point with both head and eyes when it was moved in front of their heads. When taken from the nest and put upon the floor, the birds showed a marked tendency to shrink when the shadow of the hand was thrown across the head. As I write with one upon the table before me, it is pecking vigorously at my fingers. Pecked once at a large spot on the table, then at its toe. They show no fear at being handled. They can not swim at the end of the first day. They were tried first in fresh water. No codrdination of limbs was present. Head could not be held up and birds began to sink rapidly. Water caused defeca- tion in one case and disgorgement in another (they had been left with the parent for 5 hours). The movements used in disgorgement were as rhythmical and perfect almost as is the case with the adult. The birds yawned a great deal (the same is true of adult noddies and sooties). This reaction is quite different from that used in opening the beak for food. While yawning they will not accept food. The note of the young noddy is very different from the hoarse, rattling sound of the adult. It is a soft, liquid, slow, plaintive ‘ querk-querk-querk.” The huddling reaction’ is present the first day. The birds all huddle together after being separated. They lower their heads and attempt to nestle under the bodies of their companions. At the end of the first twelve hours these birds are the superiors in the point of development of the sooties, but the sooties very rapidly outstrip the noddies. All of these reactions which are present in the young bird on the first day are of vital importance to them, with the possible exception of yawning (even this reaction, apart from its possible value in respiration, may be of value in strengthening the musculature of the jaws, etc.). It is abso- lutely imperative for them to have the free use of the head and eyes and to be able to stand erect and to peck during the first day. The feeding parent on returning alights near the young bird, puts down its beak, and suc- cessively touches and taps the beak of the young bird, then its part of the re- action is at an end, provided by successive disgorgements it keeps its beak and throat filled with small minnows. The young bird must stand up and strike the beak of the parent until the parent opens its beak sufficiently wide to admit the beak of the young bird. When the fish in the mouth of the parent come into contact with the buccal cavity of the young, the swallow- ing reflex follows perfectly. The pecking of the young birds at the objects in their surroundings is not at first a pecking in the sense that the little chick pecks. It is rather a striking reflex. By means of it the young bird gains access to the mouth and throat of the parent. The huddling reaction mentioned aboye is of value by reason of the fact that it gains for the young bird the protection of the parents’ body from the cold of the northeast trade winds which set in at night. It is nota protective reaction in the sense of hiding from an enemy. Second day: The young noddies were kept day and night in a box which was covered by a cloth. They are extremely quiet all the time. At first they are small eaters, CURSOS rarely more Lee three minnows of small size at a feed- * Not the true gregarious reaction in all crababiive It was the normal meee of reaction to the parent. The Behavior of Noddy and Sooty Terns. 239 ing. With birds in captivity the visual stimulus alone is not sufficient to pro- duce all the movements necessary to feeding. They do not raise their heads and open their beaks wide as do the young sooties. The food has to be placed on a level with the head, then at gentle tapping, the beak begins to open. There grows up a very rapid tendency on the part of the bird to work a fish into such a position that it can be swallowed head first. I found that their aim in pecking was not very accurate. Even after they are stimulated to begin striking at the fish with open beak they will strike above or below it, nor do they open the beak commensurately with the size of the fish. Apparently both sensory and motor sides of this visual-motor reaction need strengthening. Whenever the bird is disturbed, or when it is being fed, it continually emits the above plaintive little note. It is instantly hushed if the hand is placed over it. It huddles under the hand, turning, however, so as to keep the head and beak out (a characteristic reaction when under the breast of the parent). Codrdination in walking is apparently little further advanced. Third day: The birds eat more freely, but their appetites are in marked contrast to the sooties of the same age. The movements of the noddies in taking food from me are still not perfectly coordinated. They are better able, however, than a sooty correspondingly young to swallow a minnow taken crosswise into the mouth. The birds are still very quiet. In the early morning, while the young sooties kept in captivity are raising a noisy chorus, not a sound comes from the noddies. The birds were tried again in the water. This time the leg movements were fairly perfect. The head and neck were kept well above water and they managed to swim until the down became soaked. When the bird began to sink and water was taken into the mouth, disgorgement took place. A good deal of increase in strength was noticed in the movements of the wings, neck, and legs. Birds were observed pecking both at the ground and at one another, but as yet no fighting was noticed. Fourth day: The birds are eager for their food. Feeding from me, however, is still not perfectly done. My finger is often pecked at in place of the fish. They peck more frequently at one another, and I noticed for the first time one pecking at the fish another was trying to swallow. When one bird is fed at the edge of the box, the others immediately begin to crowd up near my hand. Either there is a following instinct being exhibited here, or else an association has been formed. The birds were taken to the sand and put down. They moved about very slowly. I separated them about 15 inches. In 10 minutes all had collected into one group and stood huddled together. Up to the present time there is no association established between a call on my part and a hastening toward me or an answering call on theirs. The sooties very soon learned to answer my call by running toward me and giving vent to a lusty “ peep.” Their sleeping attitudes were observed for the first time. They lie with the ventral surface of the breast down; head stretched out and turned to one side, sometimes both legs stretched out, sometimes only one. They were noticed to-day preening their feathers by movements characteristic of the adult. This was a very complete act. 240 Papers from the Marine Biological Laboratory at Tortugas. Eighth day: The 5th, 6th, and 7th days are characterized by the appearance of the fighting reaction. This is full-fledged and well codrdinated. From this time on I had to exert great care in keeping the young birds from injuring one another. Fighting is indulged in at any and all times unless the birds are covered. Apparently there is never any cause to evoke this reaction other than the mere visual stimulus of a bird nearby. On the 8th day the birds began to hop down from the box (a height of 3 inches) and to run to meet me. Ninth day: The birds have formed a great attachment for me. They will follow me all around the room. It is becoming more and more difficult to keep them in any box. They will clamber up to the top of the box if any means are at hand and will jump down from a height of 8 to Io inches. Eleventh day: Birds are now forced to eat without any assistance from me. They learn this more rapidly than do the sooties, for the reason that the adult noddy in feeding its young sometimes disgorges the food upon the rim of the nest. The birds soon learn to pick up the fish from the floor, regardless of whether the fish are squirming or still. Twelfth and thirteenth days: These days show a notable strengthening in all of their instinctive re- sponses. They peck frequently at small objects and have much more accurate aim. They rarely take the object into the mouth. The fights are now prolonged and furious. From further observation of the birds in the field I find that the pecking reaction is utilized by the young birds in controlling the movements of the parents. By means of it the young birds force the parents to cover them. When I cover the young birds with the hand they stop pecking. If I so arrange my fingers that they can not huddle underneath my hand, they continue striking it until the proper move- ment on my part is made to admit the bird under the cover of the hand. Once it receives the extended contact of the hand it gives a contented little “ querk,” which causes the others to hasten up and to huddle under- neath the hand. The nest habits of the birds of this age and older were noticed to-day for the first time. The nest is kept clean by means of their peculiar habits of defecation. The bird usually remains in the center of the nest. When defecation becomes necessary the bird backs quickly to the rim of the nest, stops suddenly, and forces the fecal matter far out over the rim (some- times 8 to 10 inches). The interior of the nest is never soiled. The birds in captivity, even when on a perfectly smooth floor, held to the same habits of defecation. The birds were turned loose on the floor at this age. They began pecking very definitely at all small objects in sight, for example, my tan shoe-lace, a spot on the toe of my shoe, grains of coral sand, bits of feathers, strings, small sticks, and matches. No attempt was made to swallow the small articles except in the case of the match. Three birds did attempt, however, to swallow this. *T have seen an adult noddy when defecated upon by another bird fly immedi- ately to the water and begin bathing. WATSON PLATE 9 Group activity of Sooties. Favorite sunning-place. « r rj : : ' i . » ‘v Inge. ru EAR a oc poh ae a aye Af ge ma ae ) hap ere ‘ce : The Behavior of Noddy and Sooty Terns. 241 One interesting fact, which bears out the contention of Lloyd Morgan, became increasingly apparent. The birds at first show no discrimination with reference to the objects at which they peck. I cite one instance in detail: One bird defecated upon a white surface. A second bird ran up, struck at the fecal matter, and got a quantity of it in its mouth. Movements characteristic of “ disgust” took place. The bird finally shook the material out. A little later the same bird came up to the fecal matter again, struck it this time lightly, and again shook its head and wiped its beak. About 30 minutes later this bird approached it a third time. Bird shook its beak violently before it came within half an inch of the object and turned away without further noticing the fecal matter. Two days later, in a similar situation, bird again was stimulated by fecal matter, but turned away before striking it. Many others of the birds were stimulated by this material, but while they were often at the point of striking at it, they always inhibited the movement so that the fecal matter was never taken into the mouth. They had in all probability already made the discrimination. On the thirteenth day a rather interesting association was set up by one of my birds. In the bottom of the box (14 inches in height) in which I was now keeping the birds, an old pair of trousers had been thrown with one leg extending over the edge of the box. One bird climbed up the trouser-leg and jumped down to freedom. This bird began scrambling out of the box as fast as I could replace it. None of the other birds imitated this one or ever escaped from the box by this method. Sixteenth day: A rather interesting development of the pecking reaction was noticed to-day. Small twigs of cedar, matches, etc., were pecked at by the birds and retained in their beaks. They made no efforts to swallow the sticks, but would walk around the room with them with head extended, a complete replica of the adult noddy throughout its nesting season (see p. 200). The birds were tested on cooked ‘‘ Cream of Wheat.” They struck at it eagerly at first, but soon learned to reject it in a manner quite comparable to the learning of the rejection of fecal matter. Nineteenth day: There is a very noticeable increase in the ability to use their wings. They are beginning to flop the wings and to use them in obtaining equilib- rium. They use them quite noticeably now when jumping down from a height. The characteristic movements of wing and leg of the adult when dozing on a limb in the sun were noticed for the first time. The wing turned to the sun is drooped so as to shield the body, while the opposite leg is stretched out, the body being supported by one leg. Beginning on the 16th day' and during the rest of my stay on the island the birds were fed only in the maze. At this age it becomes possible to use food as a stimulus in controlling their reactions. The statement of the be- havior of these birds in learning the maze will be given later on. The young birds in the field are slightly more developed at 19 days of age than the birds held in captivity. I captured 3 of these birds at about this age, *The nodding reaction (p. 196) was noticed first in a perfect form on the 27th day. Two birds were fighting ; after a vicious thrust No. 1 backed away from No. 2 and nodded vigorously. Fighting adults often do the same. 17 242 Papers from the Marine Biological Laboratory at Tortugas. but found them quite intractable. They were thrown into a state of absolute terror every time I approached them. I kept them in captivity for 72 hours, but they would neither eat nor drink. They fought me with all the vigor of an adult. Fearing their death from starvation, I returned them to their respective nests. The instinct of fear develops very early in these young birds in the field. Even at 3 days, if one approaches a nest containing a young noddy and attempts to pick it up, it will first disgorge and then strike vigorously with its beak. It is almost impossible to tame either the young noddy or the young sooty unless it is reared by hand from the first day. Lloyd Morgan, Spaulding, and others are unquestionably right when they affirm that young birds if taken early enough and reared by hand exhibit little signs of fear. THE DEVELOPMENT OF THE YOUNG SOOTY IN CAPTIVITY. The young sooty is born in a very helpless state.1_ On the 7th of June I took 8 young sooties, all born on that day. They differed slightly as regards their development. This is due to the fact that the birds dwell for varying lengths of time in the shell after it has become pipped. I have been passing through the bushes when the eggs were beginning to hatch and have heard quite lusty “peeping” and on looking about to discover the young bird have found that the noise came from birds still in the shell. Whether they are fed at this age or not I do not know. Sometimes they live as long as 2 days in the shell with only the beak protruding. At this stage in their development they make no response to the warning cry of the adults as they do later on. They go on “peeping” lustily after the adults have flown. The peculiar protective attitude of the young birds has been mentioned already; that is, the ability to lie outstretched and perfectly motionless. If left for a time in this attitude, they begin to “peep’’ as soon as the sun’s rays become oppressive. This means of protection persists until locomotion makes possible a more effective method. When the protective attitude disap- pears and locomotion becomes possible, the bird runs to cover when dis- turbed. Hiding is never in any sense complete. Indeed, the reaction seems to be almost thygmotactic. The moment the young bird can put its head in the crotch of a limb or get its body in contact with some solid object, loco- motion ceases. First day: The young birds which I captured showed apparent signs of fear. Movy- ing the hand rapidly near them, as in offering a piece of fish, caused them to dodge quite noticeably. The instinctive cry is a lusty “ peep-peep.” They are well developed, but clumsy. The wings droop and the birds have difficulty in standing; coordinated sitting positions likewise are almost impossible. The birds are somewhat hard to feed during the first day, but ‘For supplementary description of field behavior of young and parent see p. 215. The Behavior of Noddy and Sooty Terns. 243 after this age this process is easier with them than is the case with the noddies of a corresponding age. In the case of the sooties, as was not the case with the noddies, the mere sight of the food will cause the bird to open its beak. Light contact will likewise cause beak to open. Once the food gets into the mouth the rhythmical movements of swallowing follow per- fectly. Some locomotion is possible. They take a few wobbly steps with wings down and legs wide apart. Fairly well coordinated swimming movements are present, but the heavy down with which their bodies are covered soon causes the birds to become water-logged and they consequently sink. As the salt water begins to enter the mouth the bird raises its beak higher and higher and shakes out the water vigorously, crying lustily the while. One of these young birds had not emerged from its shell; only its beak and the base of the beak protruded (eyes being covered by a membrane). Almost any call from me would cause this bird to “ peep.” Small minnows were offered and swallowed. The contact of the minnows caused the beak to open. I removed the shell; the bird made adaptive movements as if com- pleting the process itself. After removal the bird was very insecure in its movements. Second day: The difference in growth is remarkable. Locomotion must have im- proved 100 per cent in the 12 hours. Birds can waddle around rapidly and maintain upright position fairly well. They follow moving objects with the head and eyes quite easily. Will dodge very quickly if hand is suddenly ex- tended toward them: Some tendency present to nestle under one another, under my hands, and under the folds of the cloth. Birds are beginning to peck at one another, the raw wings being the spot usually attacked. Feeding is quite easy. Moving the finger, fish, or bits of straw rapidly across the beak will cause the beak to open wide. Whole minnows are swallowed with ease. The birds sleep a great deal of the time. The adult birds flying over the experimental cage and calling are invariably answered by a “ peep” from the young birds. Even on the second day feeding is an active process with them. A fish dangled in front of them will cause them to strike at it. If the beak fastens upon the fish in the center of the body the young bird by a peculiar shake and twist of the head will suddenly right the fish and swallow it head first. It is safe to say that at the end of the second day the young sooty is sufficiently developed to take any fish from the beak of the parent which is likely to be found there. Even at this early age the birds are found pecking at one another's mouths or at a piece of food which another is swallowing. Pecking at one another in a way forceful enough to suggest the beginning of the enor- mously important fighting instinct was, however, not noticed up to this time. Any signs of fear which might have been present the first day have entirely disappeared by the end of the second. Third day: The birds have begun to run toward me when [ approach them and call. They will lustily answer my “peep” at any hour of the day or night. The fighting instinct appeared to-day in almost completed form. Two of my young birds faced each other and began striking simultaneously. A hold is taken on the body and maintained with grim determination, the 244 Papers from the Marine Biological Laboratory at Tortugas. victor all the time shaking his opponent as does the adult. Not the slightest sign of play has vet manifested itself. The birds as vet show no discrimina- tion as regards what objects they peck at. They are as likely to attempt to swallow my finger as the fish. Very minute objects, however, such as specks and hairs (as is the case with chicks, according to Lloyd Morgan) are not noticed by them. Fourth day: I was awakened at dawn by the lusty “ peeps” of the birds (I slept in a hammock about 10 feet away from them). When I “peeped” back the birds answered in a persistent chorus. Other sounds, which I made to test whether the above note was a discriminated auditory stimulus, were not answered when I made them higher or lower, or if they differed widely in clang quality. The birds, however, would answer the “peep” of other men if it were given at the usual pitch. The birds are healthy and larger than those in the field. Three birds, 3 and 4 days of age, were taken from the field to compare with the birds reared in captivity. These birds were exceedingly wild, and although I kept them in captivity for 2 or 3 weeks, while they “showed improvement in this respect, they never wholly lost their fear of me. Their sleeping and lounging attitudes are very striking. In most cases, after a full meal is taken, the birds go to sleep. When lying down one leg (at times both legs) is stuck out so as to give the ventral surface of the body contact with the ground. Head is completely outstretched and laid on one side; eyes are closed. At other times the bird rests its rump on the ground, using legs as a prop, and dozes in this attitude. As it dozes its head falls to the right or to the left or vertically downward between the legs. When the beak strikes the ground the eyes are partially opened, the head is raised again and the process is repeated. Fifth day: I began training the birds to-day, to get them to form the habit of com- ing to a certain place for food and eating without assistance from me. The test was very simple. A dry-goods box 8 inches in height had an inclined plane, 2 feet in length, leading from the top to the floor. The birds were to be fed on top of the box. I first put a bird upon the inclined plane and dangled fish in front of it, attempting to toll it upward. At first the bird would stand in its tracks, extending neck and head to strike at the fish. After trying this upon all of the birds I would pick them up, place them on top of the box, and there feed them. At the end of the third feeding the birds had begun to clamber up the plane. Walking upwards at this angle gave them some difficulty. They looked intently alw ays at the fish and paid little attention to their own movements ; occasionally mishaps such as falling from the plane occurred. At the end of the third feeding it was still necessary to start each bird by placing it on the inclined plane. No attempt was made by the other birds to follow the bird which was being tolled to the top of the box. The birds still show no discrimination as regards their food. They will swallow almost anything. I tested them with several different kinds of fish, small crabs, strings, and sticks—whatever was presented was swallowed unless care was taken to prevent it. The Behavior of Noddy and Sooty Terns. 245 Sixth day: There is slight improvement in getting to the top of the box. The birds are still very clumsy, but are becoming more and more eager. Seventh day: Birds eager for breakfast. Two walked up the platform before the fish were shown. Two others came up as the first two were being fed. Live fish were given them to-day. The birds’ movements are wonderfully fast. The extremely rapid movements of the fish are not faster than the action of their beaks.t_ There is little improvement in discrimination to be noticed. To-day, while I was feeding some of the birds, 1 felt a tug from behind, and on turning around, I found that one of them had swallowed about 3 inches of my handkerchief and was straining every muscle to force the rest of it down! At the second feeding on this day one bird ran immediately up the plan and three others followed. When fish were presented these four birds be- gan fighting vigorously. By the time they were fed, all of the other birds had clambered to the top. The young birds were offered lemon peel to-day. It was snapped up by all the young birds eagerly. They continued to snap and swallow the lemon rind for three to four trials, and refused thereafter to open the beak for it. Fighting is furious and prolonged. The young birds are exact replicas of the old in this respect. All that is necessary to start a fight is for one bird to come within striking distance of another. Eighth day: The association of walking up the plane to feed was perfect in all of the birds by the end of the last feeding on this day. They, however, had not yet learned sufficiently well to feed without assistance from me. The young birds began to-day to dig holes in the sand, using exactly the same movements which are employed by the adults in digging the nest, except that the young birds do not shape the hole with the breast as the adults do the nest. By digging such a hole, the bird secures a surface which is damp and cool. The holes are usually dug near some solid object, whether because of greater coolness there, or through some thygmotactic tendency, I am not able to state. A new instinctive reaction was observed to-day which is continued from this time on. A bird standing still will suddenly hop an inch or two in the air and come down in the same spot. As it descends it flaps its wings. The value of this reaction in strengthening the wing muscles is apparent. * An old sooty with a broken wing, which had learned to come to the experimental cage when the young birds were being fed, was carefully watched to-day. Live fish were put down before it in a shallow dish. The bird would catch the fish near the middle of the body and by a quick flirt would bring the fish around head end first before swallowing it. This bird’s movements were extremely quick. If I offered it a fish after it had been sufficiently fed, it would strike at the fish, take it in the beak, and then let it fall. Young birds were seen to do this early. The old bird would eat any variety of small fish but rejected sticks, grass, etc., after taking them into the beak. Lemon rind was offered. The bird took it into the beak but quickly rejected it. Would not take it a second time. 246 Papers from the Marine Biological Laboratory at Tortugas. Ninth day: I began feeding the birds in Porter’s' simple maze. The description of their behavior in learning this maze will be given further on in the paper. Detailed records of the development of these birds were kept until they were 30 days of age, but on account of the great similarity of these later records to the above, I shall not cite them in detail. Experiments on learning to discriminate were made from day to day, but all such tests gave results very similar to those which have already been described. After the birds had learned to pick up the fish from a dish, it was possible to put seaweed, grass, bits of coral, etc., in with the fish. At first these were taken eagerly, but after a few trials the birds would learn to take the fish and leave the débris. Discrimination, however, never became very accurate. When very hungry, the bird would attack objects other than the fish. I think this is but natural, since the young bird feeds entirely from the beak of its parent until it is able to fly and fish for itself. They certainly get nothing there which would tend to develop discrimination. It may be of interest in this connection to say a few words about the way the young birds in the field and those in captivity spend the greater part of their time. I have already remarked upon the fact that there are no signs of play. This was as nearly true of the birds in the field as of those ob- served in captivity. The only reaction which it is at all possible to consider a playful one is the one already mentioned, namely, the frequent hopping up and flapping of the wings. There is a question in my mind here as to whether after all Gross’s theory of play fits the facts in any other genus of animals so well as in the case of mammals. Certainly the facts which I could gather with reference to both the noddy and the sooty terns do not lend support to his theory. These birds certainly lead an instinctively com- plex life. Surely the picking up of a live fish darting over the surface of the water is as complex an act as the catching of a mouse. The greater part of the time is spent in doing absolutely nothing. The birds will lie outstretched in their sand-holes, getting up at times to stand stockstill for an hour or more, or to doze with head bent down. At inter- vals they peck at their feet and occasionally preen their feathers. In the field the routine is broken by the call from the parents which have returned to feed the young. At other times the young bird gets hungry; at such times it begins to cry lustily and to go up to the parent (which is likewise standing or fighting in the neighborhood) and begs for food by “ peeping” and by striking at the parent’s beak. Getting nothing, the bird will wander off for 1o to 15 feet into the shade and lie down again. Frequently it engages in fights with other young birds. Occasionally it will attack an adult noddy which has dropped down to gather sticks. The young sooty never attacks an adult sooty. *Op. cit., p. 253. The Behavior of Noddy and Sooty Terns. 247 As the birds get older they begin to fly short distances. At the end of 35 days’ they can cover 30 to 40 feet very rapidly by a series of flying hops. THE BEHAVIOR OF YOUNG BIRDS IN CONTROLLED SITUATIONS. After the young noddies and sooties had learned to eat without assistance from me they were allowed to learn Porter’s simple maze. A cut of this maze is appended. In place of admitting the birds to the maze at the entrance O, as Porter did, they were put down at H, and allowed to come out at O and walk up, Fic. 2—Ground plan of Porter’s simple maze. by means of an inclined plane, to the top of the box from which they had previously been taught to feed. The dimensions of the maze were the same as those used by Porter. The 'This was the most mature young bird on the island. I had taken the precaution to mark the time of the appearance of the egg in several nests early in the season, in order to have the reactions of the birds in the field to compare with those obtained from birds in captivity. 248 Papers from the Marine Biological Laboratory at Tortugas. maze as a whole was 4 feet square; the alleys were constructed of white bristol board and were made 5 inches high and 5 inches wide. The top was covered with 0.375-inch wire mesh. The bottom of the maze was of wood. In order to determine more accurately the number of useless move- ments, fine white coral sand was sifted over the bottom of the maze to a depth of 0.25 inch.t/ The movements of the birds could be traced com- pletely by the tracks which they left in the sand. By means of a small whisk broom these tracks were obliterated after each trial. Tt will be seen that the maze offers both a long and a short way to O. A few of the birds learned the maze at first by the long way; all adopted the shorter way before the experiment was over. The maze was placed in a large room, which was well lighted by 2 win- dows, one to the west of the maze, the other to the north. Since the birds were perfectly habituated to handling they were placed by hand at H and their time in reaching O, together with the number of errors made, were recorded by me. Thinking that.my presence might influence the reactions of so keen-sighted a bird, I watched their reactions through a small auger-hole bored through the south wall of the room. Two trials per day were given, one at the morning feeding and one 6 hours later at the noon feeding. The birds were allowed fully to satisfy their hunger after each successful trial in the maze. The night feeding was not made in connection with the maze, but in the large experimental cage which remained out in the open. Four sooties, of which 3 were males, and 3 male noddies learned this maze. THE BEHAVIOR OF THE SOOTIES IN LEARNING A MAZE. I began feeding the sooties in the maze by blocking off the runway G from the rest of the maze and by closing the entrance to S. The cover of the maze was raised and the bird was then put down in this runway and allowed to come out at O and walk up the inclined platform to the food which was contained in a dish resting upon a box 8 inches high.2. Two feedings were sufficient to accustom them completely to this situation, but since the noddies were not sufficiently developed to begin work upon the maze, I fed the sooties in this way for 7 days. The sooties began learning the maze properly when they were 15 days of age. At this age they are as fully developed, so far as codrdination of movement is concerned, as the noddy at 19 days. Table 3 and the graph (fig. 3) constructed from it show the time-records and the general features of the learning process. *This method is fully as accurate and far more cleanly than that of using smoked paper. * See p. 244 for tests on learning to walk up an inclined plane to the top of a box. The same plane and box used there were employed in this connection. WATSON PLATE 10 Fig. 26. Group activity of Terns. Fig. 25. Noddies attacking a strange bird. Fig. 26. Peculiar behavior of the Terns when one of their number has fallen into the water. species swims with any success.) The Behavior of Noddy and Sooty Terns. 249 TABLE 3.—Individual and average time-records of four sooties in learning Porter's simple maze, in minutes and decimals of minutes. No. of | BirdI. | Bird II. |Bird III.|Bird 1V.| Average|, No. of | Bird I. | Bird I]. |Bird III.|Bird [V | Ave trial. trial I 15.00 | 16.00 | 16.00 | 25.00 | 18.00 18 | 066| 075 | 0.83 | 0.16 | 0.60 2 | 20.00 | 4.00 | 14.00] 5.00 | 10.75 19 233) leo 2u 38 .20 .70 3 | 20.00 | 6.00 | 13.00 | 2.00 | 10.25 20 238 61 .06 .30 55 4 9.00 | 7.00] 10.00} 2.78| 7.18 | 21 ii 1.33 08 50 75 5 83 1.50 | 14.00 1.33] 4.41 || 22 30 88 | 2.12 .20 A 6 83 1.16 | 2.25 75 Te2A|\e23) -30 50 26 25 .32 7 2.33 | 4.80 | 1.93 | 1.18] 2.56 || 24 235) 2:10 50 .30 82 | 8 08 OI 45 78 78 || 25 28 155 58 25 41 98 1.28 1.55 30 1.02 26 13 -33 1.50 20 55 10 E75" Tees 61 1.00 QI) ||| 2 .20 58 81 .20 46 II .96 | 1.05 1.00) ||, “1250)||| I-02 28 .20 .30 58 30 +34 12 1.28 | 1.35 .46 .95 | 1.01 29 58 -73| 1.00 .20 87 13 -36 Ber || aes 33 S535 30 .26 .30 | 1.20 ht .51 14 2.61 61 P35 203 1.10 31 23 7) 35 .25 26 15 68 | 1.12 50 41 67 || 32 21 || 3.09 38 aun} 45 16 1.75 1.38 | 1.05 33 1.12 33 -20 -30 58 23 232 | 3% Gis) ||| Bio) 55 .23 1.13 34 2 25 .31 .16 2. The error record is practically valueless because of the habits of the birds and of the unsatisfactory nature of the maze.* The general features of the learning process are apparently not very different from those obtained by Porter on the pigeons, English sparrow and other birds. The time of the sooties, however, is markedly longer *T have been criticized both by Yerkes (Jour. of Phil., Psychol., and Scientific Method, vol. tv, p. 585) and by Miss Washburn (Jr. Comp. Neurol. and Psychol., vol. XVII, p. 532) for not presenting the error record in the case of the normal and defective rats which learned the modified Hampton Court maze (John B. Watson: Kinesthetic and Organic Sensations: Their Role in the Reactions of the White Rat to the Maze. Psy. Rev., Mon. Supp. No. 33). I wish to say in this connection that mazes constructed along the lines of these two do not permit of a satisfactory error record without an infinite amount of time being consumed in the process. With the above (Porter) maze a satisfactory error record would have to be made on a basis of the exact number of inches traversed in cul-de-sacs, the number of returns made, the number of false turns made, etc. There would have to be some way of indicating the value of a hesitancy at a blind alley, of a full turn into a blind alley, and the difference in error value between traversing, say, 5 inches in an alley and going its full length, etc. If one attempted to present an accurate record of the errors of a test, one would consume several pages in the description of each of the first few trials. If, on the other hand, one does not present such a descriptive record, and chooses arbitrarily to call any random movement an error, giving all errors equal value, how- ever much the random movements may differ in kind and extent, as Porter (op. cit., p. 253) did in the case of the sparrows, pigeons, etc., and as Kinnaman (Am. Jr. Psy., vol. 13, I, 173-218) did in tests on the monkeys, the record becomes valueless as a basis of comparison with the work of others. Errors in this sense mean nothing except possibly to the man who records them. Our technique in the field of animal psychology is so crude at the present stage of the development of the science that the problems which we present to our animals are not of the kind which easily permit the recording of “errors.” I shall welcome as eagerly as anyone a reconstruction of the field in such a way as to permit such records. (Since writing the above I have suc- ceeded in devising a satisfactory method for testing animals in the maze which per- mits us to give both the time record of each trip and a record of the total distance traversed by the animal at each trip.) 250 Papers from the Marine Biological Laboratory at Tortugas. than is the case with the birds observed by Porter. The reason for this is quite clear when one recalls that the sooties will stand perfectly still sometimes for an hour or more. Most of the time consumed by the birds was spent in idling at H or in standing in some other part of the maze. I have compared my record of “errors” with that given by Porter, and find that if anything the number of errors made by the terns is less than Minutes EERees see a i a Nea |_| = |_| LA | 7 ae ESE aNnae air BSas Cisbso acc aGieiNattar pits (23456 7 8 9 1011 [2 13 1415 16 17 18 19 20 2 22 23 24 25 26 27 28 29 30 31 32 33 34 Trials Fic. 3.—Curve showing average time of four sooties in leaving Porter’s simple maze. that made by many of Porter’s birds. Porter records 3.5 errors for the vesper sparrow; QI errors for the cowbird; 7.5 for English sparrow F 5; 58 to 45 errors for English sparrow M 6; 121 errors for English sparrow F 7; and 11 errors for English sparrow M 8, as being the total number of errors made by his birds in their first two trials It would be abso- lutely impossible for the young terns, with their idling habits, to make, within the time-limit shown above, anything like the number of errors which some of Porter’s birds made in the time shown in their records. li the time for idling, which is a characteristic mode of behavior and is not due to a lack of hunger, were taken out the curve as shown above would be much smoother and would lie very much nearer to the base line. If the latter part of this curve be compared with the corresponding portion of a similar curve obtained from the rat in learning the Hampton Court maze,” its extreme irregularity as compared with the rats’ curve will at once become noticeable. The terns never become the automata which the rats * What do “errors” mean in such cases as these! * Watson, J. B. Op. cit., p. 100, curve Im. The Behavior of Noddy and Sooty Terns. 251 become. The jerky, uneven reactions of the adult sooty are already showing themselves in the early behavior of the young sooty. THE BEHAVIOR OF THE NODDIES IN LEARNING A MAZE. The same routine of learning the maze was adopted in the case of the noddies as was described above for the sooties. The noddies, however, were 19 days of age before they were comparable as regards development with the sooties, 15 days of age. Table 4 and the graph constructed from it (fig. 4) show the time-records and the characteristic features of the formation of this association. TaslLe 4.—Individual and average time records of three noddies in learning Porter's simple maze in minutes and decimals of minutes. No. of BirdI. | Bird II. | Bird III. | Average. || No-of | BirdI. | Bird II. | Bird 1II. | Average. trial. | trial. | ae 16.00 7.00 8.00 10.33 ug} 0.43 2.12 0.18 0.91 2 17.00 7.50 | 9.00 11.16 14 41 .20 28 -29 3 19.00 83 22.00 13.04 15 43 33 .43 -39 4 7.00 3-00 6.00 5.33 16 46 +33 33 Ai; 5 7.00 5.00 3.00 5.00 17 25 45 28 32 6 7.50 66 2.25 3.47 18 SI .60 30 sy) 7 3-50 7-50 95 3.08 19 +33 +33 -16 27 8 7.00 2.00 1.00 3-33 20 30 33 30 i | 9 8.50 4.00 1.50 4.66 21 33 45 .20 *a28| 10 3.00 5.00 1.00 3.00 22 25 58 46 | 43 Tr | 7S 4.00 3.00 2157. |) 9:23 38 -51 233 || 4I | 12) |) 3:03 83 2.16 1.04 24 33 55 25 B77 | The noddy is very much slower in learning the maze than the sooty: (1) because of less general activity, and (2) because of longer periods of standing. Mention has already been made of the habit of the adult noddies of standing motionless for hours upon the top of any object which projects from the surface of the water. When the young noddy is placed at the unfamiliar position at H, he shows the tendency to react in this way to a marked degree. Although I exerted every care to keep the food conditions constant, I found that the first few trials of the noddy were quite unsatis- factory. So persistently would they stand in their tracks at H that I finally had to arouse them by tapping on the wall. The first three trials recorded above are valueless for this reason. From the fourth trial on, however, the records were made in the usual way. The errors are not markedly different in point of numbers or in kind from those of the sooty. The above curve, with the limitation above noted, is quite characteristic. It will be observed that 13 trials were necessary to bring the curve down to the point repre- senting one minute of time. This point was reached by the sooty at the eighth trial. This brings out the general fact which I have noted else- where,’ that the young sooty is much more active than the young noddy. *See p. 243. 252 Papers from the Marine Biological Laboratory at Tortugas. In any given problem, the successful solution of which depends upon random activity, the animal which has the greatest activity is likely to show a shorter time record for the early trials. In the second place, the curve for the noddy drops to a minimum at the fourteenth trial, which is maintained almost uniformly for the rest of the tests. The variation occurring throughout the curve for the sooty has already been remarked upon. A comparison of the two curves from the fourteenth yw = 4 Minutes aS SRGBe EEZeGEaSESaREot LL a2) be |_| Ee) fee) ale Para aie valle | Aaa —s = aa = S| BEEDME2ERRERSZel2e2 eee = 4567 8 9 10 tl l2 13 14 15 16 17 18 19 2021 22 23 24 Trials Fic. 4.—Curves showing average time of three noddies in leaving Porter’s simple maze. trial on brings out what I consider the second great difference between the two species: The sooty is highly excitable and nervous; the noddy is stolid and indifferent. The sooty, even though going through habitual reactions, is disturbed by slight changes in intra and extra-organic stimulations; the noddy on the other hand is indifferent to slight changes in stimulation. While these conclusions are based, it is true, upon few records, I feel after much observation of the birds that the amassing of a larger number of statistics would establish even more firmly the conclusions given above. EXPERIMENTS WITH THE MAZE IN FAINT ILLUMINATION AND IN TOTAL(?) DARKNESS. The trained animals used in the above experiment were tested at night with the maze faintly illuminated. The situation was as follows: For two or three days I had accustomed the birds to a late night feeding. I then made the room in which the maze was kept completely dark by boarding up the windows. I next mounted a commercial candle of ordinary size (standard candles were not available) upon a glass plate. I then suspended the candle so mounted over the center of the maze at a height of 50 cm. I found it absolutely impossible under these conditions to avoid shadows WATSON PLATE 11 Fig. 28. Group activity of Noddies. Sunning reaction. The Behavior of Noddy and Sooty Terns. 253 being cast over many of the pathways; consequently the results are more or less vitiated by this circumstance. I determined, however, to test the birds under these conditions. The birds were kept in the room for 45 minutes before being tested in the maze. They were quite lively and hungry. All other conditions of the test were the same as in daylight. The be- havior of the sooties is shown in the following: Sooty I: Started the short way as usual, but the deep shadows seemed to deter him. Turned into E and came to O by the longer route. Time: 1.50 minutes. Sooty II: Bird ran quickly completely around H. Came out and ran rapidly to L. Was deterred as in above case, ran into I and on into Z, came back to L and ran the full length of L and M and then went out the short way through S to O. Time: 2.66 minutes. Sooty II: Ran out rapidly from H into I and then into Z, then turned and went into R and then into E and ran up and down the alley R and E three times. On the last trip passed on through F and G to O. Time: 2.80 minutes. Sooty IV (female): This bird behaved very peculiarly. She spent 15 minutes in H. Came rapidly out to L, turned in there and ran the full length of M, came back and ran up I into Z, then back down I and poked head into R. Shadows there apparently disturbed her and she turned back into L. Dug a little hole there in the sand and after standing furtive and alert for 10 minutes, went to sleep. Total time under observation: 45 min- utes. No success. It is perfectly clear from the above test that the birds can run the maze in a faint light. Whether they can run it without error in the given illumi- nation can not be decided until we can eliminate the shadows which were mentioned above. On account of the unsatisfactory nature of this test the noddies were not tried in the same way. The next night I tried the birds with the maze in total darkness. The birds were carried in as usual and left in the maze room for half an hour before being tested. An electric contact and signal had been arranged to warn me when the bird stepped upon the inclined plane at O. Great care was taken to remove all tracings from the sand so as to get a complete record of the trial movements of the birds. The results of this test are curious. Each bird was left in the maze at its respective trial for 30 minutes. No signal reaching me, | then went in with my lantern. Not a single bird had moved from its tracks. Each had stood motionless at the point in H at which I had set him down. As soon as light was brought into the room the bird began “ peeping ”’ and moving. It is clear from these tests that the light conditions activity in these 254 Papers from the Marine Biological Laboratory at Tortugas. birds.t Under natural conditions in the field they are never in total dark- ness. Starlight on the island is exceedingly bright. We can not conclude from this test, however, without additional control tests (which I had no time to make), that the terns run the maze wholly by means of visual data or that they even learn it wholly in terms of visual data. EXPERIMENTS UPON THE ROTATED MAZE. One test was made upon the maze after it had been rotated 90 degrees to the north. The birds, for example, instead of traversing S—O in a south- erly direction, must now traverse this alley in an easterly one. The rela- tions of the turns, however, were not altered by this change ; only the cardinal directions were altered. On account of my having to leave the island, much to my sorrow, on the following day, this experiment could not be extended to the other directions. The results of this test were as follows: SOOTY IV. First trial: Bird acted very peculiarly at H. Came out of H, but ran tion L. Became confused, went up I, then into F and on around the long way. Time: 0.91 minute. (Normal time for control test: 25 minutes.) Second trial: Bird badly confused. Ran full length of L, then up I, then full length of Z, back down I, turned into R, then went long way via E, F, and G. Time: 2.43 minutes. SOOTY III. First trial: Perfect. Time: 0.20 minute. Second trial: Stumbles about at H, making circus motions, apparently trying to find entrance. Becomes discouraged, goes into L and runs full length. Comes out and runs up I. Into F, then to O by long route. When he reaches the east corner of the maze, he stops and tries to poke head out through the meshes of the wire. This direction had formerly led to the food. Time: 2.75 minutes. A third trial was given the following morning. Bird ran round to S, became confused, turned into E and came to O by long route. Time: 0.81 minute. SOOTY II. First trial: No error. Time: 0.31 minute. Second trial: Bird badly confused. Ran full length of L. Came out, turned into S. Movements hesitating and slow. Time: 0.75 minute. SOOTY 1: First trial: Perfect. Time: 0.20 minute. Second trial: Bird badly confused. Every error in the maze was made again and again. Finally reached O by the long way. Time: 2.91 minutes. 1Both species were tested to see if they would feed (without aid from me) in a photographic dark-room. The test was very carefully made. Neither the noddies nor the sooties would stir from their tracks. All sounds were hushed. The sooties would no longer reply to my “peep” as they would customarily in starlight. The Behavior of Noddy and Sooty Terns. 255 Third trial: Given this bird on the early morning of the following day. No full errors were made, but the bird showed a clearly marked tendency to turn in the old direction. Time: 0.25 minute. NODDY I. First trial: Perfect. Time: 0.16 minute. Second trial: Is confused and hesitant. Tries at two points to turn in the old direction. Time: 0.43 minute. NODDY II. First trial: Hesitancies, but no error. Time: 0.41 minute. Second trial: Confusion evident. Tendency to turn in the old direc- tion very marked. Time: 0.56 minute. NODDY JI. First trial: Badly confused at H. Tries in vain at first to find the exit from H. Runs full length of L, then out and up I into F and comes to O by long way. Time: 0.91 minute. Second trial: Runs full length of L, starts to run into S, but withdraws head, turns and goes up I, turns into Z and goes full length, comes back down I, runs into R, turns and goes up E on wrong way to O. Time: 2.43 minutes. From these experiments we are justified, I think, in concluding that there are some data used by these birds in this reaction which are not clarified by applying the term “ visual” to them.’ Porter’s results on the English spar- rows and other birds, obtained by testing trained birds in the reversed maze, are very similar to the ones reported here. My own experiments on normal, blind, and anosmic rats when tested in a similar way gave results which corroborate the above.” An explanation of the disturbances in the reactions of the birds which ensue upon changes in the directions of the alleys of the maze can not be made until a more complete and better controlled method of experimentation is at hand. 1There are two possible sources of error in this experiment. In the first place, the “environment” was not rotated with the maze; in the second place, the rotating of the maze changes its position with respect to the two sources of light. Owing to the conditions under which I worked, I could not control these defects. Porter in his work on the reversed maze (op. cit., p. 256) makes no mention of either of these two possibilities of error. *@p* cits, p; 0: IX. AN EXPERIMENTAL FIELD-STUDY OF WARNING COLORATION IN CORAL-REEF FISHES. By JACOB REIGHARD, Professor of Zoology in the University of Michigan. 5 plates. a OvIZHAW AO YOUTS-QJ9F1 IAT Aaa etn ee caNeT ABIAIANCD. VE VOPTAROIOO me PARAMORE FOUAU TR Bist lovin hel oH nr Yes te ee | Page. I. Introduction—conspicuous and inconspicuous fish of the region and their lla), Bran oes orien CRA AORIG CODSnCUc HT ncn OSC NODC C Oran On OO COR OUDD 261 II. The problem stated—as the interpretation of conspicuousness in coral-reef EISELE Pee voPe ere ea eer aeT oT NY eveTTCUSEE Tee he Siete Sia tice ovals akc lieve infele Se (avexchers Isieie 262 et Generalemiethod meniployecti:ryerercrcrerets laters fete are) sorverosss clave ckofers<¥=\o/a/als <¥eVe"el 1eielsvorereralere 264 IV. Reactions of the gray snapper toward unfamiliar qualities of color, odor, PASTCS MOT wh OLMM MIT etl O LIT all NOOO eeteretete epetere tetera ctelat syavereiesaveie savers) vee icleneiaieeeieioiets 266 ENDITGHITIAleCO LO letrtarrerrerera citer chats siaterctecerer\erelafelo) eleratelel elms viet eveiscs eleievaxsreierrers 206 Abi trial lta Shee OL O GOL. roc eiatelerefenes tebsiaee cain eicrere” sata =: otal ofeavolalaveatals: opaletoretaere 267 JANG RAMEN! asfosl Gas oouUooo00 eo cog Ce oU CoE ne eo au cues caniccnoneCG. 268 Wa Color discrimination mathe seray, SiMApp ety a.scetocsiennteysieaiiny-1o1a/ehololavor-1s) eistaiee:sl el 269 BEX penmientall smethOdS Emp OY.ed tore.e cieieieice susiciel ia) sinus « sieieresoisis et e%eee1s ve 269 Color discrimination experiments, 1907, methods...............2.0000> 270 Binewandiewhiteudiscriminationsccri cs cis cee reacildsaieee cee s catasiers 271 Bliewanderedediscrimrta ts Ors: eperctarsrs tele eralerele|ererateleiarcyerete e'evenaleyereiere ores 273 LEY ehaval Teper Cab oiribeathet-h0lOea man Anno on oD Ora rONOUCOOnE DT aooo0 275 Biwevand yellows discernment attorney yaicrerletaaeis seine eiaeiaye = cers auete elses or 276 Color discrimination experiments, 1905, methods...............000e000e 27 Bitiemandm witttemats crim imap Olle amtinr ese e erecta leis seca 279 Binexandered discriminationssce..s5 acer scence ieee nee: eee 281 Discussion of color discrimination experiments................20.00005 284 VI. Establishment in the gray snapper of a “ warning-color” reaction, involy- ing an association between color and unpalatability...................... 284 VII. The retention of the red-unpalatability warning association in the gray Snappers ((AMeEMOny er ee cee co to eer eisteieue els n a oniere in See re eh ia See ie ee ate diese 205 VIII. Results of feeding conspicuously colored coral-reef fish to gray snappers.. 206 IX. Rapidity and nicety of adjustment of the gray snapper to its food.......... 303 X. General discussion of conspicuousness in animals.................00eee00: 307 The significance of conspicuous coloration in coral-reef fishes.......... 307 History of the theory of warning coloration............ceeccsenscsseee 312 Analysis of the theory of warning coloration..............20+.eee0ee: 312 Criticism of the theory of warning coloration.................2.0+0++: 314 Immunity coloration, a substitute for the theory of warning coloration.. 316 BOS Tag TATA ED eA eye (=, dcaxas eect oihs svoyey eve cedetone val oven skeceure tek atovci ar Seale ea le/stmlatsy ste fa Reto cisyaereio werete 320 RATES I DLO STAD Yn tate nyetae wisi Careve levator tetas vor abevsvevere caste rate iste elenetetsistereie bisivuai nl steveseyeraneymcis 322 MIME planation: Otuplates cr preccner er electors cisieloiore ce icts distaste oo cicves Sisley pis.ort/orasctane 324 CONTENTS. 259 AN EXPERIMENTAL FIELD-STUDY OF WARNING COLORATION IN CORAL-REEF FISHES." By JAcos REIGHARD. I, INTRODUCTION. The islands of the Tortugas group rise gently from a submerged plateau many square miles in extent. To the southeast, at a distance of about 3 miles, is the outer reef, and beyond this deep water. Immediately about the islands the bottom is overlaid by the gray-white or yellow-white coral sand and, except where traversed by the few tortuous channels, it is covered with water from 8 to 15 or 20 feet in depth. From this bottom the inner reefs rise sheer. These vary in size from isolated coral heads of a square yard or two in area to small reefs of a few square rods or a fraction of an acre. The upper surfaces of these reefs are sometimes exposed at low tide, and at all tides one may readily wade over them. (See plates in Saville- Kent, 1893, of the Great Barrier Reef of Australia.) Although nearly level on top, the reefs are penetrated in every direction by chasms, fissures, and tortuous covered passages which intercommunicate so as to form veritable labyrinths. Structurally the whole is not unlike a mass of small boulders carelessly piled and roughly cemented together (plate 1; plate 2, fig. 3; plate 4, fig. 8). On its upper surface and for a short distance downward into the larger fissures each reef is clothed with living coral, massive or branched ; beneath this is dead coral and coral rock. The crevasses of the coral reefs harbor many species of teleostean fishes. The smaller species and the younger individuals of the larger species are never seen at a distance from this shelter. On the reef and in its imme- diate neighborhood they find their food; when disturbed they scurry to the reef and vanish into its protecting mazes. As the disturbance subsides they slowly emerge, in the inverse order of their timidity, and gradually resume their wonted activities. The larger individuals of those species that reach a considerable size may wander from the reef to a distance of some rods, protected from piscivorous fish of their own size apparently by their bulk. Thus Abudefduf *Contributions from the Zoological Laboratory of the University of Michigan, No. 116. 261 262 Papers from the Marine Biological Laboratory at Tortugas. and Hepatus may reach the shore and mingle there temporarily with the colonies of predaceous gray snappers. They invariably return in a short time to the reefs. The colors of the reefs themselves are subdued. The coral rock is gray, in places nearly white. The mantle of coral polyps seen in mass varies from cream to light-brown or delicate pea-green. It is almost a monochrome. Only the dark shadows in the fissures or an occasional purple sea-fan lends contrast to the picture. The coral-reef fish are nearly all conspicuous (plates 1 to 5, except fig. 9) either because of bright colors, patterns of contrasting colors, bizarre form, erratic movements, or because of some combination of these qualities. They often combine with conspicuousness disagreeable qualities in the form of defensive spines. The characters of some of these fish are given in table 13, page 299, and are discussed in section VIII of this paper. The conspicuous- ness is that commonly found in insects said to be warningly colored and the combination of this with defensive spines suggests the combination of con- spicuousness and disagreeable qualities of many warningly colored insects. In contrast to the fishes of the reefs are those commonly seen over the bottom of coral sand near shore. These are inconspicuous. The gray snapper, Lutianus griseus, as it appears over the coral sand (plate 4, fig. 9), is an example. It is piscivorous and its close agreement in color with its environment appears to be a case of general aggressive resemblance. Its common prey during my stay at the islands was the so-called sardine or hard-head, Atherina laticeps. This also is inconspicuous. It is almost invis- ible when seen from below against the surface film, as it ordinarily appears to the gray snapper. The fish is a plankton feeder. It does not find its food on the reefs nor seek their shelter. Its color is an instance, apparently, of general protective resemblance. Photographs were obtained of this species in its natural environment. Although the photographs were excellent the fish were so inconspicuous that the loss of contrast in the pictures, inevitable in the process of reproduction, made it inadvisable to make plates from them, as they would have shown practically nothing. Il. THE PROBLEM STATED. The conspicuousness of coral-reef fish appears to me to exclude any explanation based on protective or aggressive resemblance (cf. Wallace, 1891, p. 266). Semper (1879, p. 386) says, speaking apparently of coral reefs in general: The surface of a reef lying just under water has often been compared to a gay garden of flowers and the splendour of such a “bed” of animals is in fact quite astonishing. It is as though mother nature had here given free play to the fancy she is elsewhere compelled to restrain in some degree, by indulging her delight in lavish- ing all the colors of the rainbow and by inviting a motley company of creatures to disport themselves among the flowers and fruits of her submarine garden—blue and red star-fish, Holothuriz, of every hue, and gaudily painted fishes. Statement of the Problem. 263 The gaudily-painted fish occur on the inner reefs of the Tortugas, but the “ brilliant colors of the rainbow,” “the flowers and fruit’ I am unable to see in their environment. The brilliant fish are therefore conspicuous against a dull and nearly uniform background. I know of nothing in the environment which they resemble even remotely. Striking sexual differences have not been noted in these fishes by syste- matists, and I have observed none. I am unable to tell the sex by any ex- ternal character. Sexual selection seems therefore to be excluded as a possi- ble explanation, although it is to be expected that when we know the mating habits of the fish we shall find that they include displays of color. The theory of warning color often applied to certain of these fishes, most recently by Bristol (1903), seems then to afford the most plausible explana- tion of their conspicuousness. This theory attempts to account for con- spicuousness, more particularly in insects, by its association with some dis- agreeable quality. The vertebrate enemies of the conspicuously colored animal are believed to be warned of its disagreeableness by its color. They thus learn to avoid it. In this way, in each generation, the most conspicu- ous are preserved and through this selection conspicuous coloration is be- lieved to have been perfected. The theory thus attaches to warning colora- tion a biological significance—a present function—while at the same time it affords an historical explanation of the coloration, by asserting that it has been perfected through past selection of the functionally best adapted. The present paper embodies the results of a search for the biological function of the brilliant colors and striking color-patterns of conspicuous coral-reef fishes. If it appear that a warning function does not at present attach to this conspicuousness, then it becomes extremely improbable that it could have arisen in connection with such a function. The theory of warning colors embodies certain fundamental assump- tions which it seemed possible, in the particular case in hand, to test experi- mentally. (a) As stated by Beddard (1892, p. 155), “The theory of warning color implies not a special recollection of any type of insect, but a general association of bright colors with poisonous or dangerous qualities.” A like idea is expressed by Poulton (1887). After pointing out the few colors and limited number of patterns among warningly-colored insects, he says: “It is to be noted that advantage would accrue in the greater thorough- ness of the education, no less than by shortening the process, for a few colors with a few simple patterns would be remembered more easily than a larger number with a separate pattern in nearly every species.” Again, following Meldola (1882), he says: “All the conspicuous and dangerous or distasteful species in any country will be found to share between them a few strongly contrasted colors, arranged in few and simple patterns again and again repeated.” 264 Papers from the Marine Biological Laboratory at Tortugas. (b) The theory of warning color assumes that the enemies of warningly- colored species are capable of discriminating colors. (c) The theory assumes that the enemies of warningly-colored species are able to form associations between the conspicuous colors and patterns of their prey and their disagreeable qualities. (d) The theory assumes that the associations thus formed between con- spicuousness and disagreeable qualities are indefinitely retained (associative memory). It seemed possible to test these four assumptions by experiments on the predaceous fish that occur about the coral reefs or along the nearby shore. Do certain colors in themselves convey a warning so that prey showing these colors is avoided? Wave the predaceous fish color vision? Do they form associations? Have they associative memory? Experimental evidence on these points is presented in sections IV, V, and VI of this paper. Should the predaceous fish be found to have the qualities assumed by the theory, the warning color explanation of the conspicuousness of coral- reef fishes becomes more probable. But it would still need to be learned by feeding experiments whether any of the species of conspicuous fish are avoided, whether any show that combination of conspicuousness with quali- ties sufficiently disagreeable to render them relatively free from the attacks of predaceous fish under the conditions that normally prevail. Experiments involving the feeding of conspicuous coral-reef fish to predaceous forms are described in section VII. Through the courtesy of Dr. A. G. Mayer, Director of the Marine Bio- logical Station of the Carnegie Institution of Washington, I have been able to spend parts of the months of June and July at the station during the seasons of 1905 and 1907, while acting as assistant of the U. S. Bureau of Fisheries. I am indebted to the U. S. Fish Commissioner, the Hon. George M. Bowers, for permission to publish the results of my work. Ill. GENERAL METHOD EMPLOYED. The gray snapper, Lutianus griseus (Linnzeus), was chosen as the sub- ject of the experiments. This, the commonest predaceous fish of the region, averages 12 to 15 inches in length. At the Tortugas it is found about the inner reefs, but occurs also along the shore of Loggerhead Key wherever there is shelter. Nearly all of the individuals along the shore were, during my stay, aggregated in three colonies. The largest of these colonies was under the Laboratory dock and consisted in the summer of 1907 of from 150 to 175 individuals, but was smaller in 1905. The two other colonies were found under the two docks belonging to the lighthouse, one on the east side of the island and the other on the west. Each was about a quarter of a mile distant from the Laboratory colony. The personnel of each of these colonies seemed to be fairly constant. During the day individuals rarely REIGHARD PLATE 1 Fig. 1 R ledges near beach, showing three Abudefduf marginatus banded with black and yellow-white, and two yrunts |/ nulon sp., probably Sciurus triped with blue and yellow. Methods Employed. 205 wandered more than a short distance from the dock and always returned to it. I do not of course know that no changes took place in the colonies at night, but the results of the experiments as detailed below indicate that such changes must have been few. The gray snapper feeds in the daytime and by sight. It usually ap- proaches its prey slowly from below. When near enough (1 to 3 feet) it strikes quickly and, judging its distance with great accuracy, seizes the food, turns sharply, and returns to near its starting-point. The whole movement greatly resembles that of the end of a whip-lash when the whip is cracked ; hence probably the name “ snapper.”” The fish were entirely at liberty during the experiments and were taking their normal food and leading their normal life in an unmodified environment. The Laboratory colony was habituated to the frequent presence of people on the dock and paid little heed to them so long as they did not approach nearer than about 15 feet. This colony occasionally received food thrown from the dock and the fish were then accustomed to assemble, but the food received in this way was not enough to affect their normal appetite. They were always hungry. This feeding was, moreover, discontinued during the period covered by the more im- portant experiments. That it did not affect the experiments appears from the fact that the fish at the other docks were not fed, and yet their behavior was identical. The commonest food of the gray snapper in June and early July is the so-called sardine or hard-head (Atherina laticeps), a silvery-white fish about 2 inches long. It occurs in immense schools along the shore and is con- stantly pursued by gray snappers. To test the power of the gray snappers to discriminate colors, from associations, and retain them, atherinas were dyed various colors, as described below, and thrown from the dock to the snappers. The experiments were therefore mass experiments, in which an entire colony of gray snappers participated without being removed from their accustomed habitat. The water is very clear at the Tortugas, and it is usually calm on one side or the other of the island, so that the observer on the dock sees the snappers clearly and is able to record the results of an experiment without difficulty. 266 Papers from the Marine Biological Laboratory at Tortugas. IV. REACTION OF THE GRAY SNAPPER TOWARD ATHERINAS TO WHICH HAS BEEN GIVEN WARNING COLORS AND TOWARD THOSE OF ABNORMAL FORM, ODOR, OR TASTE. ABNORMAL COLOR. As the atherinas used in the feeding experiments had usually been colored by an aniline dye and were therefore dead, a test was first made of the re- actions of the snappers toward dead and living atherinas. After a single normal uncolored atherina had been thrown to assemble the snappers, 18 other normal atherinas were thrown dead and living alternately. One was taken as quickly as the other. In no case was there any observable hesita- tion in making the first snap at either dead or living fish. Tas_Ee 1.—Results of feeding gray snappers of the laboratory colony on “ warningly- colored” atherinas in July, 1905. Exp.| No. of : ae No, atherinas Color. Method of staining. Results fed. I 6 Bright ver- | Diamond package dye fast scar- Thrown in succession. All taken, | milion. | let, 5 minutes; weak acetic! No. 1 at once, while experi- | | acid, 2 minutes. | menter was 15 feet distant. | Nos. 2 to 6 with hesitation | | while he was but 6 feet dis- | tant. Z 7 |Deep ver- As under 1, but stained 10) Thrown alternately with normal milion. minutes. fish from distance of 15 feet. | | All taken instantly. 3 ff | MONB@ TD, ARERR || FE TN sccecestioseoncbo necoacaebendear | As in 2. | milion. | 4 Wa Wellow:... <0 Stained 20 minutes in saturated | Thrown alternately with normal solution of picric acid in sea- fish, from a distance of 15 feet. water and then rinsed. All taken instantly. 5 Si \Greenberea.=: Stained in weak methylene-blue | Thrown consecutively, from dis- and then in picric acid. | tance of 15 feet. All taken as | | quickly as normal fish. 6 7 Dark blue ..., Stained in methylene-blue in| Thrown alternately with normal sweet water, then rinsed in fish from a distance of 15 feet. | sea-water. All taken at once. 7 7 Dark blue ...| As under 6. ......... shuseventoseeestes| Thrown consecutively from a dis- | tance of 15 feet. All taken at | once. 8 7 Sky blue...... | As under 6, but for a shorter | As under 6. time. | 9 | 7 Sky *blueste-t|PAsitindenieee ee nceeccencemsgeeeceses Thrown consecutively from a dis- | | tance of 15 feet. All taken at once. | | 10 17 _ Brilliant pur- Diamond package dye violet, Thrown alternately with normal ple. | used as in experiment I. | fish from a distance of 15 feet. | All taken instantly. Atherinas were then given various colors by staining them as shown in table 1. These were thrown to snappers with the results shown in the table. It appears from the table that the snappers take atherinas colored bright vermilion, deep vermilion, yellow, green, dark blue, light blue, and purple, and all without hesitation and as readily as they take normal or uncolored fish. Not only were fish of all colors taken at once, but they were taken as readily when offered in succession as when offered alternately with normal fish. The snappers showed no hesitation in taking atherinas of any of the Reactions of the Gray Snapper toward Abnormal Atherinas. 267 colors used, except when the experimenter was too near them. ‘Thus in the first experiment the first red fish thrown from a distance of 15 feet was taken at once, while the remaining 5 thrown when the experimenter was only 6 feet from the snappers were taken with some hesitation. In subse- quent experiments the experimenter did not approach nearer than 15 feet and the colored atherinas were then taken readily. In only a few indi- vidual instances was any hesitation observed, and it was then found that this was coincident with the moving about of some person on the dock and that it concerned uncolored fish as well as colored. It is an interesting fact that this hesitation is greater if someone is fishing from the dock with a rod. The colors used ranged through the spectrum, although none of them, so far as I know, is a pure spectral color. Since the snappers gave no evidence that any of these colors served to warn them, it becomes highly improbable that any color in itself has a warning meaning to them. Had any color had such meaning they should have refused it or taken it with great hesitation. This conclusion seems to me warranted, even though the food offered them is, except for its color, one which they take at this season with great fre- quency. ABNORMAL TASTE OR ODOR. In order to study in a preliminary way the behavior of the snappers toward substances of disagreeable odor or taste, atherinas stained red were treated with various substances and thrown to the snappers with the results TABLE 2.—Reactions of Laboratory colony of gray snappers toward atherinas treated with substances LEB ARNE odor or taste, July, 1905. Seu Nowof | oe How prepared. | Results. in)! §ro Formic acid | Stained in fast scarlet containing | | Thrown after 3 red fish not g per cent.| 9 percent formicacid; rinsed | treated with formic acid ; some in sea-water. | hesitation over first 2 of these. All formic reds taken at once. 12 | 8 Formalde-| Red fish eviscerated and body | Thrown after 3 normal fish; all hyde 40. cavities filled with absorbent taken at once; none seen to per cent. cotton, which was _ then be rejected. saturated with formaldehyde, | 13 10 Formalde-| Asin 12, but surface of body also As in 12; all taken at once; 1 | hyde 40° wet with formaldehyde. formalin red swallowed with per cent. some hesitation, 14 10 Red pepper..| Red fish wiped dry, smeared with Thrown after 3 plain reds ; taken | vaseline, rolled in red pepper ; at once ; one swallowed slowly. | |__ very hot to tongue, 15 10 | Quinine ...... | Prepared as in 14, but with qui- Thrown after 3 normal fish; all | | | nine ; very bitter to tongue, taken at once, | 16 10 | Red pepper As in 14, but with mixture of As in 15. andquinine:| red pepper and quinine. 17 15 | Red pepper, | Prepared as in 11, then as in 16, Asin 15. quinine and, formic acid. | 18 10 Ammonia for-, Red fish wet with ammonia for- As in 15. | tior. | __ tior. Ig | 10 Carbon di- Red fish thrown directly from | Asi in 15. sulphide. | bath of carbon disulphide. 268 Papers from the Marine Biological Laboratory at Tortugas. shown in table 2. To call together the snappers 3 atherinas were thrown at the beginning of each experiment. These were either normal or stained red, but were not rendered disagreeable. The first 2 of these (experiment 11) were taken with some hesitation, but all other atherinas offered in the nine experiments were taken as readily as normal fish. If any of them had been subsequently rejected their red color would have made them conspicu- ous objects on the sand bottom. None were seen to be rejected. In two cases the fish, after being taken, were swallowed slowly. They could be seen protruding from the mouth of the captor, which, pursued by other snappers, slowly swallowed them. The substances used are as disagreeable to man as any that I know, some by reason of their taste, others on account of their odor or because they act as irritants. Probably none of them occur normally in the environment of the snapper. ABNORMAL FORM. To test the behavior of the snappers toward food of abnormal form, normal atherinas were modified in form as shown in experiments 20 to 25. Experiment 20: By removing the body between the anal and pectoral fins and sewing together the head and tail so as to produce a very much shortened form. Experiment 21: By removing the heads of two individuals and sewing the bodies together end to end so as to produce a very much elongated fish with a tail at each end. Experiment 22: By sewing together the bodies of three fish after removal of the heads and tails and thus forming a flat rectangular piece. Each of the modified forms was thrown to the snappers separately after a number of normal atherinas had been thrown to assemble them. The first (experiment 20) was taken with a little hesitation, but the rest were taken as readily as the normal fish. Separate heads and separate tails of atherinas were taken with the same readiness. Experiment 23: A piece of white twine 12 inches long and about 0.04 of an inch in diameter was tied by one end about the head of an atherina. It was thrown after one normal fish and taken at once. Experiment 24: A brown cord about 15 inches long and a sixth of an inch in diameter was tied by one end about the head of an atherina. This was thrown after a single normal atherina. There was a moment’s hesita- tion, three snaps which fell short, and then the fish and cord were taken. Experiment 25: An atherina was wrapped in cheesecloth and this rubbed with another atherina. It was thrown after a single normal atherina and was at once taken and was not seen to be rejected. From the preceding experiments it appears that modifications in the color, form, and chemical properties of the normal food do not prevent the gray snapper from taking it. The number of snappers involved in the ex- periments is so great that it is improbable that any individual had more Color Discrimination in the Gray Snapper. 269 than a single experience with any one sort of modified atherina. The evi- dence, therefore, does not show that the snappers do not learn by experience ; it shows merely that they try every new possible-food object that comes into the environment. ‘This fact that the behavior of the snappers has yet to be adjusted to each particular kind of new food is itself of importance in con- nection with the theories of warning color and mimicry. The snappers have, for instance, formed no habit of rejecting food of a particular color. So far as concerns new possible-food they are still in the condition of the young chick which pecks at all sorts of small, near objects. That the be- havior of the snapper may be rapidly adjusted to new qualities in its food appears in a later section of this paper. The necessity of such constant food adjustments is clear when we remember that the snapper increases many thousand-fold in bulk in its growth from the egg to the adult condition. Between the minute forms which must serve as the food of the very young fish and the larger forms upon which the piscivorous adults feed there is a wide gap, which must be bridged in the food of the growing individual. In the process of adjustment of the individual snapper a warning-color reaction may conceivably be established and may then be utilized by a mimicking form. V. COLOR DISCRIMINATION IN THE GRAY SNAPPER. EXPERIMENTAL METHODS EMPLOYED. It has been shown (table 1) that gray snappers take without hesitation red, yellow, green, blue, and purple atherinas, one with as much readiness as another. There is thus suggested the possibility that these fish may be unable to discriminate colors,’ an inability which would, 1f common to the predaceous fish of the reefs, be fatal to the theory of warning coloration. To test the color-vision of the gray snapper the laboratory colony was fed on dead atherinas that had been artificially colored. The natural silver-white color of the atherinas and the almost complete lack of pigment in the dermis makes it easy to dye them of a nearly uniform color. In addition to the normal white atherinas there were used red, yellow, blue, and green fish of light and dark shades. The snappers were first fed on atherinas of one color until they had become familiar with that color and were then offered a choice between that color and another and unfamiliar color. Choice was offered by throw- ing 10 atherinas together from the dock to the assembled snappers, 5 of the familiar color and 5 of an unfamiliar color. The order in which they were taken was recorded, so as to learn whether the snappers showed power of color discrimination by taking first the fish of the familiar color and last those of the unfamiliar color. Each throw of 10 atherinas is called * At the time these experiments were begun I knew of no experimental evidence that fish distinguish colors. Since then Washburn and Bentley (1906) have published an account of experiments on the minnow Semotilus atromaculatus, which appears to show color discrimination in a single individual of this species. 270 ~=6 Papers from the Marine Biological Laboratory at Tortugas. a trial; each experiment consists of a number of successive trials, often divided into several series. In the preliminary trials the atherinas were taken so rapidly that it was im- possible to keep an accurate record. It became further evident that normally hungry snappers might distinguish food of different colors, but might yet, on account of their hunger, fail to discriminate between one color and another. To reduce the rate of feeding to such a point that an accurate record could be kept and to make evident a possible power of color discrimination, it became necessary to prolong the preliminary feeding until the appetite of the snappers had been sufficiently dulled. The number of fresh atherinas taken by a colony of about 100 snappers was usually about 100 during the first minute. The preliminary feeding was continued until this rate had been reduced to 15 or 20 per minute. The discrimination experiments fol- lowed immediately thereafter. When, as in 1907, the atherinas used in the experiments had been preserved in formalin they were taken more slowly; the preliminary feeding was then not so long continued and the discrimi- nation experiments began as soon as the rate of feeding of the snappers was slow enough to permit an accurate record to be made. The atherinas used in each discrimination experiment were of the same average size. Those of the two colors were given a like taste by treatment with acetic acid or formalin as described below. Thus errors due to size or chemical properties are believed to have been eliminated. In each trial the 10 fish were taken together in the hand and thrown by a single movement of the arm so that they fell upon the water spread over an area of I or 2 square yards and intermingled at random. They were seized by the snappers from beneath. The random intermingling of the two colors in each trial, unlike in successive trials, is believed to have eliminated errors of position in the horizontal plane. In the experiments of 1907 all atherinas used were made to float (except in experiment 26, q. v.), so that all lay on the water in one plane. Fish which sink are brought nearer to the snappers and are taken more readily than those which float. Thus if the atherinas of one color should sink more readily than those of the other color there would be introduced an error of vertical posi- tion. This is avoided by making all the atherinas float, so that all are equally accessible to the snappers. COLOR DISCRIMINATION EXPERIMENTS (1907). While 4therina was abundant at the Tortugas in June of each year of my stay, it became scarce in July. On my arrival on July 8, 1907, the fish were already so scarce that it was impossible to obtain on any one day the large number needed to conduct discrimination experiments with fresh fish. I had, therefore, to accumulate a supply of atherinas by preserving in forma- lin those obtained each day. The discrimination experiments (except experi- Color Discrimination in the Gray Snapper. 271 ment 26) were conducted with these preserved fish. After the formalin had been as far as possible removed from them by soaking them in fresh water they were dyed and used precisely as though they had been fresh. The snappers took the formalin fish at first greedily, but after a little experience their rate of taking them became gradually much slower than with fresh fish. This rendered it possible to make a more accurate record of observa- tions, while at the same time it greatly reduced the number of atherinas necessary for the preliminary feeding. In all the experiments of 1907 the record was made by means of the following device: Two blocks of wood, each 8 inches long, 2 inches wide, and rt inch thick are held with their broader faces together by a metal pin at each end. This pin is firmly fixed in the lower block, but is received into a hole bored in the upper block, so that the blocks may be readily separated. Through the upper block are bored two parallel rows of holes, 10 in each row, and these holes are continued for about 0.125 inch into the lower block. Into each hole is loosely fitted a flat-headed copper nail an inch and a half long, which is filed round and brought to a blunt, conical point. The observer places one end of a sheet of commercial note paper between the blocks, so that the points of the nails rest on its upper surface. If he is observing the order in which red and blue fish are being taken in a trial of to fish, the upper row of nails (nearer the end of the paper) may represent the red, and the lower row the blue. The device is operated by touch, so that the experimenter may watch the experiment while making the record. Begin- ning at the left, he perforates the paper by pushing down a nail in the upper row when a red fish is taken, a nail in the lower row when a blue fish is taken. After pushing down any nail he shifts his hand one nail to the right. When a record of one trial has been made, those of several suc- ceeding trials are made on the same sheet by separating the blocks and shift- ing the paper. If the records of successive trials are kept in alignment vertically the vertical columns may be footed for each order and each color. A reduced transcript of such a sheet is given in figure 1, which shows a series of 4 trials with light red and blue. The record of each trial is to be read from left to right. In the first trial the first fish to be taken was red, the next 5 were blue, and the last 4 red. The totals at the bottom of the sheet show in how many of the 4 trials red and blue fish were first taken, second taken, etc. The numbers are the footings of the vertical columns for each color. Thus in 3 of the 4 trials blue fish were taken in the first order or place and in the other red was taken in first order or place. On the other hand, red was taken in the last place in all 4 trials. Experiment 26: Blue and white color discrimination—tThe fresh ather- inas were divided into two equal lots. One of these was placed for 20 minutes in a saturated solution of methylene-blue in 0.4 per cent glacial acetic acid; the other lot was placed for an equal time in 0.4 per cent solution 272. Papers from the Marine Biological Laboratory at Tortugas. of acetic acid without the methylene-blue. Fresh solutions were used for each experiment. The fish when taken from the solution were quickly rinsed in sea-water, and all had then the same sour taste. Before beginning the experiment the snappers were given a preliminary feeding of normal atherinas until their rate of taking them had been reduced to between 10 and 20 per minute. The discrimination experiment was begun immediately thereafter. Trial 1 5 light-red atherinas and 5 a blue atherinas were thrown in each Trial 3 trial, so that the four trials in- cluded go fish, 20 of each color. The circles at extreme right and left of the sheet are perforations made by the pins which hold the blocks together and are not in- cluded in the footings, Totals Ot Fic. 1—A reduced copy of a single record-sheet, con- taining first four trials of experiment 27. The blue and white (or normal) atherinas were divided into lots of 10 (5 blue and 5 white), and these lots were thrown to the snappers in succes- sion, a new lot as soon as all the fish of the preceding lot had been taken. The order in which the blue and white fish were taken was recorded by means of the device described above and in the form shown in fig. 1. Twenty-two trials were made. The records of the first 14 were discarded because it was observed that the white fish sank more readily than the blue, so that a position error was included. In the remaining 8 trials all the atherinas were made heavier than sea-water by laying open the air-bladder and emptying it by compression. All sank when thrown. These 8 trials appear in table 3, which shows the total number of blue and white atherinas Fig. 3. French grunts, Heweulon flavolineatum, blue and yellow striped, among coral rocks at base of reef. Atrighta jer (Lutianus apodus ?) banded with black id yellow. gonian, Plexaura. Fig. 4. An undetermined labroid feeding on rock bottom ; at its left a branching g Color Discrimination in the Gray Snapper. 273 taken in each order by a colony of 100 gray snappers, at the west lighthouse dock, July 21, 8" to 8" 30™ a. m., when thrown in lots of 10, 5 white and 5 blue; the last 8 of 22 trials; 80 atherinas thrown; all taken. TABLE 3. Order or place. | | | Grand total. } oo — —- > as I. | U1. |r. 1v.| v.| Totalfirst | vy.) vit. | vill. 1X.) X, | Total second} | 5 orders. | | 5 orders. | es (See —- |= eS) |e 8) eae PE 2 ——— | Blue SSeaceeOe Coy oP] Se 1 Coa |) | 6 l2| 8 | s|ala| 34 40 | | NAAN Seepsanced Se |) RSet Gp Ny EE} 34 16] 0 o 0 |.o |} 6 40 | | H | = | | | | | | | 80 The Roman numerals indicate the orders or places for the fish of each color. The Arabic numerals show the number of times that fish of each color were taken in each order; they are the footings for each color of the vertical columns of the original record for the 8 trials included in the table. From table 3 it appears that in all of the 8 trials the first and second and fourth fish to be taken were white. In 7 trials the third fish taken was also white; in 1 trial blue was third. In every trial of the 8 the seventh, eighth, ninth, and tenth fish taken were blue. In the first 5 orders white fish were taken 34 times out of a possible 40, that is, 85 per cent of the fish taken in the first 5 orders were white, while but 6, or 15 per cent, were blue. In the second 5 orders these numbers were reversed; 85 per cent of the fish taken are here blue and 15 per cent white. Of the blue fish 32, or 80 per cent, remained untaken until the last white fish was taken. The atherinas used in the experiment were of practically uniform size; they had like taste (sour) ; those of the two colors sank and were inter- mingled at random in each trial. The only constant difference between them was one of color. In taking 85 per cent of white fish in the first 5 orders, while blue fish are still present, and in leaving 80 per cent of the blue fish until all the white had been taken, the snappers show clearly that they dis- criminate between the blue and white fish. It does not follow that they dis- criminate between the colors blue and white. The blue fish are darker than the white ; brightness may therefore be the basis of discrimination rather than color tone. This point can not be determined in experiments involving blue and white only. It is discussed in another place in this paper and reasons are given for regarding the blue-white discrimination as one of color. Experiment 27: Blue and red color discrimination.—In this and the fol- lowing experiments of 1907 the atherinas used had been preserved in forma- lin and subsequently soaked in fresh water to remove the formalin. It is be- lieved that enough formalin remained to give to all the atherinas like quali- ties in respect to taste or odor. The atherinas used were all females of nearly the same size and were made to float by the following procedure: The viscera were removed through 19 274 Papers from the Marine Biological Laboratory at Tortugas. a longitudinal incision made with scissors through the body wall on one side from in front of the shoulder girdle to behind the vent. The body wall, stiff- ened by formalin, sprang back into place and the air inclosed by it made the fish float. The incision was made in a plane extending latero-ventrally, so that its edges were as oblique as possible. The ventral edge was thus over- lapped and held in place by the dorsal and acted as a valve to retain the air. Table 4 gives the result of an experiment in which both dark and light red were offered with blue to a colony of about 100 snappers at the east lighthouse dock. The snappers were first fed slowly for about 15 minutes on blue atherinas, both to dull their appetite and to familiarize them with the blue color. About 80 blue atherinas were thus fed before the blue- light-red trials, which began at 10 a.m. The 5 trials of blue and light-red atherinas followed at once on the preliminary feeding. An interval averag- ing about 5 minutes was allowed to elapse between the trials and during each interval 15 or 20 blue atherinas were fed. In the blue-light-red part of the experiment the 100 snappers thus had offered them 80 blue fish in the preliminary feeding, 9 & 15 135 blues in the interval-feeding, and 25 blues included in the 5 trials, or a total of 240; while but 25 reds were offered. All the 265 fish offered were taken. The six blue-dark-red trials were begun at 45 12™ p. m. of the same day, after a preliminary feeding of 30 blue fish. The last trial included but 4 fish of each color, so that the total number of blue fish offered in this blue-dark-red part of the experiment was 194 as against 29 reds. The method of staining the atherinas was that employed in experiment 26. The reds were stained as in experiment 1. The light red was obtained by a shorter stay in the stain. In order to determine the relative bright- ness of the colors used the colored fish were matched with cardboard disks stained with the same dyes and these were compared on a color wheel with grays in the usual manner and with the following result: The blue was equivalent in brightness to a gray containing 65 per cent black; the light red to a gray containing 50 per cent black; the dark red to a gray containing 75 per cent black. The experiment with the snappers was conducted in bright sunlight. A considerable portion of this light does not penetrate into the water, but is reflected from its surface. The snappers, moreover, view the colored fish from beneath. They see the side which is illuminated by light reflected upward from the white sand bottom. The intensity of this illumination was estimated by examining the floating colored atherinas from below by means of the reflecting water-glass described in the footnote on page 208. The brightness of the colors used was then measured in a light which was judged to be of the same intensity as that which illuminated the colored ather- inas in the actual experiment. Table 4 shows the total number of blue and red atherinas taken in each Color Discrimination in the Gray Snapper. 275 order by a colony of 100 gray snappers at the east lighthouse dock, July 24, 1907. The experiment includes a blue-light-red series of 5 trials and blue- dark-red series of 6 trials—each trial of 10 fish—5 blue, 5 red. The last blue-dark-red trial of 8 fish only. Blue-light-red trials began at 10 a. m.; blue-dark-red at 4" 12™ p. m. TABLE 4. | - | ia ; Order o or place. peer | Color. al eal loon Tile: | | Grand totals. | I. | U1. |11n.)1v.| v.| Dotalifirst |/vi.) vit. | vill.) 1x.| x. |Potal second | | | | gees orders. | | 5 orders. Fatae: alsi3lals| 21 |3|/ o| r lolo 4 Ae |i Light'red.*r pees 4 Bi Gell gh ee S | apn 25 f9 | WifpBlier:..<:.s 6|6|6]5/ 3} 26 }3| 9 | © | o]0| 3 29} 58 | { Dated ° 0/0 Tesi 4 \s || SO SAESs 25 29 5 | seal | | 108s * This fish sank. + All 5 reds of the first trial and 3 of the second were untaken; they are included in the table as though taken after the blues. In this experiment errors due to possible differences in the chemical properties of the fish of the two colors seem to be eliminated by the preser- vation in formalin and by the use of acetic acid in staining; errors of size are eliminated by use of mature female atherinas of practically uniform size ; errors of position are eliminated by the random mingling of the blue and red atherinas all floating at the surface. The snappers have clearly chosen the more familiar blue in preference to both a lighter and a darker red. Thus 21, or 84 per cent, of the blue fish are taken in the first 5 orders when offered with light red, and but 4, or 16 per cent, in the second 5 orders; 26, or about 9o per cent, of the blue fish are taken in the first 5 orders when offered with dark red, while but 3, or 10 per cent, are taken in the second 5 orders. I can interpret the result only as showing blue-red color-vision. To an observer the behavior of the snappers is quite as convincing as the tabulated data. Repeatedly when a blue and a red fish floated near one another, but at some distance from other atherinas, a snapper after swimming about beneath them for considerable time, as though examining both, took the blue. If in any such case during this experiment the red fish was the first of the two to be taken the fact escaped me. Repeatedly, too, snap- pers approached red fish and then jerked back and did not take them, but a similar behavior toward the blue was not observed. Experiment 28: Blue and green color discrimination.—By July 25, 1907, all 3 colonies of snappers had become so experienced in atherinas preserved in formalin that they took them slowly, and after taking a limited number in any one experiment, refused them altogether. The remaining experi- ments were thus necessarily each restricted to but few trials. On the afternoon of July 25 blue, prepared as in experiment 26, was offered with both light and dark green, all made to float. The greens were 276 Papers from the Marine Biological Laboratory at Tortugas. obtained by the use of Diamond package dye, fast dark green, used in accordance with the directions on the package. The brightness of the 3 colors was found on the color-wheel to be as follows: the blue equivalent to a gray containing 79.5 per cent black; the light green to a gray con- taining 51 per cent black ; the dark green to a gray containing 79 per cent black. Aiter the snappers had been fed on blue atherinas for about 10 min- utes 2 trials were made with the blue and light green, followed by one with the blue and dark green. The results are given in table 5, which shows the total number of blue and green atherinas taken in each order by a colony of 100 gray snappers at the east lighthouse dock, July 24, 1907. The experi- ment included a blue-light-green series of two trials, and a single blue-dark- green trial, each trial of 10 fish—s5 blue, 5 green. TABLE 5. Order or place. ~ ] a Tse a ek i . ae I. IL. 1. 1V.) v. | Total first | yy) vir. | vint.| rx.) x. | Total second ea 5 orders. a 5 orders. | E: : -| |, ——_ |__| f Blue......... ENED, ou|co|.2 5 + ao) Deon gu 5 eas | \ Light green’) o | 1 | 2| 2/0 5 \| x 2 EG) sr 5 10 j | SE rosea (AE | CP) a far 3 | x ° T0320 2 aera \ Dark green, 0 =r) rj) 0 0 | 2 || 0 | | oO) |r jor] 3 5 |_| |_| | shat 30 Table 5 gives no evidence of a discrimination between the blue and green colors used. These colors were, however, impure, the green includ- ing some blue and the blue including some green. To my own eye they were widely different, but other members of the Tortugas staff found diffi- culty in distinguishing between the blue and dark-green fish as they floated over the snappers. It is quite possible that the snappers would discriminate between a spectral blue and a spectral green if it were possible to experiment with these colors by this method. The blue-green trials are, however, of interest as showing lack of discrimination on the basis of brightness, for here the blue and dark green are of the same brightness, while the light- green is much brighter, and yet there is no evidence of discrimination. Immediately after the trial of blue and dark green recorded in table 5 a single trial was made of red and blue floating atherinas. These 1o fish were taken in the following order: R, B, B, B, B, B; 4 red fish remained untaken. The record, as well as the difference in the behavior of the snappers toward the two colors, shows clearly that they discriminate between them. Experiment 29: Blue and yellow color discrimination—On July 26 snappers in all three colonies took blue formalin atherinas so slowly at 7 a. m. that discrimination experiments could not be undertaken. At II a. m. the snappers at the Laboratory dock were found to take both floating and sinking blue fish with some interest. After feeding them on blue for about Color Discrimination in the Gray Snapper. 2047 4 minutes, 3 trials were made in which a choice was offered of blue and yellow. The blue atherinas were prepared as in experiment 26 and were of a shade equivalent to a gray containing 65 per cent black. The yellow fish were stained in Diamond package dye, fast yellow, according to direc- tions on the package. It was equivalent in brightness to a gray containing 53 per cent of black. The fish of both colors had been treated with acetic acid in the staining and were therefore of like taste. They were of. the same size and all, except one yellow, floated. The result of the experiments is shown in table 6. I was unable with the materials at command to obtain a yellow darker than the blue. The snappers in all three colonies had, moreover, now become so familiar with the unpalatability of colored formalin fish that further experiments with them could not be made. When it is re- membered that in the red-blue experiments the snappers discriminated be- tween the atherinas on account of their color tone, not on account of their brightness, and that in the green-blue experiments they fail to discriminate on account of brightness, it seems to me that in spite of the small number of atherinas included, table 6 shows clearly a power of discrimination between the blue and yellow used. The experiment includes 3 trials of 10 fish each— 5 blue, 5 yellow. TasLe 6.—Total number of blue and yellow atherinas taken in each order by a colony of 150 gray snappers at the Laboratory dock, ae 26, 1907. | Ores or Place. my : — i — | vals I mu) III. av. ly. Total first vn. Vu. vin IX aE Se Total second| nae | Bs 5 orders. . 5 orders, Ae) i) Bluesecscs: 2 Bike | II I I F |o |'o} *T Suis iho | 5} 4 aaeere saws o;o/;1 eee all wed a ae 22) 8 | *12 | | | | | | oe he | peniist. 4 | ¥26 *y blue and 3 yellow remained untaken in the last throw and are not included in the table, which therefore includes 26 fish instead of 30. 1 yellow remained untaken in the first throw and 4 in the second; these are included in the table as though taken after the blues of their respective trials. Here again the behavior of the snappers was even more conclusive, as the following extract from my notes, referring to experiment 29, shows: After feeding on blue for some time (3 or 4 minutes) they were offered blue and yellow in lots 10 (5 each). A distinct preference developed for the blue. The snappers showed little interest in the yellow, leaving fish of that color largely unnoticed, but swimming about under the blue until finally taken. One yellow sank and was examined perhaps 50 times by various snappers, but remained untaken on the bottom. The yellows were left for the most part apparently unnoticed, but when one was approached the fish jerked back from it. All besides myself who watched the experiment agree that the behavior of the fish shows more clearly than the actual record that they discriminate yellow and blue. This colony had had no previous experi- ence with yellow atherinas. we ~I Ce Papers from the Marine Biological Laboratory at Tortugas. COLOR DISCRIMINATION EXPERIMENTS (1905). I am unable to interpret the experiments of 1907, except as showing color-vision in the gray snapper. The experiments of 1905 are here added because, while following a somewhat different method, their results are in accord with those of the later experiments. These experiments differ from those of 1907 as follows: (a) The atherinas used were fresh, as in experiment 26 of 1907. (b) The preliminary feeding was therefore larger and usually included 300 or 400 atherinas. (c) The series of trials was much longer, because the fresh fish were taken more readily. (d) Some of the atherinas sank, while the others floated. If those of one color had sunk more frequently than those of the other color there would have been introduced an error of position. This was not known to happen except in the first 14 trials of experiment 26 above, which were therefore discarded. It is unlikely that in a long series of trials fish of one color sank more frequently than those of the other. (e) The order in which the atherinas were taken was called out by the observer and recorded by an assistant. The observer called, for instance, white, white, white, blue, white, blue, blue, white, blue, blue (see under trial I, table 7). The recorder gave serial numbers to the colors as called and entered them opposite the name of the color in a previously prepared blank. A portion of such a record representing 5 successive trials of a single experiment is reproduced in table 7, which is part of an original record show- ing the order in which blue and white atherinas were taken in 5 successive trials in each of which 5 white and 5 blue fish were thrown. TABLE 7. — Trial I. Trial II | Trial III. | Trial 1V. | Trial V. = _ ————$$______—.|— _ — — — —— Sj ————— BuO se. wos 4679) 10 | 3°68 9 z0 1789 10 236910 | Wihite®...<:.2: HEBSH 8 Tee GV 2345 6 Wap oi} 468 9 The results of each experiment were then tabulated to show in how many trials fish of each color were taken first, second, third, etc. Table 8 shows the result of two series of trials made on the morning of July 20, 1905. The table shows the same discrimination between blue and white that appeared in table 3, but less marked. I have no doubt that this dif- ference in results when using fresh and preserved atherinas gages accurately the discrimination of the snappers in the two cases. The greater palatability of the fresh atherinas results in those of the two colors being taken more in- discriminately, even though the color differences are perceived. With the less palatable formalin atherinas the snappers not only distinguish the colors, but discriminate more sharply between familiar and unfamiliar. There is this further difference between the experiments of 1905 with fresh atheri- nas and those of later date with formalin atherinas. Some of the fresh fish float; others sink, and are thus brought nearer the snappers and taken Color Discrimination in the Gray Snapper. 2479 first. There is no reason to believe that those of one color sink more fre- quently than those of the other, so that the sinking introduces no constant error. It does increase the difficulty of making an accurate record, but here again errors doubtless occur as frequently with one color as with the other. The greater discrimination apparent in tables from experiments with formalin fish I believe, then, to be an expression of the reality, rather than of a differ- ence in accuracy of method. Table 8 shows the total number of blue and white atherinas taken by 100 gray snappers in each order, when thrown in lots of to—5 blue and 5 white—during 26 successive trials ; 260 fish thrown; 5 remained untaken; July 20, 1905. TABLE 8. paler or _ Color ian} ] l | ! al Grand totals. if ie Il.| tv.| v,| Total first vt VII. ee 1X.| X, |Total second | | 5 orders. | 5 orders. - ———— | — | —_| —_|—— |= | eed | Blige 7| 7\ 15) 16011 56 | g| 411 | 1 12% 71 127 | WVDItG) ease aoe 19/19] 11| 10] 15 74 | 17 | 15 | 6 Ir | 54 128 Waa Lae Mare ieee | Lee eee Experiment 30: Blue and white color discrimination—This experiment was performed four times (July 19, 20, 21, 22), but the results were in- variable, so that only the series of July 20 and 21 need be considered. In each series the trials were preceded by a feeding of fresh atherinas and were continued so long as the snappers took the food readily. The series followed one another from time to time during the day, as the appetite of the snappers permitted. The results of the four series of July 20 are plotted in table 9. The data were first arranged in a table of the form of table 8, which combines the first two series of July 20. The number of times that blue and white fish were taken in each order was then calculated from the table in per- centages of the maximum number of times that it might be taken in that order, that is, in percentages of the total number of trials involved in each series of trials of the experiment. The diagram (table 9) shows in percent- ages the frequency with which a colony of 100 gray snappers took blue or white atherinas in each order in 4 series of trials including 460 atherinas. The atherinas were thrown in lots of 10—5 white and 5 blue together. The percentage of fish of each color taken in each order is calculated for each series separately on the basis of the total number of times that fish of each color might have been taken in each order, 7. e., on the basis of the number of trials in each series. This basis is in series I, 7; in series II, 19; in series INU, ES, theless WWE, se The solid line represents the white fish, the broken line the blue fish. The divisions on the ordinate at the left represent percentages; while in 280 Papers from the Marine Biological Laboratory at Tortugas. each series the columns numbered from I to 10 represent the orders. This table shows in percentages for each of the four series of trials made on this day what is shown in table 8 for series 1 and II] combined. If the snap- pers had, in each trial, taken all the white fish before taking any of the blue, then the solid line would run along the 100 per cent level through the first 5 spaces of each series of trials, while the broken line would con- tinue along the same level through the second 5 spaces. Probably if a single snapper could be experimented upon with so large a number of atherinas such a result would be reached. In an experiment involving 100 snappers they are crowded together, so that the effort of each to get first at the food stimulates the others. Apparently as a result of this competi- TABLE Q. SERIES | | SERIES 11 SERIES III SERIES IV July 20, 11h 37™to 11h 45™A.M)} July 20, 12615" to 1 P.M. July 20, 26 to 2h10™ P.M, July 20, 520™ P.M. 7 Trials 70 Atherinas | 19 Trials 190 Athevinas 6 Trials 50 Atherinas 15 Trials 150 Atherinas ORDERS ORDERS I ORDERS L ORDERS 1 : rool 2131415 6]7/8)9|tol1/2]3]4]5/6]7/8]9 10||1|2]3)4)/5]6|7]8|9 follt |2/3/4]5/6]7/8 |9 lto i + i + ir L — = ie | tt | tu an Soho I ime | | hal | ! | i ‘a CHITA Tt i ral all Wy ' HT i al 1 CHAE “GREECE Palla FA Cae A i i ri 1 ih TT} raw: LAA Et | | | and Me 5} ! iV! iat a) TN A 0 ele + i | | inal | 7 } 7 d} —— \ \ ‘ 1 val 1 TaN IND IAW AVIS 1 VTi 7 Ca a i Vea | Pe Z| | his iT f 1 =| eed AC) 1 WS | (ME V1 MS i ed | {| Vena ee SAVE Ti IW 25 H | | | | 7 2 | ! ill i I NI eae aS HT Ty lel ! | it 1| | L 1] 1 | | [Ps | [ | a ee | | tion a snapper frequently takes the nearest atherina, whatever its color, so that blue fish are often taken while white fish are still present. It results that the solid line starts usually somewhere near the 75 per cent level and zigzags downward toward the zero level, while the broken line starts usually near the 25 per cent level and zigzags upward toward the 100 per cent level. If the snappers had taken the blue fish as frequently in each order as the white, that is, had not distinguished at all between them, the solid and broken lines would coincide with one another and both would run on the 50 per cent level through the entire 10 spaces of each series of trials. If the snappers tended in the successive series of the experiment to take white and blue fish more and more indiscriminately, then in the later series there should be evident an increasing tendency for the broken and solid lines to r 4 m REIGHARD 6. Fig. 5. Two atus coeruleus, and a Chatodon ocellatus in front of branching gorgonian, Plexaura. Fi A ( apistratus over incrusting coral, Maandrina, and in front of large Gorgonian. Color Discrimination in the Gray Snapper. 281 coincide at the 50 per cent level. No such tendency appears in the diagram. Such a tendency of the two lines to approximate each other and the 50 per cent level more in the final series of the experiment than in the earlier series would then be evidence that the snappers rapidly learn the equal palatability of the blue and white atherinas. In only one respect do the lines seem to afford possible evidence that the snappers tend to discriminate less between the white and blue fish toward the end of the day’s experiment. This is in the crossing of the solid and broken lines at a point which is nearer the ordi- nate of percentages in series III and IV than in series I and II. The two lines come together more rapidly in the third and fourth series than in the first and second. The series of July 19, 21, and 22 have also been plotted. The series of July 19 and 21 do not afford the evidence apparently shown in that of July 20 of a decrease of discrimination during a single day, nor does the series of July 22 when compared with those of earlier date show any decrease of discrimination from day to day. If the snappers could be iso- lated from the living atherinas which they have constantly in view, and if they could be fed continuously on blue and white fish in equal numbers, they might soon take the one in each place as frequently as the other, but with the living white fish constantly in sight as part of the natural environment, and with the blue supplied only at intervals and for a short time, this result is not to be expected. Experiment 31: Blue and red color discrimination—On July 21 two series of blue and white discrimination trials were made, the first at 8" 7™ a. m., and the second at 2" 4o™ p. m. Three hundred and sixty atherians were used in addition to those involved in two preliminary feedings of normal fish. The results, as plotted in table 10, series I and II, do not differ from those shown in table 9. Immediately after the second series of 12 trials with blue and white fish there followed a series of 10 trials (in- volving 100 atherinas) in which red fish were substituted for the white, so that, whereas the snappers had before had a choice between white and blue and had become familiar with blue, they now had a choice between red and blue. The red atherinas were prepared as in experiment 2 and had there- fore the same acid taste as the blue. The results are plotted as series II] of the experiment of July 21 and are shown in table 10. The blue fish are represented as before by the broken line, while the red fish are represented by a line of dots and dashes. It is seen that the broken line occupies in series IV the position of the solid line in series I and II. The line of dots and dashes representing the red fish occupies, on the other hand, the position taken in series I and II by the broken line which represents the blue fish. In other words, the blue fish are now taken more frequently than the red in the first four orders. Blue is taken first in 70 per cent of the trials, while red is taken first in but 30 per cent. It is further to be noted that not only is the position of the broken line reversed in series III as compared with 282 Papers from the Marine Biological Laboratory at Tortugas. series I and II, but it first intersects the line of dots and dashes at a point farther from the ordinate of percentages than in series II and III, a further evidence of discrimination. Of the 100 fish thrown in the 10 trials of series III, 21 remained untaken or were eliminated from the experiment through an error of the record. Of the 79 fish taken, 44 were blue and 35 were red; 30 blue fish were taken in the first 5 places as against 20 red, while 14 blue fish were taken in the second 5 places as against 15 red; 16 fish remained untaken at the close of the experiment, and of these 12 were red, while 4 were blue fish to which a small amount of red stain had been accidentally transferred. The diagram (table 10) shows in percentages in each order the TABLE Io. SERIES | SERIES Il July 21, 8) 7™ to S841" A.M. || July 21, 2523 to 2546™ P.M. Blue and White | Blue and White 25 Trials 246 Atherinas * | 12 Trials 120 Atherinas SERIES III July 21, 2446™ to 3 P.M. Blue and Red 10 Trials 100 Atherinas frequency with which a colony of 100 gray snappers took atherinas in each order in three series of trials. The atherinas were thrown in lots of 10. In the first two series 5 white and 5 blue were thrown together; in the third series 5 blue and 5 red were thrown together. Percentages are calculated as in table 9; 21 atherinas remained untaken in the third trial. The plotted data seem to show conclusively that the fish discriminate be- tween the red and blue atherinas. The approximation of the two lines in the second half of their course in series III and their descent toward the zero level is due to the fact that the snappers were at this time no longer hungry, took any but normal fish very slowly, and left many untaken. The behavior of the snappers toward the red and blue fish is further proof of discrimination, as shown by the following extract from my notes: Color Discrimination in the Gray Snapper. 283 The difference in the behavior of the fish toward red and blue was very noticeable. They were from the first distinctly afraid of the red and took them only when they sank so that they could put the snout directly against them as though smelling. Even then they often jerked back, as does a horse under like circumstances, and did not take the fish at all, but left it to be taken by another, or perhaps took it after one or two attempts. The blue fish were all taken at once, and without preliminary “smelling.” They were taken quickly. The red were taken slowly, the snappers often running a little distance with them as though gingerly tasting and then finally swallow- ing them. Such hesitation as the snappers showed in taking the red fish in this experiment is to be expected even toward normal fish at the close of any experiment involving liberal feeding. The evidence of discrimination lies in the very much greater hesitation shown toward the red fish than toward the blue at the beginning of series III. The hesitation itself is merely the normal behavior of a satiated fish toward any new object in the environment, but the degree of hesitation is significant. Since it is possible that the snappers discriminated between the red, white, and blue atherinas by reason of a difference in their luminosity or brightness and without distinguishing the colors themselves, it became neces- sary to determine the relative brightness of the colors employed. The fol- lowing tests were made: (1) The red, white, and blue fish were looked at against a black back- ground in a light so feeble that colors could not be distinguished. Their brightness diminished in the order white, red, blue. (2) Disks of cardboard were colored with the red and blue stains to match the red and blue fish and the brightness of each of these disks was then determined by the same person by comparing it on a color-wheel to a gray produced by blending black and white in known proportions. Ninety- seven degrees of white to 263 degrees of black matched the red in bright- ness, while 37 degrees of white and 323 degrees of black matched the blue. The luminosity of the blue compared to that of the red is therefore ex- pressed by the fraction 37 X 60 + 323 _ 2543 blue 97 X 60+ 263 ~=—-6083, red (3) A spectroscopic examination’ was made by transmitted light of the dyes used and also by reflected light of pieces of cardboard stained to match the atherinas. For the latter purpose the apparatus of Mayer (1897) was used. The red was found to include light of wave-lengths between 675py to 5905p, while the blue included in moderate illumination light of wave- lengths between 590pup and 45tpy. A comparison of these wave-lengths with Koenig’s diagram as given by Howell (1906) shows that the red falls in the most luminous part of the spectrum, while the blue falls in the least luminous part. or approximately 5 to 12. *T am indebted to Dean John O. Reed, Professor of Physics in the University of Michigan, for calibrating the spectroscope used. 284 Papers from the Marine Biological Laboratory at Tortugas. Since in the color-discrimination experiments 30 and 31, the snappers preferred the white to the blue and the blue to the red, they choose in the first case the brighter fish and in the second case those less bright. They were therefore guided not by the brightness of the fish, but by their color. Had time permitted and material been available the experiments of 1905 with fresh atherinas would have been extended so as to include a red-blue series in which the red was the darker color. Other colors would also have been used. DISCUSSION OF COLOR DISCRIMINATION EXPERIMENTS. The colors used in the experiments described were obtained by the use of dyes. They are impure, and I know of no dyes by which pure spectral colors may be obtained. The red used shows with the spectroscope rays of all wave-lengths below the green, while the blue shows waves of all lengths above the yellow. The yellow used is nearly pure, the green includes both yellow and blue rays. We may therefore conclude that the snappers dis- criminate between a mixture of colors of that part of the spectrum which lies below the green and a mixture of the colors of that part which lies above the yellow, and that they discriminate yellow from a mixture of the colors of the part of the spectrum above the yellow. Whether the particu- lar mixtures used appear to the snapper as they do to us, we can not know from the experiments. On the other hand, these mixtures may appear to the snapper precisely as they do to us and his power of color discrimination may be as accurate for all colors as our own. The only paper known to me on color discrimination in fishes is that of Washburn and Bentley (1906), where results were obtained on a single individual of Semotilus atromaculatus. The fish was kept in an aquarium and the methods were those of the labora- tory. The results are in accord with my own. The literature of the subject is discussed by Washburn and Bentley. VI. ESTABLISHMENT IN THE GRAY SNAPPER OF A WARNING COLOR REACTION, INVOLVING AN ASSOCIATION BETWEEN COLOR AND UNPALATABILITY. In the experiments described in this section red atherinas were rendered unpalatable and were then fed to the Laboratory colony of gray snappers, in order to learn whether the snappers would form an association between the unpalatability of the atherinas and their color, of such a sort that they would refuse the atherinas at sight, 7. e., by reason of their color alone. If this should prove to be the case the color red would have come to have for the gray snapper a warning significance experimentally established under normal conditions. In some of the preliminary experiments the red atherinas were rendered unpalatable by the use of substances not normal to the environment of the A Warning-Color Reaction Experimentally Established. 285 snapper, such as quinine and red pepper. So far as the experiments went the snappers gave no evidence that these substances were distasteful. A few attempts to feed to the snappers the tissues of the medusa Cassiopea xamachana showed them to be distinctly unpalatable. However disguised in form or color they were rejected after a few trials, presumably on account of the contained nettle-cells. The red atherinas were then made unpalatable by sewing into the mouth of each the branching tip of a tentacle of this medusa. The tentacle had the appearance of food projecting slightly from the mouth. The atherinas thus prepared were thrown from the dock, one at a time, to the assembled colony of about 150 snappers. An additional atherina was thrown as soon as its predecessor had been taken or as soon as the behavior of the snappers showed that it would not be immediately taken. The snappers thus had at least one atherina always before them. A record was kept of the behavior of the snappers toward each atherina offered. Experiment 32.—Occupied 3 days, July 16, 18, and 19, 1907, but is best regarded as a single experiment. The atherinas used had been preserved in 2 per cent formalin. They were rinsed and then allowed to soak for about 2 hours in a large quantity of sea-water. A few of these when offered to the snappers were taken as though fresh. Tentacles of Cassiopea were sewn into the mouths of others in such a way as to leave the fringed end of the tentacle projecting. The tentacled fish were then stained in fast scarlet in the manner described for experiment 1, until they were of a bril- liant red color. The effect of this treatment upon the nettle-cells was not positively determined, but the result of the experiment indicates that they remained active. Atherinas prepared in this way are referred to as formalin tentacled reds; fish similarly prepared but without tentacles are referred to as for- malin reds; those prepared in formalin and washed but not stained are re- ferred to as formalin normals; while formalin fish provided with tentacle but unstained are referred to as formalin tentacled normals. The behavior of the snappers toward each atherina offered them was recorded under one of the four heads (table 11) ; taken at once indicates that the atherina was seized as soon as it struck the water; taken with hesitation, that 1 to 3 sec- onds passed before the atherina was seized; taken with much hesitation, that more than about 3 seconds passed before the atherina was seized; refused indicates that the atherina remained untaken during the time of the series to which it belonged, or else that after the snappers had had abundant oppor- tunity to take it, the tide carried it away, so that it could no longer be observed. Several minutes were required for an atherina to be thus carried away, and this happened only during the first two series of the experi- ment, in which some of the atherinas floated while others sank. In the third series the atherinas were made to sink by slitting the air bladder so that all that remained untaken were accessible to the snappers during the series and for some time afterward. = Papers from the Marine Biological Laboratory at Tortugas. 286 TABLE II. Series I, July 16, 1907, 1130" a. m. to 1205" p, m.; 35 minutes. “pesny “ay | *u01}8} “1894 yon ‘uoney Teor “uoney “99H "9900 W e------- "SON, [elas -pasny yy LAA AAAAAAZAZAAAAAAA Ilo 120 | woe} *U01}e) -189H HN Ty “2000 WwW "SON, Tepes 60 287 A Warning-Color Reaction Experimentally Established. TasBLe 11—Continued. Series II, July 18, 1907, 1230" to 12'go™ p. m.; 10 minutes. *pasny -2y MOIEG) We ea ys -1S94 Oise 8 (UAL ‘uoney | “19H 140 ‘yore? | “189 R R R R R R R R R R R R R i - Dame ee R 7o Ses ae “aoney -1say Ri ath wn Ba ie “u0ne} AARAAAZAAZARR RRR io) Ree eee jeuag | 2 20 9° se} i -- DAM RRMA iam ee _ 60 Biological Laboratory at Tortugas. ine ers from the Mar Pap 288 TABLE 11—Continued. Series III, July 19, 1907, to*4o™ to r1*30™ a. m.; 50 minutes. *uone} sissy | WW | *u01qe) “89H “uone) ea uw "u0ne | 189 3900 W Dea 80 “money “1894 yonyy “uoney Shs "aou0 W AZAZARRE RRR Meee “SON | [elas | I ° a] T 150 90 | ut) Da [HHH HR HHH 160 100 au & Ho SRR RRR Ree 120 ac 30 teaae an Two lridio bi calinus oceanops, Expanded Spirobraz Gray snapper, Lutfanus Gr attus..< attus,s pery dick, apparently us tricornis wer right hz Fig. 9 ick and white striped. corner clinging to massive coral, Orbicella, Over coral sand near s h grunt beneath one in backg on outer face ¢ A Warning-Color Reaction Experimentally Established. 289 TABLE II, Series I1[J—Continued. = rit = ra 7 1 . | 4 = ri re . Tae a 4 1 ie wal ag \| 3 Wresyse | eee ee | amet [eet Fd eee Falk weet aca) (et | gS| 28| $2) S52! £2| 52] <2) Bo! see | 2 63] <2) es) see | oF We) 6) AS ass | el ae) S| is | ag |e ne] 6 | | eis |e 181/ N | || 201) N 221 | N | N | N , N || al : | N } Ke beset N 1 N | | N 1 N | Pees N 190 N S| WEA ong | ee eee =. ||-230 N eel : N | see | erase] N | s+ ise | a3 | N | a fe eee N | se | ee | S35 | N eee | oG coe. | | N | see] | N = | 2ZO0)| NTI) ea Wigaeea (e220 hese ||penleeeee Tee all an Table 11 shows behavior of the Laboratory colony of gray snappers toward each individual of 3 series of atherinas. N — formalin normal; R — formalin red; T— formalin tentacled red. At the beginning of each series formalin normals were thrown to see whether the snappers were feeding with their usual avidity. These were followed, except in series II, by formalin tentacled reds and these in turn by formalin reds. The last atherinas thrown in each series were formalin normals. Table 11 records the behavior of the snappers toward each individual of three series of atherinas. Here the normal fish are represented by the letter N, the formalin red by R, and the formalin tentacled red by T. Opposite the serial number of each atherina thrown it is represented by its appropriate letter placed in one of four columns; in the first if taken at once, the second if taken with hesitation, the third with much hesitation, fourth re- fused. The record of series I starts with 3 formalin normals all taken at once. These are followed by formalin tentacled reds. Of the first 13 of these, 9 are taken at once, 3 with much hesitation, and 1 refused. As the record continues it is seen that but 1 of these tentacled fish is afterwards taken at once, while the number taken with much hesitation and the num- ber refused both increase toward the end. Of the 15 tentacled fish that were untaken, It are in the second 40 of the 8o tentacled fish thrown. The time required by the snappers to take tentacled fish thus increases as the series lengthens, so that the letters representing these fish shift more and more to the right in the 4 vertical columns. The untentacled red fish (R) follow the tentacled and it is seen that they are all refused or taken with 20 290 ~=Papers from the Marine Biological Laboratory at Tortugas. much hesitation. They are treated like tentacled reds. They are followed in turn by formalin normals and these are all taken at once precisely as at the beginning of the series. In series I] formalin normals offered at the beginning are all taken at once. These are followed by formalin reds, a large number of which are taken at once, more especially in the first half of the series. Nearly all those taken with hesitation are in the second half of the series. Few are taken with great hesitation, and none are refused. Formalin normals offered at the end of the series are all taken at once. No tentacled fish were offered in this series and it has little bearing on the general result. In series III formalin normals and formalin reds were both taken at once at the beginning. These were followed by formalin tentacled reds, only 8 of which were taken at once and these all in the first 15. Those after the first 15 were taken with hesitation, which increased as the series lengthened. All but 2 of the last 50 were either taken with great hesitation or refused. The last 16 were refused. Formalin reds offered immediately after the tentacled reds were all refused, while formalin normals following these were all taken at once. The behavior of the snappers toward red and white fish, which differ from each other only in color, is thus in striking contrast at the beginning and at the end of series III. At the beginning of the series the formalin reds are taken as readily as the formalin normals (white), but at the end, after the snappers have had experience of ten- tacled reds, formalin reds (untentacled) are refused, while formalin normals (white) are still taken. The fact that formalin normals are taken at once at the end of the series shows that any hesitation shown toward formalin tentacled reds earlier in the series is not due to loss of appetite. The snap- pers were hungry throughout each series. My notes record that formalin normals were taken at the end of series I, “and with the greatest avidity.” If either the first or third series could have been carried through on a single snapper, then presumably all the tentacled red fish taken would have fallen at the beginning of the series, the hesitation would then have gradually increased, and all the tentacled fish refused would have fallen at the end of the series. In an experiment dealing with a large number of snappers ten- tacled fish may be taken in any part of the series by snappers that have not yet had experience of them. Nevertheless, as shown in table 11, tentacled fish taken at once fall in the first part of each series. Thus in series I, all but one of the tentacled atherinas recorded as “taken at once” are in the first fourth of the tentacled part of the series. In series IIT, all the tentacled fish recorded as “ taken at once”’ fall in the first tenth of the tentacled fish. On the other hand, tentacled fish “ refused ” fall chiefly near the end of each series. Thus in series I eleven fifteenths of those refused fall in the second half of the part of the series which they compose, while in series III about one-third of those refused fall at the end—a most significant fact. A Warning-Color Reaction Experimentally Established. 291 The three series shown in detail in table 11 are summarized in table 12, to show the total number of atherinas of each sort offered and the total num- ber falling in each category. Series I requires no further comment. Series II has little bearing on the general result. It was introduced to determine whether the experience of the snappers with tentacled reds in series I had been of such a character as to lead them to refuse red atherinas even when not tentacled. This did not yet prove to be the case and all the red fish offered were taken, though less readily than the uncolored formalin normals. TABLE 12. 3 | No. | Taken Taken Taken with | Series and time, Atherinas offered. - at with much Refused.) offered. ins woaee | once. | hesitation, | hesitation, | | =e She Series I, July 16, 1907, | Formalin normals........ Seales Ss fo) fo) fo) 11® 30™ a. m. to 12" | Formalin tentacled reds.| 80 10 28 27 | *15 o5™ p. m.; 35 min- | Formalin reds............. 20 fo) fo) 13 * 7, utes. | Formalin normals........ fo) ° jor] “URS SON Aeanoceosasootodond| (22S tia al |neeacerul ll esses Tel | aessa5 | ssssee¢ —— | Series II, July 18, 1907, | Formalin normals.. fo) | Oo | 122 30™to 12> 40™p, | Formalin reds...... 0 6 ° m.; 70 minutes. | Formalin normals........ fe) fo) UGA G-pndotnapencsuponsod|| HO. || 5g050 MI" ceosen e S]| 560336 oer peaties III, July 19, | Formalin normals....... S| Oe aia LO 1907, 10" 4o™ to 115 | Formalin reds............. | 10 10 fo) fo} fo) 30" a. m.; 50 min- | Formalin tentacled reds, 158 | je 20 85 1@t45 ules. | Hormalinitedsi..cnces..s: Lp A Xo) fo) fo) a7 | Formalin normals......... 30 30 | fo) fo) ante) | UGB sccssaaooposeosann aastel® I) Sesaee I osc este Wt west | * The fish that remained untaken floated; although none of the formalin normals that floated aaa 6; 7; 0» 13, 15: t Including Nos. 143 to 158. § All these fish sank, All the atherinas used in series III were made to sink, so that tentacled fish that had been refused sank slowly through the colony of snappers and lay afterward on the bottom untouched. Seven formalin reds (without ten- tacles) were then offered and all were refused and remained untouched on the bottom. Immediately afterward 30 formalin normals were thrown. My notes record “all were taken as fast as fed, the first few with slight hesita- tion, the others immediately.” The hesitation referred to was so slight that it does not appear in table 11. An association between red and unpalatability thus appears to have been established in the individuals of this colony of 150 snappers. It we exclude series II this has resulted from 178 experiences of swallowing the ten- tacled fish. This is an average of but little more than one experience to each snapper. Series II closed at 11° 30" a.m. At 12" I5™ p. m. a few formalin normals were thrown. The first were taken with a little hesita- tion, the others with a rush and without hesitation. Three formalin ten- tacled reds were then thrown; the first was taken with some hesitation, while 292 Papers from the Marine Biological Laboratory at Tortugas. the second and third were refused. Three formalin reds were then thrown and all refused. The 3 sorts of atherinas, 3 of each, were again offered immediately and in the same order. The formalin normals were taken, the reds, both tentacled and untentacled, were refused. Since the tentacles project slightly from their mouths, formalin tentacled red atherinas differ from normal fish not only in color, but also in form. The snappers might therefore recognize them as unpalatable either by reason of their color or form or by both characteristics. That the snappers actually form an association between the disagreeable quality of the red tentacled atherinas and their color is shown by their refusal to take red atherinas which are not tentacled. As a further test, 20 formalin tentacled atherinas (uncolored) were offered to the laboratory colony of snappers at 1? 15™ p.m. on July 19, 1907. Five formalin normals were first thrown and these were followed by the 20 formalin tentacled normals. All were taken with- out hesitation. Several red fish, some tentacled, others untentacled, were then thrown and remained for some time untaken, but were finally taken. The fact that white fish with tentacles are taken at once is evidence that the discrimination is by color, not by form. That red fish are still taken, though after much hesitation, is probably due in part to their having been offered immediately after a considerable number of formalin normals, in part to the presence in the colony of snappers whose individual experience of the un- palatable red was not yet sufficient to give to the color a warning signifi- cance. Such irregularities are to be expected in dealing in mass experi- ments with an entire colony of snappers. That experiment 32 gave to the color red a warning significance for the snappers, so that red atherinas were afterward for some time protected from their attacks by reason of their color, is shown in the next section of this paper. The details of the behavior of the snappers toward the various sorts of atherinas used is of interest as enforcing the conclusion drawn from the tables and is here abstracted from my notes. (a) Formalin normals at the beginning of the series—The atherinas are taken when they strike the water. All the snappers rush at them and the successful fish snap so vigorously that they send spurts of water above the surface as they strike it with their tails in turning. (b) Formalin tentacled reds—These were taken quickly at first, but in a different manner from the formalin normals. The snappers did not rush at them with so much vigor as to produce a splash at the surface. As the feeding continued the tentacled reds were taken with increasing hesitation. Most of those taken were then approached deliberately. If the approach brought the snapper into contact with the tentacles he at once jerked back. This he sometimes appeared to do without contact with the tentacles. Some- times the snapper jerked back upon contact with the sides of the atherina. This approach and retreat was often several times repeated by the same snapper. If an atherina was taken after being thus approached it was taken gingerly by the side or tail in such a way that the snapper did not come into A Warning-Color Reaction Experimentally Established. 293 contact with the tentacles. It was carried a little way, protruding from the mouth, and then slowly swallowed. After approaching an atherina and jerking back from it one or more times it often happened that a snapper swam away. The atherina might then be approached by many snappers in turn before being finally taken. The time taken by one or more snappers in approaching an atherina and retreating from it is that which appears in the record as “ hesitation.” Early in a series many snappers rushed toward each tentacled red as it was thrown. Later in the series the number of snappers that responded by a rush grew progressively less, until toward the end of the series no snap- pers rushed forward when a tentacled red was thrown (evidence that the splash of the falling atherina is not a sufficient stimulus to cause the rapid approach of the snappers). The atherinas that were taken late in the red series were taken by the few snappers that happened to be nearest the point at which the fish struck the surface of the water. There was no general rush toward that point. The nearest snappers approached slowly and be- haved in the manner already described. In short, the extent to which the individuals of the colony take part in the rush at any atherina appears to be determined by the behavior of those snappers that happen to be nearest the atherina when it strikes the water. If these nearest snappers rush vigor- ously there is general participation by the other individuals of the colony. If the nearest individuals approach more slowly, no such general participa- tion takes place. A localized stimulus applied to one of the higher animals may, if weak, produce only localized response, but may, if strong, throw the entire organism into vigorous response. Similarly the vigor with which a colony of snappers responds to an atherina appears to depend on the extent to which the atherina serves as a stimulus to the individuals nearest to it. (c) Formalin reds following formalin tentacled reds——These were treated precisely like tentacled reds. Probably the snappers would have behaved no differently had all the reds been tentacled. (d) Formalin normals following formalin reds (tentacled or unten- tacled) —The behavior of the snappers is best shown by an extract from my notes on series I. ‘* The change in the behavior of the snappers was most marked. All of them rushed at once at each fish, so that the water fairly boiled. A moment earlier very few snappers had paid any attention to the reds and those that were taken were taken usually by the snapper that hap- pened to be near.”” Sometimes when a change is made from reds to formalin normals the first few of these are taken by the nearest snappers and with slight hesitation. This hesitation rapidly disappears as more normals are thrown, and the rushes quickly become as vigorous as possible and are participated in by the whole colony. The increasing vigor with which the nearest snappers respond to successive atherinas makes itself rapidly felt in the rest of the colony. Experiment 33——As a control on experiment 32, in which the laboratory colony was used, the following experiment was tried on July 19, 1907, at 1 36™ p. m., on the west lighthouse colony, which was without experience of tentacled fish. Atherinas were offered as follows: (1) 12 formalin nor- mals; (2) 44 formalin reds; (3) 30 formalin tentacled reds; (4) 30 forma- lin normals. All were taken at once without hesitation. No difference 294 Papers from the Marine Biological Laboratory at Tortugas. could be observed in the way in which red and white fish were taken by individual snappers or by the colony as a whole. Absence of hesitation in taking tentacled fish is attributable in the beginning to inexperience. That 30 tentacled fish should be so taken is no doubt due to hunger, since this colony had received less food than any other. The hesitation and refusal shown toward red by the laboratory colony is therefore attributable to expe- rience gained during experiment 32. It is neither instinctive nor the result of a previously formed habit. In 1905 an experiment was carried out differing from that described in the present section only in that the atherinas used were fresh, not pre- served in formalin. Excluding some very brief preliminary trials the ex- periment of 1905 extended over but a single day. Between 9° 48™ and 3° 11™ p. m. there were fed to the laboratory colony of about 100 snappers 117 fresh tentacled red atherinas, divided into lots of from 8 to 20, offered at intervals of 30 to 60 minutes; 20 of these, including the last 8, remained untaken, although normal fish were taken readily at all times. The experi- ment of 1907 is an extension and confirmation of that of 1905. The results of these experiments may now be briefly stated as follows: When red atherinas, rendered unpalatable by attaching to each a part of a tentacle of Cassiopea, are offered to a colony of gray snap- pers, they are at first taken instantly, later taken after the lapse of a longer or shorter time, and finally refused. In 1907 (experiment 32) the final refusal resulted (omitting series II) from feeding to about 150 snappers, 238 formalin tentacled red atherinas, of which 178 were taken and 60 refused. The feeding required 1 hour and 25 minutes, divided into 2 periods of 35 and 50 minutes, separated by an in- terval of 3 days. In this time there was an average of but little more than one unpalatable atherina taken for each snapper. This sufficed to form for the snappers an association between red and the quality or qualities which rendered the atherinas unpalatable of such a sort that formalin red atherinas were thereafter refused. Formalin red atherinas were refused even when not tentacled, while formalin normal atherinas (uncolored) were taken whether tentacled or not, hence the association is between color and the quality which renders the atherinas unpalatable, not between the form of the tentacled ather- inas and that quality. In a second colony of snappers formalin red atherinas were taken readily, in considerable numbers (experiment 33) whether tentacled or not, hence their refusal by the first colony can not be attributed to an instinctive or habitual avoidance of red. Their acceptance by the first colony at the beginning of experiment 32 points to the same conclusion. The statement seems therefore warranted that in experiment 32 the color came to have for the snappers a warning meaning. A warning color was artificially established. This result was reached in spite of the fact that atherina is the normal food of the gray snap- per at the time of year at which the experiments were made. By changing the color of this food and rendering it unpalatable the Persistence of Experimental Warning-Color Reaction. 295 natural, positive response of the snappers toward it was inhibited. This positive response, in so far as it involves the taking of small, near, possible-food objects, is doubtless instinctive; in so far as it in- volves taking the specific food atherina it is habitual. Had the food offered been such that the snappers had had no previous experience of it, their positive response would have been instinctive only. To in- hibit such a general instinctive response requires, as shown elsewhere in this paper (p. 307), a much smaller number of experiences than are necessary to inhibit a response that has become habitual. That an habitual response, toward a particular food commonly present in the environment should be as readily inhibited as in experiment 32 was unexpected and shows a high degree of modifiability in the behavior of the snappers. Vil. THE RETENTION OF THE RED-UNPALATABILITY WARNING ASSOCIATION IN THE GRAY SNAPPERS (MEMORY). The color red had come to have a warning significance for the gray snap- pers as a result of their experience with unpalatable red atherinas—an ex- perience which closed on July 19, 1907. This colony was not afterwards offered red atherinas, except at the time indicated below and for the purpose of testing the retention of the red-unpalatability association. On July 23 the colony was fed from 200 to 300 formalin atherinas, newly rinsed in sea- water to remove the formalin. These were taken, but many of them were afterwards ejected and lay on the bottom unnoticed. On July 24 an attempt was made to feed these snappers on formalin blue atherinas, but after tak- ing 30 or 40 they began to eject them and would then take no more. On July 26 the colony took 21 out of 30 blue and yellow formalin atherinas offered them. On July 27, 8 days after the red warning association had been estab- lished, I tried all the morning to get the snappers of this colony to take formalin red atherinas (untentacled). They took a few very slowly and then stopped. They took formalin normals more readily and also took blue, but if red was offered immediately after blue or white, it was, even then, taken very slowly. Apparently the snappers still retained the red unpala- tability association. The red used was a cardinal red—much purer than the red used in the association experiments. On August 8, 20 days after the close of the association experiment with formalin tentacled red atherinas, the last attempt was made to feed the labor- atory colony on red fish. All the red and blue and yellow atherinas used had been removed from formalin, washed and stained some 10 days before, and had since laid in a moist atmosphere, so that the last trace of formalin had probably evaporated from them. Formalin blue atherinas were first thrown and 4 or 5 of these were taken very carefully. Red was then thrown. Some of both blue and red sank, others floated. The reds re- mained entirely untouched, even while they sank among the snappers and 296 Papers from the Marine Biological Laboratory at Tortugas. lay on the bottom. They were not even snapped at or closely examined ; in fact they excited no visible interest. Formalin normals (uncolored) were then taken directly from 2 per cent formalin, rinsed, and thrown while still saturated with formalin. They were taken at once, and about 20 were thus taken. Additional reds and some yellows were then offered, but re- mained practically unnoticed. The stimulus of an immediately preceding feeding on white did not result in the taking of red or yellow. The behavior of the snappers showed quite conclusively their avoidance of red, which was even more marked than immediately after the establish- ment of the red-unpalatability association. At that time and also on July 27 an occasional red fish was taken, but on August 8 they remained not only untaken, but appeared to excite no interest. Although the snappers refused red on August 8 they also refused yel- low, while they took blue less readily than before. The warning mean- ing conveyed thus appears to have been transferred from the red to the yellow, and to a less extent to the blue. Probably atherinas of all colors would have been avoided, and possibly in proportion to the likeness of the respective colors to red. Nevertheless formalin normal atherinas (white) were taken, apparently as readily as ever. The facts stated in this section seem to me to warrant the conclusion that the red-unpalatability association established on July 19 was still effective on August 8.- The associative memory of the snapper has at least this duration. I leit the island on my return north on that day. Vill. RESULTS OF FEEDING THE LABORATORY COLONY OF GRAY SNAPPERS ON CONSPICUOUSLY COLORED CORAL-REEF FISH. In the preceding sections of this paper evidence is given to show that the gray snapper distinguishes colors, forms with great rapidity associations involving color discrimination, and retains these associations for a consid- erable time. It has been further shown that it is possible, in a short time, to establish in the gray snapper a warning-color association, of such a sort that its natural prey is protected from attacks when artificially warningly colored. All this lends support to the theory of warning coloration as applied to coral-reef fishes, yet it remains to be learned by experiment whether any of these fish are actually protected from the attacks of the gray snapper by the assumed combination of conspicuous coloration with un- palatability. To test this assumption as many as possible of the coral-reef fishes were collected and fed to the Laboratory colony of snappers. These fish were usually thrown living from the dock, so that they fell with a splash into the water near the snappers. A few were slipped in quietly near the shore, so that, as they swam seaward, their approach to the snappers was more normal. A few of the fish were dead when offered to the snappers, but most of them were very active and made every effort to escape. A few individuals of Adubefduf marginatus were rendered immobile by pithing. REIGHARD 1p of coral-reef hes in front of branching gorgonians and sea fan, Rhipidogorgta, Fig. 11. Grup of large coral-reef fishes, chiefly Hamulon sciurus. Purple sea fan, Rhipidogorgsa, at left. Conspicuousness of Coral-Reef Fishes. 297 The results of these feeding experiments are shown in table 13, which also includes one amphibian. The fishes fed were, with one exception (Chetodon ocellatus), small enough to be taken by the snappers. Adult individuals of the larger species would have been protected by their size. The salient features of the coloration are given in table 13. For details the reader is referred to systematic treatises (Jordan and Evermann, 1896; Evermann and Marsh, 1900). The color and patterns given in the table are those of the fish seen by the writer in their natural habitat at a little distance, as the fish appear to the snappers. The descriptions have been checked by comparison with captured fish. Preserved fish or those in aquaria (upon which systematic descriptions are usually based) are less brilliantly colored. Of the species listed, only Sparisoma flavescens is un- questionably protectively colored, though Caranx crysos may be so. The others are in varying degrees conspicuous. Jridio bivittatus (plate 4, fig. 7) and the two species of Thalassoma vary considerably. While less conspicu- ous than the others they are to my eye conspicuous. Even casual observa- tion of the remaining species in their natural habitat shows that they are highly conspicuous. Black is a very common color; bright orange and bright yellow are common. Bright metallic blue and green are also common. Scarlet is not rare. All these colors are in strong contrast to those of the reefs. Not only are the colors conspicuous, but they are combined in pat- terns of contrasting colors. Black and white, black and yellow, or black, white, and yellow in alternating bands or stripes are frequent patterns (Abudefduf marginatus, plate 1; Iridio bivittatus, plate 4, fig. 7; Lutianus griseus juv.; Chetodon ocellatus, plate 3, fig. 5; Chetodon capistratus, plate 3, fig. 6; Anisotremus virginicus juv.; Pomacanthus arcuatus juv.; Angelichthys ciliaris; Eques pulcher). Light metallic blue is found alternating in stripes or bands with blue so dark as to be seemingly black (Elacatinus oceanops, plate 4, fig. 8), and with these is sometimes associated a metallic green (Thalassoma bifaciatum). Orange may be combined with metallic blue or black so that each covers uninterruptably about half the surface of the fish (Pomacentrus leucostictus, Pomacentrus planifrons) or yellow and blue may occur in alternating stripes (Hemulon sciurus, Hemulon flavolineatum, plate 2, fig. 3). Scarlet may cover the whole surface or be combined with black (Amia (Apogon) selli- cauda). A uniform black is common (Pomacentrus leucostictus at times, Hepatus hepatius juv., plate 5, fig. 10, at a little distance). The colors and patterns are those typical of warningly-colored insects (cf. Poulton, 1887, for a list of such colors in insects). They make their possessor conspicuous in its normal environment (see plates 1 to 5). When the coral-reef fish are seen by an observer in air he looks through the surface film of the water and sees the fish usually in sharp contrast against the gray-white sand or rock. It seemed to me worth while to find out whether the fish are equally conspicuous when seen by one beneath the sur- 298 Papers from the Marine Biological Laboratory at Tortugas. face of the water, when seen as their enemies see them. They are then viewed at a different angle. They appear at times against the reef or bot- tom, at times against the totally reflecting surface film, and at other times against the blue translucence of the more distant water. I have been able to assure myself of the conspicuousness of the fish under these circumstances by two methods. (1) By the use of a reflecting water-glass,! I have succeeded, without using a diver’s suit, in seeing fish as they appear to other fish. (2) By means of a submerged camera (Reighard, 1908) I have photographed them while they were engaged in their usual activities in their normal en- vironment. Several of these photographs are reproduced in plates 1 to 5, and are sufficiently described in the explanations of the plates. They show certain of these fish as they appear to a submerged observer and seem to me to afford a sufficient demonstration of their conspicuousness. (For further evidence see the plates in Evermann and Marsh, 1902; Jordan and Ever- mann, 1905; Jordan and Seale, 1906; and Saville-Kent, 1893.) Many of these fish are rendered still more conspicuous by their form and movements. The great compression of the body (Chetodon, plate 3; Poma- canthus, Angelichthys, Hepatus, plate 3, fig. 5; plate 5) exposes to view a greater surface which may be further increased by the expansion of the dorsal and anal fins. The movements of the same fish are peculiarly slow and erratic and suggest those of a butterfly on the wing. This peculiarity arises from the use of the pectoral fins rather than the caudal in ordinary progres- sive movements. The caudal appears to be held in reserve for emergencies, as when the fish are forced to flee to shelter. This erratic, jerky method of locomotion by means of the pectorals is not confined to forms with com- pressed bodies and expanded fins, but is found also in all the Labride ob- served (Thalassoma, Iridio, Sparisoma, Scarus, et al.). “This is a rectangular box of galvanized iron, 2 feet long and with ends 6 by 8 inches. In the interior of the box at each end is a mirror firmly fixed in a metallic set- ting and placed at an angle of 45° with the long axis of the box. The reflecting faces of the two mirrors are consequently parallel and they» are directed toward each other. One end is heavily weighted with lead, so that when placed in water the box floats in an upright position with about 10 inches of the upper end projecting above the surface. Opposite the lower mirror is an opening 6 by 8 inches filled with plate glass bedded in aquarium cement. Opposite the upper mirror are 2 tubes soldered to openings in the side of the box and so spaced that the observer may look through them at the upper mirror. The tubes are lined with chamois skin and so constructed that the objective ends of a pair of field glasses may be inserted into them and firmly clamped in place. A handle on either side of the box enables it to be held steady. The observer, while wading, holds the box in front of him with the lower end im- mersed. He may then see objects beneath the surface reflected in the two mirrors, just as they would appear to him if his head were beneath the surface. In using this apparatus I finally dispensed with the field glasses, chiefly because I was able to get so near the fish that their use was unnecessary, but partly because the double images formed by the glass mirrors interfered somewhat with their use. The apparatus would be more efficient if double images were avoided by the use of metallic mirrors. These could be kept from tarnishing by sealing the box hermetically by means of glass plates cemented over the inner ends of the tubes for the eyes. With such an apparatus field glasses could be used and, within the limits set by the opacity of the water, the fish could be studied with them as birds are studied in air. Behavior of Gray Snapper when Fed Conspicuous Fishes. 299 TABLE 13.—Result of feeding 22 species of teleostean fishes and one amphibian to the Laboratory colony of gray snappers in July, 1905, and July, 1907.7 | Name. | Length. | No. How fed Coloration and armature. Result | No | fed. | Inches | | | se | 1 Abudefduf mar-| 1 to2 1x | 4 living, 7 with | Yellow and black banded...... ....... 2 living taken, 3 ginatus, | spinalcordcut, | livingescaped, | | | from dock. 6pithed taken. 2 | Angelichthys| 5 x | Dead from | Body chiefly bright blue, fins Taken at once. (Holocanthus) | dock, chiefly bright yellow, Areopercu- | ciliaris. | | lar spines. | 3 Amia_ sellicau- | 1.5 to 3.5 2 | Living, from | Scarlet, 2 black spots.............s00000 Taken at once, | | da * | dock. | | | 4 | Anisotremusvir-| 1to2 2 | Living, from | Yeliow, with 2 black stripes, black 1 taken after | ginicus. dock. caudal spot. ; short pursuit, | | xrimmediately. | 5 | Caranxcrysos... 2.5 2 | Living, from | Silvery, yellow stripe........ceeesceers 1 taken at once, | shore. 1 escaped, 6 | Chztodon ca- 1.75 x | Living, from | White,biack bands and black spot; | Taken at once. } pistratus. | dock, large dorsal and anal spines. | 7 | Chzetodon ocel- | 4 x | Living, from | Silvery ground, 2 wide black | Attacked, but | latus. dock, bands. Fins yellow in part. | escaped, | | | Large dorsal and anal spines, | | 8 | Chylomycterus 2to4 | 4 | 3living,1dead,| Greenish, heavy black stripes; 2 fishes taken, | schoepfi. | fromdock. | body armed with spines. | 2 and 3 inches | | | long; 2 fishes | escaped, 3and | | 5 inches long. | 9 | Echineis nau- | 6 |, | Living, from | Broad stripes, nearly black and | Taken at once. | | crates. | dock. white. | to | Elacatinus oce- 2 | 2 | Living, from | Broad light and blue-black stripes. 1 escaped, 1 | | anops. | dock. taken at once. | xr | Eques pulcher... 25 | x | Living, from| Black and white conspicuous Taken at once. | | | dock, stripes. | 12 Hemulon sci-| 3.5to4 | 2 | Living, from | Blue and yellow stripes................ | Taken at once | urus. | dock, | 13 Hezmulon flayo- 2.5 2 | xdead,1 living, | Blue and yellow stripes ...............- | Taken at once. | lineatum. from dock. | 14 | Hepatus hepa-| 3 to 4.5 4 Eayaigs from | Black; Zancet on PO SES) Taken at once, tus. ock, 15 _Iridio (Hali-| 3to8 1x | Living, from | Gray, dorsal line and sides with Taken at once. cheeres) bivit- dock. broad dark brown stripes. tatus. | 16 | Lutianus _ gri- 2 3 | Living, from | Yellow-and-black banded.............. Taken at once. | seus, Juv. | | dock | | «7 Pomacentrus | 2.5 to 4 12 | 8dead, 4 living, | Half orange, half blue or black...... Taken at once | leucostictus. from dock. | 18 | Pomacentrus= 2.5 to4 xo | 2dead, 8living, | All black ........ .c.ccseosescsesssecoseosess Taken at once | _leucostictus. | from dock, | 19 | Pomacentrus 2 | 2 Dead, from | Half orange, half blue or black......) Taken at once. | _planifrons. dock. 20 Pomacanthus 1.25 | x | Living, from | Black, 5 conspicuous yellow bands, | Taken slowly. arcuatus. dock, bluespots. Preopercular spine. | 2X Sparisoma fla-| 2.25 | 1 | Living, from | Mottled, olive and brown (protec- | Taken at once. | vescens. | shore. tive colors). 22 Thalassoma bi- 4to6 | 4 | 2living, 2dead, | Blue head, green body, black band | Taken at once. | fasciatum. | | | from dock. between 23 Thalassoma ni- 4to6 13. | gliving, 4 dead, | Very variable, usually green with| Taken at once. tidum | from dock. lateral purple stripe. | | 24 Diemyctilus vir- 4 eerie | sleararie petro rt WER OCe sc eccenncctscansecactatninextaneascncences Takenat once. | | idescens. | | dock. 94 a be = = _—s * 2 living specimens fed to a goby in an aquarium were at once taken. + The authorities for the scientific names in the table are to be found in Jordan (1905). They are here omitted through lack of space. and Thompson Table 13 shows that with few exceptions these conspicuous fish were at Some of them were taken in spite of possess- once taken by the snappers. ing both conspicuousness and unpleasant atributes. Thus Pomacanthus has a formidable spine on the preopercle, while Angelichthys has this and smaller spines besides. and anal. many strong spines scattered over the body. Chetodon has strong, erectile spines at the front of dorsal Chylomycterus inflates itself when disturbed and thus erects the Hepatus is armed with the 300 Papers from the Marine Biological Laboratory at Tortugas. well-known and formidable lancet at the base of the tail (plate 3, fig. 5; plate 5, fig. 10). These are all weapons which the fish habitually use. There is no evidence that the color of any of the species afforded them the least protection against the snappers. These fish, with the exception of Caranx, Chylomycterus, and Leptechineis, are reef-fish. The snappers which frequent the inner reefs are familiar with them; they are familiar with the snappers. That their colors or patterns have no warning signifi- cance for the snappers is shown by the following facts: (a) The snappers took them at once. They showed no trace of the hesi- tation or refusal shown, after experience, toward tentacled red atherinas. (b) The fish attempted to escape, sometimes toward the shore, some- times toward some other near shelter, such as is afforded by the spiles of the dock or by the floating live-boxes beneath it. In this respect their be- havior contrasts with that of many typical, warningly-colored animals (e. g., the skunk and “ Belt’s frog,” Belt, 1874), which do not flee from their ene- mies, but depend for protection on their color. The fish which escaped were either speedy and not conspicuous, or were hindered rather than helped by their conspicuousness. The following details of these escapes illustrate this: (a) Caranx crysos is silvery and yellow, and not conspicuous when seen on the yellow-white sand. One individual, 2.5 inches long, was released 2 or 3 feet from shore. It was pursued by snappers and fled along shore as near as possible to the water’s edge. It was several times captured with a net and again released near the snappers, but each time it behaved in the same way and finally escaped by following the shore. A second individual, thrown from shore to a distance of 6 or 8 feet, was at once taken. (b) Chetodon ocellatus (plate 3, fig. 5).—An individual 5 inches long was thrown from the dock after the snappers had been brought together by feeding them several atherinas. It was at once viciously attacked by a large number of snappers together, and was lost to sight for an instant. As the snappers separated, the Chetodon was seen apparently uninjured and with its spines fully erected. It immediately started seaward and was not further molested. An individual 1.75 inches long, of a rather less conspicuous spe- cies (C. capistratus, plate 3, fig. 6), was at once taken. The C. ocellatus, 4 inches long, was about 4 inches deep from tip of longest dorsal spines to tip of longest anal spine. Such a fish is clearly protected by its size and arma- ture, but not by its coloration, while a smaller individual of a closely related species (C. capistratus), conspicuous and provided with means of defense, is readily taken. (c) Chylomycterus schepfi—(1) A specimen 3 inches long was thrown from the dock. It swam vigorously seaward at the surface, followed by a dozen snappers, which swam about below, approaching and receding. It was finally lost to sight and not taken. (2) An individual about 5 inches long was thrown. It swam toward shore and was picked up and thrown farther seaward. The snappers col- lected beneath it and it inflated itself at sight of them, so far as could be seen without being touched by them. It then deflated itself and swam more rapidly so near the surface that the movements of its pectorals disturbed the Behavior of Gray Snapper when Fed Conspicuous Fishes. 301 surface water. Several times it was seized by the snappers and again re- leased. It could not be seen whether it again inflated itself, but it was finally lost to sight seaward and probably escaped. (3) An individual 3 inches long was thrown outside the dock. The snappers followed it for 5 or 6 rods, swimming below it and watching it, but did not seize it. When it had gone out over the dark vegetation-covered bottom, I followed in a row boat. It was swimming at the surface, making with the pectorals a little eddy visible at a considerable distance. It moved so slowly that a 16-foot dinghy could be brought to it by paddling stern fore- most with a single oar. Nevertheless I had difficulty in capturing it in a small tin can, because, when approached, it changed its course and dodged about. It was finally captured and thrown from the dock on the shore side. It was already inflated. The short tail, projecting like the neck of a jug from the spherical body, was seemingly the only part of the fish that could be seized. After a moment’s hesitation the fish was taken by the tail by an unusually large snapper. He took it to the bottom and mauled it back and forth there as though to force it into his mouth or burst it. He finally lost his hold on it and a second fish seized it by the tail. It passed then to a third and a fourth and each in turn mauled it on the bottom. The fourth fish finally forced it into his mouth and it disappeared. (4) An individual a little less than 3 inches long was thrown dead from the dock and was at once taken. It was carried 5 or 6 rods down the shore, apparently dropped, and then picked up by another fish. A careful search failed to find it, and it was recorded as finally taken. Chylomycterus is an occasional visitor in this region and does not occur normally on the reefs, but is often found in the open water in consider- able numbers. It is conspicuous in open water. On account of its relative unfamiliarity it appeared to offer a new problem to most of the snappers, a problem involving erratic movement, together with inflation of a heavily- spined body. The problem was nevertheless speedily solved. In the case of No. 3 a single large snapper, possibly with previous experience of Chylo- mycterus, led in the solution; the others quickly followed. Conspicuousness was here a disadvantage. (d) Abudefduf marginatus.—This very conspicuous fish is often found with the snappers when full grown. Its yellow and black bands suggest a hornet (see plate 1, figs. 1 and 2). It is deep-bodied, but its dorsal and anal spines are relatively weak (cf. Chetodon). It swims slowly, but is very agile, as shown by the following experiences: Two young individuals, 0.5 inch and 1.5 inches long, were placed in a cylindrical aquarium 7.5 inches in diameter and could then be captured only with great difficulty by the use of a conical net with a 4-inch opening. Two young Lutianus griseus about 1.66 inches long could be readily picked up from the same aquarium with the same net. It is exceedingly difficult to capture Adudefduf with a net 3 feet in diameter in open water. It usually escapes over the rim. Eleven individuals were thrown from the dock to the snappers, 4 of these being alive and active. The results were as follows: (1) Very active, was taken at once. (2) Went about 15 feet pursued by snappers, but finally reached one of the live-boxes and was left unmolested. (3) Was pursued by snappers which came within 8 or 10 inches of it. It swam to a spile underneath the dock and was left unmolested. (4) Swam to the live-box without being pursued. 302 =Papers from the Marine Biological Laboratory at Tortugas. The 3 individuals which escaped joined others of their species which had been for some days living apparently unmolested against the flat bottom and sides of a large, floating live-box. This little group of perhaps a dozen small Abudefduf was within a few feet of 100 or more voracious snap- pers. They were never seen to go more than a foot or two from the surface of the box when feeding. When approached they retreated close to the box, which was smooth on the outside and had no crannies in which they could find shelter. The snappers were never seen to attempt the difficult task of capturing one of these agile little fish against the smooth, hard sur- face of the box. Not their coloration, but nearness to a large, hard sur- face protected them. (5) Was nearly dead when thrown and was at once taken. The remain- ing 6 Abudefduf were made immobile by cutting the spinal cord. (6) and (7) Thrown in the usual way. They were not at first noticed, but as they sank and made no effort to escape, they were seized. (8) (9) (10) and (11) Offered to the snappers in the following man- ner: Bread crumbs were thrown to the living Abudefduf at the side of the live-box. They fed on these at and near the surface, but did not venture to follow them as they sank. Beneath the Abudefduf the snappers assembled to feed on the sinking bread crumbs. The immobile Abudefduf were dropped one at a time past the edge of the live-box, so that they fell among the living ones, which were only 2 or 3 feet above the snappers. As the first immobile (No. 8) Abudefduf sank it was followed for a foot or so by one or two living individuals. The snappers gave no heed to either. The living fish returned to the surface, the motionless one sank, still un- heeded, among the snappers. When it had reached the bottom and lain there for a moment it was quickly seized. (9) (10) (11) were thrown in precisely the same way as (8), and with the same result. Clearly the Abudefduf near the live-box are treated by the snappers as inaccessible, while individuals of the same species at a distance from the live-box are eagerly pursued and often captured. Immobile Abudefduf that sink from the live-box group among the snappers are taken only after a short time, time enough to bring to the snappers a perception of their distinctness from the immune live-box group. (e) Elacatinus oceanops—Vhis slender fish, 1.5 to 4 inches long, is found on the surfaces of the massive coral (Orbicella), close pressed to the living polyps. I have found it fully exposed, not ‘endeavoring to shelter itself in bottom of grooves,’ as stated by Jordan and Thompson (1905). In this position, where it is conspicuous by reason of its strongly contrasted stripes, it is shown in plate 4, fig. 8. When dislodged it takes a zigzag course to a new resting-place, as described by Jordan (1904). With the greatest difficulty it may be captured from the coral with a hand net. - Two individuals were thrown from the dock to the snappers. The first was pursued, but went at once toward a spile beneath the dock. It was lost to view, but later found clinging to the spile. A second individual thrown in further from the dock was at once snapped up. Here, as in the case of Abudefduf, coloration appears to be no protection, but erratic move- ment aids escape, and nearness to a large, hard surface affords protection, more effective, perhaps, in the case of Elacatinus, because of the nettle-cells of the coral polyp to which the fish clings. The case of Leptechineis naucrates is of interest in connection with that Adjustment of the Gray Snapper to Its Food. 303 of Abudefduf. Its broad contrasting stripes render it conspicuous. When in its natural position, attached to the body of a shark by the sucker on the top of its head, it is not molested by the snappers. An individual 6 inches long was thrown from the dock and at once taken by the snappers. Its coloration gave it no protection. Just as the snappers treat Abudefduf and Elacatinus as accessible when not close to a massive coral or large hard sur- face, so they treat Leptechineis as accessible when detached from its host. The wholly unfamiliar bright red salamander Diemyctylus viridescens' was taken with no trace of hesitation, and thus is illustrated again the failure of a warning color to afford protection. The results of the feeding experiments may be briefly summarized as fol- lows: Gray snappers attempted to capture all the 22 species of fish and 1 amphibian thrown to them. They actually took all the species but one (Chetodon ocellatus), which escaped on account of its large size and defensive armor. No hesitation was shown in seizing any of the fish offered, except in the case of the larger individuals of Chylomycterus schepfi, which are formidable by reason of their erratic movements, power of infla- tion, and defensive spines, and are probably new to most of the snappers. The species taken were of a variety of colors and color patterns and were nearly all conspicuous. They included the colors and patterns considered as typically warning. In Angelichthys, Chetodon, Chylomycterus, Hepatus, and Pomacanthus conspicuousness is combined with unpleasant attributes in the form of defensive spines, the typical warning combination, yet these fish were all instantly taken. Individuals of certain species escaped from the pursuing snappers, (a) because of inconspicuousness combined with speed (Caranx crysos); (b) because of erratic movement combined with power of inflation, size, and defensive spines (Chylomycterus schepfi); (c) because of erratic flight to a near object with large, hard surfaces, against which a small agile fish is practically inaccessible (Abudefduf marginatus, Elacatinus oceanops). The escape of these individuals was therefore not due to their conspicuous color- ation, but was rather in spite of it. IX. THE RAPIDITY AND NICETY OF THE ADJUSTMENT OF THE GRAY SNAPPER TO ITS FOOD. In this section there are brought together a few instances of behavior adjustment which appear to bear on the theory of warning coloration. On July 14, 1905, there were fed to the laboratory colony of gray snap- pers 151 atherinas, the last 28 of which were blue. Immediately afterward there were offered 3 blue atherina-shaped pieces of the arms of the medusa Cassiopea xamachana. These pieces induced no reaction whatever in the snappers, although they greatly resembled a blue atherina which was thrown immediately before them and at once taken. When it was clear that the fish would not take the blue cassiopea pieces, they were offered a composite of *T am indebted for this specimen to Mr. Davenport Hooker. 304. Papers from the Marine Biological Laboratory at Tortugas. atherina and cassiopea, made by shaping the cassiopea to the form of an atherina body and sewing to this the head and tail of an atherina. The whole was blue and at a little distance hardly to be distinguished from an atherina. It was at once taken. The next day (July 15) after a previous feeding of but 18 uncolored atherinas, 2 pieces of the rim of cassiopea were stained blue and thrown to the snappers after a single blue atherina. The pieces of cassiopea were about I inch wide and 3 inches long. One continued to pulsate; the other did not. Both were at once taken. Experiment 34: Shortly afterward pieces of the arms of cassiopea were trimmed as near as possible to the size and form of atherina. They were TABLE 14.—Behavior of the laboratory colony of gray snappers toward blue and red atherinas and toward blue and red atherina-shaped cassiopea arms, when offered, in alternation after normal atherinas. ["a=normal atherina; ba=blue atherina; ra—red atherina; bc—=blue cassiopea; rc—red cassiopea.] 7 ] July x5, | Serial | Object | | Jeken Re- | Jul 6,| Serial | Object | ,,., *|| Zaken Re- 1905. No. offered! Taken.| and fused. ee No. | offered. | Taken.| and fused. | | | rejected. | rejected. aie } aA IF | ‘ ly ee I na 2 | na rel PRE 3 | na 3 na | ba pene. | ba ‘ 3 bc 5 be re | 6a E | 6 ba a fee’ bc I, eer be a 8 | ts |] 3 | 8 us & th 19) ve es ee! Ae 2 Io | ba oo 4 | Io ba "Sp Ir | bc 3 II bc 3 12 | ba | s 12 ba 13 bc | a | bc 14 ba ioe 14 | ba 15 be 15 be 16 | ba 16 ba 17 be 17 bc Per t, cassiopea nS Be cent, pea. 19 is | | l i I na ‘Peat na 2 na | 2 na 3 na in 3 na 4 ba | AD Wada ‘ } 5 bc 5 | 5 bc 5 6 ba s 6 ba | a "7 be 4 YUE \ x 5 8 ba | ray 8 ba fe m 9 = |= 9 a x ™ 10 ha a 10 at) | RR rescence | ecm 2 II be bp II bc x is 12 ba ‘5 12 PHN DC MWiccevew ey larstecae = ens bc & 13 re x 5p 5p © 14 ba | @ 14 i PEW. Bi heeebeeeel ante: ane = I be A re | or | Ba ra tes I a | I oN || O<- Weerercecs| beressene } LY, bc 17 re PP SS 18 ba ba = | 18 ra SOSH EE ORS (CCE | | 19 (Tian ee 19 re Sen Per cent, cassiopea. 87.5 | Per cent, cassiopea. | 1 6.25 | 93.25 Adjustment of the Gray Snapper to its Food. 305 smooth, jelly-like masses without tentacles or other projections, and on one surface of each the epithelium with its nettle-cells remained intact. These pieces were stained blue and thrown to the Laboratory colony of gray snappers alternately with blue atherinas, in the manner shown in table 14 at the left. After 3 normal atherinas, 7 blue atherinas and 7 cassiopea pieces were thus thrown at noon. The atherinas were all taken at once; 4 of the cas- siopea pieces were taken but rejected, while 3 remained untaken. These 3 sank among the snappers and remained on the bottom. At 1 30™ p. m. 8 more blue cassiopea pieces were thrown, alternately with blue atherinas and following 3 normal atherinas. The atherinas were again all taken, while 7 of the 8 cassiopea pieces were refused and 1 was taken and then rejected. On the following day (July 16) at 11 a. m. blue cassiopea and blue atherinas were again offered, 8 of each, and with the result that but 1 cassiopea piece was taken and this was afterward rejected (table 14, at right). At times the snappers showed no reaction toward a cassiopea piece. At other times they swam slowly toward the piece as though to take it, but stopped when within a varying distance, never less than 6 or 8 inches, and then turned away. Usually the movement toward the piece was but 2 or 3 inches; often there was no forward movement. The atherinas, on the other hand, were seized with a rush and unerringly at the first rush. The attempt was repeated forty-five minutes later (at 11> 45™ a. m.), but with this dif- ference: that after 4 blue atherinas and 4 blue cassiopea pieces had been thrown, the remaining 4 of each thrown were red (table 14, right, below). Both blue and red atherinas were at once taken, but all the cassiopea pieces were refused. Although none of the cassiopea pieces were taken the snap- pers rushed at the first piece thrown and nearly seized it. They behaved in the same way toward the second and third pieces, but with each succeeding piece their interest lessened, until they paid little attention to the last two or three. In this experiment the atherinas and cassiopea pieces were handled and thrown by different individuals so that there was no transfer of the odor or taste of one to the other. During the 3 days this colony, consisting at this time of about 100 snap- pers, actually swallowed 2 pieces of cassiopea and 2 of cassiopea combined with atherina, while they took into the mouth and rejected 6 other cassiopea pieces. In addition to this they examined, without touching, many of the other pieces of cassiopea. As the result of this brief experience the last 13 cassiopea pieces remained untaken, while the percentage taken and rejected declined from 57 at the beginning through 12.5 to 6.25, as shown in table 14. The colony thus adjusted itself with great rapidity to a new possible-food element, of unfamiliar color and form, but with the familiar unpalatable qualities of the medusa.? In doing this the individual snappers showed con- *That in this adjustment some of the snappers profited by the experience of others is likely. We have here probably a form of imitation—a following instinct— but the data at hand do not warrant a critical discussion of the subject. 21 305 Papers from the Marine Biological Laboratory at Tortugas. siderable power of discrimination. The atherinas and cassiopea pieces were much alike. They were identical in color, but differed in details of form and in translucence, as well as in palatability. In spite of their likeness the snap- pers discriminated accurately between them. That this discrimination was based on form or translucence rather than on color appears from the fact that the snappers were not deceived when red cassiopea was substituted for blue, and from the further fact that they took at once the combined atherina- cassiopea. It is clear that no conspicuous difference, no warning coloration, was necessary to enable the snappers to rapidly differentiate the two sorts of objects. Experiment 35.—On July 23 two small aquarium jars of uncolored ather- inas preserved in formalin were fed to the laboratory colony of snappers. No record was kept of the number fed, but it was estimated to be between 200 and 300—or 2 atherinas to each of the 150 snappers. The atherinas were taken from the 2 per cent formalin, rinsed in sea-water, and thrown at once, while still saturated with the formalin. They were all taken, but very slowly toward the end. Soon snappers were frequently seen swim- ming about with heads of atherinas projecting from their mouths; 15 min- utes later many (probably 50 or 100) atherinas that had been disgorged by the snappers could be seen on the bottom under the dock. On July 24 blue formalin atherinas were offered to this colony. They were taken very slowly, and after 30 or 4o had been taken, they were re- fused. Many were then disgorged. In the midst of this feeding of blue atherinas some perfectly fresh uncolored individuals were offered. Some of these still had enough life to wriggle feebly. These were cautiously ap- proached by the snappers, which often jerked back from them, but finally took them. The living fish, while still wriggling, were thus treated by the snappers in a manner wholly unlike that usual to them when fed on fresh atherinas. It was clear that the snappers had retained since the day before an association between the atherinas colored or uncolored and the dis- agreeable qualities of formalin. Atherinas are inconspicuous when seen against the surface film or the sand bottom, as the snappers see them. They may be regarded as pro- tectively colored. If all the atherinas about the island were suddenly to be- come highly unpalatable, I do not doubt that the snappers would learn after a brief experience to let them alone. I do not doubt that they would be effectively protected by unpalatability alone, without the addition to it of a warning coloration—so rapid and so nice is the power of discrimination in the gray snapper. This ability of the snapper to discriminate with nicety has led to its adjustment to all the food elements of its environment. Frag- ments of a large coral polyp which I threw to them were taken, but at once rejected. Pieces of the arm of a brittle-star were examined and left un- touched. Both were inconspicuous. A palinurus, 3 inches long, was at Significance of Conspicuousness in Coral-Reef Fishes. 307 once taken. It also is inconspicuous. Medusz and Ctenophores of vari- ous sorts, usually inconspicuous, are not eaten. We may say by way of summary that the gray snapper discriminates with great rapidity and delicacy between the various possible-food elements of its environment, which are not conspicuously different from each other. X. GENERAL DISCUSSION OF CONSPICUOUSNESS IN ANIMALS. THE SIGNIFICANCE OF CONSPICUOUS COLORATION IN CORAL-REEF FISHES. If the foregoing account is correct the gray snappers, the commonest predaceous fish of the coral reefs, possess all the qualities required by the theory of warning coloration. They distinguish colors, form associations with great readiness, and retain these associations for a considerable time. Presumably other predaceous fish of the region do the like. By these qualities the snappers adjust themselves continually to their environment. Their capacity for behavior adjustment is indeed so great that a familiar disagreeable quality added to their wonted food is enough to render that food immune from their attacks, and this happens after approximately a single experience of taking the food into the mouth (experiment 32, p. 285). An adjustment even more rapid takes place toward food of unfamiliar ap- pearance, but with well-known unpalatable qualities (atherina-shaped colored cassiopea pieces, experiment 34, p. 305). Their capacity for adjustment is such that the snappers have learned all the food possibilities of the environ- ment—what is good for them to eat and what is not. They refuse at sight jelly-fish and brittle-stars, and in spite of the fact that these forms are very inconspicuous, they distinguish them as not good to eat. A warning colora- tion is quite unnecessary for the protection of these unpalatable forms There can be no doubt that if any animal in the environment of the snappers, whether conspicuous or inconspicuous, should develop highly disagreeable qualities, it would, after a brief experience, be unmolested by them. When, under quite normal conditions, the gray snappers were given an opportunity to feed on conspicuously colored coral-reef fishes of suitable size they took without hesitation all the species offered them. Most of these 22 species are highly conspicuous fish; several have both conspicuousness and formidable means of defense. That they are greedily taken by the gray snappers and that the fish themselves make every effort to escape are facts which seem to admit of no other interpretation than that their conspicuous- ness has no warning significance. We must, then, seek some other meaning of the conspicuousness of these fish. Reasons have been already given (p. 263) for the belief that we are not dealing here with sexual selection. That the conspicuous coral-reef fish are not instances of aggressive re- semblance either general or special is evidenced by their conspicuousness, as shown in the photographs (plates I to 5, except plate 4, fig.9). That an aggressive resemblance (enabling them to approach their prey) is unneces- 308 Papers from the Marine Biological Laboratory at Tortugas. sary is due to the nature of their food, which consists, so far as is known, of invertebrates, most of them fixed. This is seen from their mode of feeding by browsing from the surface of the coral rock or from living corals. It is further shown by the tooth structure of many of the forms, a structure adapted for nibbling from hard surfaces and for crushing hard, fixed forms TABLE 15.—Stomach contents, so far as recorded, of species of coral-reef fishes known to occur at the Tortugas, Florida. | | No. of | Stomach contents. From notes Species examined. | fish ex- | ———__— ——__—_— ——-| by writer and | amined. From Linton, 1907. From Linton, 19074. Dr. Linton, 1907. | | Abudefduf margina- Io Vegetable débris. | tus (=saxatilis). | | Angelichthys (Ho- DTH | wasncccsvemenuntenneseaceeccsnesree ares Alimentary canal filled with a red | locanthus) ciliaris sponge, a few annelids, bryozoa, a small mollusk shell, and sea- | | | weed. | Angelichthys (Ho- t Alimentary canal filled with | locanthus) isabel- | gorgonia, sponges, etc. lita. | | Balistes carolinensis Aad Nenetncsanccuetsecatussecccevstersstaccenss Alimentary canal crowded with | broken mussel shells, setz of large | | annelid. | Balistes vetula ........ x’ «* Lavauadesanvaaveceswebocssavustsencne bane Fragments of adductor muscle of | a bivalve mollusk. Cheetodipterus faber 1 | Very long intestine filled | with material browsed | from the reef, mainly gor- | gonia and sponges. | Chztodon ocellatus | 2 = .| Alge. Hzmulon flavolin- 16 Crustaceans, annelids, green alge | _eatum. | and broken shelis | Hemulon macro- (2) .| Annelids and ophiurians. stomum. | Hzemulon sciurus...) 35 ..| Crustacea and annelids. Hepatus (Teuthis) | 4 .| Broken shells, mainly serpula tubes | coeruleus. | and small gasteropods, bryozoa, | sponge, foraminifera, seaweed | and sand. Hepatus (Teuthis) | MH: lf ecacecssvaccteceacsesccavccassasccoseeccce Ascidian (Botrylloides), algz, and | hepatus. | sand | | Holocentrus ascen- | 4 | ..ssssssaceessssenseevenserenssoseennnenss Small crustaceans. sionis. Iridio bivittatus....... Bia | a caeectcnwcckheesnneaancerconcensscssener Shells and byssus of mussel, anne- lid, spine of sea urchin. Ocyuras schirysurus:|| 1) (xi) | !ocva.cecccetctecncetenicszeveus tatesees| | aadetvcseone eee ae Ree RCE ee ree 1 goose barna- (Bloch). cle, 1 isopod, many annelid spines, Os- tracoda, bi- valve mol- lusks. Pomacanthus arcu- ! 2 -,/Compound as- | * atus. | cidians, red fi- | brous sponge, | fragments) gray calcare- ous sponge, | | green pycno- | | gonid, red al- | | ge, conical | | bryozoan col- | | | _ | onies. [ePomacentras fscrst ieee enue scteceanenetseraaseseescteneeee ses cs Small crustaceans, bryozoa, forami- | nifera, algz, sand. ) Sears i(Callyodon)iin) | ice mull cc-censncssseacenneanceaaangencenancc Stomach and intestine filled with vetula. —— crabs, univalve shelis, sea-urchin | spines, seaweed, and sand | motalinssccccsr ony 108 | | 1 Additional notes on stomach contents of fish from Beaufort, North Carolina, of a few species occurring at Tortugas (Kyphosus sectatrix, Monocanthus hispidus, Chylomycterus schepfi, Chetodip- terus faber) are to be found in Linton (1905). ‘They are confirmatory of those given in this table, but are not included because, although a reef occurs there, the region is not a typical coral-reef region. Significance of Conspicuousness in Coral-Reef Fishes. 309 rather than for holding active prey. The only records of examinations of stomach contents known to the writer are embodied in table 15. This includes 18 species represented by at least 108 individuals and shows the food to in- clude the following fixed or very slow-moving forms: Algze, ascidians, bryo- zoa, sea-urchins, gorgonians, mollusks, sponges. The annelids are in one case sedentary or tube-inhabiting forms and are possibly so in all cases. Of active forms, crustacea are mentioned but four times and ophiurians once; both may have been taken after death. The sand is probably adventitious and the foraminifera may have been included in it. The conclusion seems to be warranted that the food of these coral-reef fishes consists of in- vertebrates, the bulk of it of fixed forms. There appears to be no evidence that any of these aquatic invertebrates discriminate colors (Washburn, 1908, Chapter VII). Aggressive color-resemblance can then afford no advan- tage to fish approaching such prey and selection could not have operated through the food to hold in check the development of their brilliant colors. Protective resemblance seems to have been equally unnecessary for the conspicuous coral-reef fishes. Protected by their agility and their nearness to the coral-rock labyrinths, they readily elude their enemies. The method of escape and subsequent behavior of small Abudefduf marginatus and of Elacatinus related on p. 302 show that these fish are immune from attack when near a large flat surface, whether that surface is or is not clothed with coral polyps. That the conspicuous coral-reef fishes are pursued and captured by the gray snapper when they venture away from the reefs is shown by the feeding experiments already described. One of the fish thus instantly devoured by the snappers was a young Anisotremus virginicus, or porkfish, yet I have twice seen this conspicuous fish emerge from crev- ices in the coral rock and go to a distance of 6 to 8 inches to nibble at the surface of a gray snapper. It avoided the head of the snapper and appeared to be seeking food in the region of the anal opening. Instead of attempting to seize it the snapper lowered the dorsal, wriggled as though annoyed, and then swam away. ‘The instance illustrates the immunity en- joyed by such fish when on the reefs and the recognition by the snapper of their inaccessibility under these circumstances. The gray snappers were never seen to attempt to pursue the conspicuous fish into the recesses of the coral reefs. In the gloom prevailing there the colors of the fish would be indistinguishable and could not then, even should we assume them to be associated with disagreeable qualities, serve to warn their foes. Protective resemblance is unnecessary for such fishes. For them coloration has no selective value. Even were they protectively colored the sharp sight and fine power of discrimination of the snappers would probably enable them to cap- ture the fish at a distance from the reefs, but it is only when they reach a size which renders them immune from attack that they venture to a distance. Since these fishes, if removed from the reefs, would, on account of their 310 Papers from the Marine Biological Laboratory at Tortugas. conspicuousness, be quickly exterminated, it may be objected that their colors are in fact warning colors, that the reefs with their nettling corals furnish the disagreeable qualities which protect the fish and which corre- spond to the unpleasant attributes of warningly-colored insects. It may be urged that, just as insectivorous foes, after attacking warningly-colored insects and experiencing their unpleasant qualities, afterward let them alone, so piscivorous fish, after being bumped or stung in their efforts to capture coral-reef fish, subsequently refrain from such efforts. Coral-reef fish, removed from the reefs, are stripped of their disagreeable attributes and are at once attacked. In like manner warningly-colored insects, if de- prived of their unpleasant attributes, would doubtless be soon attacked and exterminated. Thus, it may be said, the two cases are quite parallel and the conspicuousness of the coral-reef fish is as much an instance of warning coloration as that of warningly-colored insects, with the difference that the association formed by their enemies is in the case of insects with an unpleasant attribute inherent in the insect, and in the case of coral-reef fish with a similar attribute belonging to their environment. There is a meas- ure of truth in this contention, for no doubt the sight of the reefs “ warns” predaceous fish of the futility of pursuing prey into them, and no doubt also the coloration of certain insects may “ warn” their foes of their unpala- tability. That the coloration of unpalatable insects is unnecessary as a warn- ing and that it has therefore not developed under the action of selection are propositions which are discussed elsewhere in this paper. The evidence that the conspicuousness of coral-reef fish does not warn the gray snapper has been already presented, so that it need be here only added that in any pursuit by enemies, whether outside the reefs or within them, the conspicuousness of these fish is a disadvantage to them. Their capture by the gray snapper can serve only to strengthen in him a mode of behavior which would be corre- lated in human consciousness with the proposition—*that gaudy morsel is good to eat.” A gaudiness which serves to advertise palatability is surely in this case disadvantageous. It could not, therefore, have developed through selection, which, had it acted on these fishes with sufficient intensity, should rather have brought about protective coloration. That this result has not been reached may be due in part to the lack of intensity in the selective process, since selection is held in abeyance by the reefs, but more probably results from the inability of the fishes to vary in the necessary direction. If the conspicuousness of the coral-reef fish is not necessary in its court- ship, and does not serve to warn enemies of unpleasant attributes, and does not aid its possessor in eluding enemies or approaching prey, it can, so far as I can see, and whatever be the physiological uses of the chemical sub- stances involved, have no biological meaning. It has arisen not through selec- tion of any sort, but because the conditions of life permit a suspension of Significance of Conspicuousness in Coral-Reef Fishes. 311 selection so far as concerns the color characters of the fish. Selection has neither produced nor perfected the color characters; they have, on the con- trary, arisen in the absence of selection and may be regarded as expressions of the individuality of the species unhampered by selection—as expressions of #&e action of internal factors, possibly orthogenetic. In the reef environment the chemical composition, temperature, and illumination of the water show a high degree of uniformity both sea- sonal and regional (Hickson, quoted by Packard, 1902). In this environ- ment many of the fishes that live habitually in the reefs are highly conspicu- ous on account of their coloration or other characters, while other fishes that live habitually about the reefs but not in them are inconspicuous. Con- spicuous coloration can not, then, be attributed to the influence of light, temperature, or chemical composition of the water, which are the same in the reefs and about them (Hickson). Possibly the food of the reef fishes or un- known external factors tend to stimulate in them the development of a bril- liant coloration, and this might be experimentally tested. That the environ- mental factors peculiar to the reefs do not necessarily produce such colora- tion is shown by Ayphosus sectatrix, a reef-fish in the stomach of which Linton (1905) found crabs, lamellibranchs, and vegetable débris. This fish is dull-colored and inconspicuous, yet is found with the conspicuous fish and has like food. But even should it be shown that the reef environment in- cludes factors which tend to produce brilliant colors the many color-patterns characteristic of species would remain unexplained ; for the number of color- patterns among coral-reef fish is very great, while the environment is one of great uniformity and the food of the species is little varied.1. The character- istic colors and patterns I can regard only as due to internal factors, unchecked by the selection which renders them impossible in the immediate neighbor- hood of the reefs. The reef fauna may include fish conspicuous or not, according to their nature; the water immediately about the reefs can harbor only relatively inconspicuous fish. Nor does it seem to me possible that reef fish could have developed conspicuousness outside the reefs and then sought their shelter (Davenport, 1903, segregation in the fittest environ- ment), for conspicuousness at a distance from the reef shelter would be fatal. On the other hand, inconspicuous fish may have appeared in the reefs and then made their way out from them. Coral-reef fishes are not conspicuous because they are in the reefs; they are in the reefs because they are con- spicuous and can not therefore leave the reefs, and because, being in the reefs and taking the food as they do, there is no reason for their being incon- spicuous. The reefs condition their conspicuousness; they are in no sense its cause. __ * The conditions resemble those found by Gulick (1905) in the Hawaiian Achatinel- lid, but as my own observations are insufficient to warrant the discussion of divergent evolution, I do not here consider Gulick’s work. 312 Papers from the Marine Biological Laboratory at Tortugas. HISTORY OF THE THEORY OF WARNING COLORATION. Darwin, unable to explain by natural selection the conspicuous colors of many animals, suggested the theory of sexual selection, but encountered then a difficulty in certain caterpillars which, though conspicuous, do not show sexual dimorphism. This difficulty he referred to Wallace, who sug- gested (Wallace, 1867) the ingenious theory to which the term “ warning coloration” was later applied. The history of the theory up to the year 1887 is to be found in full in Poulton (1887) and need not be here re- peated. It is also to be gathered in part from the later work of Poulton (1890) as well as from Beddard (1892) and Wallace (1801). Notable contributions of facts in connection with the theory have since been made by Finn (1895, 1896, 1897, 1897a@), by Marshall and Poulton (1902), and by Pritchett (1903). The observations of Finn are of especial importance, since many of them were made on birds at liberty. Originally applied to the immature stages of Lepidoptera, the theory has been ex- tended to imagos of this group and to other groups of insects, notably to Coleoptera, as well as to various other invertebrate groups. It has now been applied to all the groups of vertebrates. Among mammals the skunk was first instanced by Wallace (1891) and is a classical example. Wallace (1891) and Marshall and Poulton (1902) have cited instances among birds and the latter authors have supported their position by experi- mental evidence. The striking colors of certain poisonous snakes are com- monly cited as examples of warning coloration. Among Amphibia “ Mr. Belt’s frog” (Belt, 1874, p. 321), and the European Salamandra maculata are accorded the places of honor, as striking instances of warning colora- tion. Among fishes Garstang (quoted by Poulton, 1890, p. 165, footnote) has suggested the black dorsal fin of Trachinus vipera as warningly col- ored. Wallace (1891, p. 266) has added certain coral-reef fishes. Hick- son (quoted by Packard, 1904) considers the patches of color about the tail spines of certain surgeon-fishes and similar markings in certain trigger- fishes to be warning in function. Bristol (1903) in a very brief note has suggested a classification of the coloration of coral-reef fish based on its biological significance and has assigned a warning meaning to certain types of coloration. The brevity of Bristol’s statement and the absence from it of supporting evidence make it inadvisable to discuss it, but the final paper may well be awaited with great interest. ANALYSIS OF THE THEORY OF WARNING COLORATION. The observations in support of the theory of warning coloration are so largely confined to the group of insects that the discussion may, for the moment, be conveniently restricted to this group. If we separate from other matter the observations upon which the theory rests they are, I think, accurately summarized in the following brief statements: Analysis of the Theory of Warning Coloration. 313 I. (a) Certain insects are readily eaten by insectivorous vertebrates and these are, with few exceptions, inconspicuous (protectively colored). It may be remarked apropos of this statement that it is an obvious neces- sity. A palatable insect without means of defense could not be conspicu- ous and exist where persecuted by vertebrate foes. The palatable forms which remain and are accessible to vertebrate foes are therefore those which are inconspicuous in their natural setting. (b) Other insects are (in one or another stage) either refused by insec- tivorous vertebrates or tasted and then rejected or eaten with more or less evident “ signs of disgust.’ These are, with few exceptions, warningly col- ored. Insectivorous vertebrates learn by experience to avoid them. (See especially Finn, 1897a. ) (c) According to the recent observations of Marshall (Marshall and Poulton, 1902) insectivorous invertebrates (mantids, spiders, dragon flies, etc.) do not refuse conspicuous insects, but usually eat them, although Acreine are usually refused by spiders. Insectivorous invertebrates need not, therefore, be further considered in discussing the theory of warning colors. These statements are based chiefly on feeding experiments in which captive vertebrates have been fed upon insects, but without being offered a choice between conspicuous and inconspicuous forms. The results of different observers are not wholly in accord. Cf. Poulton (1887) and Mar- shall and Poulton (1902), with Beddard (1892), and Pritchett (1903). As Finn (1897a) has pointed out, experiments should be made on vertebrates at liberty and they should be offered a choice. Those observations of Finn that were carried out in this way are in accord with the statements made above under (a) and (b), and are the most conclusive known to me. It is highly desirable that there should be further observations of the same sort as well as studies of the stomach contents of insectivorous vertebrates. Nevertheless, the foregoing statements may be provisionally accepted and it is unlikely that future work will essentially modify them. Upon the observations summarized above certain inferences have been based and these may in turn be conveniently summarized in the following statements : Il. (a) That the conspicuousness of those insects that are provided with means of defense serves not merely to warn insectivorous vertebrates, but is advantageous or necessary for that purpose, as the insects are thereby protected from attack, while their insectivorous foes are saved unpleasant experiences. To conspicuousness is thus assigned an advantageous or neces- sary biological function. (b) That warning coloration has been developed by selection, through the continued destruction by insectivorous vertebrates of the least conspicu- ous of the insects provided with means of defense. These tenets are so obviously a necessary part of the theory of warning 314 Papers from the Marine Biological Laboratory at Tortugas. coloration that they are often implied rather than explicitly stated by writers on the subject. I therefore quote the following detailed statement by Wallace (1891, pp. 242-243) (the italicizing is mine) : But when they (the Heliconidz) first arose from some ancestral species or group which, owing to the food of the larva or some other cause, possessed disagreeable juices that caused them to be disliked by the usual enemies of their kind, they were in all probability not very different either in form or coloration from many other butterflies. They would at that time be subject to repeated attacks by insect eaters, and, even if finally rejected, would often receive a fatal injury. Hence arose the necessity of some distinguishing mark, by which the devourers of butterflies in general might learn that these particular butterflies were uneatable; and every variation lead- ing to such distinction, whether by form, color, or mode of flight, was preserved and accumulated by natural selection, till the ancestral Heliconide became well distin- guished from eatable butterflies, and thenceforth comparatively free from persecution. CRITICISM OF THE THEORY OF WARNING COLORATION. (a) With reference to tenet II (a) above.—Since palatable forms can not be conspicuous and at the same time accessible to their vertebrate foes, it follows that unpalatability must (assuming it to have appeared after in- sectivorous vertebrates) have arisen among inconspicuous insects and these inconspicuous forms must subsequently have become conspicuous. This accords with the statement of Poulton (1890, p. 176): “it must be remem- bered that an unpleasant attribute must always appear in advance of the warning color” and also with the existence at the present time of un- palatable insects that are protectively colored (Mana typica, Beddard, 1892; Manestra persicariae, Beddard, 1892, and many others). That there is no necessary physiological relation between the unpleasant properties of the body juices and conspicuous coloration is shown by the existence of these alleged unpleasant juices in protectively-colored forms, and their ab- sence in certain conspicuous forms. Thus Wallace (1891, p.. 254) says: “The eatable butterflies comprise not only brown or white species, but hun- dreds of Nymphalide, Papilionide, Lyczenidz, etc., which are gaily colored and of an immense variety of patterns.” (See also the table of Beddard, 1892, pp. 164-165.) It is further shown by the coexistence of conspicuous coloration with disagreeable attributes other than unpleasant juices (stings, pricking hairs, Poulton, 1891). The manner in which warning colors are held to have been developed in unpalatable insects is indicated by the above quotation from Wallace. This mode of origin implies the possession by insectivorous vertebrates of a considerable nicety of discrimination, concerning the existence of which we do not appear to have experimental evidence. But the facts of mimicry seem to afford indirect evidence of well-developed powers of discrimination in insectivorous vertebrates. We have no other explanation of mimicry among Lepidoptera than that afforded by the theory of natural selection, and a careful examination of the evidence on which the theory of mimicry Analysis of the Theory of Warning Coloration. 315 rests leaves one impressed with its soundness. The resemblance between mimicked and mimicker have been characterized by Bates as at times “ per- fectly staggering.” Wallace states (1891, p. 245) that “in almost every box of butterflies brought from tropical America by amateurs are to be found some species of the mimicking Pieridae, Erycinidee, or moths, and the mimicked Heliconidz, placed together under the impression that they are the same species. Yet more extraordinary, it sometimes deceives the very insects themselves.” The differences between mimicked and mimickers may afford us a measure then of the power of discrimination of insectivorous vertebrates, for according to the theory those mimics which differed notice- ably from their models have been continually destroyed by their vertebrate foes until the survivors have come to be so like their models that they can no longer be distinguished from them. Any greater differences must attract the attention of these foes. Insectivorous vertebrates thus show a power of discrimination exceeding that of many amateur collectors. That the gray snapper possesses some such power of discrimination may be inferred from the experiments already described. Bateson says that the wrasse “finds a shrimp if the least bit exposed, in spite of its protect- ive coloration” (Poulton, 1890, p. 204). If this power be conceded to fishes, insectivorous reptiles, birds, and mammals may well possess at least equal nicety of discrimination. Recognition marks, if they are recogni- tion marks, afford another instance of discriminative power among verte- brates, while protective and aggressive resemblance are quite generally ac- cepted as having arisen through selection which has depended on the powers of discrimination of vertebrate foes. The theory of warning colors requires us, on the other hand, to assume that these same insectivorous vertebrates possess so little power of discrimi- nation that unpalatable insects need to be marked in conspicuous patterns of contrasting colors. May we not with more reason assert that, if insectivor- ous vertebrates have pushed the resemblance between mimickers and their models to the point of apparent identity, warning colors are for them quite unnecessary? The ordinary specific differences should suffice to “ warn” them of the unpalatability of prospective and familiar insect prey. If this be true warning coloration has no more title to be so called than have other distinctive characters of insects. That it actually serves to “ warn” insec- tivorous vertebrates can not be doubted, but that it is necessary to that end may be seriously questioned. With reference to tenet II (b) above.—lf insectivorous vertebrates pos- sess the nicety of discrimination required by the theories of mimicry and protective resemblance, then it is difficult to see how selection can have produced warning coloration. Selection can no more account for incipient warning coloration than it can account for other incipient characters, for like other characters, warning coloration must have developed to the point 316 Papers from the’ Marine Biological Laboratory at Tortugas. of utility before selection can have operated on it. As a condition ante- cedent to selection, incipient warning coloration must have reached a point at which insectivorous vertebrates could distinguish warningly-colored in- sects from those which did not possess that character. It is conceivable that selection then operated to intensify warning coloration, but can we believe that it carried it to its present stage—far beyond the point at which discrimination must have been easy? Should it not rather have stopped so soon as warningly-colored insects were readily discriminated, long before the present very conspicuous differences had appeared? Are we to believe that when an insectivorous vertebrate encounters together two color varie- ties of a familiar, unpalatable species which differ slightly in conspicuous- ness, it will take the less conspicuous and leave the other? Does it seem still further possible that this result will follow if the two varieties are en- countered in succession? Should the varieties differ from one another in conspicuousness to the extent of mutations, can one believe that a verte- brate foe with experience of the equal unpalatability of both would take the one and leave the other? Yet this must have been the case if warn- ing colors were perfected by selection. Is it not more probable that if any discrimination occurs, the vertebrate foe would attack the more conspicuous insect because it is less familiar and that his attacks would thus tend to retard the development of conspicuousness rather than to accelerate it? If warning coloration may be initiated in insects without the aid of selec- tion, as indeed it must be, and if the later stages in its development may not be satisfactorily accounted for by selection, then we need not invoke its aid at all. The entire development of warning coloration may well have been due to the action of the same forces that initiated it. That these forces may be orthogenetic is obvious, but my own work does not seem to afford a sufficient basis for the discussion of this subject. Mayer (1902) has concluded that in butterflies of the genera Papilio and Ornithoptera and in Hesperide the color-patterns have been mainly determined by internal factors (race tendency), not by external influences or natural selection. A SUBSTITUTE FOR THE THEORY OF WARNING COLORATION. If the views already expressed concerning the development of conspicu- ousness in coral-reef fishes are well founded, this character has resulted from the action of internal forces in the absence of counteracting selection. Selection in the direction of protective resemblance has been held in abey- ance by the coral-reef habitat which has effectually limited the attacks of predaceous fish. Selection in the direction of aggressive resemblance has not taken place owing to the nature of the food. As a result coloration has developed unhampered by selection and this development has resulted in definite colors and patterns, constant and characteristic of each species. Among these are the colors and patterns typical of warningly-colored insects. A Substitute for the Theory of Warning Coloration. Bi What has happened with coral-reef fish may well have happened with insects. With them also, with few exceptions, the nature of the food pre- cludes selection in the direction of aggressive resemblance. The develop- ment of unpalatability bars in large measure the attacks of vertebrate foes and holds in check selection in the direction of protective resemblance. As a result unpalatable insects have, probably with few exceptions, been free to develop color characters, each according to its kind. This development has, in most cases, resulted in conspicuousness, and to this conspicuousness the term “ warning coloration” has been applied. For the term warning coloration it would, it seems to me, be better to substitute the term immunity coloration. If the foregoing account is well founded the so-called warning coloration of insects has not been developed by selection, as hitherto believed, and it is not necessary to insure discrimination by insectivorous vertebrates. The term, moreover, covers only certain classes of conspicuousness, while the term “ immunity colora- tion” covers all cases of conspicuousness not attributable to selection. Im- munity coloration includes on the one hand the conspicuous color characters of coral-reef fish where it has as its conditioning feature inaccessibility; it includes also the conspicuous color-characters of those insects which are, for one reason or another, unsuitable as food for vertebrate foes, and here it has as its conditioning feature wnpalatability (whether due to stings, pricking hairs, disagreeable juices from the body cavity or glands, hard outer covering, etc.). It probably includes cases of so-called warning coloration among invertebrates other than insects (c. g., Coelenterata, nettling nudi- branchs; see Hargitt, 1904) as well as among vertebrates. That immunity coloration may exist without the screen of unpalatability is shown by the following instances: (a) According to Poulton (1887) : Wallace has shown that the shape and colors due to sexual selection run riot in localities (certain islands) where enemies are largely excluded by barriers, and in the same way the brilliant colors of nauseous or dangerous insects may perhaps be ex- plained by equal immunity, although due to other causes. The work of Mayer (1900) throws serious doubt on the inference that colors such as those that have here run riot are sexual colors in the sense that they have a sexual function, and the whole condition is probably the result of immunity. According to Packard (1904) butterflies are not com- monly eaten by birds. Their conspicuousness may therefore be wholly due to immunity. (b) Marshall and Poulton (1902) make the following statement and illustrate it by instances: Swift-flying butterflies are not likely to be caught by birds. The latter learn the futility of pursuit. The butterflies have therefore been able to acquire brilliant colors above, particularly those species having protectively-colored undersides. 318 Papers from the Marine Biological Laboratory at Tortugas. (c) Saville-Kent (1893) describes three species of a huge anemone of the genus Discosoma, which occur on the barrier reef of Australia. One of these species “not infrequently measures as much as 2 feet in diameter.” In the gastral cavity of each species of these anemones is commonly found a small fish of the genus Amphiprion. The three species of Amphiprion are all most conspicuously banded with orange-vermilion and white, but they differ in the number of white bands and in the fact that one of them, A. percula, has the white and vermilion bands everywhere separated by nar- rower black bars. Saville-Kent suggests that— The brilliant colors of the commensal guests attract the notice of other predatory fish, which, hastening to seize an apparently easy prey, are themselves entrapped within the outspread tentacles of the passively expectant sea-anemones. (d) Belt (1874, pp. 196-197) says: Three gaudy macaws were wheeling round and round in playful flight, now show- ing all the red on the under surface, then turning all together, as if they were one body, and showing the gorgeous blue, yellow and red of the upper side gleaming in the sunshine; screaming meanwhile as they flew with harsh, discordant cries. This gaudy-colored and noisy bird seems to proclaim aloud that it fears no foe. Its for midable beak protects it from every danger, for no hawk or predatory mammal dares attack a bird so strongly armed. Here the necessity for concealment does not exist. and sexual selection has had no check in developing the brightest and most conspicu- ous colors. If such a bird was not able to defend itself from all foes, its loud cries would attract them; its bright colors direct them to its own destruction. The white cockatoo of Australia is a similar instance. It is equally conspicuous amongst the dark-green foliage by its pure white color, and equally its loud screams proclaim from afar its resting-place, whilst its powerful beak protects it from all enemies excepting man. In the smaller species of parrots the beak is not sufficiently strong to protect them from their enemies, and most of them are colored green, which makes them very difficult to distinguish amongst the leaves. I have been looking for several minutes at a tree, in which were scores of small green parrots, making an incessant noise, without being able to distinguish one; and I recollect once in Australia firing at what I thought was a solitary “green leek” parrot amongst a bunch of leaves, and to my astonishment five “green leeks” fell to the ground, the whole bunch of appa- rent leaves having been composed of them. Newton (1893-1896) says of the macaw, of which there are about a score of species: “Their food . . . consists of various kinds of fruit... The sexes appear in all cases to be alike in colouring.” We can, then, hardly attribute their colors to sexual selection. The food precludes aggres- sive resemblance. Belt’s account makes it clear that we are not dealing with protective resemblance. Leunis (1883) says of the Psittaci: “Das Fleisch vieler arten gilt ftir zart und wohl-schmeckend,” so that warning coloration on this score seems to be also precluded. (e) Brown (1903, page 297, footnote), writing of the garter snakes of the moist region from latitude 40° in northern California to British Colum- bia, says: A Substitute for the Theory of Warning Coloration. 319 The liberty to indulge in the striking colors developed in the garter snakes of this region is partly due to the protection afforded by abundant vegetation, and perhaps in some degree to the absence of the three snake-eating genera, Spilotes, Ophibolus, and Elaps. All these classes of cases may be brought under the head of immu- nity coloration and may be attributed to inaccessibility. There are doubt- less many other similar cases to be found in the literature and many cases of conspicuousness in insects now attributed to unpalatability may be due to other conditions. That other circumstances than unpalatability and in- accessibility may condition immunity coloration is probable. Immunity coloration may now be defined as “ coloration, not se.rually dimorphic, which renders an organism in its natural environment conspicuous to vertebrates; which has no selective value, since it does not aid the organism in escaping vertebrate enemies by concealment (protective coloration), nor in approaching its accustomed invertebrate prey (aggressive coloration), and when associated with disagreeable qualities is unnecessary as a warning to vertebrate foes of the existence of such qualities (warning coloration) ; it 1s conceived to have arisen through internal forces under immunity of the organ- ism from the action of selection on its color characters.” The exclusion of all sexually dimorphic coloration from the definition is provisional. The obvious relation usually observable between completed animal char- acters and their function or utility is reflected in the curious anthropomor- phic feature of those modern theories of evolution that are founded on a relation between the evolution of these structures and utility. This view- point finds interesting expression in theories of animal coloration. Sexual coloration, warning coloration, and possibly recognition marks may have obvious, though perhaps not necessary, uses; their evolution is therefore assumed to have taken place in relation to these uses. The utmost in- genuity has been exercised to discover plausible utility in every fleck of color with the conviction that thereby evidence was being accumulated as to a mode of evolution. The theory of orthogenesis alone is free from this limitation, since it holds that characters may arise and be without utility or their utility be determined afterward. The present paper holds that immunity coloration has developed in no relation to utility, but it does not discuss the method of that development. The view is presented as a working hypothesis which it is hoped soon to further test. 320 ~=69Papers from the Marine Biological Laboratory at Tortugas. XI. SUMMARY. This paper embodies a search for the biological significance of the con- spicuousness which it attempts to show characterizes many of the coral-reef fish of the Tortugas region. After showing that this conspicuousness is not a secondary sexual character and that it serves neither for protective or aggressive resemblance its value as a warning character is subjected to experimental test. Experimental evidence is presented to show that the gray snapper, the commonest predaceous fish of the region, discriminates certain colors, forms associations with great rapidity, and retains these associations for a consid- erable time (memory). If any of the coral-reef fishes possess a combina- tion of conspicuousness with such unpleasant attributes as render them unpalatable, the gray snapper should have learned to avoid them at sight and their conspicuousness would then have a warning significance. It is shown that when atherina, an inconspicuous fish which serves normally as the food of the gray snapper, is given an artificial warning color and at the same time rendered unpalatable, it is, after a brief experience, no longer taken as food by the gray snapper. Artificially colored atherinas thus come to have a warning significance for the gray snapper and are avoided, even when not unpalatable, although normal atherinas are still readily eaten. The conclu- sion is thus reached that the existence of a warning coloration or of warning conspicuousness in coral fishes is easily possible. This possibility was tested by feeding to gray snappers in their natural environment but at a distance from the reefs, conspicuous coral-reef fishes both living and dead. Of the 21 species thus fed all were taken by the gray snapper, with the exception of one which escaped by reason of the large size of the individual used. It is concluded that these coral-reef fishes do not possess that combination of conspicuousness with unpleasant attributes neces- sary to the theory of warning coloration. Their conspicuousness has no warning significance. Certain further evidence is presented to show that the rapidity and nicety of adjustment of the gray snapper to its food is such-+that it learns to avoid what is unpalatable without the necessity of its being warn- ingly colored or otherwise conspicuous. The conclusion is reached that the conspicuousness of coral-reef fishes, since it is not a secondary sexual character and has no necessary meaning for protection, aggression, or as warning, is without biological significance. The coral-reef fishes have no need of aggressive inconspicuousness because their food consists of invertebrates, chiefly fixed. They have no need of protective inconspicuousness because the reefs and their agility afford them abundant protection. Selection has therefore not acted on their colors or other conspicuous characters, but these have developed in the absence of selection and through internal forces. They are the result of race tendency, unchecked by selection. Summary. 321 An attempt is made to apply this conclusion to the “ warning coloration ” of conspicuous insects. Evidence is collated to show that vertebrate foes are able to discriminate between palatable and unpalatable insects without the aid of a distinguishing conspicuousness. If such power of discrimina- tion exists, then since unpalatability must have preceded conspicuousness in insects, this conspicuousness can not have been initiated by selection. If it began without the aid of selection it may well have continued to develop without it. The conspicuousness of warningly colored insects is then at- tributed to the avoidance of them by vertebrate foes at a time when they were still relatively inconspicuous. This avoidance rendered protective incon- spicuousness unnecessary to them, while the nature of their food rendered aggressive inconspicuousness unnecessary. Their unpleasant attributes have protected them from their enemies, as the reefs have protected the coral-reef fish. Under this immunity from selection they have been free to develop conspicuousness, which is therefore regarded as an expression of race ten- dency, of internal forces, in the absence of selection. The theory of immunity coloration is proposed as a substitute for the theory of warning coloration and it is shown that it covers certain cases not covered by the theory of warning coloration. Immunity coloration is de- fined on page 316. BIBLIOGRAPHY. BeppARD, FRANK E. 1891. Warning Colors. Nature, x_v (1152), 78. 1892. Animal Coloration. London, 8vo, pp. viii 288. 4 colored plates, 36 wood- cuts. Bett, T. 1874. The Naturalist in Nicaragua. London, 8vo, pp. xvi-+ 403, map and 26 text-figs. Bristot, C. L. 1903. On the Color Patterns of Certain Bermuda Fishes. Science, n. s., XVII (430), 492. Brown, A. E. 1903. The Variations of Eutenia in the Pacific Subregion. Proc. Phil. Acad. Nat. Sci., 1903, 286-297. Butter, A. G. 1890. Notes made during the Present Year on the Acceptance and Rejection of Insects by Birds. Ann. and Mag. Nat. Hist., 1890, 324-327. Davenport, C. B. 1903. The Animal Ecology of the Cold Spring Sand Spit. Decennial Publica- tions, University of Chicago, x, 157-176. Distant, W. L. 1891. Warning Colours. Nature, xLv (1156), 175. Ersic, Hueco. 1887. Monographie der Capitelliden des Golfes von Neapel. Fauna u. Flora des Golfes v. Neapel, xv1, 2 vols. xxvii-+ 906; 37 pls. EverMANN, B. W., and Marsu, M. C. 1902. The Fishes of Porto Rico. Bull. U. S. Fish Comm. xx (1902), 49-350, pls. I-49, 3 maps, I12 text-figs. Finn, FRANK. 1895. Contributions to the Theory of Warning Color and Mimickry. Pt. 1. Common Babblers. Jour. As. Soc. Bengal, Lxiv (pt. 11), 344. 1896. Contributions to the Theory of Warning Colors. Pt. 1. Mammals. Jour. As. Soc. Bengal, txv (pt. 11), 42. 1897. Contributions to the Theory of Warning Colors. Pt. 11. Reptiles. Jour. As. Soc. Bengal, Lxv1 (pt. 1), 528. 1897a. Contributions to the Theory of Warning Colors and Mimickry. Pt. tv. Experiments with Various Birds. Summary and conclusions. Journ. As. Soc. Bengal, Lxvi (pt. 1), 614-688. Gutick, JoHN T. 1905. Evolution, Racial and Habitudinal. Carnegie Institution of Washington, Publication No. 25; xii-+ 269, 2 maps, 2 pls. Hareitt, CHARLES W. 1904. Some Unsolved Problems of Organic Adaptation. Proc. Am. Assoc. Ady. Sci., Luz (1903), 511-534. Hickson, S. J. (The animal life on a coral reef, a lecture delivered at the London Institute, known to me only from Packard, 1904.) Howett, Wo. H. 1906. A Text Book of Physiology. Philadelphia, 8vo, 905; 271 figs. Jorpan, D. S. 1904. Notes on Fishes Collected in the Tortugas Archipelago. Bull. U. S. Fish Comm. xxi (1902), 539-544, 2 pls. Jorvan, D. S. and EvermMann, B. W. 1896-1900. The Fishes of North and Middle America. Bull. U. S. Nat. Mus. No. 47, 4 vols., 8vo, 3313, 391 pls. 1905. The Aquatic Resources of the Hawaiian Islands. Part 1. The Shore Fishes. Bull. U. S. Fish Com. xx (1903), pp. xviii 574, pls. 1-Lxxi, colored; pls. 1-65 black and white; 2209 text-figs. Jorpan, D. S. and Seateg, A. 1906. The Fishes of Samoa. Bull. U. S. Bureau of Fisheries xxv, 173-455, pls. XXXIII-LIUI; ITr text-figs. 322 Bibliography. ae Jorpan, D. S. and THompson, J. C. : 1905. The Fish Fauna of the Tortugas Archipelago. Bull. U. S. Bureau of Fisheries xxiv, 229-256, 6 figs. Leunis, J. 1883. Synopsis der Thierkunde. Hannover. 2 vols., 8vo, pp. xiv-1083 and xv- 1231; 1160 text-figs. Linton, Epwin. 1905. Parasites of Fishes of Beaufort, North Carolina. Bull. U. S. Bureau of Fisheries xxIv, 321-428, pls. I-XXXIVv. 1907. Preliminary Report on Animal Parasites collected at Tortugas, Florida, June 30 to July 18, 1906. Carnegie Institution of Washington Year Book No. 5. 1906. 112-117. 1907a. Notes on Parasites of Bermuda Fishes. Proc. U. S. Nat. Mus., xxxIII, 85-126, pls. I-xv. Marsuatt, Guy A. K., and Poutton, E. B. 1902. Five Years’ Observations and Experiments (1896-1901) on the Bionomics of South African Insects, chiefly directed to the Investigation of Mimicry and Warning Colors. Trans. Entom. Soc. London, 1902, 287-541. Maver, A. G. 1897. The Color and Color Patterns of Moths and Butterflies. Proc. Bost. Soc. Nat. Hist., xXxvit, 243-330, pls. I-10. 1900. oy the Mating Instinct in Moths. Ann. and Mag. Nat. Hist. (vm), v, 183-190. 1902. Effects of Natural Selection and Race Tendency upon the Color Patterns of Lepidoptera. Science Bulletin, Mus. Brooklyn Inst. Arts and Sci., 1, (2) 31-86, 28 tables, 2 pls. Me pora, R. 1882. Newron, A. 1893-1896. A Dictionary of Birds. Cheap issue, London, 1 vol., 8vo, pp. x1I-++ 1088, map and many text-figures. Pacxarp, A. S. 1904. The Origin of the Markings of Organisms (Poecilogenesis) due to the Physical rather than the Biological Environment. Proc. Am. Phil. Soc., XLII, 393-450. Poutton, E. B. 1887. The Experimental Proof of the Protective Value of Color and Markings in Insects in Reference to their Vertebrate Enemies. Proc. Zool. Soc. London, 1887, 191-274. 1890. The Colors of Animals. London, 8vo, xv + 360, I pl., 66 text-figs. 1891. Warning Colors. Nature, XLv, 174-175, 2 figs. PritcuHetTT, ANNIE H 1903. Some Experiments in Feeding Lizards with Protectively Colored Insects. Biol. Bull. v, 271-287. REIGHARD, JACoB. 1908. The Photography of Aquatic Animals in their Natural Environment. Bull. U. S. Bureau of Fisheries xxvit (1907), 41-68, plate 1-v, 9 text-figs. SaviLLE-KEntT, W. 1893. The Great Barrier Reef of Australia: Its Products and Potentialities. London, pp. xviii 387. Photo-mesotype plates I-XLviI, chromo plates I-xVI. SEMPER, CARL. 188r. Animal Life. New York, International Sci. Ser., 8vo, XvI-+ 472, 2 maps, 106 woodcuts. Wattace, A. R. 1867. (Theory of Warning Coloration.) Trans. Ent. Soc. Lond., ser. 3, vol. v, 80-81 (proc.). 1875. Contributions to the Theory of Natural Selection. London, 8vo, pp. xvi + 384. 1891. Darwinism. London. 8vo, xvi- 494, frontispiece, map, 37 text-figs. Wasueurn, M. F. 1908. The Animal Mind. New York, 8vo, pp. x + 333, 18 text-figs. Wasupsurn, M. F., and Benttey, M. B. 1906. The Establishment of an Association involving Color Discrimination in the Creek Chub, Semotilus atromaculatus. Jour. Comp. Neur. and Psych., XVI, 113-125. EXPLANATION OF PLATES. [All figures are reproductions of instantaneous photographs made in the natural environment with a submerged camera on orthochromatic plates with the aid of a color screen.] PLATE I. Fic. 1. Rock ledges near the beach, showing 3 Abudefduf marginatus banded with black and yellow-white, and 2 grunts (Hemulon sp., apparently sciurus) striped with blue and yellow. A spherical “head” of living coral (Porites) in the middle background. Note the upper Abudefduf swim- ming by the use of the pectorals. When these fish are seen endwise against the surface film the bands blend and the fish are then almost invisible. Fic. 2. Abudefduf marginatus seen in crevasses of a reef of massive coral. Note everywhere over the reef the mantle of living coral polyps (Orbicella). PLATE 2. Fic. 3. At least three Hemulon flavolineatum, French grunt, among coral rocks and Millepora at the base of a reef. The stripes are silvery lavender-blue and chrome yellow. At the right a large snapper (apparently Lutianus apodus) conspicuously banded in black or brown and yellow. Fic. 4. A small labroid fish of an unidentified species feeding on the rock bottom; at its left a branching gorgonian (Plexaura). PLATE 3. Fic. 5. Two blue surgeons (Hepatus cereuleus) in front of a branching gorgonian (Plexaura) on bottom rock. Note the lancet in the caudal peduncle of the one at the right. In front of the surgeons a Chetodon ocellatus, the tail of which moved during exposure. The entire dorsal, anal, pelvics and caudal peduncle are spectral yellow. Fic. 6. A Chetodon capistratus over an incrusting coral (Maandrina) on bottom rock and in front of large gorgonian on which expanded polyps may be seen immediately over the fish. The fish is nearly gray, with much yellow on snout and fins. PLATE 4. Fic. 7. In the foreground Jridio bivittatus (slippery Dick) feeding on bottom. The stripes at a little distance appear black and white. In the background a second Jridio bivitattus above, and a Hemulon flavolineatum or French grunt below. Fic. 8. Massive coral (Orbicella) on the outer face of a reef. In the lower right- hand corner an Elacatinus oceanops clinging to the coral. The stripes are very light and very dark blue and appear nearly black and white. At the left an expanded tube-inhabiting annelid (Spirobranchus tricornis). Fic. 9. The gray snapper (Lutianus griseus), as it appears over the coral sand near the shore; an instance of aggressive resemblance. 324 Explanation of Plates. 325 PLATE 5. Fic. 10. A group of large coral-reef fishes over a rock cleft in front of branching gorgonians. At the left two surgeons (Hepatus hepatus). The lancet is visible in the one at the right. A third surgeon at right of middle below also shows the lancet. These fish are practically black. In the center a French grunt (Hemulon flavolineatum), striped with yellow and blue. At the left of this a blue parrot-fish (Callyodon ceruleus) ; at its right a green parrot-fish (probably Callyodon vetula). Beneath the green parrot- fish a mottled parrot (Sparisoma?). Above the grunt a second mottled parrot and to the left of this a third. Behind the green parrot a purple sea-fan (Rhipidogorgia). Note the position of the pectoral in the par rots and the left-hand surgeon. Fic. 11. A group of large coral-reef fish, chiefly yellow and blue grunts, mostly Hemu- lon sciurus. Among these at the left one or two gray snappers (Lutianus griseus). At the right are purple sea-fans (Rhipidogorgia) and to the left of these two large parrot fish. In front of the lower parrot fish a surgeon (Hepatus hepatus). A second surgeon farther to the left shows the characteristic mode of locomotion by the pectoral fins. This is seen also in the lower parrot. The camera was pointed toward the light, so that the markings of the fish are not as distinct as they would other- wise be. a ge Ay i. Mahe, |} I = ee al AA apna a Ana 4% s5.a% M AGRaag, arse 7eee 4 ; anyhghs* 4, Aa aan An an Aan Aa- aa eo AO DQ eR MA 7 f acitaeko 0 Maas an* 2 pannerii A. ‘ a La 1a 40 nr. A shunt AP papa ® r A B- i eigen) ag i-- By! 8 Tre \, é panne i] , PTY Y Wl Yale nat pe {hy a AAR a A ners ernel Ane, Nal 7 43 a”;," malt la 2 mae 2aGn Oa, aeeeN per Obed LTC TTY Wil 11} Lo Liy Saat a CHOI Oh ee caer CLL eae TPA TAYALL TR ane Q,, “hoa TNR AMAR zases Tl aaeet «mF aL Ra 5 - ' ¢ a Sealte ‘ Re . A _ —- ‘2 5 * re i: a4 Armaan, TE Fr: Se ate.e Aa at ‘+5? -\1 D oo. “" of J aA, AAA “ARAL “, So uatanana, Sef; Be Geeess a) “ Saar” Pt ih oA, TYR eface j AR r TT A - am JAA 5 >) a a ast a udp / : A 8 pm “\ UR& a Saye ar . CN Pe ier deh mal ’, Ls eae 1 pete rtnnuaaenr uly aA sane ‘ae Rian sR J “uy : . - aa Ne ann. 7" a | " ‘Ra, a*® fz a a em Pas "yy AA 7 ry A ata Gass colette 1) ine PRATT iysantiel r Rsdrnan 1 an rot - aN - : lt i Ay | nw Nah Ay. aN . gaa lt" A> ‘if ih a Biko ¥ a gue Vaal Rga ‘ ‘, Po” Waa pat, eee f wi ry rr Pe Ll ded 5" Ae Bs TY) b f- A ageh PAaQaaeu" A TH « a. yi - { Tee Lg "Y > Bate oer A 2 z oA : ARLE Ramer e ey Waa ane Meaytata aa | ie aaa AA +99. 054 J aeM ARAL RQ bias jh gyadn, iy SAA Nua stAQBannya”are.n, wate an a Prtga th 7 A canteen binges eA nai EEL PTTYY FUP AR > A Wn agAamenm ., rv a Ne | 2 PA LY RAF | |) ai Neer > { U rt ig aA RA < rte pamsnpn~ URS neon 2 28) A) |] 5 9 ag . “a A ill a o *. Joan '~ + ; . pera Aime a8 | vn wasn aaaid 95 A, THT | ahaa per sss la Raul + [! ce ss 2cyianan tiny,” ie A bee roe Ol aaa DArxn st . _ > ree 5 a ‘<) ae. ‘aCh Pee eeeaeael ae = aArA aaD T Ane ah 0 ean TLE Ras we Ana, A ah Nos aan ALS) ae , yi oo b | = a 1 Ga. 6 SAAS lotebal dE. NA \ Ry i\ i UT dedhene ID : Neate! DARAARAS ot APAr*aey ea T3377 retenentte |. “9 Antero. | AnOPT AM, al ‘sb Mrs, Aaa ") Ren ARAM. Baan neg ryt “Afana._.--*7 **” WAR e pr uihtiNRn Selene TTT SE sb Pre att Re hae AR de Qa SAA Rohan ee >’ i] Rena tadaaeD | altel | aah -naatananaAballle | ‘ 5 ao NANA ar =m saa r aye tach iol ae: Star de ala Ap mamas 9 we oan Ae JA Al Dp = & rr nN & ope AA 5 a z we 4 A, a thd ty cipk yanks - bit Dy aad. ae =P “ amAAAAAS Seeceeentteeoee | ieee ‘,apata ai s r rent Rm yes Th || HHT | A ah hee a antebettth aula Pen 1% 5 Aga, ry 2... 0a8 pant “ Na! | ae RL | My tt ALLE conpenottnet” BP gaan aa, A id Aer =a cs} aAai*, 44 4g a er olay ial oh 4 ™ lh, ~. - TARA Ree Th onl P ae : ee, RA Ay - oflase bee onamit ae neme oe ms San. SERA a, “NUK Aay ahaa aia aA? 7 5 | mast mine j Lol ¥ ey alr « x ‘ & Na at aA ' Ne 4 2 «Ore aa ale j AA asaahs Pee rheseec ahha’: 2220 Grae, | a a Paha Pee alse Aa NyAtnss lee WV anesant! une pe AltAdastenuai 1 ii AEN as ae a = | aS AA RSS ARE x ob | aa s A shat) pita uae BAA AAAs Parner an sasebayassaes syiae ghd “a> hg a <. Baggs lane a r yyy «Sip yp HBA NAY tener aan Nab . Sua wr Ahi agus “aah : Bp Wlyy : vA iol fed Leh nA jaa RARSN AP 1a WOR a‘ | AAAS AAA a anRaaAn. ~ SA, y, jad eS 4a maaiee yy Aina, 2 P* S smaaubnAg inp Nha jtaay AL a AAA, A AaNNA; SANNA riche he ATP NRAEN SBARL OSG SLID: AA DRone oan. Aga Arta ,vsalta sabt on ay. APPAR SUTTTTRMRARADI I AG ‘4 “RASA, esa ay . Apeeeauee (PPADS ~~ haan nd Nm ol SO os “ipgsRAAe AY UUAEAER Tha, 4 Dtrenerons fe- > anaasga ResBnRy ganna m0 lath | Aa i) eo a ; ty A Lael : Ah. “gait aga °. ee aS = e blty Fidel 11 ii pAna Rial t sant Lan 7 Anan AOOAIARAA : j rryy outage tan mii ty THLE Pe ~ any hh) eg! Fe gases fee'6 arimahenct HATTA IY 4 van PLN OO one atin “Ae aX | ee ns na ! Miaas, aarp ~~ Sac eae® be Tord s Wa AM aauas aie \- a2 =" pr Aaranrdr a Apatioth - Ree ; AS Pies Aaa VASA aarsAA a. i, Pema alr Ni eee poh MMOD ih ibrar opin Se --A 5 a5 none Ay. ~ Paces ; was 9! SS saa nnne2s09-77 THD A), ja me ry Wea. : pa ya09) ASS A aye anbayrvettha, eDeibe. bh hes MATTE LH NqgiP* WAR ane Aan MAAR eaaana, | ARagih TN! La HY Se aiaae “4 ranean Beata TMM HA gy vee . 4 UN& BNs4a8 apannnrAnn