cf* ^ Wv-rf 0p% gr-: ^r^4# ►* (mm °< v \\$sr y %// ^ A „sP W'5'1 g5j^^ fSc T Ml v '^i.!1 O, \ ©;,; % j\ > vr * ^ m^m °' , ^ fev-a&M iSSSW^ ^€#5 |pft||, iASf .Jv Cm I A ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME 46 • 1961 • NUMBERS 1-16 PUBLISHED BY THE SOCIETY The ZOOLOGICAL PARK, New York NEW YORK ZOOLOGICAL SOCIETY GENERAL OFFICE 30 East Fortieth Street, New York 16, N. Y. PUBLICATION OFFICE The Zoological Park, New York 60, N. Y. OFFICERS PRESIDENT VICE-PRESIDENT SECRETARY TREASURER Fairfield Osborn Laurance S. Rockefeller George W. Merck David H. McAlpin SCIENTIFIC STAFF: John Tee-Van General Director William G. Conway. . Director, Zoological Park Christopher W. Coates . . Director, Aquarium ZOOLOGICAL PARK Joseph A. Davis, Jr. . . Associate Curator, Mammals Grace Davall Assistant Curator, Mammals and Birds William G. Conway. . Curator, Birds Herndon G. Dowling. Curator, Reptiles Charles P. Gandal. . . Veterinarian Lee S. Crandall General Curator, Emeritus William Beebe Honorary Curator, Birds AQUARIUM James W. Atz Curator Carleton Ray Associate Curator Ross F. Nigrelli Pathologist & Chair- man of Department of Marine Biochem- istry & Ecology C. M. Breder, Jr Research Associate in Ichthyology Harry A. Charipper. . . Research Associate in Histology Sophie Jakowska Research Associate in Experimental Biology Klaus D. Kallman. . . . Research Associate in Genetics Louis Mowbray Research Associate in Field Biology GENERAL William Bridges . . Editor & Curator, Publications Dorothy Reville . . Editorial Assistant Sam Dunton Photographer Henry M. Lester. . Photographic Consultant DEPARTMENT OF TROPICAL RESEARCH William Beebe Director Emeritus Jocelyn Crane Assistant Director David W. Snow Resident Naturalist John Tee-Van Associate William K. Gregory .... Associate AFFILIATE L. Floyd Clarke Director, Jackson Hole Biological Research Station EDITORIAL COMMITTEE Fairfield Osborn, Chairman James W. Atz William G. Conway William Beebe Lee S. Crandall William Bridges Herndon G. Dowling Christopher W. Coates John Tee- Van Contents ✓ • 4^ Part 1. April 28, 1961 PAGE 1 . A Study of the Biology and Behavior of the Caterpillars, Pupae and Emerg- ing Butterflies of the Subfamily Heliconiinae in Trinidad, West Indies. Part I. Some Aspects of Larval Behavior. By Anne J. Alexander. Plate I; Text-figures 1-8 1 2. Hybridization Experiments in Rhodeine Fishes (Cyprinidae, Teleostei). An Intergenetic Hybrid between Female Rhodeus ocellatus and Male Acanthorhodeus atremius. By J. J. Duyvene de Wit. Plate 1 25 3. The Natural History of the Oilbird, Steatornis caripensis, in Trinidad, W.I. Part 1. General Behavior and Breeding Habits. By D. W. Snow. Plates I & II; Text-figures 1-6 27 4. Fatty Degeneration, Regenerative Hyperplasia and Neoplasia in the Livers of Rainbow Trout, Salmo gairdneri. By Ross F. Nigrelli & Sophie Jakowska. Plates I-VI 49 Part 2. September 25, 1961 5. Morphological Effects of Low Temperatures during the Embryonic De- velopment of the Garter Snake, Thamnophis elegans. By Wade Fox, Charles Gordon & Marjorie H. Fox. Text-figures 1-4 57 6. The Orang-utan in Sarawak. By George B. Schaller. Text-figure 1.. . 73 7. Observations on the Feeding, Shedding and Growth Rates of Captive Snakes (Boidae). By A. J. Barton & William B. Allen, Jr 83 8. The Feeding Mechanism of Fiddler Crabs, with Ecological Considerations of Feeding Adaptations. By Don Curtis Miller. Plate I; Text-figure 1. 89 9. Hybridization Experiments in Rhodeine Fishes (Cyprinidae, Teleostei). Intergeneric Hybrids Obtained from Acheilognathus lanceolata X Rhodeus amarus and Rhodeus amarus X Acheilognathus tabira. By J. J. Duyvene de Wit. Plate 1 101 10. Some Observations on the Metamorphosis of the Frog Rana curtipes Jerdon. By Lucy Lobo. Plate I; Text-figure 1 103 Part 3. November 24, 1961 11. A Study of the Biology and Behavior of the Caterpillars, Pupae and Emerg- ing Butterflies of the Subfamily Heliconiinae in Trinidad, West Indies. Part II. Molting, and the Behavior of Pupae and Emerging Adults. By Anne J. Alexander. Plate I; Text-figures 1-3 105 12. Melanoma, Renal Thyroid Tumor and Reticulo-endothelial Hyperplasia in a Non-hybrid Platyfish. By Pamela A. Mac Intyre & K. France Baker-Cohen. Plates I & II; Text-figure 1 125 13. Eastern Pacific Expeditions of the New York Zoological Society. XLV. Non-intertidal Brachygnathous Crabs from the West Coast of Tropical America. Part 2: Brachygnatha Brachyrhyncha. By John S. Garth. Text-figures 1&2 133 14. Nematodes and Cestodes from the Australian Monitor, Varanus gouldii. By Horace W. Stunkard & Charles P. Gandal. Text-figures 1-6 161 15. Urinary Amino Acids of Non-human Primates. By Jack Fooden. Plates I-III; Text-figures 1-4 167 Part 4. December 15, 1961 16. The Role of the Thyroid in the Development of the Platyfish. By K. France Baker-Cohen. Plates I-IX; Text-figures 1-4 181 Index to Volume 46 223 $40. 5 73 ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME 46 • PART 1 • APRIL 28, 1961 • NUMBERS 1 TO 4 PUBLISHED BY THE SOCIETY The ZOOLOGICAL PARK, New York Contents PAGE 1 . A Study of the Biology and Behavior of the Caterpillars, Pupae and Emerg- ing Butterflies of the Subfamily Heliconiinae in Trinidad, West Indies. Part I. Some Aspects of Larval Behavior. By Anne J. Alexander. Plate I; Text-figures 1-8 1 2. Hybridization Experiments in Rhodeine Fishes (Cyprinidae, Teleostei). An Intergenetic Hybrid between Female Rhodeus ocellatus and Male Acanthorhodeus atremius. By J. J. Duyvene de Wit. Plate 1 25 3. The Natural History of the Oilbird, Steatornis caripensis, in Trinidad, W.I. Part 1. General Behavior and Breeding Habits. By D. W. Snow. Plates I & II; Text-figures 1-6 27 4. Fatty Degeneration, Regenerative Hyperplasia and Neoplasia in the Livers of Rainbow Trout, Salmo gairdneri. By Ross F. Nigrelli & Sophie Jakowska. Plates I-VI 49 1 A Study of the Biology and Behavior of the Caterpillars, Pupae and Emerging Butterflies of the Subfamily Heliconiinae in Trinidad, West Indies. Part I. Some Aspects of Larval Behavior1,2 Anne J. Alexander Zoology Department, Rhodes University, Grahamstown, South Africa (Plate I; Text-figures 1-8) [This paper is one of a series emanating from the Tropical Field Station of the New York Zoological Society, at Simla, Arima Valley, Trinidad, West Indies. This station was founded in 1950 by the Zoological Society’s Department of Tropical Re- search, under the direction of Dr. William Beebe. It comprises 200 acres in the middle of the Northern Range, which includes large stretches of undisturbed government forest reserves. The laboratory of the Station is intended for research in tropical ecology and in animal behavior. The altitude of the research area is 500 to 1,800 feet, and the annual rainfall is more than 100 inches. [For further ecological details of meteorology and biotic zones, see “Introduction to the Ecology of the Arima Valley, Trinidad, B.W.I.,” William Beebe, Zoologica, 1952, 37 (13): 157-184. [The success of the present study is a large meas- ure of the cooperation of the staff at Simla, espe- cially of Jocelyn Crane and Constance Carter, the former contributing much of her knowledge of the animals, the latter helping with recording of obser- vations]. I. Introduction 1 II. Material and Methods 2 III. Feeding Rhythms 3 IV. Feeding 5 a. Leaf Patterns 7 b. Furrowing Behavior 10 c. Channeling Behavior 11 d. State of Abandoned Leaves 12 e. Feeding Positions 12 1 Contribution No. 1,007, Department of Tropical Research, New York Zoological Society. 2 This study has been aided by a grant from the National Science Foundation, G6376. Thanks are also due to the Royal Commission for the Exhibition of 1851 and to the South African Council for Scientific and Industrial Research for financial support. f. Eating Actions 13 g. Eating of Egg-shells 14 h. Eating of Cast Skins 14 i. Drinking 14 V. Defecation 15 VI. Resting 15 VI. Resting 15 a. Resting Position 16 b. Resting Posture 17 VII. Weaving 17 VIII. Locomotion 19 IX. Social Behavior 19 X. Defensive Behavior 20 XI. Phylogenetic Discussion 21 XII. Summary 24 XIII. References 24 I. Introduction The phylogeny of the neotropical subfam- ily of butterflies, the Heliconiinae, is interesting, partly because most species are distasteful and aposematically colored, partly because of the possibility of Mullerian mim- icry. The determination of relationships within the group must clearly utilize information of be- havior, physiology and ecology of the butter- flies as well as their anatomy. Indeed, work has already been going on along former lines (see the study of comparative ethology of the adults, Crane, 1955 and 1957) and that of breeding experiments and wing patterns (Beebe, 1955). It is obvious, however, that all such studies must relate not only to the adult stages but also to egg, larva and pupa. The present paper repre- sents an attempt to study comparatively the lar- val behavior of as many species of heliconiines as were available over a period of four and one- half months spent in the laboratory at Simla. Because of limitations of material and time, 1 2 Zoologica: New York Zoological Society [46: 1 the project was necessarily of a preliminary nature and there are many unavoidable gaps in information. The significance of some late larval behavior was clearly to be sought in the pupa and possibly the emergent adult; limited observations were thus made on these stages as well. These, together with information on molting of the larvae, will be presented in Part II. Implications of the phylogeny of the species observed are discussed tentatively but in rela- tion to this study only, and it is hoped that a later paper will correlate evidence from the several fields and workers on the subjects of relationships among the Heliconiinae. For descriptions and illustrations of the ex- ternal characteristics of the larvae to be dis- cussed, see Beebe, Crane & Fleming (1960) and Fleming (1960). II. Materials and Methods Of the 14 species of heliconiines known to occur in Trinidad, three were not available at all during the time of this study and any refer- ence to their behavior is from the notebooks of the Simla staff. These species are Heliconius wallacei wallacei Reakirt, Heliconius numata ethilla Godart and Philaethria dido dido Clerck. Of the others, observations on Heliconius doris doris (Linnaeus) were limited to two days, while only a single specimen of Dryadula phaetusa phaetusa (Linnaeus) was obtained. Dione juno juno (Cramer) is gregarious and a single group of 37 healthy individuals was ob- served; these were, however, already in their second instar when found. Lastly, information on Heliconius sara thamar Hiibner and, to a lesser extent Heliconius erato hydara Hewitson, was limited, as healthy stocks were unobtain- able for much of the period. Observations on the emergence of H. sara are due entirely to Constance Carter to whom I owe many thanks. The remaining six species consist of Heliconius melpomene euryades Riffarth, Heliconius ricini insulana Stichel, Heliconius aliphera aliphera ( Godart ) , Heliconius isabella isabella ( Cramer) , Dryas iulia iulia (Fabricius) and Agraulis van- illae vanillae (Linnaeus). After initial observations of caterpillars on single leaves in glass jars, it was found more satisfactory to keep the larvae on lengths of vine, one to three feet long, the ends of which were thrust into narrow-necked bottles of water. Fresh vines were added to the bottles every few days and the caterpillars were free to move onto them, or were very occasionally transferred by hand. In such conditions the caterpillars re- mained on the vines, apparently content. With the exception of the two species mentioned earlier, all larvae were very healthy. Well- formed butterflies emerged from the pupae and individuals of those species whose normal court- ship is known behaved as would be expected of healthy adults. Most of the observations of larval behavior were made on vines in an isolated room on the fringe of the forest around Simla. Conditions of light, humidity and wind were therefore close to those of the natural habitat of at least some of the species outside (see Beebe, 1952, for de- tails of ecological conditions in the Arima Val- ley). Late larval or prepupal behavior was watched outside the laboratory among rows of Passiflora vines planted by Simla staff. Where the host vine was not available in the vine rows, pre- pupational larvae were kept on extra large vines stuck into bottles in the laboratory. Observations of feeding patterns and choice of pupational site were also made in the field where possible. In no such case was any discrepancy found be- tween these observations and those made under laboratory conditions. At night caterpillars were watched beneath either a red or very dim white light. Individual animals were identified by peculiarities of an- atomy or color or, in cases where they were very similar, by keeping them on separate vines. Numerous mirrors had to be used to render the activities of the caterpillars visible without dis- turbing them. III. Feeding Rhythms It has generally been recognized that most, if not at all, lepidopterous larvae show rhythm in their behavior patterns. Thus Crowell (1943) reported that a large number of species have feeding periods of about 20-30 minutes alterna- ting with rest periods of similar duration. The caterpillars of at least ten species of heliconiines from Trinidad are no exception (see Text-fig. 1 for examples). As Ford (1945) has recorded for caterpillars such as the Silver-studded Blue ( Plebejus argus ) and the Black Hairstreak ( Strymonidia pruni), most of these species show rhythmic bursts of feeding evenly throughout both day and night. An examination of the activity patterns sug- gests that there are specific differences in the details of this rhythm, but the picture may be blurred by changes in the pattern during the course of an instar. Thus the duration of feeding periods immediately after a larval molt is shorter than of those immediately before (Text-fig. 2) . Again, the rhythm differs between instars, while just before pupation there are long, almost un- interrupted periods of feeding. 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 3 Text-fig. 1 . Feeding rhythm in caterpillars the day after their fourth molt. Thick straight lines = feeding; thin straight lines = resting phases; thin wavy lines = walking or other movement. The time intervals on the upper line = 30 minutes, a, D. iulia; b, H. isabella; e, H. aliphera; d, H. melpomene: ©, H. erato. i i i j i i ! i i i t A — — - BMBBBBMi ■ — ■ . fgj 1 — — ... — . . - — — — goggfrjaga ■ ■ -■* Text-fig. 2. The feeding and resting periods in a caterpillar of H. melpomene: a, the day before the fourth molt; b, immediately after eating the cast fourth exuvium; c, the following morning. The convention used is as in Text-fig. 1. The line drawn down through the first feeding period of 2b indicates an interruption. i ! I ! I '—*V» I •Mi Text-fig. 3. Synchronization of feeding and resting in a group of four H. doris caterpillars. While the majority of the species feed through- out the 24 hours, H. doris does not normally feed at night, a habit which Ford (1945) re- cords for the Dark Green Fritillary ( Argynnis aglaia ) and the Swallow-tail ( Papilio machaon) . Furthermore, two species, D. iulia and H. mel- pomene3, show a tendency, during their fifth instar, to rest throughout the day4. Ford has recorded a similar change of feed- ing habit in the last larval stage of the Scotch Argus ( Erebia aethiops ) which feeds only dur- ing the night at this period, although earlier it was not so restricted. 3 From the laboratory notebook of Barbara Young (1957), it seems probable that P. dido, a species which I have not observed, is similar in this respect to D. iulia and H. melpomene. The characteristic may, however, appear earlier in the life of the caterpillar, possibly by the third instar. 4 Using the amount of dung deposited during a cer- tain time as a measure of the intensity of feeding, it seems that, in these latter cases, light may be a direct inhibitory stimulus. Thus a caterpillar of D. iulia, if kept in darkness during the day, eats more than twice as much as it does in daylight, producing 2.4 pellets per hour, as opposed to 0.9. In naturally gregarious species, such as D. juno, H. doris, H. sara and to a lesser extent FI. ricini, the caterpillars synchronize their feeding and resting periods (Text-fig. 3). The beginning of the feeding period is somewhat less strictly coordinated than its end; this re- flects the possibility that synchronization is con- trolled by at least two factors. An internally controlled rhythm could initi- ate feeding periods. In caterpillars which have been deprived of food or for some other reason are not eating, e. g., the nocturnal phase of H. doris, there are indications of a persistence of rhythmicity expressed as alternating periods of rest and locomotor activity. If indeed the loco- motor activity corresponds to the feeding activ- ity, it will follow that the onset of the activity period is determined not simply by a reflex due to lack of food, as suggested by Crowell (1943), but by some endogenous pattern. The rest periods are usually induced by factors which affect all the animals simultaneously, possibly such extraneous stimuli as a sudden wind, the passing of an ant or the touch of an observer. The last-mentioned stimulus has in fact been noticed to end the feeding periods in various 4 Zoologica: New York Zoological Society [46: 1 Table I. Relationship between Larvae of the Subfamily Heliconiinae and of the Family Passifloraceae in Trinidad Plants Key: * — Host plant. -f— Will accept if put on plant. — — Refuses if put on. ( *) — Very occasionally found on. ( * — )— Eggs found on it but larvae refused it. (S.T) — South Trinidad. Species of Passiflora Species auriculata HBK vespertilio Linnaeus tuberosa Jacquin rubra Linnaeus quadriglandulosa Rodschied serrato-digitata Linnaeus laurifolia Linnaeus lonchophora Linnaeus foetida Linnaeus 1. Dione juno + - + * (*-) + 2. Agraulis vanillae + + + (*) * 3. Dryadula phaetusa + (*) + - (*-) 4. Dryas iulia + * + _ _ — 5. Philaethria dido * 6. Heliconius isabella + *(S.T) * 7. Heliconius aliphera + - * ? *(S.T) + * + 8. Heliconius melpomene + (*) + * + 9. Heliconius numata * 10. Heliconius erato + *(S.T) * (*) + 11. Heliconius ricini * 12. Heliconius sara * + + 13. Heliconius wallacei * 14. Heliconius doris * (*) species. If this stimulus is given early in the feeding period, it will have no effect or merely occasion a momentary halt. Towards the natural end of the period, however, it usually causes premature resting. These considerations may also apply to the habits of the non-gregarious caterpillars, for these tend to fall into a syn- chronized rhythm of eating and resting when kept on the same or neighboring leaves— a phenomenon which has been observed especially in H. aliphera and H. melpomene. Regardless of how such synchronous eating is controlled, any selective advantage of such behavior would surely come from the limit which it sets to the time during which any cat- erpillars are moving. Movement of the prey is 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 5 Table II. Vegetative Characteristics of the Vines on which the Caterpillars Feed Species Vine Size Texture of the Stem Leaf Texture 1 Covering of Leaf Abundance of Leaves Leaf Shape Passiflora auriculata Slight, Slender, maximum 10 , smooth climbing Tender- medium Smooth Fairly scarce Simple, entire Pas si) lor a rubra Medium, bushy, climbing Medium, branches more than other vines Medium Medium Abundant Widely bi-lobed Passiflora tuberosa Slight, effervescent Slender Tender- medium Mat Scarce Bi-lobed, narrow-wide Passiflora quadrigranaulosa Climbing, reaches height of 30' Tough- medium Tough- medium Rough- mat Fairly scarce Bi-lobed, unequal Passiflora serrato-digitata Medium, sprawling Smooth, sturdy Tender- medium Smooth Fairly abundant 5-7- palmate Passiflora laurifolia Climbing, reaches up to 30' Tough, frequent branching Tough- medium Smooth Abundant Simple, entire Passiflora foetida Sprawling, medium Tough- medium, hairy Medium Hairy & glandular Fairly abundant Tri- foliate Passiflora lonchophora Slender, climbing up to 40' Tough Tender- medium Smooth Fairly abundant Tri- foliate important to insect and reptilian predators and, furthermore, would assist in calling the atten- tion of any bird to a caterpillar. Thus with the exception of “agonistic” movements when disturbed, the caterpillars would do well to re- main as motionless as possible during non-feed- ing periods. Without synchronization of feed- ing movements, some animals could always attract the attention of predators to the resting caterpillars, as well as to themselves. IV. Feeding Although heliconiine imagines feed on nectar from a variety of flowers, their larvae feed only on various species of passion vines (Family Passifloraceae). Usually feeding is confined to the leaves, but in some cases tendrils, stalks, flowers and hairs are also eaten. Ten species of vine growing in Trinidad have been found to support one or more of 14 species of heli- coniine caterpillars (see Table I). These vines vary fairly widely in their vegetative character- istics (see Table II) and are, moreover, some- what variable within each species. One factor which determines which species of vine a caterpillar eats is, of course, the fairly high specificity in respect to oviposition site shown by the heliconiines; the eggs are almost invariably laid on the “natural” food-plant of the species. Occasional “mistakes” are, however, found in the field (see Table I). Tolerance of an “unnatural” host vine may alter according to what a caterpillar has eaten previously, the stage and number of its instar and possibly its water load immediately before a test. Thus H. aliphera raised to the fifth in- star on P. lonchophora accepts P. rubra almost immediately and with no change of its previous patterns of eating and resting. The converse is not true. H. aliphera, raised on P. rubra to the equivalent stage and then transferred to P. lon- chophora, wanders and rests without eating for several hours before beginning to feed. In this species there is a tendency for later instars to show less specificity in their food preferences than the earlier ones. The matter has not been investigated in detail among other species, but tests made on the day before pupation indicate that host preference is then less strict. Sometimes when a butterfly oviposits on an abnormal species of vine, the caterpillars refuse to eat altogether, e.g., D. juno laying on P. laurifolia (Table I). In other cases they may accept the abnormal food plant but their growth may be reduced and/or retarded. Text-fig. 4 shows the latter effect on a group of D. juno fed on leaves of P. auriculata and P. rubra as opposed to one fed on the natural food-plant (P. serrato-digitata) . The number of instars is greater, molting is more protracted and most growth periods are longer. It has not been established whether they actually eat less in 6 Zoologica: New York Zoological Society [46: 1 > i i i < i i e i i i i i i i i i i i i i i i i i i i Text-fig. 4. The effect of food plant on instar number and length in D. juno. Growth and feeding are indicated by the black lines, molting by the white lines. Crosses indicate the time of hanging up and tri- angles the shedding of the last larval skin. Time interval = 24 hours, and the arrow indicates when the second group were collected as they molted into the second instar, a, Four caterpillars on P. auriculata and P. rubra: b, 13 caterpillars on P. serrato-digitata, their normal food plant. Information for the upper picture was kindly supplied by Jocelyn Crane. unit time under the unfavorable conditions. There are, moreover, behavioral changes in such cases; for instance the synchronization of feed- ing is lost. This may reflect a more basic physi- ological disturbance, for the synchronization of molting and pupation also disappears. It is not clear how the caterpillars distinguish between a vine which they accept and one they reject. Some rejections occur after the material has actually been chewed. Thus H. ricini refused to eat P. tuberosa after tasting it and rejected an aquaeous extract of P. tuberosa leaves, though it would drink similar extracts made from its food-plant P. laurifolia. In other cases caterpillars can certainly make a distinction without having to taste the material. Thus neither H. isabella nor D. iulia will attempt to eat P. foetida; in fact the former cannot be persuaded to remain on this vine, commonly dropping off as soon as it is put on a P. foetida leaf. It seems possible therefore that in this spe- cies of caterpillar, behavior is controlled by more than one factor, in keeping with Dethier’s (1937) findings on larvae of the gypsy moth. Such an effect might explain conclusions such as those of Merz (1959) who declares that among the larvae of monophagous species of Lepidoptera many are less specialized than had been previously supposed. The technique she used in testing the larvae was to present them Table III. Information Relating to Crowding of Eggs and the Site where They Are Normally Laid by the Heliconhne Butterfly Species Gregariousness Leaf Size Leaf Surface Tendril Dione juno 60 /90 in raft almost touching Medium (but tender) Under Agraulis vanillae Single Leaves indiscriminately Upper Dryadula phaetusa Single Medium Dryas iulia Single Medium Upper Fresh, dry Philaethria dido Single ? ? Thick Heliconius isabella Single or few scattered Medium-large Under Heliconius aliphera Scattered 5 /6 or single Medium-large Under Heliconius melpomene Single Subterminal leaflets, medium Upper Young Heliconius erato Single Subterminal leaflets, medium leaves Upper Heliconius ricini 4/12 loose cluster Among leaf buds Heliconius sara About 25 in tight cluster Among leaf buds Heliconius wallacei 25 /30 fairly loose cluster Among leaf buds Heliconius doris 36 /52 in raft almost touching Medium leaf Upper 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 7 Text-fig. 5. Patterns left on leaves by the feeding of caterpillars, the stippled part being that eaten. The caterpillars concerned are shown in their typical resting positions and postures on the appropriate leaves, a, H. Isabella on P. laurifolia; b, H. melpomene on P. laurifolia; c, H. ricini on P, laurifolia, chewed to give the ragged effect rather than the straight across, shown for H. melpomene and often produced by H. ricini; d, D. iulia on P. tuberosa; e, A. vanillae on P. foetida. with dried leaves moistened with sugar water. Regardless of additional complications of the possible attraction of sugar water itself, the physical properties of the leaf are vastly changed by desiccation. a. Leaf Patterns Since the butterflies further show consistent preferences for oviposition in particular sites on the vine (Table III), the food first accepted by the caterpillars will be affected by this choice, for the newly-hatched animals usually eat the food nearest to them. Thus H. isabella and H. aliphera begin eating the undersurface of a leaf, although they will feed when placed on its upper surface, a position in which they are never found in the field. Again, a high per- centage of H. melpomene start by eating a tendril. However, of a batch of 20 caterpillars from eggs laid on tendrils and offered a choice of tendrils and young leaves, 55% ate leaves, disregarding the tendrils after they had inves- tigated them. Thus, in certain cases at least, the feeding sites of first instar larvae are de- termined simply by the normal oviposition site and where, as with H. melpomene, some selec- tion might be exercised, this cannot find ex- pression. Besides showing specific preferences for a single or various species of vine, the larvae leave characteristic patterns (Text-fig. 5) on material which they have been eating, a phenomenon which is well known in many phytophagous in- sects (see Bering, 1926). The following features are among those dis- tinguishing different patterns in the heliconiine species of this study. 1. The green cells alone may be scraped away from the surface of a leaf, leaving only a layer of transparent epidermal cells on the far side. This is in contrast to an actual hole being chewed in the leaf or its being eaten away from a margin so that bays or channels are left. 2. The midrib of the leaf or leaflet may re- main when the caterpillar abandons the leaf, or it may be eaten together with the blade of the leaf or, in complex patterns, it may be ignored while the caterpillar eats most of the blade but is finally eaten before the leaf is abandoned. 3. Whether it is finally eaten or not, the midrib may also be the focus for other specific attentions. H. melpomene, H. ricini and, in special cases to be mentioned later, H. erato, will chew a small chunk out of the ventral midrib of the leaf on which they are feeding. Occasionally there may be two such “furrows” chewed across the same midrib but usually a single one is cut between the base of the leaf and the level at which the caterpillar rests between feeding periods. On the other hand the midrib may be involved in complicated channeling and 8 Zoologica: New York Zoological Society [46: 1 Table IV. Feeding Behavior of Ten Species of Heliconune Caterpillars (Information relating to Philaethria dido supplied by Constance Carter) Key: -| — Eaten. Species Scraping, holes, margin chewing Midrib eaten, eaten later, or left Midrib treatment Leaf margins eaten Leaf tips eaten Amount of leaf abandoned Dione juno Scraping in first instar at least Left ? + Dropped Frill left on leaf, no petiole eaten Agraulis vanillae Holes chewed Eaten later or ignored + + Frill left on leaf, no petiole eaten Dryadula phaetusa 1st instar, 2nd, chews channels in from margin Eaten later Channeling and bridging + + Frill around base of leaf or petiole chewed Dryas iulia Chews channels in from margin even in first instar Eaten later Channeling and bridging + + Often frill around base of leaf, but more usually petiole also eaten, sometimes stem as well Philaethria dido Chews channels in from margin even in first instar 7 ? ? ? ? Heliconius isabella Scraping 1st, 2nd instar Left, occasionally eaten later + Dropped Base of leaf, midrib and tip are left Heliconius aliphera Scraping 1st, 2nd instar Sometimes eaten later, usually not + Sometimes definitely discarded Always base of leaf is left Heliconius melpomene Very occasional scraping, usually channels or bays chewed from margin Eaten together with blade Chews fur- rows across under surface + + Vi leaf or frill around base is left Heliconius erato Never holes or scraping but channels or bays from margin Eaten together with blade May chew furrow in petiole + + All leaf, and frequently petiole and often part of stem, is eaten Heliconius ricini Never scraping, but holes or channels and bays from margin Eaten, often left by groups of larvae May chew furrows across under surface Often left, especially when eating as a group + About half leaf if larvae is solitary, if in group may eat even stem Heliconius sara As in Heliconius ricini Eaten May chew furrows across under surface If larvae in a group, no leaf left; even stem chewed Heliconius doris Chewing from margin ? ? 7 ? ? bridging behavior shown by D. iulia and D. phae- tusa. 4. The margins of a leaf may be left by H. ricini as tattered edges or they may be eaten with the rest of the leaf. 5. Leaf tips may be regularly eaten, as by H. ricini and H. melpomene; if the tips are not eaten they may be allowed either to fall to the ground or to remain with the midrib. 6. When a caterpillar moves over to a new leaf, as much as half the old leaf may be left, only a small stub may remain or the caterpillar may have eaten it entirely. Some species ( e . g., D. iulia and H. erato ) normally eat the petiole, while H. erato eats the stem as well. Table IV shows the distribution of these six characteristics among the species studied. Many of these specific features appear to have no very obvious significance in the lives of the caterpillars. Some are extremely consist- ent and appear to be endogenously differenti- ated while others occur only under certain conditions and it seems that environmental fac- 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 9 tors are at least partly responsible for their appearance. The case of H. aliphera and H. Isabella will be considered first. These scrape the cells from the surface of leaves instead of chewing through as do other larvae which eat the same host vines, P. lonchophora and P. laurifolia respectively. The scraping is seen only in the first, second and to some extent third instars of both species, which gives the im- pression that it occurs when the animals are too small to do otherwise. The same idea is also gained from the fact that H. Isabella occurs on the tougher leaves of P. laurifolia while young H. melpomene and H. ricini, which feed on the same vine and do not use the scraping method, are limited to thin and tender leaves. Moreover, on several occasions a first instar H. melpomene did scrape for a short while when it was on a tougher leaf than is normal for this species. Finally, it is possible to induce a third instar H. Isabella , which has been scrap- ing at a tough leaf, to eat holes by putting it on a tender one, though it reverts to scraping when replaced on the tough leaf. These observations suggest that the scraping habit of H. Isabella is determined by purely mechanical considerations. This cannot be the full explanation, however, for the first and second instar H. Isabella still scrape when put on tender leaves through which H. melpomene will chew. Further H. aliphera scrapes on the softest of P. lonchophora leaves. Nor can the effect be attributed in any simple way to the differences which occur in the sculpturing of the mandibles for H. aliphera and H. isabella are capable of chewing through both the upper and lower surface of a leaf at which they never- theless only scape if left on one side. Thus it does not seem that these caterpillars eat only the one epidermal layer because they are in- capable of dealing with the other. Scraping would rather seem to be, at least in part, an inherited pattern in H. aliphera and H. isabella, although it is lost later and can be modified by external circumstances. It is probably present in the early instars of some species such as H. melpomene but is not expressed in the condi- tions in which these normally live. Eating or leaving the midrib is another char- acteristic for which there seems to be a simple, mechanical explanation. Although both H. isa- bella and H. melpomene feed on P. laurifolia leaves of similar texture, the appearances of the leaves left by the two species is quite different. H. isabella leaves the tip and the midrib entire and only the blade on either side is chewed away. (Text-fig. 5a). H. melpomene eats the tip of the leaf and straight across the blade, midrib included. (Text-fig. 5b). If H. aliphera is made to eat the leaves of P. laurifolia, its feeding pattern is the same as that of the H. isabella. it could be suggested that H. melpomene has more effective mandibles than H. isabella or H. aliphera. Indeed, an individual of H. isa- bella or H. aliphera does leave small veins pro- jecting along a margin where it has been eating, as if those parts which are slightly fibrous are less readily taken. Further, when H. isabella feeds on P. serrato-digitata, with a more succu- lent and tender leaf, it does not invariably re- ject the midribs. Another species which rejects the midrib, at least in the third, fourth and early fifth instars, is D. juno. This caterpillar feeds on P. serrato- digitata. If the same mechanical explanation applies here, we would expect its mandibular apparatus to be still less efficient. The only measure of efficiency of mandibular apparatus comes, at the moment, from consid- ering the sculpturing on the biting surface of the mandibles (Text-fig. 6). On this factor the argument about the ineffectiveness of chewing in H. aliphera and H. isabella seems to be borne out, for these two possess none of the ridges and cusps which are clear on the maxillary edge of H. melpomene and H. ricini mandibles and the molar process is distinctly lower, flatter and smoother ( cf . H. aliphera and H. melpomene, Text-fig. 6a and b). On the other hand, such studies fail to support the suggestion that chew- ing apparatus of D. juno is inefficient, for this caterpillar (Text-fig. 6c) has distinct cusps and ridges on the maxillary edges of its mandible although they are admittedly broader and less well formed for cutting than those of H. mel- pomene. The molar process, like that of H. aliphera, is well separated from the maxillary edge but is nevertheless somewhat cusped and there are small auxiliary cusps lying part way between molar and maxillary edges. Thus, argu- ments from simple considerations such as man- dibular sculpturing do not throw light upon feeding differences, if indeed it is legitimate to expect them alone to serve as an index of chew- ing efficiency. It is important, moreover, to recognize that the actual feeding patterns of these caterpillars are distinct. H. melpomene (and also H. ricini) eat across a leaf, taking both blade and midrib in a single action, whereas, if they accept the midrib, H. aliphera and H. isabella first eat the blade of a leaf and then, subsequently, the mid- rib which remains. Thus, whether or not mechanical factors determine whether the mid- rib can or cannot be eaten, these do not exert 10 Zoologica: New York Zoological Society [46: 1 Text-fig. 6 Biting surface of the right mandible of a fifth instar caterpillar, showing the maxillary edge along the upper margin of the drawing and the molar process in the mid foreground. The projec- tion on the lower right is the point of articulation. a, H . aliphera; b, H. melpomene; c, D. juno. Draw- ing by F. Waite Gibson. an immediate influence over the pattern of feed- ing behavior. b. Furrowing Behavior H. melpomene and H. ricini share another very distinct behavior pattern in regard to the midrib. Caterpillars of either species will be found on the undersurface of H. laurifolia leaves, eating from the tip of a leaf back toward the base. Some time after a caterpillar has be- gun eating, it will turn around, walk a short way up the midrib and chew out the furrow mentioned earlier (Text-fig. 5b). At its deepest point this furrow is one-third of the depth of the midrib. It may be chewed out immediately after a bout of eating, during a rest period or just before the animal returns to the tip of the leaf to eat. It seems improbable, therefore, that the material is eaten merely for its nutritive value. Initially it seemed possible that such furrows might serve to control the flow of water into a leaf; perhaps when the water content of the leaf blade rises too high for a caterpillar, it chews furrows across the midrib, so reducing the water flow. Experiments do not support this hypothesis; desiccated caterpillars and those in normal water balance still chew furrows both in wilting and in normal leaves. Thus it seems improbable that the primary function of the activity is that of upsetting water transport in the leaf. Further, the furrow often does not go deep enough to injure the vascular bundles in the midrib. A second possibility is that furrows repre- sent a simple way of preventing other cater- pillars’ coming down the midrib and disturbing or even attacking the larva beyond the furrow; it may serve as a form of territory marker. No unequivocal evidence has been found for or against this theory. When an intruder has turned back and away after reaching a furrow, the response might have been mediated by move- ments made by the furrow-owner. Conversely, on one occasion when the owner had been re- moved, an intruder walked on down the midrib, crossing the furrow with scarcely any hesitation. This could be attributed to the intruder’s being aware that there was no caterpillar beyond the furrow and that it could therefore ignore the “warning.” Slight indications that this may be at least part of the explanation come from the fact that both H. ricini (semi-gregarious) and H. sara (gregarious) do not produce these fur- rows when they live together in a group but if individuals of either species are kept isolated they may do so. Another point of interest in regard to this furrowing habit relates to H. erato. In many respects the behavior of this caterpillar is clearly related to that of H. melpomene and H. ricini and it might therefore be expected to show some signs of chewing furrows. Its normal food- plant is P. tuberosa, whose ventral midrib hardly protrudes at all, and on this vine it makes no attempt to chew furrows. Raised on P. rubra or P. auriculata, however, H. erato produces furrows indistinguishable from those of H. mel- pomene. This latter species will, however, attempt to chew furrows when kept on P. lon- chophora, another vine with negligible midrib. These furrows are quite recognizable, although they may go rather deep and even be developed into a hole. Thus H. erato may, on an unusual host plant, make furrows if the midrib is 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 11 strongly developed while H. melpomene will retain its normal pattern on an unusual host even if the midrib is almost absent. Compared with H. melpomene and H. ricini, H. erato displays very aggressive behavior towards other caterpillars. This can be interpre- ted in terms of the suggestions made above that one function of furrowing is as a territory marker. Possibly H. erato lived previously on a P. laurifolia- like vine with its thick veins. After the change, furrowing was difficult and its role in territorial defence was replaced by the devel- opment of more aggressive behavior. c. Channeling Behavior The channeling behavior of D. iulia and D. phaetusa (possibly also P. dido ) is in some re- spects reminiscent of the furrow-chewing of H. melpomene, H. ricini, H. sara and H. erato. Both D. iulia and D. phaetusa have been seen on P. tuberosa while the former has shown pre- cisely the same pattern on P. rubra and on the simple-leafed P. auriculata. A larva walks down the midrib to the tip of the leaf or leaflet, then it turns and walks back a variable distance, often about one-third of the length. If its body is long enough to reach from the midrib to the margin of the leaf, the caterpillar stretches out and begins to chew a channel across from the margin to the midrib of the leaf. The channel usually slopes slightly towards the leaf base as it nears the midrib but sometimes is almost at right angles to the midrib. As soon as this first narrow channel reaches the midrib, the caterpillar stretches across to the other side and begins a second which will extend inwards as did the first. When it, too, has reached the midrib, the caterpillar is on a small island of leaf, bridged merely by the midrib of the leaflet (Text-figs. 5d and 7a). Usually the caterpillar then chews at the midrib and spins some silk across the bridge. Sometimes, though not in- variably, the larva crosses back to the main part of the leaf, walks along the proximal margin of its channel and bites at it. This does not appear to remove any material but just dents the edges slightly. The caterpillar then returns to the distal side of the bridge and rests along the midrib or begins to eat the island. Although it is usual for this channeling be- havior to occur in relation to the midrib of a leaflet, it may also relate to the margin of a leaflet (Text-fig. 7b), especially in the early instars, or even to the central vein running down between the two leaflets (Text-fig. 7c). After the first island has been eaten away completely, midrib and all, the caterpillar walks further up the leaflet and repeats the procedure once or Text-fig. 7. D. iulia feeding on P. tuberosa. a, A caterpillar chewing the second of two channels which will cut the tip of the leaflet almost free of the base; b, channel chewed in relation to the mar- gin; e, a channel chewed in relation to the vein between the two leaf lobes. twice more before crossing over to the other leaflet. Channeling is shown by caterpillars of D. iulia and D. phaetusa from the first to the fifth instar, although both the first and fifth tend to show it less distinctly. Immediately before pupa- tion, larvae lose the channeling habit and eat leaves either from the tip or lateral margin, the pattern being very like that shown throughout life by PL erato on the same vine. Channeling could serve the same function of territory-marking as has been postulated for the midrib furrows shown by H. melpomene, H. ricini, H. erato and H. sara. If a number of D. iulia caterpillars are put onto a stem with an abundance of leaves, they establish them- selves, each one on a separate leaf, or at most one to each leaflet of a leaf. When a caterpillar is moved onto a leaf occupied by another ani- mal, it walks down until it reaches the bridge and then turns back. This may be due to some movement on the part of the owner, who often crosses the bridge and swings at the intruder with its head. Two observations of a caterpillar turning back although the owner had been re- moved are offset by several in which it did not. The question may finally be raised as to whether there is any direct relationship between furrowing and channeling. It seems unlikely, because D. iulia, though normally found on P. tuberosa which has only a slight midrib, will continue to channel and bridge when cultured on P. rubra and P. auriculata, both of which have thick midribs. It shows no tendency to fur- row. It seems therefore more probable that the furrowing and channeling patterns have been independently evolved. 12 Zoologica: New York Zoological Society [46: 1 d. State of Abandoned Leaves H. ricini is the only one among the species studied here which ever abandons a leaf in the state illustrated in Text-fig. 5c. Parts of the blade remain adhering to the margins, which are left almost entire. Even the midrib is not chewed level. The caterpillar lies with most of its body along the midrib as it eats, just stretching its head out towards the leaf margin. Where the leaf is too wide to allow the caterpillar to reach the margins without moving its body off the mid- rib, the animal may just leave the margins un- touched. There is some indication that this par- ticular pattern may be correlated with the gre- gariousness of H. ricini. When eggs are laid singly or larvae isolated by the experimenter, the feeding pattern of the caterpillars recalls the “straight across” action of H. melpomene. Several animals together produce a more ragged effect. Similarly H. sara, when part of its nor- mal group, shows no clear-cut feeding pat- tern, though when a caterpillar is raised alone it eats in the typically H. melpomene fashion, straight across blade, midrib and margins and even chews furrows in the midrib. The amount of leaf material left when a cat- erpillar abandons a leaf (Table IV) and moves to a new one is to some extent characteristic of the species, although external factors, such as whether or not there is more food available, do have some influence. There is also a tendency for earlier instars to abandon a leaf with a higher percentage of it remaining than in the case of later ones. Four species, D. iulia, D. phaetusa, H. erato and H. sara, habitually eat the petiole of their leaf and sometimes continue and eat the stem as well. The first three species eat P. tuberosa and it could be that the stem of this vine is more palatable than that of others. This seems somewhat improbable. Moreover, other caterpillars, e.g., H. melpomene and H. aliphera, eat neither petioles nor stems when they are fed on P. tuberosa. It is possible to relate this habit to the vege- tative characteristic of the vine on which the caterpillars normally live. Thus the natural food of H. sara is the small P. auriculata vine, while P. tuberosa is also slight and slender and its leaves are frequently very scanty (see Table II). Presumably there would be strong selective pressure for any animals living on such vines to practice economy as far as possible and this might explain this aspect of the behavior of D. iulia, D. phaetusa, H. erato and H. sara. It is relevant that, when H. erato is fed on P. laurifolia or D. iulia on P. rubra, they both still eat the petiole as well as the leaf. This facet of their feeding behavior is thus, at least to some extent, independent of environmental control, e. Feeding Positions It has already been said that positions adopted by feeding caterpillars differ specifically. The most obvious difference in this respect is whether or not the midrib is used as an orienting feature. A caterpillar of H. melpomene, H. erato, D. iulia or D. phaetusa usually has at least part of its body along the midrib while it eats a leaf (Table V). If the leaf is so wide that the animal cannot stretch to the margin, it lies parallel to the midrib (see p. 16 for the exception in the case of H. ricini ) . This adherence to the midrib may possibly have some selective advantage in relation to camouflage from predators but whether this is so or not, it certainly gives the caterpillars a more secure hold on the leaf. It is much easier to dislodge an individual of H. melpomene placed on the blade of a leaf than one that has been allowed onto the midrib. While H. aliphera and H. isabella occasion- ally orient themselves parallel to the midrib as they eat, they do not lie along it. Normally their orientation bears no relation to it. D. juno, A. vanillae, H. doris and H. sara appear sometimes to orient to it or along it but at other times to pay it no attention. D. juno, H. doris and H. sara are gregarious and the lack of a definite and con- sistent orientation in feeding may be correlated with this habit; certainly H. sara when isolated from its fellow caterpillars will lie along the midrib, while a group of D. juno walk only along the midribs of the leaflets when they search for a new feeding place. A. vanillae, however, shows no signs of any consistent orientation of its body with respect to the form of a leaf in feeding. Neither D. juno nor H. doris has been tested singly on a leaf. In the field the caterpillars of most species feed on the under or abaxial surface of a leaf (Table V), which is almost invariably the ven- tral one. This is not necessarily so in the labora- tory and it has been found that the different species tend to differ in their responses to a leaf whose position has been reversed (Table V). H. aliphera and H. isabella both return to the ventral position, even though this means that they will not be eating from the true upper sur- face. D. iulia, H. melpomene, H. erato and to a lesser extent H. ricini still choose the under sur- face though this is now dorsal. Thus H. aliphera and H. isabella seem to be using different criteria for their choice than the other four species. Feeding in first instar H. ricini and H. sara is noteworthy in that the caterpillars are found on the upper surface of a tender leaf which is still young enough to be at least partly folded. 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 13 Table V. Further Information Relating to Feeding Behavior Key: -| — Behavior in column above does occur. Behavior in column above does not occur. Species Oriented along midrib Eats standing on petiole or stalk Holds loose piece in forefeet and eats Eats from upper or under leaf surface Normal leaf Leaf inverted Dione juno Sometimes Only just before pupation -- Both Both Agraulis vanillae Sometimes — — Both Both Dryadula phaetusa Usually + ? Under ? Dryas iulia Usually + + Under Upper Heliconius isabella No — — Under Under Heliconius aliphera No — — Under Under Heliconius melpomene + — — Under Upper Heliconius erato + + + Under Upper Heliconius ricini 4- Except when grouped together & when parallel to it Upper in 1st instar, lower or both later Upper Heliconius doris ? ? ? Both Both Heliconius sara Sometimes — — Both Both They congregate between the two blades of the leaf, chewing away from inside their cover. This habit, which in H. ricini may continue into the early part of the third instar, differs from the behavior of the older larvae, which remain as far as possible on the under surface of the leaf (Table V). In H. ricini, however, the attraction does not lie in the upper surface of the leaves as such but in the fact that the blades provide a cover. If a young leaf is bent in the opposite way from normal (j. e., with two under surfaces together), young H. ricini still collect between the two folded blades, although they are now on the under surface. The surface of a leaf on which a caterpillar feeds may in some cases determine the feeding pattern adopted. When H. melpomene is placed on the under surface of a P. lonchophora leaf, it orients along the midrib of one of the leaflets, starts from the tip and eats straight across, midrib and all. If, however, it is placed on the upper surface it orients as H. aliphera and H. isabella, and eats from the margin of the leaflet, tending to leave the midrib. Thus the straight- across type of feeding is probably dependent to a large extent on the larva’s having the midrib for orientation and this feeding pattern may be a specialization of one not dependent on the pres- ence of a marked midrib. f. Eating Actions The third column of Table V shows that only two species, D. iulia and H. erato, are known to use their legs for holding a loose fragment of leaf while they eat it, though it is possible that D. phaetusa may also behave in this way. £>. iulia and H. erato are the two species which normally eat even the petioles of their leaves, so it is not surprising that they have evolved a behavior pattern allowing them to make use of small fragments which are lost to other species. On the other hand, when it is the exuvia cast after a molt that are being eaten, all species which have been watched use their first legs and usually their second in manipulating the empty skin and its scoli. This is perhaps ex- plicable if the exuvia-eating pattern is an old and stable one within the subfamily, uninflu- enced by evolutionary adaptations relating to new food plants. H. erato, D. iulia and D. phaetusa also share what seems to be a related behavior pattern, in which they sit near the base of a leaf, on the petiole or even the stem, and chew at the distal end of the leaf, holding it with their legs and bending it back towards themselves. This habit of “pulling up” the leaf would also seem to be associated with the particular vine on which these species normally feed. P. tuberosa is one of the few species with medium-sized leaves still flexible enough to allow such bending. The behavior occurs when the larvae are in the fourth or fifth instar and are already beginning to become heavy for the thin and flexible leaves. H. erato and D. iulia appear to twist their 14 Zoologica: New York Zoological Society [46: 1 heads more freely in eating than do the other species, while H. melpomene and H. ricini seem more flexible in this respect than H. aliphera and H. Isabella. These differences are also re- flected in the fact that, when they rest between a series of bites, H. erato and D. iitlia, and to a lesser extent H. melpomene and H. ricini, have their heads overlapping the margin where they have been chewing. There is no indication of “right-handed” or “left-handed” caterpillars, in that the head is never twisted to one side more than another. It goes to whichever side is the more convenient, considering the surface to be eaten and the dis- position of the caterpillar’s body as a whole. Eating movements of all these species are alike in that the caterpillar extends and twists its neck. The mandibles then bite into the leaf. The head is drawn a little closer to the body and the jaws bite a second time. This is repeated until the head is against the body or the edge of the food is reached. The number of bites made dur- ing this movement of the head varies from 3 to 15, depending upon the length of material available for eating and its nature— more than one bite is given at a point where the material is tougher, such as a leaf vein. While the head is being brought in from the extended position, the legs in all species are usually involved to some extent in shifting the thorax backwards. Sometimes all three pairs move but usually only the first two. There is a suggestion that the hind legs of A. vanillae are more active in this backward movement than those of other species. This may flatten the num- erous glandular hairs with which P. foetida is covered. As will be seen later, this species shows a locomotory specialization which may also have evolved in relation to the problem of the thick, sticky hairs on this plant. In eating, H. isabella and H. aliphera move their legs somewhat differently from the other species studied. A front leg, the one on the side to which the head is turned, beats rhythmically as the caterpillar makes its series of bites. The beats are generally made in the air but occas- ionally the foot will touch the leaf surface. No suggestion can be made concerning the signifi- cance of this movement and it is mentioned merely as a pattern shared by H. aliphera and H. isabella and appearing in no other species observed. g. Eating of Egg-shells Recently emerged larvae of the solitary H. melpomene and H. erato eat their own egg- shells within 10 to 20 minutes of leaving them. If they come across remnants of the egg-shells of other caterpillars, they chew these too. When they find an egg which still has an embryo within, they show no inhibitions about eating into the shell and devouring the embryo. Were such behavior present in gregarious species such as H. dor is, H. sara, D. juno or H. ricini, a whole batch of eggs might be destroyed by the first few larvae to emerge. If, however, those caterpillars which are gre- garious are less attracted to egg-shells they would be less likely to eat their fellow-larvae. This should be reflected in the absence of a tendency to eat their own empty egg-shells. Twelve H. sara eggs have been investigated in this respect and it was found that they had not been chewed other than at the emergence holes. H. wallacei has apparently a similar inhibition about eating its own or other egg-shells. This does not, however, hold for D. juno nor H. ricini, so that either a different method of pre- venting cannibalism has evolved here or inhibi- tion of egg-shell eating is in no way associated with protecting developing embryos from preda- tion by other caterpillars. h. Eating of Cast Skins The only question relating to the eating of cast skins that has been investigated, is whether or not foreign skins were acceptable and for how long. The answer is that any one species will accept the skins of any other if they are fresh and from first to fourth instar animals. Skins more than 24 hours old may still be eaten but they are often rejected. Similarly the skins cast by larvae molting out of the fifth instar and into a pupa are occasionally eaten (with the exception of the head capsule) but are usually rejected. In testing the edibility of these various skins, the practice was to tie a piece of cotton thread tightly around the skin to be tested and then tie or tape the thread to the stem or leaf at the point from which the test animal’s own skin had been removed. It was often the thread which the caterpillar ate when it moved to in- vestigate the skin. Part of the attraction to the skin might lie in the spikiness of the spinules of the scoli, for the cotton threads had num- erous minute threads attached to them, project- ing like the spinules on the skin. i. Drinking Caterpillars of 10 species have been seen to drink water from small droplets on their leaves and there is little doubt that the other species also do this. The mouth is applied to the water, the mouthparts move and then remain still for the rest of the time. There is no indication that the drinking pattern is derived from that of eat- ing or vice versa. 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 15 In the field even during the dry season there is abundant dew at night so that caterpillars have opportunity to drink adventitious water even when there has been no recent rain. They appear to thrive better in captivity when given free water to drink than when they are kept at a very high humidity but given no free water. A pattern which is most noticeable in H. ali- phera and H. sara, although it occurs in cater- pillars of other species as well, is that of re- gurgitating a drop of green fluid when severely disturbed. After the effect of the disturbance has passed, the caterpillar takes up the fluid again, using the same pattern as when drinking water. V. Defecation Caterpillars all defecate at intervals through- out their periods of eating and resting and dur- ing locomotion. The rate of defecation varies, as measured by the time between the production of two fecal pellets, rising during any one instar with the increase in the duration of feeding periods. On the other hand the size of the pellets increases from one instar to the next, so that the range over which the defecation rate changes is roughly the same for the different instars of any species. As Nagasawa (1957) reported for the larvae of the gypsy moth, Lymantria dispar Linnaeus, there is no stepwise increase in the size of the fecal pellets as there is in head cap- sule size. A fecal pellet is extruded as a small green cylinder, the form being very similar in all species. It dries and darkens to a brownish- black mass over a period of 2 to 3 hours. The defecatory behavior patterns of the species of heliconiine caterpillars studied here are all alike; the anal prolegs release their hold of leaf or stem, the hind end is raised into the air and the pellet expelled with more or less force from the anus. The rectal opening then closes and opens several times, moving rhythmically with the pads of the anal prolegs. Finally the hind end is once more lowered. H. isabella is the only species in which there is a noticeable wag- gle of the hind end immediately after defecation and even in this species it does not always occur. A caterpillar of any of the ten species watched (H. doris, H. numata, H. wallacei and P. dido have not been observed), will turn its head to its anus and with its jaws remove a pellet which has become stuck. Sometimes eveo after the pellet has been thrown free, the cater- pillar turns and chews in the region of its anus, presumably at particles which remain. When caterpillars are on a vine in a natural position, the fecal pellets are either shot off or roll free of the leaf. Raised in dishes in the lab- oratory, however, the caterpillars frequently come across their own cast pellets. In these conditions a larva picks up the pellet with its mandibles, lifts its head and releases the pellet so that it may fall off the leaf or roll away. There is a slight tendency to jerk the head as the pellet is released so that it is actually thrown. In a dish in the laboratory a caterpillar invar- iably comes across the same or another pellet within a few minutes. It continues to reject pellets for a variable length of time and then apparently accommodates to the situation, ignor- ing pellets which are right beside its head. Constance Carter, who has raised large broods of H. melpomene for genetical work, reports that her caterpillars finally ceased to show the pellet-throwing behavior. H. aliphera and H. isabella, when presented with fecal pellets, may attempt to throw them as do the other species. They may, however, show a variation which is peculiar to them. Instead of picking the material up in its mouth, a caterpillar bobs its head several times in the direction of the pellet, knocking it with the long head scoli. If the pellet is merely caught on some irregularity on the leaf, this treatment sometimes frees it and it falls away. In lab- oratory rearing dishes it is of course not usually effective. What is of interest is that the head bobbing or beating with the head scoli is also a response which both H. isabella and H. ali- phera give to other more general disturbances. D. juno is the only caterpillar in which an activity takes place that might be called “social defecation.” Instead of a caterpillar turning and pulling a pellet free of its own anus, D. juno may have the pellet removed and rejected by another larva which is passing at the time. Usually the second caterpillar assists when the pellet is almost free, but sometimes it is almost dragged out of the defecating animal. It is not known how early in larval life this habit appears, but it persists into the late fourth instar before disappearing in the fifth. During the period over which the behavior occurs, a caterpillar is per- fectly capable of ejecting pellets alone or of turning round and freeing a pellet from its own anus. VI. Resting After an interval of feeding, caterpillars stop eating, turn away or walk backwards a short distance, and go into a resting phase. Both the position of a caterpillar relative to the leaf and its posture are, to a large extent, specific char- acters. 16 Zoologica: New York Zoological Society [46: 1 Table VI. Resting Behavior of Ten Species of Heliconiine Caterpillars Species Main /general larval pattern of resting position in relation to leaf Any change in resting position with prepupal stage Posture of body Dione juno No definite orientation except to each other Either straightened out or with thorax elevated Agraulis vanillae No definite orientation Tends strongly to rest on stem or petiole In “J” during early instars, later straight and extended on stem Dryadula phaetusa Along midrib Usually “J” but straighter in later instars Dryas iulia Along midrib Some tendency to rest on stem Occasionally straight especially late 5th instar. Usually in “J” Heliconius isabella As in eating or facing opposite direction Invariably straight, or with very slight curve Heliconius aliphera As in eating or facing opposite direction Thorax usually curved to side. “J” less acute Heliconius melpomene Along midrib None except coming off stem onto bottle in morning Straight but usually thorax just out of line in front Heliconius erato Along midrib or stem Strongly tends to rest on stem Straight Heliconius ricini Along midrib, parallel to midrib or an oblique position Straight but often thorax just out of line in front Heliconius sara No definite orientation except to each other Straight, sometimes thorax elevated Heliconius doris ? ? ? a. Resting Position Resting positions on a leaf are indicated in Text-fig. 5 and the last column of Table VI. In the same way that many behavior patterns change during the larval period, so does the choice of resting position. These positions can be divided into three main categories: 1. Those in which the caterpillars (H. aliphera and H. isabella) are oriented neither to the midrib or margin of the leaf nor to gravity. The only describable regularity is that in the resting position a caterpillar’s body is oriented along the same line (or one par- allel to it) as that taken during the previous period of feeding. The hind end may be pointing directly to or away from the site where the caterpillar finished eating. 2. Those (e. g., D. iulia, D. phaetusa, H. mel- pomene, H. ricini and to a lesser extent H. erato ) which rest with the main part of their body along the midrib, irrespective of their last feeding position. After two individ- uals of H. melpomene have been eating side by side on a P. laurifolia leaf, they both crawl up and rest on the midrib, one behind the other. 3 . The final group contains those species whose resting position shows no consistent rela- tionship to the form of leaf. The details vary from species to species. Thus A. van- illae sometimes rests along the midrib, sometimes faces its feeding place, some- times faces the opposite direction and often lies with its body having no particular or- ientation to anything at all. In later instars, however, this species most frequently rests on the petiole or stem of its plant. In social resting among those species which are to some extent gregarious, the caterpillars usually lie side by side, frequently but not in- variably facing in the same direction. In natural conditions the orientation of D. juno is in relation to the other caterpillars of the group rather than to environmental markers. The same would appear to be true of H. doris and also H. sara. If there has ever been a tendency to orient their bodies to the leaf in a particular way dur- ing the resting phase, it seems to have been lost in these gregarious species. The same explana- tion might apply to the slight tendency which the semi-gregarious species H. ricini shows away from the H. melpomene- type orientation, for it does not invariably rest along the midrib, but may lie obliquely or parallel to it. Part of the pattern of resting during the day, or at least in the morning, shown by fifth instar H. melpomene and D. iulia, has been the selec- 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 17 tion of a particular resting position, away from the leaf on which they were eating. In many cases it has been a large leaf near the base of the vine but in the laboratory caterpillars have frequently chosen to rest on the bottle in which their vine was stuck. The same position was selected day after day. b. Resting Posture (see Table VI). The posture adopted by these various species of heliconiines during their resting phase must also be considered. A resting pose which may well be basic to the subfamily Heliconiinae is one in which the greater part of the caterpillar’s body shows a particular orientation on the leaf but the head and thorax are turned to one side, giving what may be called the “J” position. This is shown in its most extreme form in D. iulia and D. phaetusa (Text-fig. 5d) . H. aliphera usually has a far less acute bend on the “J” and H. isabella such a gentle curve (when it is present at all) that it hardly resembles a “J” (Text-fig. 5a). H. melpomene and H. ricini rest in a “J” position in which there is no sug- gestion of a hook at the base of the “J”, only a slight deviation from the straight line of the body, (Text-fig. 5b, c). When a number of H. ricini are kept together as a group, the “J” tends to straighten further. Those species which show distinctly gregar- ious tendencies, H. sara, H. doris and D. juno, all rest with their bodies straight out or with the head and thorax arched up, the latter habit being especially marked in D. juno. This would seem to be merely a modification of the straight- ened position, allowing more crowding while preventing the animals’ heads being buried. H. erato is somewhat anomalous for, although a solitary species, it rests with its body not only in a straight line but also extended; all the other species when they rest have their bodies con- tracted to some extent. A. vanillae in its fourth and fifth instars does, however, extend its body in resting on the stem. The contraction of its body is most marked in a second anomalous case, that of D. iulia (Plate I). Here the whole thorax is drawn in towards the abdomen and this results in a marked humping of the first and second abdom- inal segments. Consequently the scoli borne on this region do not keep their normal orientation but all point forward along with those of the meso- and meta-thoracic segments, giving the animal a hunched appearance. The dorsal scoli of the second and third thoracic segments min- gle with those of the first and second abdominal. The first lateral scolus, that belonging to the second thoracic segment, is always quite distinct in the resting pose, but the lateral scoli of the metathoracic segment lie with the supralat- eral ones of the first two abdominal segments. This rest position is highly characteristic of D. iulia and has not been seen in other species. It is interesting in that it is extremely like the position taken up when the caterpillar is dis- turbed by a blast of air (not a touch), the only difference being that a disturbed caterpillar contracts so that the first lateral scolus also lies in the bunch of scoli, instead of being sep- arate as it is in the rest position. VII. Weaving When caterpillars are kept on fresh vines which are acceptable to them as food, they do not walk about very much. Between eating and resting a caterpillar may turn and walk away but often H. melpomene, H. erato and H. isa- bella just shuffle slightly backwards. Character- istic of even this brief walking, however, is behavior which may be called “weaving.” The head is swung regularly from side to side and the spinneret behind the mouth spins a silk thread, touching it down to the surface so that figure-of-8 tracks are left behind the caterpillar. Even when a caterpillar walks as fast as it may, it still trails a silk thread behind it, though in such conditions it is not attached as fre- quently. This spinning of silk throughout the entire larval stage is undoubtedly of value to the cater- pillar in that it always has a safety line attaching it to the leaf or stem and there is thus less danger of its being swept away from its vine. If a cater- pillar falls, it hangs suspended on its thread of silk and subsequently climbs up this and back onto the point where it was attached. A mat of silk threads is also spun over the surface of a leaf and this provides a secure footing for the claws and for the crochets of the prolegs. By watching the individual move- ments of a caterpillar’s feet in walking, it is clear that the claws rely to a great extent on the silk trail to provide a foothold. A limb may make several movements and only when its claws make contact with the silk does it grasp. The effect can also be demonstrated by com- paring the footholds of caterpillars on surfaces upon which they have been allowed to weave. Parts of a glass plate and of a smooth leaf ( P . lonchophora) were covered by the weav- ing of H. aliphera. A caterpillar was then dropped gently onto one surface or the other and as soon as it was on its feet, the glass or leaf was turned upside down. Occasionally the animal on the glass hung down for a short time, 18 Zoologica: New York Zoological Society [46: 1 its prolegs not grasping the surface immediately, but there was never any risk of its actually fal- ling off. On the silk-covered leaf even the pro- legs were attached quickly enough to prevent the hind end from hanging loosely. Conversely the caterpillar fell off clean glass, although oc- casionally it managed to remain for some time on the clean leaf. During this time it would rapidly weave on the under-surface of the leaf and as soon as its claws and crochets came in contact with the silk attachments, risk of their slipping decreased. An ability to remain on glass plates when these are reversed may have little value in the natural life of a caterpillar of H. aliphera, yet it is a fact that H. aliphera and H. Isabella are the two species which weave most markedly and are also the animals which rest under the smooth blades of their leaves. It has already been suggested that part of the advantage of resting on the midrib is that it allows a more secure foothold, so a species living on smooth leaves and not utilizing the purchase of the midrib might well increase the extent of its weaving so that silk could be used in place of the midrib. A caterpillar of H. aliphera or H. Isabella will touch its spinneret to the substratum, then stretch its head back and then finally forward and down again. This behavior may be called “yawning” and though all species do yawn during the construction of the silk pad just before pupation, only H. aliphera and H. Isa- bella have been seen to do it during earlier larval life. In spinning the pupational pad, yawning draws the silk so that a loop is formed. H. aliphera and H. isabella yawn predominantly when they return to their resting positions after feeding and it seems that these two species rest on a mat of somewhat looped silk instead of one with plain attachments like the other species. This would presumably provide them with a more secure foothold. Weaving may be expected to occur when- ever the foothold of a caterpillar is precarious. In fact, it seems to be elicited by any surface which is strange to the caterpillar— even if it has already been covered with silk by another animal. There are exceptions to this rule, the ambulatory phase immediately before a larva hangs up to pupate being a clear example. During this period the caterpillar walks onto many strange surfaces but will not begin to weave until it reaches a potential site. Another example of the weaving behavior was first noticed in the laboratory dishes. Fresh leaves were put in for the caterpillars each day and frequently when those of the previous day were being removed they were found to be tied firmly onto the dish and/or each other with silk threads. If there were a remant of the petiole left it was almost invariably this part which was attached to the dish, though sometimes another projection might be used. This phen- omen is explicable in terms of what has already been said about weaving. The caterpillars re- main on their leaves for most of the time, refusing to abandon a leaf in order to walk onto the glass. When they encounter the glass, however, there is a bout of weaving as they touch the strange surface; thus at this point the stem and glass become attached by a series of silk threads. The species on which these observations were initially made comes onto or deserts a leaf, walking along the midrib, so that caterpillars in the dishes would usually have come into contact with the glass from one or the other end of the midrib. Thus it was here, the petiole or sometimes the tip of the leaf, where the silk attachment was formed. There are observations which suggest that such “tieing-up” behavior is not unnatural and that it occurs in the field. In a number of in- stances caterpillars, living freely on large pieces of vine, have been seen to weave steadily be- tween the petiole and the stem to which it is attached, thus reinforcing the natural junction. The behavior has been particularly noticed in H. melpomene but also occurs in other species. The same or a very closely related pattern is sometimes observed in H. sara, where a group of caterpillars will tie two or several leaves to- gether with silk. Bell (1920) records how the larvae of the oriental lycaenids, Vivachola isocrates and V. perse, bind the stalk of the fruit on which they are feeding onto the branch with silk, and such behavior doubtless occurs among many other lepidopterous larvae. Tieing the whole leaf onto the stem would seem a possible safeguard against the leaf’s be- coming detached while the caterpillar eats from it— an event which has in fact been seen on two occasions in Trinidad. The same explanation could be true of the behavior of gregarious H. sara, for P. auriculata leaves will sometimes drop off while still green and while being chewed by the caterpillar. It is also an advantage when the leaf concerned is chewed loose by animals proximal to it. Any caterpillars which are feed- ing distally are then still able to cross by the silk bridge onto another leaf and so back to the vine. This also has been observed in the labora- tory. Even if its leaf comes loose or another animal chews away the link between leaf and vine, there is still a possibility that a caterpillar may be able to climb up and regain its place. This 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 19 cannot be said for a pupa in the same circum- stances, because its power of movement is so limited. It is therefore not surprising that a caterpillar, preparing to pupate on a leaf petiole, old flower stalk, tendril or even a leaf, should frequently weave between the object on which it will pupate and the main body of the vine. Such behavior would certainly be selectively ad- vantageous and has been recorded in many lepidopterous larvae (see Ford, 1945, and Hin- ton, 1955). Indeed, in the present study there have been four instances in D. iulia, one in A. vanillae and one in H. melpomene in which the slight silk attachment alone was holding the pupal support on the vine. The consequences of one of these pupae falling to the ground will be considered later. The final point about weaving is that it may occur in situations in which the caterpillar seems merely to be generally disturbed. H. aliphera is especially prone to weave very actively on the surface of its leaf if it is poked, blown on, shaken or if another caterpillar comes near it. This could perhaps be regarded as an example of a displacement activity, though it may also be argued that a behavior pattern which ensures a more secure attachment to the substratum would be appropriate in a situation in which the animal is being attacked in any way. It seems very similar to behavior described by Dethier (1943) in lepidopterous larvae removed from their plant food and which he interprets as a “visual searching movement.” VIII. Locomotion Locomotion which occurs between feeding periods is normally of short duration. Neverthe- less some specific differences are apparent. H. aliphera, H. isabella, H. ricini, H. melpomene, H. sara, D. juno and D. phaetusa make prac- tically no movement at all, sometimes just walk- ing slightly backwards, sometimes turning and walking a few centimeters. H. erato, D. iulia and A. vanillae often walk an appreciable distance, up the leaf, onto and along the stem. D. iulia walks very much more quickly than the others, directly and without stops. Conversely H. erato barely seems to be moving at all and frequently stops altogether. A . vanillae walks slowly, eating the hairs from the leaf, petiole and stem as it goes. Both A. vanillae and D. iulia have peculiari- ties in their mode of walking but these do not always appear. A. vanillae has a strange, jerky stride; the anal prolegs are raised, carried for- ward and then oscillated back and forth just above the stem several times before they are finally put down in the new position. The head and legs also show this jerky motion but not as strongly as the hind end. It is possible that the action normally results in flattening down the glandular hairs with which P. foetida is lib- erally covered, thus producing a clear area for attachment of the anal prolegs. These cater- pillars seem rarely to leave their vines, even pupating on them; this might explain why they sometimes walk in this jerky way even in sit- uations where there are no hairs to flatten, as when walking down a piece of wire. The peculi- arity in walking of D. iulia also consists of a movement of the hind part. In this case the last segment with the anal prolegs is lifted and low- ered sharply several times during each short burst of forward locomotion. No explanation is offered for this pattern. The locomotor stage which frequently occurs just before pupation will be considered more fully later. Suffice it to say here that in most species it starts very sharply and that it may last as long as three hours, during which time the animal may cover as much as 50 meters of ground. With the exception of A. vanillae and occasionally D. iulia, the walking motion is smooth. H. aliphera, H. isabella, H. erato and D. phaetusa will seldom walk as far as a meter without stopping for one or more short rests. During this walking stage the animals show no obvious photopositive or photonegative orienta- tion. There is, however, slight evidence of a negative geotaxis, at least in H. aliphera, H. isabella and D. iulia. IX. Social Behavior Information which allows an estimate of the degree of social behavior among larvae is avail- able for only 10 of the 14 species of Heliconiinae in Trinidad. Of D. phaetusa, P. dido, H. numata and H. wallacei it can be said only that the last-mentioned alone among them is gregarious. The other 10 species can be arranged in a series from the typically aggressive, asocial cater- pillars of H. erato to those of D. juno which are not only gregarious but show signs of actual social behavior. As has already been said (Table III), the eggs of some of the species are laid together in a group so that the larvae of these species, D. juno, H. ricini, H. sara and H. doris, start living communally. The eggs of H. isabella and H. aliphera are never laid in a group but the female is not averse to laying more than one egg —up to six in the case of H. aliphera— on a sin- gle leaf. This contrasts strongly with the be- havior of the female H. erato which flies away to another vine rather than lay an egg on a set of leaflets where one is already glued. The degree of intolerance of one caterpillar 20 Zoologica: New York Zoological Society [46: 1 for others, either of its own or another species, is greatest in H. erato, which is never found sharing a leaf without fighting. D. iulia, H. all - phera and H. Isabella also remain solitary on their leaves but when more than one caterpillar are together, they do not bite at each other as H. erato would do. A specimen of either of these three species will swing its head and thorax at an intruder or will shake the whole of its anterior half but there is very seldom any con- tact at all between such caterpillars. Two indiv- iduals of H. aliphera, for instance, may try for several hours to share the same leaf and it seems to be only their mutual disturbance which even- tually results in their separation. Further, newly- emerged larvae of H. erato will attack and eat other larvae or eggs, while such behavior is comparatively rare in D. iulia, H. aliphera and H. isabella. H. melpomene, while it eats eggs and newly- emerged larvae during its first instar, appears later to become more tolerant and will share a leaf with another caterpillar of approximately its own size even when there is an ample amount of leaves available. A. vanillae shows no violent reaction to sharing its leaf with others of its own or other species and has, furthermore, never yet been found to eat eggs or newly- emerged larvae. H. rieini shows even less re- sponse to other caterpillars than does A. van- illae, not even swinging its head at them as a rule. There is nevertheless no great attraction evident between H. rieini caterpillars. Although they may eat, rest and molt together as a group, one or two will frequently remain separate from the others, an occurrence which is almost never seen in H. sara, H. doris or D. juno. Except in cases of food shortage, there is never any aggression between caterpillars of H. sara, H. doris and D. juno. In fact, individ- uals of the last-mentioned species seem distinctly unsettled when separated from others of their group. Even small groups of three or four cater- pillars do not remain discrete in the presence of a larger collection, but join it. One specimen of D. juno appeared less unsettled when allowed to share a leaf with H. rieini than when it was alone. Molting has not been watched in conditions approaching normal for H. doris. In glass dishes, however, they clearly orient to each other. Groups of H. sara and D. juno molt on the stem of their vine and are oriented to each other. D. juno form a double ring around, the stem, bodies parallel to each other and to the stem, most with their heads pointing to the center. H. sara shows a far less strict orientation although the bodies of the larvae are parallel to each other. As has been said in the section on feeding, gregarious species and in fact those which are not normally gregarious but are kept in groups by the experimenter, establish synchronization of feeding and resting. The degree of coordina- tion varies from the strict effect in H. doris and D. juno to the more ragged one typical of H. rieini and the naturally solitary species. Simil- arly the synchronization of molting and pupa- tion is more marked in the truly gregarious species. Finally, D. juno alone has been seen to prac- tice what has been termed social defecation, one larva discarding a fecal pellet from the anus of another. It has not been established whehter D. juno will remove pellets from the anus of a species other than its own, nor is it clear whether or not there is any selective advantage to be had from social defecation. Certainly it would appear an easy habit to acquire in that all species of heliconiines have been seen to throw free pellets which they come across. X. Defensive Behavior Study of these caterpillars was started with a view to using their defensive reactions as clues in physiological work on their sense organs. It soon appeared that such responses are very variable and probably depend on the basic activ- ity of the larva at the time. It was therefore necessary to expand the project so as to get some idea of these “basic activities.” The cater- pillars were disturbed as little as possible by the observer, the result being the information pre- sented in sections I and II of this study. A fur- ther result was that defensive behavior was seldom elicited except in encounters between the caterpillars themselves or in fights with insects such as ants or mantids. It should therefore be held in mind that caterpillars may be capable of far more drastic and clear-cut defense re- sponses than are described for them here. The stock response is that of turning from side to side or banging the head and thorax continually towards the side that was stimulated. In H. isabella alone the caterpillar may loosen the hold of its anal and posterior prolegs and thump its tail end up and down when disturbed. It is noticeable that this species is alone in having its posterior segments a bright contrast- ing color (see Beebe, Crane & Fleming, 1960). Two apparently otherwise unrelated species, H. sara and H. aliphera, are both prone to regurgi- tate contents of the gut, though such behavior can be elicited in other species as well. H. ali- phera and H. isabella both beat at an object or intruder with their long head scoli, a habit al- ready mentioned in the section on defecation. 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 21 Although H. melpomene and D. iulia also have long scoli on their head capsules, they have nev- er been seen to use them as do H. Isabella and H. aliphera. XI. Phylogenetic Discussion The observations recorded here were part of what was essentially a preliminary study, an attempt to expose problems which allow experi- mental analysis, and consequently any phylogen- etic conclusions must necessarily be extremely tentative. It would appear that they are worth discussing, nevertheless, if only because they define more clearly what further information is needed. The activity patterns have not revealed any striking similarities or differences between the species. While it would obviously be of great interest to investigate the control of these, it seems improbable that information on this score will contribute to knowledge of relationships. The nocturnal feeding in D. iulia and H. mel- pomene has probably been acquired independ- ently—it is clearly a later specialization, taking effect only in the fifth instar. It certainly seems worth discovering whether the method of con- trol is the same in both. On the basis of feeding behavior, the 10 species studied fall into three major groups. This is true of patterns in the first instar larvae as well as later on and may be further correlated with the resting position and posture taken up between feeding. Within these three there are other, closer associations of species. The first group is that comprising A. vanillae, D. juno, H. aliphera and H. Isabella. None of these habitually orient their bodies along the midrib of a leaf or leaflet, either during feeding or rest. It is not known for certain how the first instar A. vanillae eat but certainly some individ- uals have been seen to chew holes in their leaves. The other three species all scrape the green cells from the surface of leaves, and it seems probable that this pattern is fairly gen- erally present among heliconiines even though it is not in all cases the one naturally shown. If a leaf is thin enough, the caterpillar is likely to produce small holes where it has scraped and it seems possible that this is what occurs in the case of A. vanillae. While A. vanillae and D. juno do not orient in relation to the midrib, H. aliphera and H. isabella never lie along or on it and appear to orient away from it. Their feeding patterns, es- pecially that of H. isabella, result in the mid- rib’s being left on an abandoned leaf. This ten- dency is perhaps reflected in two other charac- teristics shared by H. aliphera and H. isabella. First, these two show a greater development of weaving behavior than any other species. This would be of importance when the foothold pro- vided by a midrib is unavailable (see p. 18). Second, there is the trend seen in H. aliphera and characteristic of H. isabella towards attach- ing the pupa to the smooth blade of a leaf rather than the midrib of some other vein or protrusion (see section II). H. aliphera and H. isabella both wave a sin- gle foreleg rhythmically during their feeding, a further indication of the sharing of behavior patterns. The second group consists of H. melpomene, H. ricini, H. erato and H. sara (possibly H. doris as well). When these caterpillars are solitary their feeding position is sharply oriented towards the midrib, in that they lie along it during the rests between chewing as well as during feeding. When alone, all of these species eat straight across the leaf, cutting through the midrib as well as the blade. When raised in a group H. ricini and H. sara both become inconsistent in their patterns but this is to be expected, for the animals now tend to orient in relation to each other rather than to the leaf. Furrowing across the midrib occurs in some circumstances in all of these four species. Until its function has been established, too much emphasis should not perhaps be laid on this. Yet it is suggestive of a fairly close relationship among the species, especially as it is in H. erato, at least, what might be called “vestigial behav- ior”—a pattern elicited only when the caterpillar is in somewhat abnormal conditions. It seems possible that H. erato comes from a stock which lived on P. laurifolia, as do H. mel- pomene and H. ricini. When it migrated to the slender P. tuberosa vine, H. erato lost the sharp orientation to the midrib from its natural reper- toire of behavior patterns and this only reappears now if it is fed on P. laurifolia. On P. tuberosa feeding and resting are still relative to the mid- rib but the habit of midrib furrowing is not displayed. Chewing straight across the blade and midrib is replaced to some extent by behav- ior such as the “pulling-up” technique. This enables H. erato to utilize even the tips of slender leaves onto which it is too heavy to climb. The flimsiness of many P. tuberosa leaves may also have led up to the evolution of the practice of resting on the stem instead of the leaf. H. sara and H. ricini share the characteristic of eating holes or channels in subterminal leaf- lets while these are still folded together. This might be taken to indicate a close relationship. 22 Zoologica: New York Zoological Society [46: 1 Another explanation is, however, possible. Their eggs are laid in batches and are smaller than those of H. melpomene or Ii. erato; the maximal diameters of the eggs of H. sara and H. ricini are 0.65 and 0.70 mm. respectively; correspond- ing minimal diameters are 0.92 mm. for H. melpomene and 0.81 mm. for H. erato. Corre- lated with the smaller eggs are smaller first instar larvae and it is possible that their size allows the caterpillars of these species to pass between the folded blades of the young leaves. The significance of egg size is not clear. It might be related to the habit of laying eggs in batches, /. e., laying a large number at once instead of a few each day. If this is indeed the case, the similar early larval habits may be more a re- flection of adult behavior and physiology than a simple case of larval similarity. The third group contains D. iulia, D. phaetusa and almost certainly P. dido. Feeding and rest- ing orientation, as in the second group, are in relation to the midrib. Bridging and channeling behavior, however, is quite distinct from fur- rowing. In D. iulia such behavior is very con- sistently present and occurs on P. auriculata and P. rubra exactly as it does on P. tuberosa. The leaves of these two vines are furrowed by H. erato and H. sara, however, so that there is no question of channeling behavior being dis- tinct from the furrowing simply because they occur on different plants. Here again the feeding of first instar larvae is fairly distinct from that of the other two groups. D. iulia, P. dido and almost certainly D. phaetusa chew long channels, starting at the margin. This pattern is in fact clearly re- lated to their channeling and bridging behavior which occurs later. Despite the fact that D. iulia and H. erato have been relegated to separate groups here, there are nevertheless a number of similarities in their feeding patterns; both will bend a leaf back and chew it, both eat the last remnant of leaf, then the petiole and often even the stem. The sculpturing of the mandibles is superficially very alike in H. erato, D. iulia and D. phaetusa, the large cusps on the mesial end of the max- illary edge being especially well developed (see Text-fig. 8a, b, c). All these seem to be adapta- tions that might well be produced in species living on vines with few, flimsy leaves and where behavior and chewing apparatus were evolved in relation to minimal wastage of plant material. The feeding pattern of the larvae of the first group, and especially A. vanillae, is considered to be more primitive than those of the others. This view is based on two facts. First, an indiv- idual from the second or third group, when forced to feed without the orienting signal of a midrib, shows a pattern which could easily have been produced by a caterpillar of the first group. Thus H. melpomene placed on the upper surface of a P. lonchophora leaf eats the margin instead of the tip of the leaf, frequently leaves the midrib and between periods of feeding orients relative to the place where it has been chewing, not the central part of the leaf. Second, P. laurifolia has a very well-developed midrib. This feature or the smoothness of the blade might influence a caterpillar towards orienting on the midrib; both together should be more effective. A . vanillae will indeed orient its body along the midrib of a P. laurifolia leaf. How- ever, it neither eats in the specialized straight- across manner of H. melpomene, isolated H. ricini and H. sara, nor does it normally orient to the midribs of other vines. Resting positions on the leaf have already been discussed. The basic pose of the body dur- ing rest seems to be a curve or bend and, with the exception of H. erato and larvae living in groups, the caterpillars of all species show this at some stage. H. isabella is less inclined to do so than H. aliphera, another fact indicating that it is somewhat further along a line of specializa- tion than H. aliphera. There are clear differences between the pos- tures of D. iulia or D. phaetusa, that of H. ali- phera and that of H. melpomene or the solitary H. ricini. The first two take up a distinct “J” position, with only the head and thorax twisted to one side. The bend in a resting caterpillar of H. aliphera is approximately at its second ab- dominal segment, while H. melpomene lies in a straight line with only its head and thorax slightly to one side (see Text-fig. 5b). All species which rest along the midrib, H. erato and in their later stages D. iulia and A. vanillae, hold their bodies straightened out but this seems to have no more significance than that most stems or midribs are straight. H. erato and the later stages of A . vanillae rest in an ex- tended position, but the fact that young A. vanillae rest in a contracted and bent posture suggests that this is more probably a case of convergence than relationship. D. iulia alone produces its peculiar humped-up rest pose, with its characteristically bunched spines, and no sign of this has been seen elsewhere, not even in D. phaetusa. The gregarious species, D. juno, H. doris, H. sara and to some extent H. ricini, usually rest with their bodies laid straight along the sub- stratum though this is less marked the fewer 1961] Alexander: Larval Behavior of Heliconiinae in Trinidad 23 animals there are in the group. In some instances the caterpillars rest with their thoraxes elevated. This usually occurs when animals would other- wise overlap each other, it may appear even when there is no suggestion of crowding and occasionally occurs when a caterpillar rests alone. It seems clear that the resting heliconiine caterpillar has its body bent laterally and that those species which rest in a straight line are the exceptions. However, there is no suggestion concerning the control of the behavior in each individual, for D. iulia will occasionally be found in a “J” pose even when sitting on a straight stem. The anterior part is in this case free in the air. Nor is there any indication of a possible selective advantage of such a bend to the larvae. The range of social behavior does not at all reflect the groupings suggested on the basis of feeding and resting. H. erato is fairly distinct in its sharp intolerance for other caterpillars of any species. On the other hand the remaining species of the group in which H. erato has pri- marily been classified vary between vague tol- erance on the part of H. melpomene to the dis- tinct gregariousness of H. ricini, H. sara and H. doris. The question of social behavior in D. phaetusa and P. dido is largely an open one; if they are at all like D. iulia, they show neither violent aggressiveness nor any signs of gregar- iousness. The third group varies from slight in- tolerance shown by H. aliphera and H. isabella through the tolerance of A. vanillae to very distinct gregariousness in D. juno. On the information available here there is nothing to distinguish the trends towards social behavior seen in H. ricini, H. sara, H. doris and D. juno. There is no indication that these have been independently achieved although this would be inevitable on evidence from feeding and resting. The matter of eating egg-shells (p. 14) has not been taken far enough to serve as more than a pointer to the need for further information. In conclusion, larval behavior as estimated from feeding, defecation, locomotion, resting and weaving is specifically distinct. It seems likely that the differences reflect phylogenetic relationships within the group. If this be so, the present study suggests that H. isabella and H. aliphera are closely related to each other, H. aliphera being nearer the other species of Heliconius. Of these other Heliconius, H. erato is specialized away from H. melpomene, H. ricini and H. sara but shows indications of rela- tionship nevertheless. Text-fig. 8. The biting surfaces of the right man- dible of a, H. erato; b, D. iulia; and c, D. phaetusa, showing the common development of the first maxil- lary cusp and the separation of the maxillary and molar parts. Drawings by F. Waite Gibson. D. iulia, D. phaetusa and almost certainly P. dido form a group which in many ways par- allels that of H. melpomene. Of these D. phae- tusa is less specialized in behavior patterns and shares to a lesser extent similarities which seem to have been independently evolved in H. erato and D. iulia. D. juno and A. vanillae are less alike but on the whole they resemble each other more than they do any other species. In most respects 24 Zoological New York Zoological Society [46: 1: 1961] these two, and more particularly A. vanillae, would seem fairly closely related to primitive helicon iine stock. XII. Summary 1. Larval behavior of 11 of the 14 species of heliconiine butterflies of Trinidad was ob- served in the laboratory and to a limited extent in the field. 2. Periods of feeding alternate with quiescent phases, the extent of each depending some- what on the species but varying in different instars and during any particular instar. 3. Feeding behavior is described in terms of preference for particular species of the food plant, vines of the Family Passifloraceae, the pattern left on the leaves and the move- ments made during eating. The first two differ from species to species but the last is very similar in all species examined. 4. Observations are recorded on how the lar- vae eat egg-shells, drink and defecate 5. Resting is described in relation to position on the leaf and posture of the body. 6. The activity of spinning silk threads is de- scribed, its significance discussed and the different emphasis on such behavior in dif- ferent species pointed out. Slight peculiari- ties of locomotion in some species are men- tioned. 7. The extent to which a caterpillar will tol- erate others of the same or other species is discussed, together with evidence for gre- garious behavior in regard to feeding, rest- ing, molting and defecation. 8. Defensive behavior is briefly mentioned. 9. On the evidence of the behavior described above, the grouping of species is discussed, It does not agree fully with that of the present taxonomy although it is itself con- sistent within its limits. XIII. References Beebe, W. 1952. Introduction to the ecology of the Arima Valley, Trinidad, B.W.I. Zoologica, 37: 157-184. 1955. Polymorphism in reared broods of Heli- conius butterflies from Surinam and Trini- dad. Zoologica, 40: 139-143. Beebe, W., I. Crane & H. Fleming. 1960. A comparison of eggs, larvae and pupae in fourteen species of Heliconiine butter- flies from Trinidad, W.I. Zoologica, 45: 111-154. Bell, T. R. 1920. The common butterflies of the plains of India. J. Bombay Nat. Hist. Soc., 27: 29- 32. Crane, J. 1955. Imaginal behavior of a Trinidad butterfly, Heliconius erato hydara Hewitson, with special reference to the social use of color. Zoologica, 40: 167-195. 1957. Imaginal behavior in butterflies of the family Heliconiidae: changing social pat- terns and irrelevant actions. Zoologica, 42: 135-145. Crowell, H. H. 1943. Feeding habits of the Southern Army Worm and rate of passage of food through its gut. Ann. Ent. Soc. Amer., 36: 243-249. Dethier, V. G. 1937. Gustation and olfaction in lepidopterous larvae. Biol. Bull., 72: 7-23. 1943. The dioptric apparatus of lateral ocelli. II. Visual capacities of the ocellus. J. Cell. Comp. Physiol., 22: 115-126. Fleming, H. 1960. The first instar larvae of the Heliconiinae (butterflies) of Trinidad, W.I. Zoologica, 45: 91-110. Ford, E. B. 1945. The New Naturalist: Butterflies. Collins, London. Bering, M. 1926. Biologie der Schmetterlinge. Julius Spring- er, Berlin. Hinton, H. E. 1955. Protective devices of endopterygote pupae. Trans. Soc. Brit. Ent., 12: 50-92. Merz, E. 1959. Pflanzen und Raupen. Biol. Zentralbl., 78: 152-188. Nagasawa, S. 1957. On the increment of size of faecal pellets following the growth of the gypsy moth, Lymantria dispar L. Botyn-Kagaku Inst. Insect Control. 22: 176-182. EXPLANATION OF THE PLATE Plate I Fig. 1. Resting posture of D. iulia, showing the contracted state of the first abdominal and last two thoracic segments and the conse- quent bunching of the anterior scoli. Pho- tograph by Russ Kinne. ;V»V\ ALEXANDER PLATE 1 FIG. 1 A STUDY OF THE BIOLOGY AND BEHAVIOR OF THE CATERPILLARS. PUPAE AND EMERGING BUTTERFLIES OF THE SUBFAMILY HELICONIINAE IN TRINIDAD, W. I. 2 Hybridization Experiments in Rhodeine Fishes (Cyprinidae, Teleostei). An Intergeneric Hybrid between Female Rhodeus ocellatus and Male Acanthorhodeus atremius J. J. Duyvene de Wit Zoology Department, University of the Orange Free State, South Africa (Plate I) IN an attempt to determine whether any genetic affinities still exist between rhodeine species, a number of interspecific and inter- generic crosses have been performed under laboratory conditions. This paper deals with results obtained by crossing Rhodeus ocellatus (Kner) with Acanthorhodeus atremius (Iordan & Thompson), both of Japanese origin. One male A. atremius and four female R. ocellatus were placed in an aquarium, together with three South African freshwater mussels ( Aspatharia wahlbergi Krauss). Although spawning behavior of these fishes differs slightly, spawning occurred. At short intervals fry were released by the mussels, and 15 arbitrarily se- lected larvae have been raised to the adult stage. A representative specimen of the adult hybrid form is illustrated in Plate I, Fig. 1. All hybrids showed a male phenotype. In the breeding season they displayed full nuptial col- ors. Tubercles were abundantly present on the top of the snout. Spermatogenesis was normal. Their taxonomic characteristics will be de- scribed. in a separate publication. Four of these hybrids were allowed to breed freely with six females of the parental species, R. ocellatus, in the presence of six South African najads ( Aspatharia wahlbergi Krauss and Unio caffer Krauss) . Offspring were produced in the course of the following six weeks. Twenty-five larvae, which had escaped from the mussels during one day, were placed in a separate aquarium and raised to the adult stage. Eighteen of them developed into functional fe- males while the remaining seven specimens showed a male phenotype. The females devel- oped a long ovipositor, and normal spawning be- havior was observed. They were uniform in body size and shape and intermediate between both ancestral forms. A representative specimen is illustrated in Plate I, Fig. 2 (top). The males, however, were not uniform in size. The largest ones attained the body sizes of adult male R. ocellatus and the smallest ones those of adult male A. atremius. Representative specimens showing these extremes are illustrated in Plate I, Fig. 2. In all males, spermatogenesis was normal. The above-mentioned back-cross generation was again allowed to interbreed freely, and an- other interfertile experimental population was obtained. Not full grown but sexually mature female and male specimens from it are illus- trated in Plate I, Fig. 3. It is still too early to decide whether the characteristics of this popu- lation are similar to those of the ancestral stock or not. The taxonomic significance of the present work will be discussed and forthcoming experi- mental populations will be reported in future publications. Summary The successful intergeneric hybridization of female Rhodeus ocellatus and male Acanthorho- deus atremius, both of Japanese origin, is re- ported. All the hybrids were males. The offspring produced by crossing the hybrids back to the female parental species consisted of functional females and males. The females were uniform in body size and shape, and intermediate be- tween both ancestral forms, but the males were not. The breeding of further generations from this back-cross stock is now in progress. 25 26 Zoologica: New York Zoological Society [46: 2: 1961] Acknowledgements The author wishes to express his sincere grati- tude to Professor Tokiharu Abe and Dr. Yoshit- sugu Hirosaki for supplying the species of Jap- anese bitterling, and to Mr. F. G. du Jardin for the photographs of the hybrid specimens and their offspring. This investigation was generously supported by the Council for Scientific and In- dustrial Research of the Union of South Africa. Addendum Since the preparation of the manuscript, the following inter- specific and intergeneric hybrids have been reared to the adult stage: Female Acheilognathus lanceolata Acheilognathus rhombea Acheilognathus tabira 46 14 44 14 Rhodeus ocellatus (Japan) Rhodeus ocellatus (Korea) Rhodeus spinalis Tanakia tanago Male Acanthorhodeus atremius Acheilognathus limbata (Japan) Acheilognathus rhombea Acheilognathus tabira Rhodeus ocellatus (Japan) Rhodeus ocellatus (Korea) Rhodeus spinalis Tanakia tanago Acheilognathus lanceolata Acheilognathus limbata (Japan) Acheilognathus tabira Rhodeus ocellatus (Japan) Rhodeus ocellatus (Korea) Rhodeus spinalis Tanakia tanago Acheilognathus lanceolata Acheilognathus limbata (Japan) Rhodeus ocellatus (Japan) Rhodeus ocellatus (Korea) Tanakia tanago Acanthorhodeus atremius Acheilognathus limbata (Japan) Rhodeus ocellatus (Korea) Rhodeus spinalis Rhodeus ocellatus (Japan) Rhodeus spinalis Acheilognathus lanceolata Acheilognathus limbata (Japan) Acheilognathus tabira Rhodeus ocellatus (Japan) Rhodeus ocellatus (Korea) Tanakia tanago Acheilognathus lanceolata Acheilognathus limbata (Japan) Rhodeus ocellatus (Japan) Rhodeus ocellatus (Korea) Ichthyologists who are interested in the description and pub- lication of the taxonomic characteristics of these specimens, which are new to science, are invited to communicate with the author. EXPLANATION OF THE PLATE Plate I Fig. 1. A representative adult intergeneric hybrid obtained by crossing Rhodeus ocellatus 9 with Acanthorhodeus atremius $. Standard length 48 mm. Fig. 2. Three representative adult specimens of the back-cross generation (see text), one fe- male (top) and two males of the same age but of different size. Standard length: 38, 51 and 39 mm., respectively. Fig. 3. Two representative, not full grown but sex- ually mature, specimens of the “interbreed- ing population” (see text). Standard length of female (top) 25 mm., that of male 38 mm. DEWIT PLATE I FIG. 1 FIG. 2 FIG. 3 AN INTERGENERIC HYBRID BETWEEN FEMALE RHODEUS OCELLATUS AND MALE ACANTHORHODEUS ATREMIUS 3 The Natural History of the Oilbird, Steatornis caripensis, in Trinidad, W.L Part 1. General Behavior and Breeding Habits1,2 D. W. Snow Department of Tropical Research, New York Zoological Society, New York 60, N. Y. (Plates I & II; Text-figures 1-6) [This paper is one of a series emanating from the tropical Field Station of the New York Zoological Society at Simla, Arima Valley, Trinidad, West Indies. The Station was founded in 1950 by the Zoological Society’s Department of Tropical Re- search, under the direction of Dr. William Beebe. It comprises 200 acres in the middle of the Northern Range, which includes large stretches of undisturbed government forest reserves. The laboratory of the Station is intended for research in tropical ecology and in animal behavior. The altitude of the research area is 500 to 1,800 feet, with an annual rainfall of more than 100 inches. [For further ecological details of meteorology and biotic zones see “Introduction to the Ecology of the Arima Valley. Trinidad, B.W.I.,” William Beebe. (Zoologica, 1952, Vol. 37, No. 13, pp. 157-184).]. Contents Introduction 27 Methods 28 Acknowledgements 28 General Appearance, Stance and Locomotion . . 29 Senses 32 General Behavior and Daily Routine 33 Social Behavior 33 The Nest 35 The Eggs 35 Incubation 35 The Young 36 Adaptations to Cliff-nesting 44 1 Contribution No. 1,008, Department of Tropical Research, New York Zoological Society. 2 This study has been supported by National Science Foundation Grant G 4385. Ecological Factors in the Evolution of the Oilbird 44 Summary 45 Literature Cited 46 Introduction HUMBOLDT’S original account estab- lished the main features of the Oilbird’s unique way of life (Humboldt, 1817; Humboldt & Bonpland, 1817). In 1799 he visited the now famous cave near Caripe in the mountains of northern Venezuela. He described how he found it filled with hundreds of scream- ing birds, of the size of a fowl but with the aspect of vultures. He reported that the birds left the cave only at night to feed on the fruits of forest trees, spending all day, and nesting, deep within the cave, where their ear-splitting shrieks and snarls made them seem, to the intruder, more like devils than birds. The scientific name which he chose, Steatornis, marked another memorable feature: that the young birds become exceed- ingly fat; he described how they were collected and boiled down by the local inhabitants to give oil for cooking and for lamps. Humboldt’s two original specimens were lost at sea and it was not until 1834 that the first specimens reached Europe (l’Herminier, 1834). As more specimens became available they at- tracted a great deal of attention from the bird anatomists of the day (especially Muller, 1842; Sclater, 1866; Garrod, 1873; Parker, 1889; see also Wetmore, 1918). These investigations showed that Steatornis is almost certainly closer to the caprimulgiform birds than to any other group, but that even to them the relationship is 27 28 Zoologica: New York Zoological Society [46: 3 very distant, while in certain characters they resemble the owls, perhaps due to convergence. Analysis of egg-white proteins corroborates the relationship to the Caprimulgiformes (Sibley, 1960). For over 100 years far more was known in detail about the Oilbird’s anatomy than about its ecology and behavior. Among a number of accounts of visits to Oilbird caves, most of which added little that was new, mention must be made of the more prolonged visit to the Caripe cave by Funck (1844) , the first naturalist to visit it after Humboldt and Bonpland, and of the observa- tions made by Stolzmann (1880) in Peru, which contain a number of points of great interest (quoted less fully in Taczanowski, 1884). Then Griffin investigated the Oilbird’s method of orientation inside the caves. He showed that they are able to avoid obstacles when flying in pitch darkness, and that they do so by a method of acoustic orientation akin to that of bats, ex- cept that the note given out is easily audible, not supersonic as in bats (Griffin, 1954). More re- cently, Pietri (1957) has given an account of the Oilbird in Venezuela containing some interest- ing observations on the birds’ behavior when feeding. A short preliminary account of the present study has already been published (Snow, 1958). Apart from these, little has been written about the Oilbird in life that is not anecdotal. The remoteness of most of the caves where they live, and the inaccessibility of the nesting ledges, have prevented sustained field study. A small colony of Oilbirds inhabits a gorge near the head of the Alima Valley in Trinidad, about three miles from the New York Zoolog- ical Society’s Tropical Field Station. This colony is the most easily accessible in Trinidad; further- more, the nests are more easily reached than those in any other Trinidad colony, and prob- ably more easily reached than in any colony throughout the bird’s range. The gorge is situ- ated on a private estate and is carefully pro- tected. A further advantage is that a good deal of daylight enters the gorge, which is only par- tially roofed over, and around midday the nests are well enough illuminated for the birds’ be- havior to be easily observed. The present paper is based mainly on observations made at this colony over a period of 314 years. Methods Much of the information gained has come from frequent routine visits to the colony, usu- ally once or twice a week but sometimes daily for short periods. A total of some 250 visits have been made, and they are being continued. At each visit the contents of the nests are checked, a food sample is usually taken, and any necessary weighing, measuring or banding of young birds is carried out. The food samples have been collected in catching trays made of fine wire mesh, slung on the slopes below the nests, and from the nests themselves. From 1958 onwards all the young reared in the cave have been banded. In addition four adults have been caught and banded. No at- tempt has been made, however, to band all the adults, since the handling of an adult bird causes great alarm among the whole colony and makes the birds shy for some time afterwards. The handling of young birds has no such effect. In December, 1957, a platform was erected, spanning the gorge at the same height as the nests and about 15-25 feet away from them. A blind was set on the platform and from it ob- servations were made on the birds’ behavior by day and night. By day most of the birds, accus- tomed to my repeated visits, returned to their nests and behaved normally soon after I entered the hide, though some of them remained aware of my presence. Between 10.00 and 14.00 hrs. all details of their behavior can usually be seen, unless the weather is overcast; before and after this time the light is dim and less can usually be made out. By night much can be learned by listening from the blind, and by occasional quick inspections of the nests with a flashlight. This has been the usual method employed. In addi- tion some observations were made at night by means of a battery-operated infra-red “Snooper- scope,” but technical difficulties have so far limited the success of this method. Acknowledgments I am grateful to several persons and institu- tions for help in this work, and especially to the following: my wife, for help with the field work, particularly in catching the adults and banding adults and young; Mrs. Frederick B. Bang, for details of the Oilbird’s olfactory apparatus; Mr. J. Barlee, for help in understanding the Oilbird’s Sight adaptations; Dr. William Beebe, for notes on the colony in the Arima gorge in earlier years, and for continuous interest and encouragement in the course of the work; Dr. W. G. Downs, for copies of photographs taken in the cave; Mr. J. Dunston, for making routine visits to the colony at times when I was unable to do so; and Professor J. K. Loosli, for analysing samples of the Oilbird’s food. Above all I am indebted to Mrs. H. Newcome Wright, the owner of Spring Hill estate where the colony is situated, for her help and hospitality during the whole of the work. It is through her efforts that this colony, which is very vulnerable to human disturbance, has been able to survive. 1961] Snow: General Behavior and Breeding Habits of the Oilbird 29 This investigation, part of a wider program of studies on the ecology of neotropical birds, has been substantially aided by a grant from the National Science Foundation. General Appearance, Stance and Locomotion Several points about the Oilbird’s general ap- pearance deserve mention. It is a large bird, about 1 8 inches from beak to tip of tail and with a wing span of 3 to 3 Vi feet. The plumage is mainly rich brown with a scattering of white spots that are especially conspicuous on the wing coverts and outer secondaries. The body feathering is short and rather soft. The beak is strongly hooked and the upper mandible is notched on the cutting edge. The gape is very wide and the tongue short. Long vibrissae sur- round the beak and project mainly forward, beyond the tip of the beak. The legs are un- feathered and very short, but not weak; the claws are not strongly hooked. The tail is rather long, ample and markedly graduated. When it is folded the arrangement of the feathers is un- usual: they form in transverse section an acute- angled inverted V. Soon after observations were begun from the hide a slight but consistent color difference be- tween the sexes was noticed. Males are a grayer, slightly darker brown, females paler and more rufous. Funck (1844) also noticed this differ- ence, which has been of value in interpreting the behavior of pairs at the nest. (The sexed mu- seum specimens that have been examined also show this difference, except that a proportion of the males tend towards the female coloring. These may be young birds, or in some cases per- haps “foxed” skins). Wing length is very vari- able in both sexes, but males average larger than females (nine Trinidad males, 307-333 mm., mean 320; eight females, 292-321, mean 307). Oilbirds spend most of the daytime perched on the more or less flat surface of their nests. On such a surface they normally rest with the head held low, the body nearly horizontal but tilted somewhat forward, and the tail pointing slightly upwards. The feet are placed far for- ward, so that the bird appears to be crouching over them (Text-fig. 1; Plate I, Fig. 1). This “down-by-the-head” position is unusual for a bird; it is due to the fact that the Oilbird’s legs are very short and come free from the body at a point rather farther forward than is usual in short-legged birds, while the center of gravity also lies well forward, the pectoral musculature being well developed, the sternum deep and the head large. Under such conditions a stable rest- ing position on a flat surface can be achieved only by straightening out the leg joints as much as possible and rotating the whole limb as far forward as possible, and tilting the body head- downward, so that the breast is just above the feet. In this position the articulation of the femur with the pelvic girdle is well above the center of gravity. The bird thus rests with its weight, as it were, slung between the two more or less upright struts formed by its legs. This arrangement, which as a resting position is per- haps unique in birds, makes it mechanically impossible for the Oilbird to stand on one leg. Ingram (1958), from an examination of specimens, concluded that the short, thick tarsus functions as an integral part of the foot, and that when the bird is perched both lie flat on the substratum. This is not so, however (Text-fig. 1; Plate I, Fig. 1). The tarsus is held within about 30° of the vertical, as in most other birds. Three of the toes point forward while the hallux projects inward approximately at right angles to the line of the body. Movement on the nest, or on any other flat surface, is effected by very short shuffling steps, a method well adapted to prevent the bird from suddenly stepping off the edge. From the mobility of the hallux it has been 30 Zoologica: New York Zoological Society [46: 3 suggested (Ingram, 1958) that when the Oil- bird clings to narrow ledges all four toes point forward, as in the swifts (the pamprodactyl ar- rangement). In fact, in such a situation the toes are spread out more than when perched on a flat surface. The second and third toes point for- ward as usual, while the fourth (outer) toe may be held more to the outside of the foot than usual. The hallux still points inward or some- times even backward (Plate I, Fig. 2). Bock & Miller (1959) have shown how a rather similar arrangement is most effective in enabling wood- peckers to cling to rough vertical surfaces, and for the Oilbird too it probably gives a surer grip on the rough ledges to which it clings than if all four toes pointed forward. For additional sup- port, the tail is fanned a little and pressed hard against the rock face. An Oilbird cannot, however, cling to a ver- tical surface, like a woodpecker on a tree trunk or a swift on a wall. Its claws are not very strongly hooked and its tail feathers are not stif- fened. As Plate I, Fig. 2 and Plate II, Fig. 3, show, when an Oilbird clings to a small ledge or rough slope the feet are held far back, not forward, as in the woodpeckers, swifts and other birds ad- apted for clinging. An Oilbird can only cling in such a place if it can bring its center of gravity inside, i.e., to the cliff side, of its feet; otherwise it would simply fall off. To do this, it must not only hold its feet well back but must also keep its head and breast well into the cliff side, an inefficient method which shows that Oilbirds are not primarily adapted for clinging to rock faces. As mentioned above, Oilbirds are unable to support themselves on one foot. This is clearly seen when they scratch their heads. As in many non-passerine families, the foot is brought to the head directly from below, not from behind the wing. When the bird scratches, it lowers the wing on the same side as the foot which is raised, so that the carpal joint takes the bird’s weight. Occasionally both wings are so lowered. It has been supposed, from the conformation of the tarsus and foot, that Oilbirds are unable to perch in trees (Ingram, 1958). However, Stolzmann (1880) reported that they do so occasionally, and Pietri (1957) gives a detailed account of Oilbirds perching on the bare branches of trees at night. They are also able to alight on quite slender perches. When the birds were suddenly disturbed in a semi-open cave with a top entrance, a few miles east of the Arima gorge, a bird perched for a few moments on a slender woody vine that hung across the cave mouth. Since Oilbirds may be absent from their caves for six hours or more at night (p. 41), it is likely that they make frequent use of their ability to perch on trees. The scene in the well-known illustration in Brehm’s Tier- leben of Oilbirds perched on trees outside a cave apparently by day, though fanciful, is not phys- ically impossible. Aerodynamically the Oilbird is highly spe- cialized for its way of life. Life in caves demands that it should be able to fly very slowly, hover, and turn and twist with agility, all within nar- row confines. Its method of feeding, on bulky fruits often collected far from caves, again de- mands the ability to hover, and also to carry con- siderable weight. I am indebted to J. Barlee for showing how well these demands have been met, and for making the calculations given in Table I. The Oilbird’s wing combines, to a striking degree, low wing-loading with an extremely low aspect-ratio. (Text-fig. 2). Low wing-loading (weight/wing-area) enables a bird to fly slowly, manoeuvre easily and carry large loads. The very low aspect-ratio (wing-span/mean width) enables the Oilbird to achieve the necessary low wing-loading without having a large wing-span. This must be of special importance in negotia- ting the narrow passages of caves. The Oilbird’s wing-loading is comparable to that of a harrier ( Circus sp.) or an owl, both slow-flying birds which carry considerable weights (Table I), while its aspect-ratio is, Mr. Barlee informs me, one of the lowest known for birds of large size. There are further refinements in addition to this major adaptation of wing shape. The wing has plenty of wing-tip slotting, to reduce stalling speed, and is deeply cambered, to give high lift at low speed. The ample tail further improves manoeuverability and gives extra supporting area when the bird is flying slowly and hovering (Plate II, Fig. 4). Barlee suggests that these adaptations may give the Oilbird a flight speed as low as one or two knots, which is in agree- ment with observation. Stolzmann (1880) explained the inverted-V arrangement of the tail feathers as an adapta- tion to hovering. Fie described how, when hov- ering, the Oilbird rhythmically elevates and de- presses the tail, as some hummingbirds do when feeding. He suggested that, owing to the in- verted-V arrangement, the downward move- ment of the tail generates lift, while the upward movement meets with little resistance from the air. I have never seen a bird hovering for long enough, in good light, to be able to see the tail movement which Stolzmann describes. He ap- pears to have been an acute observer and his suggestion deserves consideration. 1961] Snow: General Behavior and Breeding Habits of the Oilbird 31 Table I. Aerodynamic Characters of Oilbird Wing Compared with Marsh Harrier and Long-eared Owl Weight (gm.) Span (cm.) Wing-area (cm.2) Aspect- ratio (span/ mean width) Wing- loading (weight/ wing- area) Pectoralis major/ supracora- coideus Oilbird 415 96.5 1450 6.4 0.29 15 Marsh Harrier ( Circus aeruginosus) 510 124 1820 8.5 0.28 22 Long-eared Owl (Asio otus) 291 95 1270 7.2 0.23 12 High-speed flash photographs show that in slow flight the wingbeat is deep and the upstroke of the wing is propulsive (Plate II, Fig 5), as it has been shown to be for pigeons in rising flight (Brown, 1951). In hovering, too, the up- stroke (which, with the body in a half-upright position, becomes a backstroke) must generate lift as well as the downstroke. Hence it would be expected that the muscles that raise the wing would be highly developed. The supracora- coideus, which is usually considered to be the chief muscle raising the wing, is however quite small (weight 2 gm., compared with 30.5 gm. for pectoralis major) . In this the Oilbird agrees with most other birds of low aspect-ratio. Prob- ably most of the power for the upstroke comes from the deltoid muscles, which are well devel- oped (weight 2.5 gm.) and have a broad attach- ment along half the length of the humerus, rather than from the supracoracoideus, which in addition to being smaller has a mechanically less efficient attachment to the humerus. In the open, at night, the Oilbird’s flight is rather different from its flight in caves. The wingbeat is rapid and shallow. Doubtless the wings are held in a more sweptback position, and flying speed is thus increased by the re- duction in wing area. As in the caves, the flight is quite silent. Stolzmann, who clearly had excel- lent opportunities for observing them by night, saw them occasionally dive down like falcons, with wings half closed. Speed of normal flight in the open has not been determined by observa- tion, but Barlee suggests that it may be about 16 miles per hour. Flight speed may be import- ant ecologically, as it must have a bearing on the time taken to fly to the food trees, and hence on the number of times that the adults can feed the young in the course of the night. The method of feeding is not easy to observe in detail. In the Arima Valley birds have been watched feeding on trees of two kinds, Ocotea wackenheimii and Trattinickia r hoi folia. Invar- iably they have been seen to fly up to the tree, hover, and swoop away a moment later. Text-fig. 2. Outline of Oilbird with wings and tail fully spread. (Traced from a freshly killed specimen). 32 Zoologica: New York Zoological Society [46: 3 Occasionally a quick forward thrust of the head could be seen, as the bird seized a fruit. Pietri’s account of Oiibirds feeding on Persea caerulea is similar. However, Stolzmann saw Oiibirds cling momentarily, with beating wings, when taking fruit from a lauraceous tree (probably Nectandrci sp.), and Ingram (1960) has re- ported that when feeding at a palm (called Sabal, but probably Livingstona) they cling to the bunches of fruit for several seconds. In Ocotea and Trattinickia the fruits do not grow in large compact bunches, as in the palms. It is likely that Oiibirds adapt their methods of feeding to the type of tree, clinging when this enables them to pluck a number of fruit from a single bunch. An analysis of the Oilbird’s food will be given in Part 2 of this paper. Here it need only be said that the fruit taken varies greatly in size, from the small round fruits, about 4 mm. in diameter, of the palm Geonoma vaga to the relatively huge fruits, up to 60 mm. long and 30 mm. wide, of the lauraceous tree Beilsch- miedia tovarensis. All the main fruits eaten are alike, however, in having a single relatively large seed surrounded by a firm pericarp. The peri- carp is digested and the seed regurgitated. Re- gurgitation of the night’s feed is completed by about 09.30 hrs. on the following morning. Senses Although orientation by sonar (Griffin, 1954) takes the place of visual orientation within the caves when the amount of light is reduced be- low a certain point, it is not known to what extent sonar would be effective outside the caves at night; in particular, it is not known how small an object can be detected by this method. This is a question that must be settled by ex- periment. Observation shows, however, that the Oilbird’s eyes are very sensitive to light and whenever possible sight is used instead of sonar, and suggests that sonar is normally never used outside the caves. Thus in the Arima gorge, when a bird is fly- ing toward a dark recess the echo-locating clicks are uttered, but they slow down or stop as the bird wheels around toward a better-lighted part of the gorge. When the birds were watched in the evening leaving the Oropouche cave, several miles east of the Arima gorge, they clicked con- tinuously as they flew down the narrow passage toward the cave mouth and stopped clicking as soon as they emerged into the open. I have never heard clicks from birds feeding at night. It may be noted that the Oilbird’s eyes, though not very large, have a very wide pupil and a tapetum which shines bright red when illuminated with a light held beside the observer’s eye. In locating food it seems likely that the sense of smell is important, though here again experi- mental work is needed. For the following notes on the Oilbird’s olfactory apparatus I am in- debted to Mrs. Frederick B. Bang, who recently examined freshly preserved specimens from Trinidad as well as comparable material from other species of birds. The Oilbird has a rela- tively very large and heavily innervated olfac- tory organ, with one of the thickest mucous membranes of any bird examined. The nasal chamber is beautifully adapted to carry the in- current airstream, after being initially filtered by the respiratory concha and anterior concha, straight onto this mucous membrane. It is of interest that the respiratory (middle) concha is relatively enormous, but the functional signifi- cance of this is uncertain. Such a highly developed olfactory apparatus must surely have an important function. That it is used for locating food trees is suggested by the fact that many of the trees on which Oil- birds feed are spicy or aromatic, in particular members of the families Lauraceae and Burser- aceae. It seems unlikely, however, that the sense of smell can be used for locating individual fruits; for this, sight is almost certainly used. Very many of the fruits taken are green when unripe and turn dark purple or black when ripe. That the birds nevertheless sometimes make mistakes is shown by the fact that the food sam- ples collected in the caves often contain a small proportion of unripe fruits which have been regurgitated undigested. Pietri’s account is par- ticularly significant on this point. A party of Oil- birds watched feeding on Persea caerulea set- tled on trees when the moon was obscured by clouds, and began feeding again when the clouds had passed. Within the caves the sense of smell could per- haps be used for locating the nest, which with the decaying fruit on it usually has an odor per- ceptible to the human nose, but it is unlikely that this is of importance in view of the un- doubted accuracy of the birds’ orientation by sonar. Once the bird is on the nest, the sense of touch probably plays the chief part in orienta- tion with respect to the mate or young, and for this, as Ingram (1958) points out, the long, forwardly-projecting vibrissae are undoubtedly used. It may be noted that the birds themselves have a characteristic musty odor, which may perhaps play a part in individual recognition. Whatever senses are used in the various ac- tivities related to food and to the nest, an ex- tremely highly developed kinaesthetic sense may be postulated. It is therefore of interest that the Oilbird’s cerebellum is unusually large (Bang, in litt.). 1961] Snow: General Behavior and Breeding Habits of the Oilbird 33 General Behavior and Daily Routine At all times, whether breeding or not, adult Oilbirds spend most of the daytime in pairs, perched on their nests. Usually they perch side toy side, facing outwards. For most of the time they are quiet; sometimes they sleep. Occasion- ally the silence is interrupted by an outburst of harsh calls, as perhaps when two birds on ad- jacent ledges engage in a tussle, with beaks in- terlocked, or an unestablished bird tries to land on a ledge near an occupied nest. Toward eve- ning there is an increase in calling and general activity and birds begin to leave their nests and fly around. This period of restlessness lasts an hour or more before the departure from the cave begins. In the Arima gorge the evening departure is difficult to study in detail as there are four ways of exit, up the gorge, down the gorge, and by two top holes. On September 25, 1957, watch was kept in the cave from 17.30 to 19.30 hours. From 18.00 to 18.45 there was great activity which gradually decreased as more and more birds left. At 19.00 the departure seemed to be complete. Inspection by flashlight, however, dis- turbed two birds which flew around for a few seconds and then left. The cave was then empty of adults except for one which was brooding a small nestling. In another watch, on December 24, 1957, most birds had left by 18.15. At 18.30 the flashlight revealed five adults still present, of which four were probably attending nest- lings or eggs. On April 16, 1960, almost all had gone by 18.45, and by 19.00 the cave was empty of adults except for two which were attending nestlings. By contrast with these observations of the departure of a small colony from a cave with several exits, the departure of a large colony from a cave with a single exit takes much longer. A watch was kept at the mouth of the Oropouche cave on the evening of October 25, 1958, a night of full moon. This is a large cave extend- ing back about 400 yards into the hillside, with one rather small entrance hole. The first birds came out at 18.10. By 18.45, 102 had come out, and by 19.40, 62 more. In the next five minutes 11 more came out. Between 19.45 and 20.00 only one more bird came out and the departure seemed to be over. When the cave was entered at 20.05 the flashlight disturbed about 25 birds; almost certainly these were attending eggs or young. The birds had shown considerable hesi- tation in leaving the cave, repeatedly flying up to the cave mouth and turning back before finally coming out into the open. In addition, the narrowness of the exit passage had appar- ently forced the birds to “queue up” to leave and it was this, combined with their hesitation in leaving, that made the departure of the colony so slow. Most of the birds were in pairs as they emerged, with a small proportion single or in threes. The birds seek food as soon as they leave the cave. Almost certainly they fly directly to fruit- ing trees that they already know. Thus the first birds arrived at a favorite food tree about half a mile from the Arima gorge at 18.20, 18.22 and 18.40 on three successive nights in December, 1958, suggesting that they had flown straight to it on leaving the cave. I have not spent an all-night watch in the cave at a time when the birds were not breeding, in order to see whether they return periodically to the cave during the night. When they have eggs or young they return at intervals, as would be expected. In any case, the main return to the cave takes place shortly before dawn. On Feb- ruary 10, 1959, when no birds were nesting, one bird was already present when I entered the cave at 04.30. No other birds arrived until 05.35, when two or three returned. The number of birds present then gradually increased until 06.00, by which time all were probably present. On April 16, 1960, when most nests had young, the final return of the adults took place between 05.10 and 05.40. Oilbirds are occasionally found in the day- time out in the open, sometimes far from a cave. Stolzmann (1880) mentions two such instances; he supposed that the birds had not left them- selves enough time to return to their cave, and having been surprised by the oncome of day- light were waiting for nightfall before resuming their flight. But it seems very unlikely that an experienced bird would make such a mistake, and a more probable explanation is that these are mainly recently independent young birds, which have either failed to find enough food and have become weak or have become separated from their kind and lost. Social Behavior Since nesting activities occupy a large part of the year, and when they are not nesting Oil- birds still spend most of the daytime in pairs on their nests, there is little doubt that the pair bond must normally be permanent. The forma- tion of pairs has not been observed. It is prob- able that the behavior connected with pair formation takes place at night, since it is at night that the birds are active and have the oppor- tunity to meet birds from other colonies. When the birds were watched leaving the Oropouche cave in the evening, twice a trio of birds, on emerging into the open, circled around 34 Zoologica: New York Zoological Society [46: 3 each other, evenly spaced and with a swift, swooping flight. Similar behavior was seen on two occasions when birds were watched feed- ing in the Arima valley. After taking food, two or three birds circled around and around each other with low, clucking calls and a long, harsh “karrrr.” It may well be that behavior of this kind is involved in pair formation. Stolzmann (1880) observed similar behavior in Peru and interpreted it as courtship. In the daytime, one member of a pair some- times preens its mate’s head as they sit side by side on the nest. This behavior has only been seen just before the eggs are laid or during the laying period, except in one pair in 1959 which laid no eggs in that year (or lost them as soon as they were laid). The preening bird, with its eyes closed, works carefully over the other bird’s head. The latter keeps its eyes open or only half-closes them. In nearly every case when this behavior has been seen, the preening bird has been known or presumed to be the male. The exception was a single instance when the presumed female, having been preened by its mate several times, was seen tentatively to preen its mate’s head. This behavior is certainly a form of courtship and, from the times when it has been observed, must be closely associated with the period of copulation, but copulation itself has not been seen. Relations between neighboring pairs are gen- erally harmonious. When a bird alights on its nest, it sometimes provokes an outburst of ex- citement and calling from its neighbors, but such outbursts are short-lived. Occasionally, for no apparent reason, neighboring birds spar with their beaks and engage in tussles, gripping each other by the beak and twisting and pulling, with harsh calls, sometimes for several minutes on end. More prolonged fighting occurs when an apparently unestablished bird tries to secure a foothold on a ledge near an occupied nest. On May 10, 1959, two birds were watched trying repeatedly to secure a foothold in the same place, on a steep slope just below an occupied nest. Both were repelled. One of these, which was banded, was a young bird fledged in August of the previous year. These conflicts have been seen only shortly before, or during the early part of, the breeding season. That there is some sort of cohesion between the adults of the colony when they are out at night, apart from the cohesion of the pairs, is apparent from observations on the times of feed- ing of the young which will be described later. Except for three nests in which the young were small and were fed more often, returns of the adults with food were concentrated into three main periods of about 20 minutes each, spaced about two hours apart. During these periods each nest, as far as could be ascertained by listening, was visited by one or two adults and the young were fed. This could only have been possible if the adults were keeping company with each other while collecting food. The few observations that have been made on their feeding behavior also give evidence of a strong social tendency. When Oilbirds have beerr watched feeding at night in the Arima valley, up to five birds have arrived at the food tree within a minute, and after feeding for several minutes have departed within a short time of each other. Pietri (1957) also refers to their social behavior when feeding and describes how, when one bird of a feeding party was shot, the others called and swooped down low over the dying bird. Except for the echo-locating clicks, the signifi- cance of the Oilbird’s various calls has not been elucidated. When the birds are disturbed in a cave, the noise can be almost deafening; the calls range from clucks and rather low-pitched “hawk- ing” sounds, reminiscent of the last of the bath- water going down the drain, to long-drawn-out, harsh screams. The calls made by aggressive birds on their nests, when another bird ap- proaches, are similar. Birds flying in the open at night sometimes utter a shorter, less harsh “karr, karr” or “kuk, kuk,” which is probably used in maintaining contact with other indivi- duals. As Griffin (1958) points out, these is no sharp distinction between some of these shorter calls and the longer bursts of echo-locating clicks. Whatever their signal function may be, the acoustic qualities of these calls are well suited to the conditions under which they are uttered. Many Oilbird caves are full of the noise of run- ning water or breaking waves, so that, as for cliff-nesting sea-birds, very loud calls are essen- tial. The harsh guttural quality of the calls, de- pending on a rapid succession of staccato sounds of many different frequencies, probably makes it easy for the other birds to detect the position of the calling bird (Marler, 1955). By uttering long-sustained calls, birds flying in the confined space of a cave can make their course known to the other birds. Thus social contact between the individuals can the more easily be maintained. The loud harsh calls are, however, not mere traffic signals. Purely for the avoidance of mid- air collisions, the echo-locating clicks are suffi- cient. Thus large numbers of Oilbirds can fly together in pitch darkness, uttering only clicks, but any disturbance will at once elicit a chorus of screams and snarls. 1961] Snow: General Behavior and Breeding Habits of the Oilbird 35 The Nest As already mentioned, adults occupy their nests all the time, whether they are breeding or not. The nests in the Arima gorge are placed on narrow ledges, 8 to 15 feet above the stream bed. There are no suitable higher ledges. In other caves nests are normally much higher, partly be- cause the caves themselves are much larger and partly because frequent raids on the caves have caused most of the lower and more accessible ledges to be abandoned. The nest has a diameter of approximately 1 5 inches, with a shallow central depression and a slightly raised rim. It seems at first sight to be made of mud, and has been so described. It is, however, made primarily of regurgitated matter. When nest-building, the bird moves its head around the rim of the nest and with quick jerky movements plasters on semi-liquid matter which it allows to exude from the side of its beak. From the firmness of the resulting structure, it seems probable that saliva is important in bind- ing together the regurgitated pulp, but this point needs investigation. The central part of the nest is formed from the accumulation of regurgi- tated seeds which the birds let fall; nest-building behavior does not, as far as I have seen, include work on any part except the rim. The faeces of the adults contribute little or not at all to the structure, for when defecating they turn and face inwards and shoot the faeces well clear, like other cliff-nesting birds. The young, too, turn when defecating but until they are well grown the faeces are usually deposited on the nest edge, where they contribute a little to the structure. As nests are used year after year, they grow into low cylindrical mounds. Three of the nests in the Arima gorge have fallen away during the period of observation, as they became too large for the small ledges on which they were based. Two have been partly rebuilt. Parts of other nests have fallen away and been rebuilt. It is presumably by such a process of falling away and rebuilding that year after year the nests re- main more or less the same size. The birds build up and repair the nest rim most actively in the few weeks before egg-laying begins, but the behavior also occurs when there are eggs and young. The Eggs The eggs are white ovals, slightly pointed at one end. The surface of the shell is slightly rough. The average weight of ten eggs, weighed soon after laying or early in incubation, was 20.2 gm. (range 17-22.5 gm.). Most eggs become spotted and blotched with brown soon after laying; this has led to erroneous statements that the Oilbird lays spotted eggs. The normal clutch is 2—4 eggs. (Clutch size will be dealt with more fully in Part 2 of this paper). Although several daytime watches have been made at times when the birds were laying, there has been no record of an egg being laid during a period of observation. It seems probable that when about to lay, the female remains behind and lays her eggs before going off to feed. The interval between the laying of successive eggs is unusually long and very variable. Daily visits to the colony over a period of up to two weeks may be necessary to ascertain the inter- vals between the laying of eggs in only one nest. As nests are not always well synchronized, daily visits over several weeks would be necessary in order to obtain exact information for the whole colony. This has not been possible, with the con- sequence that data on this point are fragmen- tary. The most accurately ascertained intervals between the laying of successive eggs were 2-4, 5, 6-7 and 6-7 days. In addition, minimum inter- vals of 6, 7 and 9 days were recorded. Between the laying of the first and third eggs in a clutch, the following intervals were recorded: 6-8, 7-9, 9, 9, 8-10 and 9-11 days. Other less exactly re- corded intervals were consonant with these. Incubation The eggs are covered from the time they are laid. Complete clutches were never seen to be left uncovered (except after the birds had been frightened off the nest), but there were two ob- servations of the eggs in incomplete clutches being left uncovered for 4 and 10 minutes while the parents perched on the edge of the nest. Both sexes incubate the eggs. The eggs lie very far forward under the incubating bird, between the chin and the legs, the whole rear half of the bird being slightly elevated. When the bird is re- laxed the head is withdrawn between the shoul- ders and the eyes may be closed. The other bird, standing beside its mate, is usually alert, with its head held forward. The lengths of the turns taken on the eggs are usually long, but very variable. Thus in a watch lasting three or four hours in the middle of the day, when the light is good enough to see de- tails, only one or two completed spells may be observed. At nests where the sexes were known, females were recorded incubating for just over twice as long as males (1,129 as against 504 minutes), but this was due almost entirely to one nest, at which the female alone was seen to incubate. At the other nests the total time spent by the male and female on the eggs was nearly equal, and the few completed spells recorded for the two sexes were of similar length: males, 4, 36 Zoologica: New York Zoological Society [46: 3 32, 96 and 109 minutes; females, 7, 29 and 95 minutes. These spells are shorter than average, as the majority of the longer spells overlapped the beginning or end of the watch and so their complete length was not known. The change-over is effected silently and with- out ceremony. Usually the incubating birds gets up off the eggs and shuffles to one side, while the other bird takes its place. Shuffling around by one or both birds may continue for several min- utes before they settle down and become still again. Sometimes the non-incubating bird takes the initiative by becoming restless, shuffling about, and perhaps inserting its head beneath the incubating bird as though to ease it off the eggs. At the nest where only the female was seen to incubate, the male periodically became restless. Once after apparently trying to rouse the female from the eggs he performed nest-building move- ments, but without regurgitating any material, probably a displacement activity. Once he spent most of the watch on another ledge a few feet away from the nest. Two experiments suggest that the Oilbird’s egg-retrieving behavior is very poorly developed. One egg of a clutch of three was moved four inches from the other two, towards the edge of the nest. The male soon returned and incubated. Sitting on the two eggs, he looked at and occa- sionally touched the third egg with his beak but made no attempt to roll it back. After about a minute he sat quietly, ignoring the third egg. A few days later at the same nest one egg was moved three inches from the other two. The male again returned to incubate. When he had settled down the third egg was lying about one inch in front of him. He gradually shuffled it under- neath him by moving forward himself a little and touching the egg with his beak so that it rolled a little. When the egg was nearly touching him he finally poked it underneath him. The whole process took three minutes. Nests are usually surrounded by regurgitated seeds but these do not usually remain in the cen- tral depression of the nest with the eggs. This suggests that the incubating bird moves them, though this has not been seen. Sometimes, how- ever, as many as three of the seeds of the palm Jessenia oligocarpa lie with the eggs. They are easily the largest of the seeds regularly taken by the birds in the Arima gorge and are apparently near enough in size to be accepted as eggs when they are regurgitated into the nest. The eggs hatch at approximately the same intervals as the intervals between laying. For a day before the egg hatches the young bird can be heard cheeping inside. It emerges from the shell by cutting a circular cap from the broad end of the egg. The hatching process is quick; a young bird that was just starting to chip the shell at 15.30 hours was fully hatched at 17.00 hours. There were several instances of chipped eggs be- ing hatched by the time of the next visit 24 hours later, and no cases of prolonged hatching. The broken eggshells are not cleared away promptly by the parents; they often remain on the nest for a day or two before, like the regurgi- tated seeds, they are pushed, or perhaps picked up and dropped, over the edge. During a watch of nearly four hours an adult brooded the newly- hatched young with half an eggshell lying beside her; several times she fumbled with the broken shell but made no attempt to remove it. The incubation period (measured from the time of laying to the time of hatching) is 32-35 days. Table II gives the most accurately deter- mined periods for a number of marked eggs. In most cases there is a possible error of one day or a day and a half either way, as it was not usually possible to visit the colony frequently enough to give greater exactitude. It will be seen that there are no consistent differences in incu- bation period between eggs of different position in the clutch, which indicates that the eggs are not merely covered but effectively incubated from the time they are laid. Table II. Incubation Periods Egg 1 1 Egg 2 I Egg 3 | Egg 4 32 (±1) 321/2 (±li/2) 32 (±1) 34 (±1) 3 3 1/2 (±1£) 341/2 (±11/2) 33 (±1) 33!/2 (±l‘/2) 341/2 (±11/2) 331/2 (±1/2) 34 341/2 (±1/2) 351/2 (±1/2) The Young The most striking feature of the development of the young Oilbird is its extreme slowness. Young Oilbirds do not usually leave the nest un- til they are between 95 and 120 days old. During this time they lay down the great deposits of fat which have led to their being exploited for oil, attaining around the age of 70 days a weight much greater than that of the adult, and then losing weight for the last 30-50 days as their feathers develop. Growth and Development. —The young bird at hatching weighs from 12 to 15.5 gm. It is naked except for some sparse down, chiefly on the un- der side. (Text-fig. 3). The amount of down at hatching varies; it is always thickest on the un- der surface, while on the back and flanks some 1961] Snow: General Behavior and Breeding Habits of the Oilbird 37 Text-fig. 3. Distribution of down feathers on newly hatched Oilbird. Each dot represents one feather. Areas shown in detail in left-hand figure (ventral view) are outlined by broken line in right-hand figure. Area enclosed by broken line behind thigh in right-hand figure contains many feather rudiments visible below skin, but down sprouting only where shown. (Specimens vary individually). birds have a little down at hatching, or very soon after, while in others it does not burst through the skin until a few days after hatching. These first down feathers are short, pale gray and high- ly branched. In the second week after hatching a second generation of down feathers appears as black streaks beneath the skin. These second down feathers, which are darker gray and much longer, come from the same rudiments as the first and bear the first on their tips as they begin to break through the skin in the third week. Be- neath some of the first down feathers on the ven- tral surface, especially towards the posterior end, no black streaks appear, and these down feathers are not succeeded by any others. The first feathers of adult type, those of the tail, secondary coverts and scapulars, burst from their sheaths at the age of about 35 days. There- after the feathering of the wings and tail, head, back and underparts, in that order, grows stead- ily. By the age of about 70 days the nestling is quite like an adult, except that the wings and tail are very short and the body is still mainly downy below. It is noteworthy that there is not only no juvenile plumage but also no juvenile appear- ance of the head and beak. The nestling, when ready to leave the nest, is indistinguishable from the adult. Being adapted to complete darkness, the nestlings are without the visual signs which elicit parental behavior in other birds. In Table III the main changes in the appear- ance of the nestling are tabulated. Because of the very long development, an accurate knowledge of these changes is necessary if the breeding sea- son is to be dated from a single visit to a colony with eggs and young. It will be noted that there is great variability in the state of development of different birds of the same age. Text-figs. 4 and 5 show the growth in weight of a single nestling and of a family of three, with the wing lengths in the last two or three weeks. It will be seen that the young birds leave the nest at the time when the decreasing weight and the increasing wing length have simultaneously reached the adult values. The Nestling Period — Whereas small young are sometimes restless during the day, and even beg occasionally (see next section), large young are inactive. Until very shortly before they leave the nest they usually show no tendency to fly when disturbed or even handled, at least during the daytime. (The one exception, a bird 109 days old, flew quite strongly when I rather awk- wardly tried to turn it in the nest in order to read its band number) . This is probably due in part to their being less active during the day, but also. 38 Zoologica: New York Zoological Society [46: 3 300 250 2 200 s Text-fig. 4. Growth in weight and wing length of a single nestling Oilbird. The broken horizontal line in the weight diagram shows the mean adult weight (415 gm.). The two broken horizontal lines in the wing length diagram show the limits of adult wing length. The vertical dotted lines indicate the period within which the nestling left the nest. - 300 - 250 z 200 S AGE OF FIRST-HATCHED NESTLING (DAYS) Text-fig. 5. Growth in weight and wing length of a family of three nestling Oilbirds. Conventions as in Text-fig. 4. and more importantly, to their being accus- tomed to frequent harmless disturbances. Thus when a boatman climbed up to some nests in a sea cave where the birds are still exploited by the local people, two nearly fledged birds left their nests and fluttered down into the sea. For the first few days after the young birds have left the nest they sometimes return to them by day, but Snow: General Behavior and Breeding Habits of the Oilbird 39 1961] Table III. Development of Nestling Oilbird Day Description:1 feather development Weight (gm-) Weight Limits Wing (mm.) Wing Limits 1 Nearly naked above, short sparse down below; eyes closed; squeaks when handled 12 12-15.5 5 As before, but larger 22 20-30 — — 10 Wing feather rudiments visible as dark points (apparent on day 8 ) 40 30-75 _ 15 Vibrissae sprouting; rudiments of second generation of down feathers as short black streaks on body; eye slits beginning to open; bird about AVi inches long 65 50-120 24 Down feathers bursting out all over body; eyes open but not widely 145 100-230 . ___ 30 Down prominent all over body; bird about 6V2 inches long 230 120-300 _ 35 Secondary coverts and tail feathers just beginning to sprout from sheaths 275 160-440 40 Tail feathers about 5 mm., secondary coverts 3-4 mm. beyond sheaths; primary coverts sprouting 330 240-500 _ 50 Head feathers well out of sheaths (burst about day 47); scapulars, wing coverts and tail feathers forming al- most complete covering of upper side; whole of body still downy 450 290-570 60 Body still downy; feathers growing on throat 580 390-610 — — 70 Back well feathered; feathers growing on under side of body 595 560-650 190 80 Appearance like adult, but wing and tail short. 595 550-650 240 160-260 90 Appearance like adult 565 520-600 280 200-300 100 Appearance like adult 490 420-550 300 240-305 1 Description is based on a nestling whose rate of development was average. Weight limits and wing limits are limits for all nestlings which fledged successfully. as soon as they are disturbed they fly off at once. Thus the fledging period can be measured quite exactly as the interval between hatching and the time when the young bird first voluntarily leaves the nest. Only a few periods could be ascertained exact- ly. Usually they could be determined only within a few days, owing to the fact that it was rarely possible to determine both the hatching date and fledging date exactly. All fledging periods ascertained within limits of eight days or less are given in Table IV. There is a great difference between the shortest, 88 days, and the longest, 125 days, but most (71%) fall within 100 and 115 days. There is a tendency for the youngest member of a family to have a longer nestling period than its nestmates. This was certainly in some cases, and probably in all, because their early development was slowed down through competition for food with the older nestlings. Text-fig. 5 shows an example of a family in which this happened. The full length of a single nesting, from the laying of the first egg to the fledging of the last young, is commonly round 150 days. The long- est recorded, for a family of four all successfully reared, was 168 days. Behavior of the Young— For about the first 25 days after hatching, the young are brooded by the parents. They lie with their heads under the wing or breast of the parent bird; as they be- come larger the head often protrudes from be- tween the parent’s wing and body. They fre- quently thrust their heads upward toward the parent’s neck, body or wing, apparently seeking contact; having gained it, they will stay motion- less with the neck awkwardly kinked and the beak pointing upward. Occasionally they make food-begging movements, thrusting their head up at the parent’s beak and nibbling at it. But, except perhaps when they are very small, the young are not usually fed by day (see next sec- tion). The call of the young at hatching, and for a day or two before hatching, is a high-pitched cheeping. Later, by the age of 20 days, it devel- ops into a loud, rather hoarse squeak, which be- comes louder as the chick grows older. Large nestlings, when begging, utter a chorus of shrill but rather hoarse squeaks. By the time they are 40 Zoologica: New York Zoological Society [46: 3 Table IV. Fledging Periods Position in Family1 (1) (2) I (3) (4) Family of 4 111 (±4) 112 (±2*4) 110 (±3*4) 125 (±2) Families of 3 114 (±2) 111 (±3) 125 104 (±31/2) 100 (±3*4) 104 (±4) 98 (±3*4) 98 (±2) 106 (±2) 109 (± 3) 1092 114 (±2) 93 (±2) 99 (±1) 103 (±2) 99 (±2) 104 (±3) 110 (±1*4) (Died at 62 days) 112 (±2*4) >116 Families of 2 119 (±3*4) 121 (±3*4) 108 (±2) 115 (±3*4) 101 (±2Vi) 102 (±2) (Died at 42 days) 102 (±2) Families of 1 102 (±4) 112 (±2) 100 (±3*4) 88 (±3) 1 In families in which one of the nestlings died in the first few days after hatching, this nestling has been left out of consideration in placing the other nestlings in their positions in the family. 2 This bird flew on being disturbed (see text). well feathered they begin to utter, if alarmed, the harsh screams of the adult. They begin to preen themselves at the age of about 20 days. From the age of a few days, when defecating they turn, back towards the nest rim, and deposit the faeces on the edge of the nest. When they are larger the faeces are shot clear of the nest edge, as in the adult. For about the first 50 days after hatching, the young rest with the lower surface of the body, the tarsus and the foot in contact with the sub- stratum. Toward the end of this period they are able to raise the body clear of the substratum when shuffling about the nest. Later they begin to stand with the body clear of the nest, their weight supported only by the tarsus and the foot, and finally, by the age of about 75 days, they can stand on the foot only, with the tarsus held at an angle of 45-60°. Advanced young can clamber efficiently up quite steep slopes. To do this they not only grip and push with their feet but pull themselves up with the beak and dig in the leading edges of the wings. This behavior is of obvious value in enabling them to regain the nest if they are accidentally pushed out. They are very conservative in the position which they occupy in the nest. If taken out and replaced in different positions they shuffle and clamber over each other until they have regained the old posi- tions. Parental Behavior.— As already mentioned, for about the first 25 days the young are brooded by the parents. Usually only one bird at a time cov- ers the young, but at one nest with four young both parents were seen to cover them for part of a watch. The number of young probably af- fects the length of time for which they can be brooded; at two nests single nestlings were brooded by day at the ages of 29 and 30 days, a longer time than was recorded for broods of two or more. At night one adult stays with the young while they are small; larger young are left by both parents. Thus during evening watches one adult remained on each of four nests with young 3-6, 12-18, 20 and 30-40 days old, while both parents departed from six nests with young 49 days old or more. Many hours of watching have shown no evi- dence that the young are ever fed by day, except perhaps when they are very small. Very small young sometimes become restless, thrust their heads up jerkily toward the parent’s head, and utter the food-begging call. Occasionally the par- ent has then been seen to lower its beak towards the nestling and itself make slight jerky move- ments. On such occasions some semi-liquid food may be passed to the young, but the adult’s posi- tion, crouching over the young chick with low- ered head, makes it almost impossible to see the details. From the 12th day, and perhaps earlier, the young are fed, at least partly, on whole un- digested fruits. Observations on the feeding of the young were made during an all-night watch on April 16-17, 1960. On this night four nests, all ad- jacent to one another, contained three young each, aged from 49 to 58 days, while three other 1961] Snow: General Behavior and Breeding Habits of the Oilbird 41 nests contained, respectively, three young 12-18 days old, one young 20 days old and four young 30-40 days old (Nest K). When the evening departure of the adults was over, at about 19.30 hours, one parent remained at each of the three nests with smaller young, while both parents had gone from each of the four nests with large young. The first feeding was at Nest K, with four young 30-40 days old. It began at 20.51 and continued, with pauses, until 21.18. The first feed at the other two nests with small young was at about 21.00, but as these nests were more distant from the hide, and very close to each other, further detailed observations were not made on them. Although the darkness was total, it was easy to tell when a family was being fed. The adult on arriving would fly around for a little time, its position being shown by the echo- locating clicks. As it approached the nest to land the clicks would become more rapid and then suddenly cease as the bird landed. At once there would be a shrill chorus of squeaks from the chicks, which would continue at greater or lesser intensity while the feeding lasted. The other nests where no feeding was taking place would by contrast be almost or completely silent. At 21.20, two minutes after the feeding was over at Nest K, inspection by flashlight showed that both parents were still present. At 22.00 only one was present. In the course of the night there were five more bouts of feeding at this nest, and perhaps a sixth: at 22.13-22.27, 23.33- 23.44, (00.05-00.06, not certain, and in any case very brief), 01.29-01.37, 03.44-03.50 and 05.22-06.15. At the four nests with large young, feeding began much later. There was one short feed at one nest only at 23.02-23.06, after which the parent departed again. Nothing further hap- pened till 01.35 when a great burst of feeding activity began and lasted until 02.05. During this period there were at least seven landings by adults on the nests, followed by outbursts of begging calls. At 02.26 there was a single land- ing followed by a short feed. There was then over an hour without activity. The second main feeding period was from 03.42 to 04.06, when there were at least six, and probably eight, land- ings by adults followed by bursts of begging calls. There was a minor feed at 04.42-04.43, when only two adults landed, and a final main feeding period beginning at 05.35 and continu- ing until dawn. During this feeding period at least five adults landed. Thus there were three main feeding periods in the night, during which all or nearly all of the parents brought food, 01.35-02.05, 03.42-04.06 and 05.35-06.15, and three minor feeds by single birds or two birds. at 23.02, 02.26 and 04.42. The total number of recorded landings followed by feedings was 21 or 23. Probably one or two others were missed. Thus each of the eight adults attending the four nests brought food on average about three times during the night. Probably, with some ex- ceptions to account for the three minor feeding periods, each bird brought food once during each of the main feeding periods. The fact that the main feeding periods were synchronized at the four nests with large young (and also, though less well, at the three nests with smaller young) strongly suggests that the adults were foraging in company. Presumably they were feeding themselves during the six hours after they had left the cave and before the first main feeding period. As it grew light, at 06.00, the last feeding was still in progress and it was possible to see the birds at the better-illuminated nests. At Nest K all four young were seen craning their heads up toward one of the parents, squealing shrilly. When being fed, the chicks half-turn their heads so that their beaks interlock with the adult’s beak. The shrill begging calls cease abruptly at the moment the beaks interlock. Neither beak is opened very wide. As the adult regurgitates the food, its head and that of the nestling with it moves in short quick jerks. At the nests with large young the feeding was most vigorous; in the half-light these nests appeared to be filled with a heaving mass of birds. When the light improved it was possible to see that as adult and young both pushed strenuously during the feed- ing, with beaks interlocked, they reared up to- gether with the head of the chick pointing obliquely upward and that of the adult down- ward. The adults were clearly under great phys- ical strain. They could be seen pushing with their feet and would sometimes flap their wings to avoid falling backward. In this attitude both parent and chick would rear and heave together for a minute or more. It could be seen that there was keen compe- tition between the chicks of each family for the attention of the parent with food. At the only nest with four young, feeding continued longer than at the other nests, and the smallest chick of the four continued to beg for several minutes after the others, when the parents apparently had no more food left. Competition for food probably accounts for the slow early growth of the last-hatched nestling, mentioned earlier, and for the occasional death of small chicks, but there has been no evidence that nestlings have suffered from shortage of food in the later stages, when they need far more. The Food of the Young— During the first few 42 Zoologica: New York Zoological Society [46: 3 days after hatching, semi-digested food is prob- ably given to the young, as nestlings up to the age of ten days have occasionally regurgitated fruit pulp but not seeds. Also, as already men- tioned, adults sometimes appear to pass semi- liquid matter to very small young during the day. Later, whole fruits are fed to the young. The earliest age at which a nestling, when handled, has regurgitated a seed is 12 days. A nestling 15 days old, when taken from the nest after its last feed at dawn, regurgitated four seeds in the course of the day, almost certainly too small a number to represent the whole of its last feed. Thus the change from a pulp to a whole-fruit diet is probably gradual. A full analysis of the Oilbird’s food is reserved for Part 2 of this paper. Here only a few special points will be mentioned. Except that they are not given very large fruits, such as those of the palm Jessenia oligocarpa, the nestlings are fed on the same fruits as the adults themselves eat. The chief of these, during the periods when the food of the young has been studied, have been the palms Euterpe oleracea and Badris cuesa; the Lauraceae Ocotea oblonga, Phoebe elongata and one unidentified; the burseraceous trees Trattinickia rhoifolia and D aery odes sp. ; and an unidentified, probably myrtaceous tree. Like the adults, the nestlings digest the pericarp and usually regurgitate the seeds. Very small seeds, however, may be either regurgitated or passed through the intestine. Most of the seeds of the night’s feed are regurgitated by midmorn- ing. Several times nestlings have been removed from the nest soon after their last feed, kept for the day and returned to the nest in late after- noon. Of the total of 428 seeds regurgitated by these nestlings, 306 (71%) were regurgitated before 09.00 hours, and all except 12 (97%) by midday. On May 14/15, 1960, an attempt was made to estimate the amount of food given to a nest- ling in the course of the night. Two nests were cleared of all regurgitated seeds in the evening. Next morning three nestlings were taken from these nests at dawn, immediately after the last feed, and all the freshly regurgitated seeds lying by their beaks (hence almost certainly not re- gurgitated by the parents) were collected. The results, given in Table V, show that each nest- ling received approximately one-third or one- quarter of its body weight. Stolzmann (1880) kept a nestling Oilbird for about three weeks. After feeding it at first on various kinds of unsuitable food he was able to obtain fruit of a Nectandra sp. (Lauraceae), one of the Oilbird’s chief food trees in Peru. From his description the bird was then about 70 days old. Two experiments, made on dif- ferent days, both showed that the bird was able to eat 14 fruits at a time, and that it regurgitated the first seed half an hour after it had eaten and the last seed one hour after it had eaten. This bird was almost certainly undernourished, which may explain the very short time taken for re- gurgitation compared with the nestlings studied here. At the same time his experiment shows that regurgitation can begin very soon after the food has been eaten, and suggests that under natural conditions at least a proportion of the seeds from the first feeds of the night will be regurgitated before dawn. This is confirmed by the present study. About one-third of the seeds regurgitated by the three nestlings in Table V had been regurgitated before they were taken from the nest. On the seven further occasions when nestlings have been taken from the nest at dawn (without earlier clearing of the nests and collecting of the fresh seeds regurgitated before dawn), the seeds regurgitated in the course of the day have never represented a feed of more than one-sixth of the nestlings’ weight. These young birds that have been removed from their nests have regurgitated a considerable number of whole, undigested fruits, the pro- portion varying between individuals and accord- ing to the kind of fruit. Fruits of the palm Euterpe, with a rather hard pericarp, have been regurgitated whole much more often than any other kind. Proportionally more fruits have been regurgitated whole early in the morning than later. Though the disturbance of being removed Table V. Amount of Food Eaten by Nestling Oilbirds in the Course of a Night Nestling Age in Days Weight of Nestling (gm.) Number of Fruits Eaten Total Weight of Fruit Eaten (gm.)1 Oldest of 3, Nest F 56 525 86 126 Oldest of 4, Nest K 40 350 73 103 Youngest of 4, Nest K 30 120 20 35 1 Total weights of fruit calculated from the following mean weights of individual fruits: Bactris cuesa, 1.9 gm.; Euterpe oleracea, 1.3 gm.; Ocotea oblonga, 0.6 gm.; Dacryodes sp., 2.8 gm. 1961] Snow: General Behavior and Breeding Habits of the Oilbird 43 from the nest may cause some premature re- gurgitation, this is probably not the full reason for the regurgitating of whole fruits, since col- lections of food in catching trays below the nests regularly contain a proportion of whole fruits, especially during the seasons when the young are being fed. Temperature Control and the Fat Deposits.— Small young feel cool to the touch after they have been left uncovered for a few minutes, and they rapidly become cooler. In order to study nestling temperatures, cloacal tempera- tures were taken with a quick-registering mer- cury thermometer as soon as possible after arrival at the colony (Text-fig. 6). Cloacal tem- peratures of 31.8° to 35.0° C. were recorded for nestlings up to six days old, while three nestlings 15, 16 and 20 days old had cloacal temperatures of 35.2°, 35.8° and 36.2° respec- tively. For older nestlings temperatures of 37.0° and over were recorded, most being between 39° and 41° from the age of 40 days onwards. (One very low reading for a nestling 68 days old may have been due to the thermometer lodging in a mass of faecal matter on the point of being expelled). The ability to maintain body temperature appears to be acquired at the age of about three weeks. Thus the temperature of four young aged 2-6 days fell at the rate of between 1.3° and 2.3° in 10 minutes after being uncovered, while that of three young birds 15, 16 and 20 days old fell at rates of 1.0°, 0.6° and 0.3° respec- tively, and temperatures of older nestlings have usually not fallen appreciably during exposures of up to half an hour. Air temperatures at the nests are around 22° C. at midday, falling to about 18° at night. In addition to the acquisition of temperature control, the age of about 25 days marks three other important and related changes in the life of the young Oilbird. The down feathers are bursting out all over the body (Table III). The young bird’s weight is increasing rapidly (Text- fig. 4), probably due to a relative increase in fat deposits as well as to an increase in over-all dimensions. At the same time the parents are ceasing to brood the young bird by day, and at night they are beginning to leave it for several hours while they are out foraging. There is little doubt that the thick down is important in enabling the nestling to maintain body heat, and it seems probable that this is also an important function of the deposits of fat. Deposits of fat in other young buds with slow development, especially Tubinares and swifts, have usually been considered to be re- serves against periods of food shortage. These are birds in which the ability to find food is greatly dependent on the weather, and even when conditions are favorable the parents may have to travel long distances in obtaining it. For the Oilbird, however, there is no evidence that the nestlings are liable to undergo periods of fasting. The indications have been that food has been consistently abundant, and the ability of the adults to find the food does not seem to be much affected by weather. Throughout the breeding season in each year, as already men- tioned, a proportion of the fruits brought to the nests have been regurgitated intact and dropped over the edge of the nests. Text-fig. 6. Cloacal temperatures of nestling Oilbirds. 44 Zoologica: New York Zoological Society [46: 3 Adaptations to Cliff-nesting Cliff nest-sites are safe from most predators; indeed it is because of this that natural selection has favored their use in various groups of birds. If they are also in darkness, they are of course completely safe from visual predators. But suit- able cliffs are not numerous, whether in caves or not. Hence cliff-nesting birds tend to defend their nests jealously once they have gained pos- session of them, but at the same time tolerate the close proximity of birds of their own kind and often of other species. Furthermore, in order to nest safely on a cliff a bird’s behavior must be such that it does not knock the eggs or young off, and the young themselves must have behavioral adaptations preventing them from falling off. Cullen (1957) has shown how many of the morphological and behavioral characters of the Kittiwalce ( Rissa tridactyla) , which dis- tinguish it from other gulls, are attributable to its cliff-nesting habit. In the Oilbird, too, sev- eral adaptations to cliff-nesting are apparent, though in this case there are no close relatives with which it can be compared. Nest sites suitable for Oilbirds are extremely limited in number, being restricted to a rather small number of caves in Trinidad and parts of northern South America. This has probably been the chief factor in the evolution of their highly social nesting behavior. It also probably accounts for the continuous occupation of the nest throughout the year, since a pair, once dis- possessed, would find it very difficult to estab- lish themselves again. In the Kittiwake, cliff-nesting is associated with the relaxation of various anti-predator fea- tures. The same tendency is evident in the Oil- bird; in particular, there is complete relaxation of all features that help to protect the nest against visual predators. Thus the adults are rather tame when on the nest, predators are not attacked, the eggs are white, and the chicks are not camouflaged. If the nest site is safe, slow development of the eggs and young is not a serious disadvantage. Several authors have commented on the general correlation between safety of nest site and length of incubation and fledging period in birds, al- though no detailed study has yet been made. The Oilbird’s development is exceptionally slow; among land birds only the California Condor ( Gymnogyps calif ortxianus ) is known to have a longer fledging period (Koford, 1953), while that of the Bateleur Eagle ( Terathopius ecau- datus ) is almost exactly the same (Brown, 1955). Such a slow development could hardly have been evolved if the nests were subject to such heavy predation as are those of most trop- ical birds (see, e.g., Skutch, 1945); and in fact the evidence is that they are subject to little dis- turbance except from human beings. However, it is unlikely that the slow development can be attributed simply to the great safety of the nest site. It is probable that the Oilbird’s specialized diet, of low protein content and containing a large indigestible fraction (the seeds), necessi- tates a slow development. If this is so, cliff -nest- ing and fruit-eating must have been intimately bound up with one another in the Oilbird’s evolution. Egg-rolling behavior is of little use to a cliff- nesting bird, whose eggs, if they are not in the concave nest-cup, are likely to be lost over the edge. It is not surprising therefore that the Oil- bird is in marked contrast to the ground-nesting nighthawks, some of which are known to be able to move their eggs many feet, by pushing them or carrying them in the beak, and which do so repeatedly if disturbed. Likewise, when the young Oilbird has hatched, it is safe only if it remains in the nest-cup, and here again its ten- dency to stay still and maintain contact with its nest mates may be contrasted with the mobile behavior of young nighthawks. Remaining still in the middle of the nest is not due to inability to move, as even when they are quite small young Oilbirds move backward to the nest edge in order to defecate. Their ability to climb with feet, beak and wings gives them a chance to save themselves if they should nevertheless go over the edge of the nest. The parents’ contri- bution to the safety of the young, as of the eggs, is limited to their tiny shuffling steps, which prevent them from kicking anything off the nest, or indeed from effectively moving any obstacle from their path. Ecological Factors in the Evolution of the Oilbird The Oilbird’s combination, unique for a bird, of fruit-eating and nocturnal habits is undoubt- edly due to its evolution from an originally noc- turnal or crepuscular ancestor. As already men- tioned, anatomical evidence suggests that the Oilbird’s closest affinities are with the Caprimul- giformes, though it has certain characters in common with the owls. A consideration of feed- ing behavior, however, makes it almost certain that they are in fact closest to the Caprimulgi- formes. All other caprimulgiform birds seize insects or other small animals in the mouth and swallow them whole, the feet not being used at all. Owls on the other hand seize their food in the talons and tear it up with the beak. Oilbirds, as we have seen, pluck fruits with the beak and swallow them whole. It is an easy transition to this method of feeding from the typical capri- 1961] Snow: General Behavior and Breeding Habits of the Oilbird 45 mulgiform method, but from the owl’s method the transition is almost inconceivable. We may then regard the Oilbird as the most extreme product of the rather limited adaptive radiation of the caprimulgiform stock. Since many of the larger diurnal birds of tropical forest are mainly or entirely frugivorous, it is perhaps not surprising that this food supply should have been exploited by one nocturnal bird. However, for an originally insectivorous caprimulgiform bird, the change to a fruit diet must have involved a number of ecological problems. The Oilbird’s solution of these prob- lems has had effects on every aspect of its life. For the fruit-eater, forest trees in fruit are, essentially, temporary and discontinuous pock- ets of abundant food whose location is always changing. For a nocturnal fruit-eater, the short distance at which such pockets of food can be seen, or otherwise perceived, is an added prob- lem. For a strong-flying bird there is no ap- parent advantage in searching for such food singly or in pairs, or in maintaining feeding ter- ritories, and in fact the parrots, toucans and other large fruit-eating forest birds are generally social feeders, as also are the fruit-eating bats. It is probable, therefore, that the change from an insect to a fruit diet in the ancestral Oilbird stock involved the enhancement of social and gregarious behavior at the expense of territorial behavior. It is probable, too, that the Oilbird’s large size, compared with that of most other caprimulgi- form birds, was another consequence of the change to a fruit diet. It must have been a great advantage to be able to exploit the larger fruits, up to two inches or so long, of the tall forest trees, which form the main food of the large diurnal fruit-eating birds, since these not only give more nourishment for every fruit taken but are also much more conspicuous at night than the smaller fruits of second-story trees and shrubs which provide much of the food of the smaller frugivorous birds. Increasingly social habits, increased size and a diet of fruit must have eventually necessitated radical changes in breeding behavior. In particu- lar, increase in size and probably also the fruit diet (comparatively poor in proteins) must have lengthened the period of development of the young. At some point in its evolution the Oil- bird must have faced in acute form the “choice” between either making the nest extremely incon- spicuous, as do most other caprimulgiform birds and probably its own ancestors, or else making it extremely safe. Its large size, its fruit diet, which involves the accumulation of much re- gurgitated matter around or under the nest, its awkwardness in trees and lack of complex nest- building behavior (common to all the Capri- mulgiformes), must all have favored the choice of a very safe nest site. Of the two main types of safe nest site available, cliffs and hollow trees, there is little doubt that natural selection would favor the former as being safer than tree holes and less sought after by other creatures. Cliff- nesting, as already mentioned, is usually asso- ciated with social breeding behavior; hence both for feeding and for nesting natural selection must have favored gregarious as against terri- torial tendencies. (It is perhaps instructive that one large fruit-eating cotingid, the Cock-of-the- rock, Rupicola, has adopted the same type of nest site, on cliffs or in shallow caves, and breeds semi-socially). Presumably, then, the first step in the evolu- tion of cave-nesting was from the ancestral site (probably the ground or on tree stumps) to cliff- ledges in the open, and it was at this stage that the rudiments of the echo-location faculty were evolved. Pressures of predation probably then led to the seeking of deeper and deeper recesses, until the perfection of echo-location allowed the birds to occupy the deepest caves, and so opened up for them a wealth of nest sites that were completely safe (until the arrival of man) and for which no other creatures competed. Summary An account is given of the general behavior and nesting of the Oilbird, based on 3!A years’ observations. Oilbirds are gregarious cave-dwelling birds, almost certainly of caprimulgiform stock. They spend all day in caves and fly out at night to feed on the fruits of forest trees. The Oilbird’s stance is peculiar, the body being tilted forward and the very short legs ro- tated as far forward as possible. Aerodynam- ically they are highly specialized for flight within restricted spaces and for load-carrying. Sight is well developed and is used whenever possible. The sonar method of orientation, dis- covered by Griffin, is used only when there is not enough light. There is some evidence that the olfactory sense is important in food-finding. Daily routine and social behavior are de- scribed. The birds leave the cave at dusk and return before dawn. They are gregarious while feeding. The pair bond is probably permanent. Aerial displays, probably connected with pair formation, have been seen at night. Courtship behavior on the nest consists of the preening of the female’s head by the male. The nest, eggs and young are described. The breeding cycle is very slow; eggs are laid at in- 46 Zoologica: New York Zoological Society [46: 3 tervals of several days, the incubation period is usually 33-34 days, and the fledging period 90- 125 days. The young become very fat, reaching a weight half as much again as the adult’s weight at about the 70th day. There are two sets of down feathers, followed by the growth of the adult plumage. The young are fed at long intervals during the night, large young three or four times, smaller young five or six times a night. The food of the young is the same as the adult’s. Nestlings eat about one-third or one-quarter of their body weight during each night. The young acquire temperature control at the age of about three weeks. Both the thick down feathers and the fat deposits are considered to be important in maintaining body temperature. The ecological aspects of the Oilbird’s evolu- tion are discussed. It is argued that the original change from an ancestral insect diet to a fruit diet led to increased gregariousness, increased size, slower development, the adoption of cliff nest-sites, and finally, with the perfection of echo-location, to the colonization of pitch-dark caves. Literature Cited Bock, W. J. & W. de W. Miller 1959. The scansorial foot of the woodpeckers, with comments on the evolution of perch- ing and climbing feet in birds. Amer. Mus. Novit., 1931, 45 pp. Brown, L. 1955. Eagles. Michael Joseph, London. Brown, R. H. J. 1951. Flapping flight. Ibis, 93: 333-359. Cullen, E. 1957. Adaptations in the Kittiwake to cliff-nest- ing. Ibis, 99: 275-302. Funck, N. 1844. Notice sur le Steatornis caripensis (Gua- charo incolarum). Bull. Acad. Roy. Sci. Bruxelles, 11 (2) : 371-377. Garrod, A. H. 1873. On some points in the anatomy of Steat- ornis. Proc. Zool. Soc. London, 1873: 526-535. Griffin, D. R. 1954. Acoustic orientation in the oil bird, Steat- ornis. Proc. Nat. Acad. Sci., 39: 884-893. 1958. Listening in the dark. Yale University Press, New Haven. Humboldt, A. von 1817. Voyage aux regions equinoxiales du Nou- veau Continent, fait en 1799, 1800, 1801, 1802, 1803 et 1804 par Al. de Humboldt et A. Bonpland. Vol. 3. Paris. Humboldt, A. von & A. Bonpland 1817. Memoire sur le Guacharo de la caverne de Caripe. Recueil d’observations de zoologie et d’anatomie comparee. Vol. 2. Paris. Ingram, C. 1958. Notes on the habits and structure of the Guacharo Steatornis caripensis. Ibis, 100: 113-119. 1960. Guacharo feeding habits. Ibis, 102: 140. Koford, C. B. 1953. The California Condor. National Audu- bon Society, New York. L’Herminier, F. 1834. Memoire sur le Guacharo ( Steatornis caripensis (Humboldt)). Nouv. Ann. Mus. Hist. Nat. Paris, 3: 321-331. Marler, P. 1955. Characteristics of some animal calls. Nature, 176: 6-7. Muller, J. 1842. Anatomische Bemerkungen fiber den Quacharo, Steatornis caripensis v. Humb. Arch. f. Anat., 1842: 1-11. Parker, W. K. 1889. On the osteology of Steatornis caripensis. Proc. Zool. Soc. London, 1889: 161-190. PlETRI, E. DE B. 1957. El Guacharo (monografia). Bol. Soc. Venezolana Cienc. Nat., 18: 3-41. SCLATER, P. L. 1866. Notes on the American Caprimulgidae. Proc. Zool. Soc. London, 1866: 126-145. Sibley, C. G. 1960. The electrophoretic patterns of avian egg- white proteins as taxonomic characters. Ibis, 102: 215-259. Skutch, A. F. 1945. Incubation and nestling periods of Central American birds. Auk, 62: 8-37. Snow, D. 1958. Trinidad’s Oilbirds are yielding new facts. Animal Kingdom, 61: 117-121. Stolzmann, J. 1880. Observations sur le Steatornis peruvien. Bull. Soc. Zool. France, 5: 198-204. 47 Snow: General Behavior and Breeding Habits of the Oilbird Taczanowski, L. 1884. Ornithologie du Perou. Vol. 1. Rennes. Wetmore, A. 1918. On the anatomy of Nyctibius with notes on allied birds. Proc. U. S. Nat. Mus., 54: 577-586. 48 Zoologica: New York Zoological Society [46: 3: 1961] EXPLANATION OF THE PLATES Plate I Fig. 1. Typical stance of Oilbird on nest. Note position of feet far forward under breast. Fig. 2. Oilbird clinging to narrow ledge; showing backward position of inner toe, and tail pressed against rock face. Feet are held much farther back than when perching on level surface (Fig. 1). For explanation see text. Plate II Fig. 3. Oilbirds clinging to sloping ledge; showing position of legs and toes. For explanation see text. Fig. 4. Oilbird in slow flight, at early stage of up- stroke; tip of left wing still on the down- stroke. To show the extreme width of wing and fully spread tail. Fig. 5. Oilbird in slow flight; the upstroke. The bend in the wing shows that the wing beat is propulsive. SNOW PLATE I FIG. 2 THE NATURAL HISTORY OF THE OILBIRD, STEATORNIS CARIPENSIS, IN TRINIDAD. W.l. SNOW PLATE II FIG. 3 FIG. 4 FIG. 5 THE NATURAL HISTORY OF THE OILBIRD. STEATORNIS CARIPENSIS, IN TRINIDAD. W.l. 4 Fatty Degeneration, Regenerative Hyperplasia and Neoplasia in the Livers of Rainbow Trout, Salmo gairdneri Ross F. Nigrelli & Sophie Jakowska New York Aquarium, New York Zoological Society (Plates I-VI) Introduction LIVER tumors in fish were first reported in European brown trout ( Salmo trutta 4 Linnaeus) by Plehn (1909a, 1924). The first case in rainbow trout ( Salmo gairdneri Richardson) was described by Haddow & Blake (1933) in England. The disease was also re- ported in this species in the United States in 1953 by Nigrelli (1954), who indicated that the incidence of hepatomas in such trout might be relatively high. Four tumors were subsequently found in a single lot of 600 3-year-old rainbow trout in a Pennsylvania hatchery, and the histo- pathology of these was briefly described by Nigrelli & Jakowska (1955). Almost simulta- neously, similar tumors that had been reported by Scolari in 1953 in rainbow trout from all the hatcheries along the Alpine Arc and the lake region of northern Italy, were described by Cudkowicz & Scolari (1955). They reported an incidence of 50% or more in some hatcheries and described the histopathology and distribu- tion of the disease and indicated its possible cause. In the spring of 1960, rainbow trout that were to be introduced into California’s waters were examined at the state line by California Fish and Game inspectors and found to be affected with liver tumors. An embargo was immediately placed on all rainbow trout being shipped into the state. This led to the examination of trout in many hatcheries, whereupon it soon became evident that trout with liver diseases were widely distributed in the United States, with an inci- dence of hepatomas of 50% or more being re- ported in some hatcheries. In retrospect, some fishery personnel have indicated that the hepa- tomas were prevalent in rainbow trout as far back as 1937 (Rucker et al., 1961). It is highly probable that the disease went unrecognized for a great number of years, since little or no atten- tion was given to fish pathology by fishery bi- ologists in this country until relatively recently. Since no critical distinction has been made between regenerative hyperplasia and neoplasia, we cannot accept the numerous unofficial re- ports that all the “tumors” seen in rainbow trout are hepatomas, especially if the term is used to indicate a neoplastic process. The present paper is therefore principally concerned with a micro- scopic analysis of the livers of rainbow trout from several hatcheries, mainly in Idaho, Mon- tana and Wyoming, and those of a few speci- mens of brown and rainbow trout taken from natural waters. Observations and Discussion Three hundred and thirty-three livers were examined. They varied considerably in size and appearance, and those with apparent lesions were characterized by grayish or yellowish spots, streaks (Plate I, Fig. 1), numerous petichiae or larger clots, or by one or more encapsulated (Plate I, Fig. 2) and non-encapsulated nodules on the surface or deep in the parenchyma. In a few instances, livers with hob-nail or miliary ap- pearance were also seen. The so-called normal liver showed a histolog- ical picture resembling that consistently found in teleosts kept in captivity and fed artificial diets. The usual pattern consisted of lobules with a central vein in which radiating liver cells, separated by sinusoids, were arranged as a muralium1, or wall, two cells in thickness but without polar orientation of the nucleus (Plate 1 The significance of two-cell-thick muralia in hepa- tomas of mammals and birds is discussed by Elias (1955). 49 50 Zoologica: New York Zoological Society [46: 4 III, Fig. 6). The over-all picture was one of uneven staining, suggesting physiological dif- ferences in the cells of the various lobules. The liver cells were uniform in size and typical in shape, but some relatively larger cells with larger nuclei, suggestive of polyploidy, were also observed. The cytoplasm was usually lightly basophilic, but the degree of basophilia some- times varied with the lobules, apparently affected by the presence or absence of cytoplasmic par- ticulates. The nucleus was typical, with one or more acidophilic nucleoli. Although binucleate and multinucleate cells were not unusual, mito- tic figures were rare. After appropriate staining, changes indicative of disease were interpreted as fatty infiltration (Plate II, Fig. 3), glycogen infiltration or deple- tion, ceroid deposition (Plate II, Fig. 4), hema- tochromatosis (Plate II, Fig. 5), focal necrosis with lymphocytic infiltration (Plate III, Figs. 6, 7), portal cirrhosis (Plate III, Fig. 8), biliary cirrhosis (Plate III, Fig. 9), hyperplasia and neoplasia. Hyperplasia, which ranged from simple soli- tary (Plate IV, Fig. 10) or widely diffuse regen- erative areas (Plate IV, Fig. 12) to single or multiple nodules, was found in 50 of the livers. The hyperplasia was evident as islands or as ex- tensive growth of liver cells of irregular pattern in which uninucleate and multinucleate elements were densely packed, especially at the periphery of the nodules. Some of the larger nodules were completely necrotic. In some areas, clusters of larger cells with acidophilic cytoplasm and ec- centric nuclei were present (Plate IV, Fig. 11). The origin and nature of these cells, some of which contained ceroid, remain to be deter- mined. In close proximity to these, as well as in areas adjacent to regions of apparent active growth, there were considerable morphological changes in the liver cells, and these changes extended even to elements more distantly lo- cated. The lobules were often distorted and the cells showed variation in size, shape and stain- ing reactions; in many instances the tissue as- sumed an adenomatoid appearance (Plate V, Fig. 13). These effects were more pronounced in hyperplastic livers with extensive fatty infil- tration and ceroid deposition and with portal and biliary cirrhosis. Nineteen of the livers were diagnosed as neo- plastic and the changes noted in them ranged from adenoma (Plate IV, Fig. 14) to hepato- cellular carcinoma. The latter exhibited either a normal cell arrangement (but without distinct lobular pattern), or an anaplastic, highly pleo- morphic structure with considerable degenera- tion and hemorrhage. The hepatocellular lesions were frequently indicated by a solid cord-like cellular arrangement in which the hepatic cells were relatively small, the nuclei pyknotic and vascularization scanty (Plate V, Fig. 15). Cholangiomatous changes (Plate VI, Fig. 16) and cirrhosis (Plate VI, Fig. 17) were found in a few cases. Other details were similar to those reported by Nigrelli & Jakowska (1955) and by Cudkowicz & Scolari (1955). Some of the abnormal conditions described above, including adenomatous and cholangio- matous growths, were also found in livers of four rainbow trout and Loch Leven (brown) trout taken from natural waters in Idaho. The histological picture of the atypical cell growths in the livers of hatchery-raised rainbow trout is strikingly similar to that reported in mice and other experimental animals kept for pro- longed periods on a low choline diet, and in which fatty livers with little or no cirrhosis de- velop (Hartroft, 1954, 1955). In general, liver diseases in higher animals involving necrosis, fatty changes and ceroid deposition, with or without hyperplasia or neoplasia, have been attributed to deficiency or imbalance of one or more of the following substances: methionine, cystine, choline, vitamin E, vitamin B12 and riboflavin (Kensler et al., 1941; Endicott & Lillie, 1944; Engle et al, 1947; Lee, 1950; Casselmen, 1953; Daft, 1954; Drill, 1954; Schwarz, 1954; Hartroft, 1954, 1955; Salmon & Copeland, 1954; Salmon et al, 1955; see also Tannenbaum, 1953). Although most of the above essential sub- stances are incorporated into artificial trout diets in the United States, it is generally agreed that the complete nutritional requirements of trout have not yet been fully determined, a fact which is in part responsible for the wide variety of diets still in use in hatcheries here and abroad. The non-tumor and regenerative lesions de- scribed in this paper may be indicative of some nutritional faults. Some of these abnormal con- ditions, especially in relation to fatty degenera- tion and ceroid deposition, have been previously reported in experimentally- and non-experi- mentally-fed hatchery trout (Plehn, 1909b, 1915, 1924; Gaschott, 1929; Hewitt, 1937; McLaren et al, 1946 ;Mann 1952; Davis, 1953; Schaperclaus, 1954; Scolari, 1954; Faktorovich, 1956a, b, 1958, 1960). Some striking effects of purified rations on the livers of rainbow trout were shown by McLaren et al. (1946) . Yearling rainbow trout, fed a ration composed of cerelose 48%, casein 40%, fat 2%, dried liver 5% sup- plemented with brewers’ yeast, were able to maintain, for a period of time, a hemoglobin level and growth rate equal to that produced by feeding 100% dried liver or a standard hatchery 1961] Nigrelli & Jakowska: Liver Diseases of the Rainbow Trout 51 meat ration. After 8 weeks, however, the fish suddenly died and autopsies showed that they had developed greatly enlarged, yellow, lohu- lated livers. In this case, the liver damage was prevented by reducing the carbohydrate level to 20%. The most important disease in hatchery-raised rainbow trout is generally considered to be fatty degeneration of the liver, but with proper diet recovery is often complete. Thus, Faktorovich (1956b) for 11 months carefully followed the regeneration of hepatic tissue in rainbow trout during recovery from fatty degeneration, a con- dition which he later (1960) interpreted as taking the form of ceroid deposition. It should be emphasized, however, that ceroid deposition appears to be a general characteristic of teleosts, especially those kept in captivity (Pickford, 1953; Wood & Yasutake, 1956; Nigrelli, 1960), those infected with parasites (Nigrelli, 1954), or those exposed to toxic substances, e.g., cop- per (Calventi et al.t 1960). The epizootic nature of the liver diseases in rainbow trout suggests that hereditary, viral or carcinogenic agents may also play a role. The possibility that the tumors are hereditary in origin has been indicated by Cudkowicz & Scolari (1955). Evidence for this suggestion is based on the following: (1) rainbow trout in the hatcheries in northern Italy are all the progeny of fish imported from Rocky Mountain streams in 1880, (2) a high incidence of tumors was found in populations derived from a stock inbred for 20 years in one hatchery, (3) the liver tumors were not found in hatcheries in Switzerland and Bavaria, two areas that had had no exchange of fish with the Italian hatcheries, and (4) , rainbow trout from Denmark that were raised in the Italian hatcheries under identical conditions were not affected with the growths.2 Rainbow trout in the United States have been artificially bred for approximately the same length of time as in Europe. Although inbreed- ing records are not available to us, hatchery practices, such as discarding rants and selecting highly colored or early maturing fish, inevitably lead to the selection of special strains— possibly with inherent susceptibility to liver dysfunctions, e.g., inability to metabolize fats properly. Whether or not such strains can be associated with specific genetic factors remains to be deter- mined. In certain mouse colonies, mutant strains (CBA and C3H) highly susceptible to hepa- 2 Cystic degeneration of the liver has also been de- scribed in rainbow trout in an Italian hatchery by Cas- telnuovo & Rizzo (1938). It was suggested that this disease is congenital. tomas have spontaneously appeared (Ander- vont, 1950; Woolley, 1951). Virus and carcinogens as etiological agents cannot be excluded, but mammal geneticists generally agree that their principal effect is to enhance the existing intrinsic susceptibility to spontaneous tumors (Heston, 1941; Heston & Deringer, 1947, 1949; Duran-Reynals, 1953; Sangvi & Strong, 1958). The following sub- stances, occurring in nature or introduced with feeding, are known to produce liver damage in fish and other animals: zinc chloride, copper sulfate (Calventi, Jakowska & Nigrelli, 1960), carbasone, sulfonamides, antibiotics (Good- man & Gilman, 1955), bentonite (Jakowska & Nigrelli, 1956), radioactive and ionizing sub- stances, and arsenicals (see Hueper, 1953; Wilson, 1954). Their possible role in the eti- ology of liver diseases in rainbow trout must be considered. In addition, pathological liver changes may be associated with the following well-known diseases of trout: furunculosis, ulcer disease, infectious (viral) pancreatic necrosis, bacterial or viral infections of the kid- ney, mycosis, helminthiasis, cnidosporidiosis and acute or chronic anemia. Increased pressure in the venous system, resulting from these dis- eases, may severely affect the liver. The effect on the hepatic cells may result from anoxemia, a condition also associated with acute and chronic anemia Whatever the cause of the liver lesions in rain- bow trout, most of the damage is degenerative, and any hyperplasia seen represents a compensa- tory response of parenchymal tissue. The pleo- morphic changes noted in surrounding and more distant hepatic cells are the result of pres- sure from the active hyperplasia. It is highly probable that in certain instances the hyper- plasia is supervened by neoplasia, especially if the proliferation of the hyperplastic tissue be- comes excessive. Most of the so-called hepatomas in rainbow trout are hyperplasias rather than neoplasias. It has been pointed out by Hartroft (1955) that the differential diagnosis of hyperplasia and neo- plasia in the livers of both experimental animals and man has perplexed and confused patholog- ists for many years; the close resemblance of small neoplastic foci in a cirrhotic liver to foci of hyperplasia irresistibly suggests that the former may arise regularly from the latter. The sudden and widespread concern about hepatomas in rainbow trout is reminiscent of a similar situation at the turn of the century with regard to thyroid tumors in trout and other salmonids in Europe and the United States. These growths were originally described as 52 Zoologica: New York Zoological Society [46: 4 adenocarcinomas or carcinomas and, as in the present case of the hepatoma, the disease oc- curred in epizootic proportions (Plehn, 1902; Pick, 1950; Gaylord & Marsh, 1914). Most of the thyroid tumors were later recognized as simple hyperplasia or goitres (Marine & Len- hart, 1911; Marine, 1914). This interpretation has been substantiated by the fact that since iodine has been regularly added to the water or food, the incidence of thyroid tumors has been reduced to a point where they now are extremely rare, even though a relatively large number of the growths originally described were actually neoplastic. A similar interpretation may prove to be true for the liver tumors of the rainbow trout. Summary Microscopical findings in livers of more than 300 hatchery-raised rainbow trout from several states and from rainbow and brown trout from natural streams are described. Most of the livers resembled those of fish kept in captivity and fed artificial diets for prolonged periods. The lesions, when present, were mainly degenerative and characterized by fatty infiltration, glycogen de- pletion or infiltration, ceroid deposition, hema- tochromatosis, focal necrosis with lymphocytic infiltration, biliary and portal cirrhosis and other histological and cytological changes. These conditions were associated with com- pensatory non-nodular and nodular hyperplasia of the liver cells, and it is assumed that under certain circumstances hyperplasia is supervened by neoplasia. Hereditary, nutritional, carcino- genic and other possible etiological factors are discussed. References Andervont, H. B. 1950. Studies on the occurrence of spontaneous hepatomas in mice strain C3H and CBA. J. Nat. Cancer Inst., 11: 581-592. Calventi, Idelisa, Sophie Jakowska & R. F. Nigrelli 1960. Copper sulfate and liver damage in fish (laboratory records). Casselmen, W. G. B. 1953. Factors influencing the formation of ceroid in the livers of choline deficient rats. Biochem. et Biophys. Acta, II: 445- 447. Castelnuovo, G. & L. Rizzo 1938. Caso di degenerazione cistica del fegato e di altri organi in trota iridea. Boll, di Pesca, Pisciculture e Idrobiologia, 16:1-10. Cudkowicz, G. & C. Scolari 1955. Un tumore primitive epatico a diffusione epizootica nella trota iridea di allema- mento ( Salmo irideus) . Tumori, XLI (fasc. 5): 524-537. Daft, F. S. 1954. Experimental differentiation between liver necrosis and liver cirrhosis and some diet- ary factors affecting their development. Ann. N. Y. Acad. Sci., 57 (6): 623-632. (Conference on Nutritional Factors and Liver Diseases). Davis, H. S. 1953. “Culture and Diseases of Game Fishes.” Univ. California Press. 332 pp. Drill, A. 1954. Lipotropic effects of vitamin R12 and other factors. Ann. N. Y. Acad. Sci., 57 (6): 654-663. (Conference on Nutritional Factors and Liver Diseases). Duran-Reynals, F. 1953. Virus-induced tumors and the virus the- ory of cancer. Chapt. 13 in: “Physiopath- ology of Cancer.” Hamburger & Fishman, Ed. Hoeber-Harper, New York, Pp. 298- 337. Elias, H. 1955. Human hepatocarcinoma and the com- parative embryology of the vertebrate liver. In: Conference on Experimental Hepatomas. J. Nat. Cancer Inst., 15 (5), suppl. Pp. 1463-1468. Endicott, K. M. & R. D. Lillie 1944. Ceroid the pigment of dietary cirrhosis of rats. Amer. J. Path., 20: 149-153. Engle, R. W., D. H. Copeland & W. D. Salmon 1947. Carcinogenic effects associated with diets deficient in choline and related nutrients. Ann. N. Y. Acad. Sci., 49 (1): 49-67. Faktorovich, K. A. 1956a. (Disease of two year old rainbow trout raised on artificial diets). Voprosy Ikhtio- logii. No. 6: 156-173. (In Russian). 1956b. (Regeneration of the liver of rainbow trout). Dokl. Akad. Nauk SSSR, 110 (2): 301-303. (Biol. Abstr., 35 (15) Abst. No. 4191). 1958. (Disturbance in fat metabolism in the liver of rainbow trout raised on artificial diets). Proc. Symposium on Physiology of Fishes, 1956, Acad. Sci. USSR: 237- 243. (In Russian). 1960. Uber das Wesen des Stoffes, der Sich in der Leber der Regenbogenforelle bei der sogenannten lipoden Degeneration der Leber Ablagert. Zeitschr. f. Fischerei, 9 (’A): 95-99. 1961] Nigrelli & Jakowska: Liver Diseases of the Rainbow Trout 53 Gaschott, O. 1929. Die verlustreichen Leberkrankungen der Forellenmasbetriebe im Friijar 1929. Allg. Fischerei-zeitung, 1929, No. 2: 340-346. Gaylord, H. R. & M. C. Marsh 1914. Carcinoma of the thyroid in the salmonoid fishes. Bull. Bur. Fish., 32 (1912), doc. no. 790: 363-524. Goodman, L. S. & A. Gilman 1955. “The Pharmacological Basis of Thera- peutics.” Macmillan & Co., New York. 1831 pp. (Toxicity effects on liver: carba- sone, p. 1218; sulfonamides, p. 1298; an- tibiotics, p. 1386; copper, p. 1469). Haddow, A. & Isobel Blake 1933. Neoplasms in fish. A report of six cases with a summary of the literature. J. Path, and Bact., 36: 41-47. Hartroft, W. S. 1954. The sequence of pathological events in the development of experimental fatty liver and cirrhosis. Ann. N. Y. Acad. Sci., 57 (6): 633-645. (Conference on Nutritional Factors and Liver Diseases). 1955. General histophysiology and histopathol- ogy of the liver. J. Nat. Cancer Inst., 15 (5) suppl.: 1463-1468. Heston, W. E. 1941. Relationship between susceptibility to in- duced pulmonary tumor and certain known genes in the mouse. J. Nat. Cancer Inst., 2: 127-132. Heston, W. E. & M. K. Deringer 1947. Relationship between lethal yellow (Ay) gene of mouse and susceptibility to spon- taneous pulmonary tumors. J. Nat. Cancer Inst., 7: 463-465. 1949. Relation between hairless gene and the susceptibility to induced pulmonary tum- ors in mice. J. Nat. Cancer Inst., 10: 119- 124. Hewitt, E. R. 1937. Some recent work on fatty livers in trout. Progr. Fish-culturist, No. 25: 11-15. Hueper, W. C. 1953. Environmental cancer. Chapt. 26 in: “Physiopathology of Cancer.” Hamburger & Fishman, Ed., Hoeber-Harper, New York. Pp. 730-777. Jakowska, Sophie & R. F. Nigrelli 1956. Effects of diet containing bentonite on the cichlid Tilapia macrocephala. Anat. Rec., 125 (3): 655-656. Kensler, C. J., K. Sugiura, N. F. Young, C. R. Halter & C. P. Rhoads 1941. Partial protection of rats by riboflavin with casein against liver cancer caused by dimethylam inobenzene. Sci., 93: 308-310. Lee, C. S. 1950. Histochemical studies of the ceroid pig- ment of rats and mice and its relation to necrosis. I. Nat. Cancer Inst., 11 (2) : 339- 348. Mann, H. 1952. Zur Frage der lipoiden Leberdegeneration bei Forelle. Fischereiwelt, 4: 14-15. Marine, D. 1914. Further observations and experiments on goitre in brook trout. I. Exp. Med., 19: 70-88. Marine, D. & C. H. Lenhart 1911. Further observations and experiments on so-called thyroid carcinoma of brook trout, and its relation to endemic goitre. J. Exp. Med., 13: 445-475. McLaren, Barbara A., F. Herman & C. A. Elvehjem 1946. Nutrition of rainbow trout: studies with purified rations. Arch. Biochem., 10: 433- 441. Nigrelli, R. F. 1954. Tumors and other atypical cell growths in temperate freshwater fishes of North America. Trans. Amer. Fish. Soc., 83 (1953) : 262-296. 1960. Autopsy and laboratory records, N. Y. Aquarium. Nigrelli, R. F. & Sophie Jakowska 1955. Spontaneous neoplasms in fishes. IX. He- patomas in rainbow trout, Salmo gaird- neri. Proc. Amer. Assoc. Cancer Res., 2 (1): 38. Pick, L. 1905. Der Schildriisenkrebs der Salmoniden. Aus dem Laboratorium der L. and Th. Lan- dau’schen Frauenklinik, Berlin. Berliner & Klinische Wschr., 1905 (46/49): 1435- 1542. Pickford, Grace E. 1953. A study of the hypophysectomized male killifish, Fundulus heteroclitus (Linn.). Bull. Bingham Oceanogr. Coll., 14 (2): 5-41. Plehn, Marianne 1902. Bosartiger Kropf (Adeno-Carcinoma der Thyroidea) bei Salmoniden. Allg. Fisch- erei-Zeitung, no. 7: 117-118. 54 Zoologica: New York Zoological Society [46: 4 1909a. Uber einige bei Fischen beobachtete Geschwiilste und geschwiilstartige Bild- ungen. Berichte D. K. Bayer. Biol. Ver- suchs., 2: 55-76. 1909b. Uber die Leber der Salmoniden. Allg. Fischerei-Zeitung, No. 24: 525-527. 1915. Kenntnis der Salmonidenleber im gesun- den und kranken Zustand. Zeitschr. f. Fischerei, 17: 1-24. 1924. “Praktikum der Fischkrankheiten.” E. Schweizerbart’sche, Stuttgart. Pp. 301-479. Rucker, R. R., W. T. Yasutake & H. Wolf 1961. Trout hepatoma — a preliminary report. Progr. Fish-culturist, 23 (1): 3-7. Salmon, W. D. & D. H. Copeland 1954. Liver carcinoma and related lesions in chronic choline deficiency. Ann. N. Y. Acad. Sci., 57 (6): 654-677. (Conference on Nutritional Factors and Liver Dis- eases). Salmon, W. D., D. H. Copeland & M. J. Burns 1955. Hepatomas in choline deficiency. J. Nat. Cancer Inst., 15 (5) , suppl.: 1549-1554. Sanghvi, L. D. & L. C. Strong 1958. Effects of selection on chemically induced tumors in mice. Ann. N. Y. Acad. Sci., 71 (6): 839-878. (Conference on Genetic Concept for the Origin of Cancer). Schaperclaus, W. 1954. “Fischkrankheiten.” Akademie, Berlin, 708 pp. Schwarz, K. 1954. Liver necrosis versus fatty liver and cir- rhosis. Ann. N. Y. Acad. Sci., 57 (6): 617-621. (Conference on Nutritional Fac- tors and Liver Diseases). 1954. Factors protecting against dietary necrotic liver degeneration, ibid: 878-888. SCOLARI, C. 1953. Atti Soc. Ital. Scienze Veter., 7: 599. (See Cudkowicz & Scolari, 1955). 1954. Su di una epizoozia della trote iridee d’allevamento, “la lipoidosi epatica”. Clin. Veter., 77 (4): 102-106. Tannenbaum, A. 1953. Nutrition and Cancer. Chapt. 15 in: “Physiopathology of Cancer.” Hamburger & Fishman, Ed. Hoeber-Harper, New York. Pp. 392-437. Wilson, J . W. 1954. Hepatomas produced by feeding benton- ite in the diet. Ann. N. Y. Acad. Sci., 56 (6): 678-683. (Conference on Nutritional Factors and Liver Diseases). Wood, E. M. & W. T. Yasutake 1956. Ceroid in fish. Amer. I. Path., 32: 591-603 Woolley, G. W. 1953. Genetics of neoplastic diseases. Chapt. 11 in: “Physiopathology of Cancer.” Ham- burger & Fishman. Ed. Hoeber-Harper, New York. Pp. 171-224. 1961] Nigrelli & Jakowska: Liver Diseases of the Rainbow Trout 55 EXPLANATION OF THE PLATES Plate I Fig. 1. Female rainbow trout ( Salmo gairdneri Richardson) showing discolored lesions on the liver. Reduced 2.5 X. Fig. 2. Section through an encapsulated tumor mass of rainbow trout from a Pennsylvania hatchery. 5 X. Plate II Fig. 3. Fatty infiltration in an otherwise normal liver of a rainbow trout. Hematoxylin- eosin. 500 X. Fig. 4. Extensive ceroid deposition of liver. Acid- fast stain. 250 X. Fig. 5. Accumulation of siderin in areas surround- ing biliary duct. Hematoxylin-eosin. 1000 X. Plate III Fig. . 6. “Normal” lobule of liver of rainbow trout with two-cell-thick muralia radiating from a central vein. Note necrotic areas ad- jacent to the lobule. Hematoxylin-eosin. 250 X. Fig. 7. Typical lymphocytic infiltration in a ne- crotic area. Hematoxylin-eosin. 500 X. Fig. 8. Section of otherwise “normal” liver show- ing portal cirrhosis and a small necrotic area. Hematoxylin-eosin. 250 X. Fig. 9. Biliary cirrhosis and disorganization of adjacent hepatic elements. Masson’s stain. 400 X. Plate IV Fig. 10. Focal necrosis and localized hyperplasia of liver cells from rainbow trout. Hema- toxylin-eosin. 250 X. Fig. 11. Cluster of acidophilic cells with eccentric nuclei. Some cells contain ceroid. The na- ture and origin of these cells are unknown. Hematoxylin-eosin. 500 X. Fig. 12. Diffuse regenerative hyperplasia associ- ated with extensive degeneration of the liver. Hematoxylin-eosin. 400 X. Plate V Fig. 13. Adenomatoid appearance of hepatic tissue associated with hyperplastic foci in more distantly located areas of the same section. Hematoxylin-eosin. 400 X. Fig. 14. Details of adenomatous liver. Note large ceroid globule. Masson’s stain. 2000 X. Fig. 15. Cord-like arrangement of hepatocellular tumor. Hematoxylin-eosin. 600 X. Plate VI Fig. 16. Cholangiomatous changes in hepatoma of rainbow trout. Hematoxylin-eosin. 500 X. Fig. 17. Extensive cirrhosis seen in hepatoma of Pennsylvania rainbow trout. Hematox- ylin-eosin. 250 X. NIGRELLI a JAKOWSKA PLATE I FIG. I FIG. 2 FATTY DEGENERATION, REGENERATIVE HYPERPLASIA AND NEOPLASIA IN THE LIVERS OF RAINBOW TROUT. SALMO GAIRDNERI V NIGRELLI a JAKOWSKA PLATE II FIG. 3 FIG. 4 FIG. 5 FATTY DEGENERATION, REGENERATIVE HYPERPLASIA AND NEOPLASIA IN THE LIVERS OF RAINBOW TROUT, SALMO GAIRDNERI NIGRELLI & JAKOWSKA PLATE III rmmei . C W.’^ 'tJ-ii’S v FIG. 7 FIG. 9 FATTY DEGENERATION, REGENERATIVE HYPERPLASIA AND NEOPLASIA IN THE LIVERS OF RAINBOW TROUT, SALMO GAIRDNERi NIGRELLI & JAKOWSKA PLATE IV i . * V; ■ ^ . = ■- v. ;: - mMik »« ]$:A i0%0$8j$0m^m ,; L : -1' — wmmxm . ,. ~,-.h ,--'t- - - ; , .' ■ . • ^,' i; :: : : ■^■L,;*j$s?:ii;v. *■■$ • ■ „ ■•. ■. *. - / O*^ »A Ti? V-.V{ i ’;••»->.'<:• i 4< - A > > ' 0 : . '. • ' r’-., - - '..' ^ ■ .- „ 1 • ■- «&«*;•* F-VV& *vP%S5?Sfev’ FIG. 17 FATTY DEGENERATION. REGENERATIVE HYPERPLASIA AND NEOPLASIA IN THE LIVERS OF RAINBOW TROUT. SALMO GAIRDNERI NEW YORK ZOOLOGICAL SOCIETY GENERAL OFFICE 30 East Fortieth Street, New York 16, N. Y. PUBLICATION OFFICE The Zoological Park, New York 60, N. Y. OFFICERS PRESIDENT VICE-PRESIDENT SECRETARY TREASURER Fairfield Osborn Laurance S. Rockefeller George Wall Merck David H, McAlpin SCIENTIFIC STAFF: John Tee-Van General Director William G. Conway. . Associate Director, Zoological Park Christopher W. Coates. . Director, Aquarium ZOOLOGICAL PARK Joseph A. Davis, Jr.. .Associate Curator, Mammals Grace Davall Assistant Curator, Mammals and Birds William G. Conway. .Curator, Birds Herndon G. Dowling. Curator, Reptiles Charles P. Gandal . . Veterinarian Lee S. Crandall General Curator Emeritus William Beebe Honorary Curator, Birds AQUARIUM James W. Atz Associate Curator Carleton Ray Assistant to the Director Ross F. Nigrelli Pathologist & Chair- man of Department of Marine Biochem- istry & Ecology C. M. Breder, Jr Research Associate in Ichthyology Harry A. Charipper. . .Research Associate in Histology Sophie Jakowska ..... Research Associate in Experimental Biology Klaus D.Kallman, . . . , Research Associate in Genetics Louis Mowbray Research Associate in Field Biology Homer W. Smith Research Associate in Physiology GENERAL William Bridges . . Editor & Curator, Publications Sam Dunton Photographer Henry M. Lester . .Photographic Consultant DEPARTMENT OF TROPICAL RESEARCH William Beebe Director Emeritus J ocelyn Crane Assistant Director David W. Snow Resident Naturalist John Tee-Van Associate William K. Gregory. . . .Associate AFFILIATE L. Floyd Clarke Director, Jackson Hole Biological Research Station EDITORIAL COMMITTEE Fairfield Osborn, Chairman James W. Atz William G. Conway William Beebe Lee S. Crandall William Bridges Herndon G. Dowling Christopher W. Coates John Tee-Van ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME 46 • PART 2 • SEPTEMBER 25, 1961 • NUMBERS 5 TO 10 PUBLISHED BY THE SOCIETY The ZOOLOGICAL PARK, New York Contents PAGE 5. Morphological Effects of Low Temperatures during the Embryonic De- velopment of the Garter Snake, Thamnophis elegans. By Wade Fox, Charles Gordon & Marjorie H. Fox. Text-figures 1-4.. 57 6. The Orang-utan in Sarawak. By George B. Schaller. Text-figure 1.. , 73 7. Observations on the Feeding, Shedding and Growth Rates of Captive Snakes (Boidae). By A. J. Barton & William B. Allen, Jr.. 83 8. The Feeding Mechanism of Fiddler Crabs, with Ecological Considerations of Feeding Adaptations. By Don Curtis Miller. Plate I; Text-figure 1. 89 9. Hybridization Experiments in Rhodeine Fishes (Cyprinidae, Teleostei). Intergeneric Hybrids Obtained from Acheilognathus lanceolata X Rhodeus amarus and Rhodeus amarus X Acheilognathus tabira. By J. J. Duyvene de Wit. Plate I 101 10. Some Observations on the Metamorphosis of the Frog Rana curtipes Jerdon. By Lucy Lobo. Plate I; Text-figure 1. . 103 5 Morphological Effects of Low Temperatures during the Embryonic Development of the Garter Snake, Thamnophis elegans 1 Wade Fox, Charles Gordon2 & Marjorie H. Fox Louisiana State University, School of Medicine and Department of Biology, New Orleans (Text-figures 1-4) Since scute patterns and scute (“scale”) counts are among the principal taxonomic characteristics of reptiles, any experimen- tal evidence suggesting that scutellation can be altered by environmental factors is of consider- able interest to herpetologists. One of us (Fox, 1948a) reported one such experiment in which exposure to low temperatures during embryonic development resulted in decreased numbers and modified patterns of scutes in garter snakes. If such accidents of development can happen in the laboratory, conceivably they can also occur under natural conditions. Although the data presented in 1948 were statistically significant, they were not altogether convincing because of the small number of sur- viving experimental litters and the small size of these litters. Hence, the experiment was repeated in 1948 under essentially the same laboratory conditions as in 1947 (reported in 1948). Methods The subspecies of garter snake used in this ex- periment was the same as was used in the pre- vious report. This was referred to as Thamnophis elegans atratus according to the designated clas- sification at that time (Fox, 1948b). However, this population of garter snakes was of the terrestrial form which later (Fox, 1951) became known as Thamnophis elegans terrestris. None of the snakes used was of the semi-aquatic type which Fox (1951) recognized as a distinct al- though sympatric race of garter snake identical to the type specimens of Thamnophis elegans atratus (Kennicott). lExperiment conducted at the University of Cali- fornia, Berkeley. 2Partially supported by a grant from the Louisiana Heart Association. The garter snakes, Thamnophis elegans ter- restris, were collected from a 25-mile stretch of the Pacific coast which included the 1947 col- lecting site on Skyline Blvd. of the San Fran- cisco Peninsula in San Mateo County, California. Since the Skyline Blvd. area had been collected heavily during the previous three years, a more extensive collecting area proved necessary in order to find adequate numbers of gravid female snakes. All animals were collected during the last week of April and the first week of May as it had been determined from previous collecting that gravid females obtained during this period would already be inseminated and would either have ovulated recently or would probably ovu- late within a few days following capture. Immediately upon capture 18 gravid females were placed in a cool room (experimentals) in which the temperature ranged from 65° to 85° F, and four gravid females were placed in a warm room (controls) in which the temperature ranged from 75° to 95°F, the same temperature ranges utilized in 1947. These rooms were lo- cated in a small, incompletely insulated green- house and the temperature variations reflected variations in the outside environmental tempera- tures. The temperature rose and fell simulta- neously in both rooms but a marked differential was maintained between them. The small control sample seemed permissible in view of the very large control sample available from the previous experiment (1947) and the desirability of mak- ing the experimental sample as large as possible. The four control females littered 46 young in early August. By September the experimental mothers had not littered and it was decided to terminate the experiment. The mothers were killed and fetuses were obtained from 15 of them. The others contained partially reabsorbed 57 58 Zoologica: New York Zoological Society [46: 5 Table I Samples Sample Size Ventrals P3 Subcaudals P3 1 9 1948 Experimental 63 150.3 + 10.01 138 _ 1602 65.6 + 9.3> 48 -IV 2 Mothers of 1948 Experimental 12 154.3 + 3.9 149 - 160 .03 71.2 +4.0 67 -80 .005 3 9 1948 Control 20 154.5 + 3.5 150 _ 158 .008 73.7 + 3.8 68 -76 <.001 4 9 1947 Control 50 154.9 +4.1 147 - 167 .005 70.4 +2.9 62 - 77 .001 5 1948 Experimental 65 154.6 + 11.0 134 - 163 74.7 +11.2 48 -86 6 ^ 1948 Control 26 160.8 + 3.4 154 - 164 <.001 82.7 + 6.5 71 -89 <.001 7 ^ 1947 Control 58 158.7 + 3.6 152 - 168 .007 80.5 + 3.7 71 -88 <.001 1 Mean and standard deviation. 2 Range. 3 Probabilities of significance of differences between the female experimental sample and their mothers and two female control samples, and between the male experimental sample and the two male control samples. embryos and hard masses of yolk in their ovi- ducts. The experimental fetuses were in various stages of maturity but all had fully com- pleted the development of scute patterns. The young snakes and their mothers were preserved in alcohol or formalin and stored in jars until 1958. During the interval three of the litters dried out so badly that reliable scale counts could not be made. Therefore the exper- mental sample is based on 1 2 litters yielding 1 28 fetuses. In a few other instances the number of scale rows around the body was questionable due to dehydration of the specimens. These were also omitted from the calculations. The number and condition of the following scutes were observed on each specimen : ventrals, subcaudals, longitudinal rows around the body, upper labials, lower labials, preoculars, posto- culars, temporals, loreals, nasals, chin shields and anal plate. Special abnormalities were noted wherever they appeared. The significance of differences between the experimental litters and (1) their mothers, (2) the 1948 controls, and (3) the 1947 controls were calculated (Tables I, II, III, and IV) . Males were tested against males and females against females for all scute counts. Within the experi- mental samples and the control samples the lit- ters were pooled and no allowance was made for variation between litters. In cases of bilateral characteristics, left and right sides were tested independently. Student’s t test was used in cal- culating the significance of the differences in numbers of ventral and subcaudal scutes.3 All other characteristics were tested by formula for chi-square (including Yates’ correction factor). Probability values were taken from Fisher’s tables. Results Ventrals and Subcaudals.— The female experi- mental fetuses (Table I, Row 1) had signifi- cantly fewer ventral and subcaudal scutes than their mothers (Row 2), the females of the 1948 controls (Row 3) and the females of the 1947 controls (Row 4). The mean numbers of ven- trals and subcaudals of the male experimental sample (Row 5) were very significantly smaller than those of the 1948 (Row 6) and 1947 (Row 7) controls. Seven litters averaged significantly lower in numbers of ventrals and/or subcaudals than the averages of the control samples or the natural population (Text-figs. 1 and 2). Two litters averaged approximately the same as the latter averages, and three litters averaged slightly high- 3A comment is in order concerning the assumption of homogeneity of variance in the groups under study. The hypothesis that the variances were equal in the two populations of experimentals was accepted; likewise, the hypothesis that the variances were equal in the five control groups was accepted. The hypothesis that the variances were equal in all seven groups was rejected. Therefore, when comparisons are made between the experimental and control groups by means of the /-test the assumptions of the /-test are not met. In spite of the above, the approximate /-test was used and the appro- priate number of degrees of freedom estimated. 1961] Fox, Gordon & Fox: Effects of Low Temperatures on Embryonic Garter Snake 59 er than the control averages. Extreme reductions in scute numbers occurred in six of the experi- mental litters. In one of these the four males averaged 137.8 ventrals and 57.3 subcaudals, whereas the lowest averages for males in a single control litter were 155.4 and 75.6 respectively. The lowest litter averages for female experimen- tal fetuses were 143.7 ventrals and 54.3 subcaud- als, whereas the lowest averages for a control litter were 153.0 and 68.1 respectively. It is of interest to compare each litter with its respective mother. Ventrals (Text-fig. 1) in the female fetuses of eight experimental litters and subcaudals (Text-fig. 2) in six were marked- ly lower than the numbers of these scutes on their respective mothers. In four litters the averages of the females were more or less equal to those of their respective mothers, but in no case was the litter average markedly higher than that of the mother. Additionally, in litter No. 10 (Text- fig. 1) the two surviving female fetuses were normal but three of the four males possessed the smallest number of ventrals encountered. Scale Rows.— The numbers of longitudinal scale rows on the body were counted at the neck, thoracic region and at the caudal end of the body just in front of the vent. In using chi-square to test the differences between samples, frequencies of occurrence of 20 scale rows or less were tested against frequencies of 21 rows or more in the neck and thoracic regions; for the caudal end of the body frequencies of 16 rows or less were tested against 17 or more. This race of garter snake usually has 19 scale rows at the neck. No wild populations in the range of the subspecies were found with more than 21 rows and only in the northernmost pop- ulations were a few individuals found with fewer than 19 rows (Fox, 1948a). The majority of experimental males and females possessed 19 scale rows at the neck (Table II, Rows 1 and 5) but three had only 17 rows, one had 22 and two had 23. Hence, animals exposed to cool temperatures had a wider range of variation than did the control samples. The number of scale rows at the neck of the mothers of the experi- mental did not differ significantly from those of their female offspring (Table II). The six individuals with extremely low (17) or extreme- ly high (22, 23) numbers of scale rows were all born to mothers with 19 scale rows. A peculiar inconsistency appeared in the 1948 study. The controls, both males and females, showed a high tendency for 21 scale rows at the neck, whereas nearly all the controls of the 1947 experiment had 19 rows. This resulted in a very significant P value when comparing the 1948 experimentals with the 1948 controls. However, when comparing the 1948 experimentals with 1947 controls no significant differences were found between the females, and the experimen- tal males revealed an unexpected significantly greater tendency towards an increased number of scale rows (Table II). There are 21 scale rows in the thoracic region (the usual position of the maximum number) in about 75 per cent, of the wild population (Fox, 1948a). The 1948 experimentals showed, a con- siderably greater tendency toward a reduction to 19 scale rows than did their mothers, the 1947 or 1948 controls (Table II), or the wild popula- tion. A surprisingly large number of the experi- mental males had only 17 scale rows in the thoracic region, whereas no count this low was found among the controls or wild population. Nearly all individuals with 17 scale rows were born to mothers with 21 rather than 19 scale rows. This suggested that the experimental con- ditions played a more significant role in deter- mining fetal scutellation than did the maternal pattern. The 1948 male controls displayed a greater tendency toward 21 scale rows in the thoracic region than did the 1947 male controls. As in almost all specimens from the natural population in the area of collection, 17 rows of scales were present at the caudal end of the body of nearly all controls. Although there was a greater tendency toward reduction of this number in the experimentals, this trend did not test to be significant (Table II). Nearly all of the individuals showing a posterior reduction of scale rows were from four litters. One of these litters included one individual with 14 rows, seven individuals with 15, one with 16 and two with 17. The mother of this litter had a typical scale formula: 19-21-17. An unexpectedly large number of male experimentals with 18 or 19 scale rows occurred in three experimental lit- ters. Labials— In T. e. terrestris the customary number of upper labials is 8 (Text-fig. 3A), but a tendency for this number to be reduced to 7 (Text-fig. 3B) has been found in popula- tions from cooler climates (Fox, 1948a). Male and female experimental fetuses were consid- erably more variable (Table III) in this charac- teristic than the wild sample. In spite of the oc- casional occurrence of nine upper labials, the male and female experimental samples had sig- nificantly fewer upper labials (Text-fig. 3C, D, E, F) on both left and right sides than either their mothers, control fetuses or the natural popula- tion from the area of capture. Ten lower labials were present in 82 per cent, of the natural population. As with the upper labials, although this normal scale number was 60 Zoologica: New York Zoological Society [46: 5 130 135 140 145 150 155 160 165 170 I I 1 1 1 I I I I I 1 I I I'l I I I I I I I 'l"T'l I I ITT I IT r-I 1 11 I I'l NATURAL 55 POPULATION 57? I I EXPERIMENTAL LITTERS 6?| «H 2 3 7&[ 4 5 4 $ I1 JL 6d»l T 6 7 8 3<3Tf _i 5^1- 4^ 3$HH * 4<5H-1- — I 9 10 4<3,h“— 1 2 1 4 7rf4 12 l&T 4$! I — I — I — I — I — I I I I I 1 I I I I I L-J_l_i— J I I— I I I I I — I — I — I — I — I — I — I — I — I— I — I — I — I 130 135 140 145 150 155 160 165 170 Text-fig. 1. Range and mean number of ventral scutes found in the 12 experimental litters and compared with the natural population from the San Francisco Peninsula. The number of ventrals of each mother is indicated by the vertical arrows which are arranged in increasing order from bottom to top. occasionally exceeded by a few experimental fetuses, the strongest trend was towards a sig- nificant reduction (Table III; Text-fig. 3B, C, F; Text-fig. 4B). Most individuals of all but three litters pos- sessed a reduced number of upper and lower la- bials. Of these three litters, one contained an unusually high number of individuals with 11 1961] Fox, Gordon & Fox: Effects of Low Temperatures on Embryonic Garter Snake 61 45 50 55 60 65 70 75 80 85 90 m i i ii i ii i i i ir i i r i i i t m i i i rm inmiiiiiiiiii NATURAL 0 POPULATION 57 + I EXPERIMENTAL LITTERS 52^h 3? | * — — — — — 2- 2 3 4 5 6 7 8 9 10 12 9? p<5*l- 7?| 1 1 6?HH * 3d*l — f — I 2$ hH * * 5cfh ?!■ 7. TJ O « 20 .12 O T3 G 1 3 46 .56 2. Sungei Katibas area Forest Dept. J About 100 Between Sungai Knowit and Katibas Forest Dept. 1 infant collected (1956) 3. Between Ulu Batang Balui and Ulu Batang Baleh near border Forest Dept. 10-20 4. Sungei Pila, Mengiong Forest Reserve (Sungei Mengiong south of Batang Baleh) Forest Dept. Less than 20 5. Sungei Baram near Long Moh Forest Dept. 1 male shot (1958) 6. Border region near Lubok Antu Maias Protection Commission inated over 600 square miles of forest in that area. Two hundred and ninety square miles are set aside as forest reserves. The map (Text-fig. 1 ) indicates the forest reserves, which harbor a sparse but continuous population of orang-utans. Records from within the past two years outside the reserves are indicated by dots and crosses. Reports of isolated populations and stragglers from areas not visited during the study are listed in Table I.1 The numbers correspond to those on the map. Locations 3 and 4 appear to represent populations of unknown size in the border hills between Sarawak and Indonesian Borneo. Loca- tion 5 is a straggler, for orang-utans are un- known to the natives of the area (Smythies, pers. comm.). But the whole border region is nearly uninhabited and seldom penetrated, and it is pos- sible that orang-utans occur sparsely along the whole range between locations 4 and 5. Loca- tions 1 and 2 are included in the 625 square mile Lanjak-Entimau forest reserve, which is sur- rounded by shifting cultivation. Location 6 ap- parently represents a scattered remnant in an area greatly disturbed by agriculture. The only other distributional data come from infant orang-utans captured by Dayaks and con- fiscated by the government. Some of these rec- ords are unreliable, because infants are smug- gled over from Indonesian Borneo, and confis- i It was not possible, in the short time available, to check the accuracy of the various reports about isolated populations and stragglers. Sarawak lacks roads and most travel is by river, a time-consuming and expensive undertaking if uninhabited country is to be visited. Unknown territory along the Indonesian border is so extensive that censuses would require at least 6 months. Thus, the only records available are the scanty ones by forestry officers, which undoubtedly are second-hand for the most part. cations are usually made in or near towns. Thus, there are records from Lundu, Sibu and Bau near Kuching, regions from which there are no recent records of free-living animals. Numerical Status Although no population estimates are availa- ble prior to 1959, it is apparent from the earlier literature that orang-utans have declined dras- tically in the past hundred years. Hornaday (1885) not infrequently encountered animals twice in the same day while traveling along rivers on which I saw only scattered nests during the present study. Beccari (1904) noted a total of eight orang-utans in three separate encounters in one day while walking in the forest near the Batang Lupar River. In 1960 I observed animals on the average of only once a week. This population decline is not surprising in view of the hunting pressure to which the orang- utan has been subjected. James Brooke (1848) in 1840, for example, collected four animals during one days’ outing. Wallace (1869) shot or captured 17, Hornaday (1885) 43, and Beccari ( 1904) 24. All these animals were collected be- tween the Sadong and Batang Lupar rivers. In addition, numerous zoo collectors left with one or more infants, only obtained by shooting the mother. This slaughter continued until 1947 when the orang-utan became legally protected in Sarawak. Data on comparison of population levels in different parts of the range were essayed by in- dices based on the number of nests encountered divided by the number of hours walked. The most intensive transects were conducted in the lowland and hill forests of the 106.5 square miles of the Sabal and Balai Ringin reserves adjoining 76 Zoologica: New York Zoological Society [46: 6 the Indonesian border. In the former 4.3 (1.4- 8.0) nests per hour were recorded; in the latter 1.7 (0-3.2) nests. Going on the assumptions that my rate of walking is one mile per hour, that orang-utans build one nest per day, that all nests are counted in a 200-foot wide strip, and that nests remain clearly visible for six months, a total population of 36 independent animals was de- rived for the two reserves. This figure could per- haps be raised to 40 or 45 if infants are included. Assuming that a rough density of one animal per two square miles is typical, a population of only 150 orang-utans would appear to inhabit the 290 square miles of forest reserves between the Sadong and Batang Lupar rivers. The remaining forests surrounding the reserves have been ex- tensively disturbed and patches have been iso- lated by clearing and logging. Small, scattered populations persist, however, and if transect data in three locations are representative, an addition- al 200 animals could perhaps be added, bringing the number between the Sadong and Batang Lupar rivers to 350. This estimate agrees closely with that of Smythies (1960) who writes: “The only concentration of maias appears to be in the swamps between the Sadong and Batang Lupar (even here it is doubtful whether there are more than 200-300 left), and south of the Serian- Simanggang Road in the Balai Ringin and Sabal Forest Reserves.” A population estimate for the whole of Sara- wak is entirely conjectural in view of the paucity of distributional records in the unknown border regions. It can, however, be safely said that the total population is not less than 450 animals and probably not more than 700. Ecological Distribution Orang-utans are animals of primary equatorial rain forest and as such their habitat shows little ecological diversity. Much of their habitat in Sarawak lies below an altitude of 500 feet in the swampy forests bordering the coast and the major streams. To the south the land rises and becomes gently rolling, interrupted here and there by prominent hills. Along the Indonesian border of western Sarawak mountains reach a height of 3,000 feet. Above that altitude the trees become more stunted. Banks (1931) notes that the upper limit of orang-utan distribution over most of their range appears to be 3,000 feet, although he mentions a record from as high as 6,000 feet on Mount Kinabalu in North Borneo. Climatically Sarawak is characterized by rain- fall of 120 to 160 inches per year, and uniform high temperatures (72 to 88°F at mean sea level) and humidity (98% mean relative humidity at 0600). Rain falls during all months of the year, but is heaviest from October to February (Seal, 1958). In keeping with the climate, the forests consist almost entirely of evergreen trees, prob- ably exceeding 2,500 species in number, and covering some 75 per cent, of Sarawak. Several types of forest can be distinguished. Those in- habited by orang-utans are described below. Terminology and classification follows that of Smythies ( 1960) . No obvious habitat preference for certain types of primary forest could be de- tected in the orang-utan. Peatswamp Forest— Five vegetation types are recognized by the Forest Department in the peatswamp forest of Sarawak (Smythies, in litt.) , but only the mixed swamp forest type, which is by far the most prevalent, is of importance to orang-utans. Three of the forest reserves con- taining orang-utan (Sedilu-Sebuyau, Simunjan) are primarily of the mixed swamp forest type. Structurally this forest is typical rain forest with an upper canopy of big trees 120 to 150 feet high and several layers of poorly-defined lower-story trees. Groundcover is essentially absent and the peat is covered by water ankle to chest deep over much of the area. Except along the margins of the rivers where alluvium is present, the soil con- sists of over 95% organic matter in a semi-liquid, decomposed state. Drainage and irrigation are generally impracticable, and the soil is therefore of little use for agricultural purposes, other than the growing of sago palms. Although there are only about 200 big tree species, some of these are valuable for timber and are being extensively logged. Heath Forest— Small stands of heath forest occur over most of Sarawak on very poor, acid podsol soils. The trees are of relatively small size, lacking for the most part the upper story trees with big trunks. This habitat is of little im- portance to orang-utans, for only small patches exist in their present range; one visited supported a sparse population of animals. Riparian Forest— This type of forest, rarely half a mile wide, occurs on riverine alluvial soils. Where rivers flow rapidly Shorea spp. is espe- cially prominent; along slow-moving streams the valuable timber tree Eusideroxylon zwageri be- comes important. Because the soil is rich and the forest easily logged, this habitat has for all practical purposes disappeared. Dipterocarp Forest— This type is divided al- titudinally into (a) Lowland Dipterocarp forest, (b) Hill Dipterocarp forest and (c) Moss forest. Together they still cover 28,000 square miles (59 per cent.) of Sarawak. Structurally the low- land and hill forests resemble the mixed swamp forest. Both contain some valuable timber spe- cies, and large areas, mostly of the lowland type, 1961] Schaller: The Orang-utan in Sarawak 77 are being logged, (a) The lowland forest com- prises many hundred tree species dominated by such Dipterocarps as Shorea, Dipterocarpus and Dryobalanops. The topography is rolling to hilly and the ground well-drained. The soil is non- podsolic, and much is probably suitable for agri- culture. Orang-utans inhabit lowland forest in the adjoining Balai Ringin and Sabal reserves, and most likely in the Lanjak-Entimau reserve and in other border forests, (b) The hill Diptero- carp forest is confined to steeply sloping, often rocky, ground at altitudes above 1,000 feet. It is similar to the lowland forest except that the number of Dipterocarp species is fewer and the amount of big timber has decreased. Some of the slopes might be suitable for agriculture. Orang- utans inhabit this type in the Sabal reserve, (c) Moss forest occurs on the mountain ridges at about 3,000 feet in the Sabal reserve. Most of the big trees are lacking; others are of small girth and rarely exceed 80 feet in height. A few nests were noted in this habitat, but on the whole it is probably too restricted to be of much im- portance to orang-utans. Distributional Dynamics Although the major portion of Sarawak is still covered by the types of forests used by orang- utans, only a small fraction of the available habitat is frequented by the animals today. This, coupled with the fact that in the past they in- habited areas not occupied now, suggests that neither topography nor habitat had been limiting factors to dispersal and to numbers before the advent of agriculture. The high mountains (to over 5,000 feet) along sections of the Sarawak- Indonesian Borneo border act undoubtedly as partial barriers to a spread to and from the west, but numerous low gaps exist which appear to have been penetrated in the past all the way to the coast. Large rivers, like the Sadong and Batang Lupar, are definite barriers to direct movement perpendicular to their flow, but most rivers follow a course which does not hinder access to and from Indonesian Borneo. It is not improbable that hunting by man has, at least in some parts, been the major limiting factor before the expansion of agriculture. That primitive man hunted and ate orang-utans ex- tensively is shown by the remains in the Niah caves. In the past hundred years habitat destruction has been added to hunting as a limiting factor to dispersal in some areas, and the former has reached a stage where the orang-utan is effec- tively prevented from extending its range in the major area of concentration. The Sadong-Batang Lupar population is hemmed in by two large rivers and agriculture; the population in the Lanjak-Entimau forest is surrounded by shifting cultivation. It is doubtful if the orang-utan will ever adapt to habitat other than primary forest and old secondary forest. Although the animal can sur- vive in very lightly logged forest where the canopy is only partially broken, no sign was ever noted or records obtained from very young secondary forest or from forest so heavily dis- turbed by logging that large sections of the can- opy have been removed. Unfortunately the area with the greatest present-day orang-utan con- centration supports also a heavy human popula- tion. All marginal agricultural land will undoubt- edly have been put to use within the next few years, and the remaining patches of forest will be so logged and dissected with fields as to make them unsuitable for orang-utan. The only hope for the remaining orang-utans in Sarawak lies in the preservation of the forest reserves which still harbor animals. These re- serves are at present being logged in the acces- sible portions. In the peatswamp forest a 60- year cycle is used; in the lowland forest a 70-year cycle. The available undisturbed habitat in the reserves appears to be still adequate to support the orang-utan population. But logging has damaged some sectors so heavily that expansion of this activity can only be detrimental to the remaining animals. All hunting, too, must cease if the orang-utan is expected to survive. Illegal hunting for meat and infants appears to continue on a limited scale. Between 1956 and 1960 a total of 13 in- fants were confiscated by the government, some of which were, however, smuggled over from Indonesian Borneo. Behavior Orang-utans were seen four times and ob- served directly for 5% hours. Indirect data from nests were also collected. Locomotion — The orang-utan is basically arboreal, a habitat to which it is eminently adapted, for it possesses in effect four hands with which to clasp. It apparently frequents the ground only rarely. Wallace (1869) and Shelf ord (1916) received information of Dayaks injured by males which were surprised on the ground, and I was several times informed that orang-utans descend to eat the fruits and shoots of the palm Zalacca conferta in times of food shortage. My observations confirm those of Wallace (1869) and Hornaday (1885) that the orang- utan brachiates on occasion but that its usual mode of progression is by climbing. One subadult 78 Zoologica: New York Zoological Society [46: 6 was observed as it fled through the tree tops over a distance of Va mile. On large horizontal limbs it moved on all fours. It carefully but rapidly climbed into smaller branches, reached for and pulled in the foliage of the adjoining tree, secured a firm hold, and, with legs briefly free, swung across. Once it jumped three feet down to another branch, the only such instance noted. It also brachiated along a branch for about eight feet. The animal was greatly disturbed by my pres- ence, and moved along at a speed of about three miles per hour, making it barely possible for me to keep up on the uneven ground. Similar observations were made on two other orang-utans. A female with infant swung her legs free several times while holding onto a branch with her hands, primarily to reach an- other limb beneath her; but only once did she brachiate along a branch for about ten feet. A medium-sized male, who was followed by me for two hours, brachiated several times for distances of less than ten feet each. The rest of the time he climbed, always holding on to as many branches as possible. Twice he came to a gap in the canopy which prevented him from reach- ing the branches of the adjoining tree. In the first instance he inched out on a small horizontal limb until it bent under his weight, bringing him to within grasping distance of another branch. This he clutched and only when his hands were anchored securely did he release the first branch which had been retained with the feet. Another time he climbed into the crown of a vertical branch. Suddenly he threw his weight to one side, continuing his hold with one hand and foot only. The bending branch and extended arm enabled him to grasp the foli- age of the next tree. The grasping power in the feet enables the animal to hang with ease in practically any position. Hanging by one arm and foot while reaching for food is not unusual, and once a subadult hung two to three seconds by its feet alone. Group Size and Composition— The orang- utan is the least gregarious of the apes. Whereas chimpanzees and gorillas sometimes go in groups of ten or more and gibbons and siamangs occur most frequently in groups of three to six, the orang-utan is rarely encountered in aggrega- tions of more than two or three. Most of the records by early collectors are of single animals or groups of two. The largest aggregation noted by Brooke (1845), Wallace (1869) and Horna- day (1885) was three; by Beccari (1904) four; by Mjoberg (1930) six. Carpenter (1958) noted four, with a fifth animal in the vicinity. During the present study no single animals were seen. Three of the four encounters were of groups with two animals each, and only once did direct and indirect evidence point to a group of four animals. Group composition varies. Lone adult males have been encountered by nearly all explorers and collectors. Although Wallace (1869) writes; “I never saw two full grown animals together . . .,” other investigators have done so. Table II presents the composition of four groups as determined during this study as well as other compositions mentioned in the literature. The data show that the most frequent combinations are( 1 ) lone males, (2) females with one or more subadults of various ages, (3) pairs, and (4) small groups of subadults. Nothing is known about the stability of groups. Data from a group of four animals which had just moved into a section of forest suggests that groups may be quite plastic. The group consisted of a female and a small infant, which were seen, and a subadult and an adult male judging by the size of the nests and by information from natives who had encountered them. One night all four animals nested in ad- joining trees; another night the female with infant and juvenile nested together, but the male slept about 400 feet away; a third night the female with infant and juvenile again nested together but there was no sign of the male in the vicinity. The female was observed that morning, but the juvenile was not with her and searching for about two hours did not reveal it. Mother-Infant Relationship— Personal obser- vations were made on (1) a female with small infant, and (2) a female with a semi-independ- ent infant of about two years. ( 1 ) A female fed in a Durio tree at a height of about 70 feet with the baby clinging to her side, its head just below her armpit. The arms and legs of the infant were “spidery” and with but little hair, yet it was alert and looked around. The infant began to advance across the mother’s chest, and she cradled it in one arm. Restless, it clambered upward toward her shoulder, but she pulled it down. It moved up again, so the female pushed it over to its original position at her side. When it continued its wanderings, she laid it across her arm and groomed the back, using the long fingers of the same arm that held the infant. Finally she grasped a branch above her head with both hands, and the baby attached itself to her side, a place it occupied for the next hour while the female first fed and later climbed away through the trees. (2) A female fed in one tree and her large offspring in another 50 feet away. They joined 1961] Schaller: The Orang-utan in Sarawak 79 Table II. Composition of Orang-utan Groups Group Size Composition ; Source 1 1 young male Wallace (1869) 1 1 adult male Wallace (1869) Hornaday (1885) Beccari (1904) 1 1 female Wallace (1869) 2 1 male, 1 female Beccari (1904) 2 1 medium-sized male, 1 female This study 2 1 female, 1 infant Wallace (1869) Hornaday (1885) 2 1 female, 1 semi-independent infant This study 2 2 subadults (of undetermined sex) This study 3 3 subadults Wallace (1869) 3 1 female, 1 subadult, 1 infant Brooke (1848) Hornaday (1885) 3+ 1 female, 2 or more subadults Wallace (1869) 4 1 male, 1 female, 1 subadult, 1 small infant This study 5 2 females, 2 juveniles, 1 male in the vicinity Carpenter (1958) and the infant sat beside its mother, one arm across her back. The infant was about five feet from her when she spotted me. Immediately she reached over and snatched the infant to her chest. As the female climbed into the next tree, it moved and held on to her back, one arm over her shoulder, the other grasping her side. The female traveled about 500 feet, stopping in- termittently to look down at me. While moving the infant clung either to her back or to the side; while sitting it always stayed at the side. Al- though fully capable of climbing independently, the infant not once released its hold on the mother and stayed with her as she built the night nest an hour later. Foods and Food Habits— Orang-utans are primarily frugivorous. They apparently also eat the leaves of certain trees, according to Horna- day (1885) and others, but no direct evidence for this was obtained. The shoots and fruits of the palm Zalacca conferta are eaten, according to the natives. One instance was noted in which an orang-utan had gnawed the inner lining of bark from a tree (probably Ganua motleyana ) and spit out the chewed pulp. At times the animals raid cultivated fruits at the edge of the forest. Perhaps insects and other animal foods are sometimes eaten. Ryhiner (1958) observed an orang-utan “occasionally catching an insect on the bark.” Mrs. Harrisson (pers. comm.) maintains that they eat the eggs of birds in the wild. Table III presents a list of forest trees which produce fruits eaten by orang-utans. Information on food habits was supplied by natives and was reliable in the three cases when I was able to check it. Many of the fruits listed are eaten also by man, Hylobates moloch and various monkeys. Hornaday (1885) confirms that orang-utans eat the cultivated Durio and Nephelium. Animals were feeding when observed at 0845, 1100, and 1625. A male and female fed quietly on the small olive-sized fruits of Calophyllum. The only sound was the steady patter as seeds rained on the leaves below— a sound which first attracted attention to the animals. They plucked several fruits in a row by pulling them off be- tween thumb and index finger; they then popped them all into the mouth, chewed, and ejected the seeds through pursed lips. Similar behavior was noted in a female and large infant, except that the whole of the small fruit of Grewia was eaten. Feeding ceased as soon as the animals saw me. A female with small infant, on the other hand, continued to feed in spite of my presence from 0845 to 1005 on the muskmelon-sized fruit of the durian ( Durio carinatus). She had nested during the night in the adjoining tree, and was feeding when first found. The spiny durian fruits were plucked by hand, usually one at a time. Once she put her mouth close to a durian, tapped the back of the fruit with one hand and at the same time grabbed it between her teeth and de- tached it by jerking her head back. Three times fruit were transported to another branch as far as 20 feet away. Twice one was carried in the mouth, and a third time one was held in the mouth and another with one foot. When eating she held them lengthwise in the long hand and rapidly bit off chunks of tough outer skin. The soft pulp was also discarded. Apparently only 80 Zoologica: New York Zoological Society [46: 6 Table III. Fruits Eaten by Orang-utan Forest Type Species Cultivated fruits Dario zibethinus Nephelium lappaceum Artocarpus spp. Peatswamp forest Pongamia, probably P. pinnata Polyalthia sp. Quercus sp. Pometia pinnata *Durio carinatus Nephelium melanomisoum (?) Copaifera palustris Heath forest Sandoricum, probably S. emarginatum Mangifera, probably M. havilandii Lowland Dipterocarp forest Diospyros maingayi Garcinia sp. Quercus sp. *Grewia sp. Sandoricum sp. Nephelium sp. Durio sp. Hill Dipterocarp forest *Calophyllum cf. retusum * Indicates direct observation of feeding. the pecan-sized seeds were eaten, of which each durian contains about ten. During the time of observation the female consumed the seeds of at least 15 durian, wasting only half of one seed. Nests and Nesting Habits— Orang-utans build nests when wounded (Brooke, 1848; Wallace, 1869), in the evening before going to sleep and perhaps during the day when resting. St. John (1862) and Wallace (1869) mention that orang- utans hide in nests when pursued. No detailed accounts of orang-utan nests have been pub- lished. Hornaday (1885) briefly and erron- eously writes: “The nest of the orang-utan is simply a lot of small green boughs and twigs broken off by the animal, and piled loosely in the fork of a tree . . and Wallace (1869) and Beccari (1904) make essentially similar state- ments. Nests may be placed anywhere in a tree as long as there is sufficient support and a suitable quantity of small branches from which to con- struct a stable platform. The 228 nests observed in this study were found in forks made by the main trunk and smaller limbs, in the forks at the extremity of the larger limbs, along one or more horizontal branches, in the crowns and even between two close trees. Thirteen nests were examined in detail. The basic platform was con- structed of three to nine (mean 4.7) major branches broken in toward the animal. They were laid across each other without pattern or interweaving. Of 62 branches used in nest con- struction, only seven were completely detached. Most nests have a lining of small twigs about one foot in length, which the animal breaks off around the nest. In one instance the twigs were collected seven feet below the nest. The broken ends usually face outward, but in some nests they appear to be placed at random. The number of twigs in each nest varied from 0 to 34 (mean 10.0). The height of nests above ground is given by Wallace (1869) as 20 to 50 feet; Hornaday (1885) 15 to 40 feet; Banks (1931) 30 feet; and other authors quote similar heights. The heights of 228 nests in mixed swamp, lowland and hill forests was either estimated or deter- mined with a camera range finder. The results are summarized in Table IV. From Table IV it is evident that the majority of nests (82.3 per cent.) occur between 31 and 70 feet; in other words, in the middle of the lower story trees. However, a few nests are lo- cated in the lowest story and in the tops of the emergent trees. There are no records of nests on the ground. 1961] Schaller: The Orang-utan in Sarawak 81 Table IV. Height of Orang-utan Nests Above Ground Height of nest (in feet) Number of nests Percent 0-10 0 .0 11-20 2 .8 21-30 5 2.2 31-40 35 15.3 41-50 57 25.0 51-60 56 24.5 61-70 39 17.5 71-80 14 6.1 81-90 6 2.6 91-100 7 3.0 101-110 1 .4 111-120 5 2.2 121-130 1 .4 Total 228 100.0 Wallace (1869),Hornaday (1885) and others have described the building of nests in free- living orang-utans. The amount of time spent in construction and the elaborateness of the nest depends probably on the weather conditions and general disposition of the animal. Nest building was observed once when an agitated female built a nest up 120 feet during a downpour at 1730. She squatted in a fork among the smaller branches. With one hand she broke in a limb, pushed it under foot while reaching for the next one with the other hand. The latter as well as a third branch were also placed beneath her feet. She paused briefly, then tore off a final branch and laid it across the others. The time required for building this crude nest was approx- imately 10 seconds. Reaction to Man— Orang-utans usually avoid contact with man, but in areas where primary forest adjoins cultivation and the animals are not hunted, they may become used to the noisy pres- ence of Dayaks working in their fields. The reactions of orang-utans to me standing beneath their tree varied. On the one end of the scale were two subadults who fled soundlessly and as rapidly as possible; on the other a female with small infant who, except for brief glances, ignored me while feeding and finally moved off at a leisurely pace, once descending to 45 feet to urinate directly above me. This same female showed no response to natives talking at a dis- tance. During the other two encounters the ani- mals fled slowly, stopping intermittently to pro- duce vocalizations and other sounds, and to throw branches to the ground. Sounds— Sounds were emitted only when the animal was aware of my presence and presum- ably annoyed. Under similar circumstances Hornaday (1885) notes deep, guttural growls, and Wallace (1869) howling. Three types of sound were made interchangeably but continu- ously for 45 minutes by a female with large in- fant: ( 1 ) A smacking or kissing sound was made by tipping the head back, pursing the lips with the lower one protruding beyond the upper one and drawing the air in. Three times the female put the knuckles of one hand to her mouth and kissed them loudly. The kissing sound was usu- ally followed by (2) a gluck-gluck-gluck that resembled loud gulping of liquid. The lips were pursed as above, and the throat moved as if she swallowed air. After the gulps the female pro- duced several times a sound that can be best described as (3) a loud two-toned burp, starting on a long, low key to end on an abrupt higher note. These sounds were heard again during an- other encounter with a female and a male. Both gave the smacking sound, the male continuing it at the rate of once every few seconds for two hours. But only once did the male “gulp” and “burp.” Throwing of Branches— The fact that orang- utans may drop or throw branches if an observer stands beneath the tree has been recorded by Wallace (1869) and Attenborough (1957), al- though strangely neither Hornaday (1885) nor Beccari (1904) noted this behavior— probably because the animal was always shot before it had a chance to do so. I observed throwing of branches twice: (1) After fleeing from me for 40 minutes, a male stopped in a tree. He moved about and stepped on dry branches which broke off. Once he broke one off with the palm of his hand and watched the limb as it crashed to the ground. He also reached for branches around him, bent them until they snapped, wrenched them from the tree with a jerk, then threw them 82 [46: 6: 1961] Zoologica: New York Zoological Society downward. Half an hour later he hurled branches for 10 more minutes. (2) A female with a large infant spent 15 minutes throwing a total of about 30 branches varying in size from twigs to limbs ten feet long and three inches in diameter. Considerable effort was expended at times in tearing off the larger branches. Limbs were thrown in three ways: (a) she merely held the branch at her side and dropped it limply; (b) she looked down at me and swung the branch like a large pendulum, and at the peak of the arc closest to me she released it; (c) she lifted branches either as high as her chest or above her head with one hand and hurled them down forcefully. Whatever interpretation is given this behavior, there is no doubt that it in- duced me to jump nimbly at times and that it kept me effectively away from beneath the tree. Summary The distribution of the orang-utan encom- passes the northern tip of Sumatra and extensive sections of Borneo. A two-month study in Sara- wak, which comprises about v6 of Borneo, in all types of primary forest habitat, revealed that the orang-utan is limited in distribution to about 2,000 square miles and in number to not less than 450 and probably not more than 700 ani- mals. Hunting has been the major cause for their decline in the past and habitat destruction threatens them at present in their major area of concentration. Preservation of the continuous tracts of primary forest still inhabited by orang- utans is essential if the small population of this ape is expected to survive. All hunting for meat and specimens, including live animals for zoos, must remain prohibited. Behavior data is based on observations of 228 nests and four visual encounters for a total of 53A hours on eight animals. Orang-utans seldom descend to the ground. In the trees they are climbers, but brachiate on occasion. Groups number rarely more than four, and lone males and females are not infrequently encountered. Subadults appear to form their own group at times. There seems to be little group stability. Observations on a female with a small infant and a female with a large infant are described. Orang-utans are primarily frugivorous. Nests may be built nearly anywhere in a tree, but the majority are found 31 to 70 feet above ground. Basically the nest consists of three to nine branches broken in to form a platform and a lining of 0 to 34 twigs. A female built a crude nest in ten seconds. Reaction to man varies from immediate flight to seeming indifference. Sounds made by agitated animals can best be described as smacking, gulping, and burping. Throwing of branches while the observer stood near the tree was noted in two animals. Bibliography Attenborough, David 1957. Zoo Quest for a Dragon. London. Banks, E. 1931. A popular account of the mammals of Borneo. Journal Malay. British Asiat. Soc., IX (2): 1-139. Beccari, O. 1904. Wanderings in the Great Forests of Bor- neo. London. Brooke, James 1848. Narrative of events in Borneo and Celebes. Vol. 1, London. Carpenter, C. R. 1938. A survey of wildlife conditions in Atjeh, North Sumatra. Netherlands Committee for International Nature Protection, Comm. No. 12, pp. 1-33. 1958. Soziologie und Verhalten freilebender nichtmenschlicher Primaten. 10 (11): 1-32. Sonderdruck from the Handbuch der Zoologie. Harrisson, Tom 1949. Large Mammals of Borneo. Malayan Na- ture Journal, IV (2): 70-76. Hornaday, W. T. 1885. Two Years in the Jungle. London. Mjoberg, Eric 1930. Forest Life and Adventure in the Malay Archipelago. London. Ryhiner, P., with D. Mannix 1958. The Wildest Game. London. Seal, John 1958. Rainfall and sunshine in Sarawak. Sara- wak Museum Journal, VIII (11): 500-544. Shelford, R. W. 1916. A Naturalist in Borneo. London. Smythies, B. E. 1960. Annual Report on the Forest Department for the Year 1959. Kuching, Sarawak. St. John, Spenser 1862. Life in the Forests of the Far East. Vol. 1, London. Stott, Ken, Jr., & C. Jackson Selsor 1961. The Orang-utan in Northern Borneo. Oryx, 6 (1): 39-42. von Koenigswald, G. H. 1958. Remarks on the prehistoric fauna of the great cave at Niah. Sarawak Museum Journal, VIII (12): 620-625. Wallace, Alfred 1869. The Malay Archipelago. 2 vols., London. Yerkes, R. M. & A. W. Yerkes 1929. The Great Apes. Yale University Press, New Haven. 7 Observations on the Feeding, Shedding and Growth Rates of Captive Snakes (Boidae) A. J. Barton & William B. Allen, Jr. The Stony Brook School, Stony Brook, Long Island, New York, and the Highland Park Zoological Gardens, Pittsburgh, Pennsylvania Introduction Opportunities for recording the growth rates and observing certain activi- ties of the large boids are exceedingly limited. Field studies are non-existent because of the difficulties of restraining and measuring in the wild creatures of such great size and strength, and among captive specimens, those responsible for them are usually reluctant to subject them to the necessary handling for even so worthy an objective as the collecting of new data. Few observations of even the feeding schedules and shedding rates of these great ser- pents have been recorded. Thus our knowledge of many aspects of their life histories is very meager. The Highland Park Zoological Gardens in Pittsburgh has had in its care several well-ad- justed pythons and boas, snakes with which cer- tain liberties may be taken. The senior author initiated a series of observations during his tenure as Zoo Herpetologist from 1946 to 1952, observations which his successor, the junior au- thor, resumed and expanded, beginning in 1953 and continuing through the present. A summary of their findings is presented in this paper. Reptilian Growth Rates The misconception that reptiles generally grow slowly has long been widely accepted. It is surprising that this fallacy has been so ten- acious when one considers the volume of data refuting it, data that are based on several rep- tilian orders. Heller (1902) reported that a juvenile Gala- pagos tortoise ( Testudo vicina) increased in weight from 29 pounds to 130 pounds in three years. Ditmars (1933) observed the growth rate of captive alligators ( Alligator mississip- piensis) and found that newly hatched young having a length of eight inches and a weight of only 1.75 ounces could in six years attain a length of 72 inches and a weight of 72 pounds. Breeding adults continued this rapid growth for a time, so that a six-foot eleven-inch specimen grew 55 inches in six years to a length of eleven feet six inches. Mcllhenny (1935) validated Ditmars’ figures with a field study that yielded essentially identical information. Pope (1957) has reported that an Indian Rock Python ( Python molurus) in his custody was “about 24 inches” long on hatching. At the age of one year its length had reached 60 inches and its weight 3.5 pounds; at two years, 100 inches and 21 pounds; at three years, 125 inches and 38 pounds. Oliver (1952), reporting the first year’s growth of three molurus hatched in the New York Zoological Park, stated, “One year ago on this date . . . our three young Indian Rock Pythons emerged from their eggs. They were then only a little more than one foot in length and weighed under four ounces . . . On their first anniversary the largest measured four feet two inches in over-all length and weighed a pound and a half . . . The other two individuals are only slightly smaller than their brother.” Wall (1921) mentions one brood of this species in which the hatchlings were “about 2 feet.” In a second brood, the young upon hatching aver- aged two feet five inches. Some of these grew eleven inches in only four months. It is clear from all these observations that young reptiles may indeed grow rapidly. There is, however, less information available regarding the growth rate of pythons and boas once they have reached maturity. Loveridge (1945) recorded a Python reticulatus at the London Zoo that grew from a length of ten feet to twenty-one feet in ten years. “Another took fourteen years in which to grow from 19 to 24 feet. On this basis we can roughly estimate 83 84 Zoologica: New York Zoological Society [46: 7 growth as proceeding at the rate of a foot per year in the period between 10 and 20 feet, but at only about half that rate subsequently.” Hence, even among the initiated, the notion per- sists that these giant snakes grow rapidly at first, but can increase their dimensions only very slowly after they have achieved adult propor- tions. The following data lead us to reconsider this hypothesis. Python sebae Our own observations agree with earlier con- clusions that young boas and pythons grow rapidly. When an African Rock Python was re- ceived on October 3, 1951, its length of slightly over two feet indicated an age of less than one year. Thirteen months later it measured be- tween three and three and one-half feet and had a greatest diameter of three-fourths of an inch. Since the snake shed its skin seven times during this first year, there was a length increment of some two inches per shedding. No periodic measurements were taken subsequently, but on May 9, 1957, the python was found to be nine feet three inches in length and 37 pounds in weight. With only a few months of the first year unknown, the average annual increase in length for this snake’s first six years was 14 inches. On June 18, 1960, its length was 11 feet 2 inches and its weight 56 pounds. Thus during 37 months of its seventh, eighth and ninth years, it had an average annual increment of only 7.6 inches and 6.3 pounds. Its food intake during this period totaled 148 pounds, or 7.8 pounds of food for every pound of weight gained. The animal has been kept in a temperature between 73° and 78° Fahrenheit and supplied with ex- cess food all 12 months of the year. A record of the food it consumed is presented in Table I. Python molurus The Indian Rock Python, a light-colored fe- male, was within a few inches of nine feet when received on December 18, 1947. Because the snake had a tendency toward rectal prolapse, as well as for other reasons, it was measured but twice during its ten years’ tenancy in the Pittsburgh Zoo’s reptile wing. On January 21, 1955, it measured just 13 feet in total length. On January 7, 1958, it was found dead and at this time measured 13 feet 8 inches in over-all length. Its weight a few weeks earlier had been found to be 39 pounds. The slow growth during its last three years and the decid- edly low ratio of weight to length are clearly not normal, and thus limit the value of these data. Pope’s observations (1957) on the same species, already noted, should be compared. The present specimen increased its length by only 56 inches in ten years, despite a total food intake of 654 pounds. Forty-eight inches of this total was gained during its first seven years, or an average of seven inches per annum. Records of this speci- men’s food intake and shedding frequency are presented in Table II. Python reticulatus Perhaps the most spectacular snake now on display in the United States is “Colossus,” the male Reticulated Python received at the Pitts- burgh Zoo on August 10, 1949. The locality from which it was taken is unknown, but the fact that it was shipped to the United States from Singapore suggests that it was probably of Malayan origin. On arrival, its length was just 22 feet. After having rejected the fowl and rab- bits offered it during its first two months, it ac- Table I. Food Ingested by a Captive Python sebae, 1951-1960 Year Feedings Food Taken Total Weight of Food (lbs.) Mice Rats Rabbits Other 1952* 17 32 8 juv. — _ 2.5 1953 33 110 1 1 juv. - 2 English sparrows 11.0 1954 20 91 4 juv. - 2 pigeons 8.3 1955 26 46 1 19 2 starlings, 1 guinea pig 47.7 1956 10 - 2 8 2 guinea pigs 27.5 1957 12 - — 11 3 ducks 48 1958 13 - — 12 3 squirrels 52 1959 8 - — 11 — 48 1960 9 - - 6 2 chickens, 1 duck 44 83A yrs. 148 289 ♦First eight months. 1961] Barton & Allen: Feeding, Shedding and Growth Rates in Captive Snakes 85 Table II. Food Consumption and Shedding Frequency in a Captive Python molurus, 1948-1957 Year Feedings Food Taken Total Weight Dates of Shedding Rats Rabbits Other of Food (lbs.) 1948 14 45 — 1 pigeon, 1 jungle fowl, 3 ducks 59 Apr. 16, Aug. 5 1949 13 69 - — 67 May 4, Aug. 3 1950 8 26 4 — 36 Mar. 7, May 19, Sep. 23 1951 12 — 18 1 duck 80 Mar. 4, Aug. 7, Oct. 26 1952* 6 1 11 — 59 Apr. 5, June 10, Aug. 7 1953 12 — 7 6 ducks, 2 guinea pigs 68 Mar. 29, Oct. 5 1954 6 — 1 8 ducks 52 Mar. 12, Sep. 27 1955 4 — 5 — 34 No record 1956 11 — 12 3 ducks 85 ” 1957 13 — 15 7 ducks. 114 9% yrs. 99 3 pigeons 654 *First eight months. cepted the first pig offered. This was on October 14 when it took a 15-pound suckling. Thus be- gan a pattern of regular feeding that has con- tinued until the present time (Table III). This Reticulated Python has accepted nothing but pigs; rabbits, chickens and ducks have been refused on several occasions. When hungry, the snake strikes and seizes the prey with its mouth immediately upon the animal’s introduction into the cage. It is instantly enmeshed in one or one and a half coils. This grasp is usually retained for about 15 to 25 minutes although the prey rarely survives longer than three minutes. Stunned food-animals receive the same treatment as active ones. After the long waiting period, the mouth grip is freed and the coils are relaxed without being released, as the snake begins a thorough investigation of its food, the tongue flicking leisurely in and out throughout the process. If the snout and ears of the prey chance to be buried in the gravel of the cage floor, the snake prolongs its search and finally grasps the Table III. Food Consumption and Shedding Frequency in a Captive Python reticulatus, 1949-1960 Year Feedings Pigs Total Weight of Food (lbs.) Dates of Shedding 1949* 4 4 98 Dec. 2 1950 9 9 287 Feb. 11, Apr. 29, July 1, Sep. 12, Nov. 10 1951 6 6 203 Feb. 1 1, Apr. 18, July 10, Oct. 6, Dec. 20 1952t 7 7 202 Mar. 30, July 6 1953 6 6 169 Feb. 10, Aug. 25 1954 5 5 161 Feb. 13, Dec. 22 1955 5 5 154 Apr. 16, Oct. 23 1956 4 4 108 Mar. 8, July 3, Sep. 30, Dec. 20 1957 5 5 136 Mar. 10, May 13, Aug. 5, Oct. 8 1958 5 5 128 Apr. 1, June 8, Nov. 4 1959 6 6 184 July 8, Aug. 31, Nov. 19 1960 6 6 161 Feb. 11, Aug. 2, Nov. 1 1 1 years 68 1991 *Last five months. tFirst nine months. 86 Zoologica: New York Zoological Society [46: 7 animal again, nearly always by the shoulder, and drags it a foot or two before resuming its inves- tigation. Should this procedure fail to disclose either the snout or an ear, the snake may lose interest and abandon the attempt to feed, or it may repeat its effort to locate one of these cru- cial points by again moving the prey. Apparently this snake will secure its mouth grip preparatory to swallowing at only one of two points: the snout or an ear. Indeed, the identification of one of these two points by the snake’s tongue may possibly be thought of as a releaser mechanism which triggers the engulfing behavior. The largest pig offered to this python weighed 54 pounds. It in no way taxed the snake’s capac- ity and was swallowed in 64 minutes. This Reticulated Python was 22 feet long when received on August 10, 1949. On June 4, 1951, it was approximately 23 feet 3 inches long, hav- ing increased about 15 inches in 22 months. Per- haps our most accurate size data for this speci- men were obtained on February 24, 1954, when it was found to weigh 295 pounds. The junior author noted the weight in his daily journal on that date and stated: “This was shortly after its winter fast of AV2 months. The snake was then returned to its transfer cage. I opened the trans- fer door, which is 14 X 18 inches, and the python began to crawl back into its main cage. By placing a measuring tape on its back and work- ing along hand over hand so that it would not slip, I got a measurement of 27 feet, two inches, which I believe will be as accurate a measure- ment as we shall get until the snake dies.” The specimen had grown 47 inches in 3214> months. On November 15, 1956, it was found to measure 28 feet 6 inches, having grown 16 inches in the intervening 33 months. The snake has shed from two to five times a year. We cannot offer these length data as exact measurements, because of the way in which they had to be collected, but we are certain that they are accurate within a few inches. They show that large snakes can continue to grow at an appreciable rate after they have achieved adult dimensions. Having attained 85 percent, of the maximum length known for its species, this specimen is still growing at an average rate of 10.75 inches per year. This figure is in good agreement with a recent report from the New York Zoological Park of a 19-foot 4-inch, 170- pound Reticulated Python that gained 10 inches and 26 pounds in a single year. A. C. Stimson’s observations (in lift.) on a male Reticulated Python in the Houston Mu- seum of Natural History show how great may be the range of variation in growth rates under different conditions: “We have had it for 22 years and it was approximately 18 feet long when received. It now measures 24 feet, five inches and weighs 247 pounds.” Hagenbeck’s prize Regal Python was 28 feet in length, and weighed 250 pounds (Pope, 1937). This was the heaviest snake for which we find a definite record, so that the Pittsburgh specimen may be the heaviest snake ever weighed under reliable conditions. It would be unreasonable to suppose, however, that this captive would outweigh the record 32-foot specimen, for which no weight was recorded. We may assume that a snake of the latter dimensions greatly exceeds 300 pounds in weight. Eunectes murinus Our larger Anaconda measured 16 feet 4 inches and weighed 108 pounds when it was received from a shipper in Belem, Brazil, on June 13, 1950. After shedding on June 24, it re- fused living pigs, rabbits, ducks and carp through an eight-week period until, on August 10, it accepted a small mallard duck. Since that date only waterfowl— mallard ducks, white pekin ducks and a snow goose— have been accepted (Table IV). The Anaconda has grown more slowly than the pythons. On February 26, 1954, it was found to be 18 feet 7 inches long and to weigh 160 pounds. This represented a gain of 27 inches and of 52 pounds in 44 months. By March 21, 1957, it had attained a length of 19 feet 6 inches and a weight of 200 pounds, having increased its length by 1 1 inches and its weight by 40 pounds in 36 months. Most recently, measurements were re- corded July 10, 1960, when the snake’s total length was found to be 20 feet 7 inches. This represents an increment of 13 inches in 40 months. It was not weighed. The average annual length increment for this specimen during nearly eleven years in captiv- ity was only five inches. It may be that this growth rate would have been increased if more food had been accepted, but the snake refused to in- crease its intake despite every opportunity to do so. During the 81 months between the earliest and most recent weighings, a net gain of 92 pounds has been recorded. During this same period it consumed 539 pounds of waterfowl, or 5.86 pounds for every pound of weight gained. Despite our Anaconda’s very lethargic behavior, it cannot compare in efficiency with Pope’s (1957) phenomenal young Python molurus which reportedly gained a pound for every 1.76 pounds of food consumed. Pope (1955) men- tions a gravid female Anaconda having a length of 19 feet which weighed 236 pounds. Soon thereafter she gave birth to 72 young, each of 1961] Barton & Allen: Feeding, Shedding and Growth Rates in Captive Snakes 87 Table IV. Food Consumption and Shedding Frequency in a Captive Eunectes murinus, 1950-1960 Year Feedings Ducks Total Weight of Food (lbs.) Dates of Shedding 1950 8 20 54 June 24 1951 14 24 114 Feb. 7, May 30, Sep. 13 1952* 8 17 88 Jan. 25, June 13 1953 8 17 83 Mar. 3, Sep. 17 1954 7 13 65 Jan. 8, July 1 1955 7 16 80 Jan. 8, July 18 1956 6 11 55 May 8, July 14, Dec. 16 1957 7 12 72 May 16, Nov. 29 1958 9 12 68 May 2 1959 8 11 65 No record 1960 7 8 52 May 1, Aug. 1 1 lOVi years 89 796 *First eight months. which was about 38 inches. Assuming that each of these young weighed about a pound, the mother’s residual weight would have approxi- mated that recorded for our specimen of com- parable length. The average growth per ecdysis for the muri- nus is again seen to be a little more than two inches. This figure is notably constant in the four snakes reported herein. Literature Cited Ditmars, Raymond L. 1933. Reptiles of the world. Rev. ed. Macmillan, New York, xx + 321 pp. Heller, Edward 1903. Papers from the Hopkins Stanford Gala- pagos expedition, 1898-1899. XIV, Rep- tiles. Proc. Wash. Acad. Sci., vol 5, pp. 48-59. McIlhenny, Edward A. 1935. The alligator’s life history. Christopher Publ. House, Boston, vii +117 pp. Loveridge, Arthur 1945. Reptiles of the Pacific world. Macmillan, New York, xii + 259 pp. Oliver, James A. 1952. Our Indian rock pythons grew three feet in a year. Anim. Kingdom, vol. 55, p. 132. Pope, Clifford H. 1937. Snakes alive and how they live. Knopf, New York, xii + 238 pp. 1947. A python in the home. Chi. Mus. Nat. Hist. Bull., vol. 18, no. 4, pp. 4-5. 1955. The reptile world. Knopf, New York, xxv + 325 + xiii pp. 1957. Reptiles round the world. Knopf, New York, xv + 194 + xii pp. Wall, Frank 1921. Snakes of Ceylon. Cottle (Govt. Printer). Colombo, xxii + 581 pp. 8 The Feeding Mechanism of Fiddler Crabs, with Ecological Considerations of Feeding Adaptations1,2 Don Curtis Miller Duke University Marine Laboratory & Department of Zoology, Duke University (Plate I; Text -figure 1) Introduction THE ability of an organism to obtain nutri- tion from its environment is one of the basic requisites for survival and is thus a factor governing the distribution of animals. Should an animal become uniquely adapted to obtain food in a specific manner, the variety of habitats in which it can live becomes accord- ingly limited. This study of the feeding mechan- ism of three species of fiddler crabs, Uca pugila- tor (Bose), U. pugnax (Smith) and U. minax (Le Conte), was undertaken to observe feeding adaptations which may contribute to limiting their distribution. Ecological studies have elucidated the grosser conditions prevailing in typical habitats of Uca. Pearse (1914), working at Woods Hole, Massa- chusetts, and Schwartz & Safir (1915) at Cold Spring Harbor, New York, described the habitat of U. pugilator and U. pugnax. Both species live centrally between the tide marks, where the substrate consistency permits burrowing. How- ever, U. pugnax is limited mainly to the mud or clay areas, while U. pugilator lives on a sandy substratum. Gray (1942) studied U. minax at Solomons Island, Maryland, and noted that the substratum may vary from sand and mud, to clay, though this crab is mainly limited to hab- itats of lower salinity. The more comprehensive study of distribution of Uca, done in the Georgia 1 A portion of the field studies included here were supported by a grant from the National Science Foun- dation (G-5577); grant administered by Dr. F. John Vernberg. Complete reports of these studies will be made in future papers. 2 This work was submitted in partial fulfillment of the requirements for the Master’s degree in the Graduate School of Arts and Science, Duke University. salt marshes by Teal (1958), has generally sup- ported these observations. The same workers have also considered the food material of Uca. Both Pearse (1912), then working in the Philippines, and Schwartz & Safir found algae and vascular plant tissue to make up the bulk of the stomach contents in the Uca examined, though decayed animal matter and inorganic material were also present. By ex- perimental feeding, Gray found U. minax to in- gest a wide variety of foods, with the notable ex- ception of putrified material. In controlled experiments, Teal found U. pugilator and U. pugnax able to live on cultured marsh bacteria. In regard to the manner in which Uca feeds, Pearse (1912) described rather completely the action by which the minor chela passes material into the buccal cavity where the food to be in- gested is selected by the mouth parts. Working in Brazil, Matthews (1930) observed U. lepto- dactyla, a small species, to take single grains of sand into the buccal cavity, where adhering or- ganic matter is scoured off and the cleaned grain rejected. This scouring action is achieved by the movement of the second maxilliped en- dopodites across the bristled endite lobes of the first maxilliped protopodite. Matthews also ob- served spoon-shaped tips on some medianly- projecting bristles of the second maxillipeds. These bristles have become known as spoon- tipped hairs, mainly through the descriptive work of Crane (1941, 1943). Examining trop- ical Uca, as well as the three species of the northern temperate region, she observed a great number of spoon-tipped hairs in the sand-in- habiting species, while crabs living in muddy habitats have an abundance of woolly hairs on the second maxilliped. She suggested the con- 89 90 Zoologica: New York Zoological Society [46: 8 nection of such hair modifications with feed- ing. Altevogt (1957) has compiled a list of crabs which show a similar type of hair modifi- cation, which supports Crane’s generalization correlating the modification of the second max- illiped hairs with habitat. He observed that the spoon-tipped hairs aided in the ridding of coarse particles from the buccal cavity. He has also contributed to the understanding of the feeding mechanism with his description of the washing of coarse material from the buccal cavity by water from the branchial cavity, while the lighter detritus is suspended in the fluid and ad- heres to the mouth parts for eventual passage to the mouth. In this study of the feeding mechanism in Uca, the morphology of the mouth parts is described, with composite figures showing their relative positions within the buccal cavity. The various aspects of the sorting process within the cavity are considered in detail. Habitat limita- tion created by the mode of feeding is examined and species differences in mouth-part structure are analyzed, so that the ecological consequences of such specific adaptation may be considered. I wish to express my appreciation to Dr. F. John Vernberg, who directed this study, and to Dr. R. H. Siepmann for his translation of the German paper by Altevogt. Methods of Study While the grosser aspects of the feeding mech- anism were observed in the field with the aid of binoculars, or in aquaria in the laboratory, the more detailed processes occurring within the buccal cavity were observed with a dissecting microscope. Details of the mouth parts and as- sociated hairs were most easily discerned when the crabs were submerged in a pan of water. Since Uca is very hardy, the third maxilliped could be excised to permit observation of the actions of the other mouth parts. Careful re- moval of one or both of the first maxilliped endite lobes permitted similar observations of the appendages more immediate to the mouth. The manner in which the mouth appendages manipulate particles was seen by placing sand grains or pieces of modeling clay on the ap- pendages, while the direction of action of the mouth parts was observed by stimulating them with a probe. Semidiagrammatic figures of the mouth parts, drawn to show their relative position within the buccal cavity, are included to aid in the under- standing of the interrelationship of the append- ages. Though the hairs which cover the surfaces of the mouth parts have been generally deleted from the figures for simplicity, these hairs are described in detail in the text. Observations of the habitats in which the crabs live, as well as the studies of their general distribution, were made during the summers of 1957 and 1958 at Beaufort, North Carolina. The habitats have been characterized by the condition of the substratum, and by the to- pography and vegetation. Percentage of sand content of the substratum was determined by the hydrometer method as outlined by Bouyoucos (1928). Appendages and Mouth Structures of the Buccal Cavity The terminology adapted for description of the mouth parts follows that used by Snodgrass (1950, 1952). However, reference was also made to the general descriptions of crustacean appendages by Siebold (1874), of the Brach- yura by T. H. Huxley (1878), and of Callinectes sapidus by Lochhead (1950). The appendages are described here with reference to their rela- tive position within the buccal cavity while the crab is in a typical feeding position. At this time the buccal cavity is at a 45° angle in respect to the horizontal, which means that the upper edge of the third maxillipeds can also be termed the anterior edge, and the lower edge, the pos- terior. These terms are used interchangeably in this discussion. The third maxillipeds enclose the buccal cav- ity from beneath. The basal portion of the max- illiped, termed the protopodite, is composed of a coxopodite and a basipodite. An epipodite arises laterally from the coxopodite and extends along the base of the branchiostegite, where it func- tions as a valve to regulate the intake of respira- tory water before it turns into the gill cavity proper. The endopodite and exopodite extend from the basipodite. On the endopodite, the is- chiopodite and meropodite are greatly enlarged and join to form a broad plate. The endopodite is bent medially at the fourth joint, to fold the three terminal segments across the meropodite, and finally to extend downward along its median edge as an endognathal palp. The palp is flexible and is used to clean the eyes, antennae and an- tennules, as well as functioning in the feeding process. The exopodite of the third maxilliped borders the endopodite as a long segment, which terminates as a slender, many-jointed flagellum which extends medially across the buccal cavity. The peduncle of the exopodite shares with the broad plate of each endopodite and their ter- minal segments to form an operculum across the buccal cavity when the third maxillipeds are closed (Text-fig. 1A). Plumose hairs (setae) fringe the edges of the third maxillipeds, making the appendage more effective as an operculum. 1961] Miller: Feeding Mechanism of Fiddler Crabs 91 Text-fig. 1. The buccal cavity of Uca, ventral view. A. The maxillipeds, with the left third maxilliped re- moved. B. Right first maxilliped, maxillae and man- dibles. C. Right second maxilla, first maxillae, with mandibles in detail. 1 , third maxilliped endognathal palp; 2, third maxilliped exopodite; 3, third maxilli- ped endopodite; 4, first maxilliped flagellum; 5, first maxilliped endopodite “flap”; 6, second maxilliped exopodite; 7, second maxilliped endopodite (mero- podite segment); 8, first maxilliped coxal endite; 9, first maxilliped basal endite; 1 0, first maxilliped ex- opodite; 1 1, first maxilliped endopodite; 12, mandi- ble; 13, second maxilla basal endite; 14, second maxilla coxal endite; 15, first maxilla; 16, anterior The segmentation of the second maxilliped is similar to the third, though the two basal seg- ments are now united. The endopodite is com- posed of the five typical segments, but in con- trast to the third maxilliped the ischiopodite is fused with the basipodite, and the meropodite is greatly enlongated (Text-fig. 1A). The ter- minal segments of the endopodite extend med- ianly from the meropodite, with flexibility similar to those forming the palp of the third maxilliped. The exopodite of the second max- illiped also corresponds with that of the third maxilliped. The first maxilliped deviates from the general form of the other two maxillipeds as large in- ward-projecting lobes, or endites, extend from both the coxopodite and basipodite (Text-fig. 1 A, B) . Short bristles completely cover the outer surfaces of the endites to form a continuous brush-like surface across the lower half of the buccal cavity. The segments of the en- dopodite extend more along the sides of the cavity, with the meropodite elongated, as in the second maxillipeds, but narrower. A comb, made up of closely set, stout hairs, ex- tends from this segment to lie beneath the bristled endite lobes. The hairs achieve the ap- pearance of a comb due to their close and reg- ular spacing, which is maintained along much of their length by numerous setules along each hair that serve to hold them together on one plane. The fourth joint of the endopodite is again bent, making the terminal segments ex- tend across the buccal cavity. However, in the first maxillipeds, the terminal segments of the endopodites are broad and flattened, so that they appear as flaps. Defining the top of the buccal cavity, the flaps serve to concentrate mineral and food material in the vicinity of the mouth parts. Also, being heavily fringed with plumose hairs, they prevent particulate matter from entering the gill cavities by way of the ex- current openings, which are located immediately above. The exopodite of the first maxilliped is parallel to the endopodite, as described for the other maxillipeds. The terminal flagellum lies, along with the other flagella, across the upper surface of the wide distal portion of the endo- podite (Text-fig. IB). Together, the three pairs of exopodite flagella function to prevent mater- ial suspended in the water from entering the excurrent respiratory opening, for by rapid, upward-flicking motions, the flagella can set up lobe of labrum; 17, mandibular palp (endopodite); 18, first maxilla basal endite; 19, first maxilla coxal endite; 20, month region. 92 Zoologica: New York Zoological Society [46: 8 currents in the water which carry away such suspended material. The second maxilla is distinct in form in that a pair of bifid endites extends medianly from the coxopodite and basipodite. The lower, or coxal, endite takes the form of narrow finger-like pro- jections which are relatively devoid of hairs, while the upper, or basal, endite appears as one large lobe, though its primitive bifid condition is noted by a short slit which extends into the endite from its median edge (Text-fig. 1C). A comb, very similar to that of the first maxilliped, arises from the arched base of the coxal endite and extends beneath the bifid endite. The comb, with the endite, underlaps a portion of the first maxilla, while the upper endite lies under the first maxilla basal endite. The endopodite of the second maxilla is present only as a tiny spiked remnant extending from the proximal part of the upper endite. The exopodite has combined with the epipodite to form the long, flattened lobe called the scaphognathite. The lobe extends laterally into the gill chamber, where it acts as a pump, moving respiratory water as the second maxilla oscillates laterally. On the first maxilla, a relatively large lobe extends medianly from an arm of the coxopo- dite (Text-fig. 1C). The lobe is covered with fine, short hairs, except at its median edge where the lobe curves strongly upwards toward the mouth; there it is fringed with a few rows of stubby, stout bristles. These bristles are set so that they point upwards at a slight angle towards the mouth. Another arm extends from the cox- opodite of the first maxilla, which is regarded as the basipodite. This segment also bears an endite, which underlaps the mandible. The en- dite is slightly curved to fit the contour of the mandible, and its median edge is armed with stubby, stout bristles similar to those of the coxal endite. The small endopodite appears to be rudimentary here, while the exopodite and epipodite are absent from the first maxilla. The mandibles extend across the upper por- tion of the buccal cavity with broad quadrate basal parts carrying the large gnathal lobes which project underneath the mouth. In Uca, the biting edges are toothless, giving the gnathal lobes a blunt appearance. However, the lobes narrow down to thin, sharp edges as they curve inwards and the upper side becomes somewhat hollowed. Thus it is possible that the mandibles may be utilized for mastication. A three-seg- mented palp, made up of the terminal joints of the endopodite, extends along the anterior edge of the gnathal lobe and then turns downward to lie in the hollow above the biting edges of the mandible (Text-fig. 1C). Siebold has sug- gested that the palps serve as tactile organs for the mandibles. Neither the exopodite nor the epipidite are represented in this appendage. A pair of paragnaths lies above and close to the mandibles and immediately over the mouth. They are flat elongated lobes which project for- ward from the posterior ends of the lateral mouth folds. They have no musculature, accord- ing to Snodgrass (1950), though with their fringe of short plumose hairs they appear as rudimentary mouth parts. The mouth is a distensible opening above the lower portion of the mandibles. From the top of the mouth arises a triangular - shaped fleshy upper lip, called the labrum, which comes to lie beneath the mouth. A portion of the labrum also projects anteriorly above the mandibles as a small triangular lobe. The two lower sides of the lobe are grooved in such a fashion that the mandibular palps fit into it as they curve around the anterior edge of the mandibles (Text-fig. 1C). Thus, any vertical movement of the palps will cause a corresponding movement of the labrum, an upward movement lifting the labrum away from the mouth opening, while a down- ward movement serves to partially close the opening. Otherwise the mouth is without modifi- cation, but the elastic tissue quickly coalesces to form the tubular esophagus. The Feeding Mechanism In feeding, the fiddler crab scrapes the surface of the substratum with the minor chela. Among the mineral particles scooped up are detritus, algae, bacteria and perhaps nematodes. The minor chela carries this material up to the buccal cavity and passes it between the slightly gaping third maxillipeds. Within the buccal cavity, the food material is selected for ingestion and then passed on to the mouth. Unsuitable material is passed to the bottom of the buccal cavity and permitted to fall out between the third max- illipeds. This is mainly large inorganic particles. Ingestion of a small amount of coarse mineral material is apparently not harmful to the crab, for it can be rejected by the cardiac stomach if need be. However, the presence of a large quan- tity of coarse material, such as sand grains, may interfere with the preparation of food material for digestion by the small, closely-set teeth of the gastric mill. Such a condition does not ap- pear to be present in crabs feeding on a muddy substratum, for it is unlikely that fine, ingested silt particles would interfere with the functioning of the gastric mill. Thus, although there is little selection against such fine inorganic particles during feeding, as evidenced by the large per- centage of silt in the fecal pellets of mud-feeding crabs, selection does occur among crabs inhab- iting areas having a coarse substrate. 1961] Miller: Feeding Mechanism of Fiddler Crabs 93 One method by which food material is sorted from the coarse mineral fraction is described by Altevogt and termed the flotation process in the present discussion. It involves introduction of water into the buccal cavity to float light food and silt particles free from the heavier material. The lighter fraction is suspended in water and held between the mouthparts, especially between the maxillipeds, by capillary action, while the heavier particles are washed to the lower part of the buccal cavity. At the base of the third maxillipeds, outside of the buccal cavity, the rejected material forms a fluid ball which is subsequently removed by the minor chela. The water required for the flotation process is pumped from the gill cavity out through the excurrent opening by action of the scaphogna- thite. Flooding the buccal cavity, the water flows by gravity around the first maxilliped endopo- dite flaps and over the mouth parts. The reverse beat of the scaphognathite pulls a portion of the water back from the buccal cavity, while the re- mainder can return by way of the incurrent opening at the base of the branchiostegite. As the water is withdrawn, many of the fine, sus- pended particles adhere by capillary action to the mouth part surfaces, particularly to those covered with plumose hairs. Also, fine particles adhere to the two pairs of comb surfaces within the buccal cavity. The remainder of the sus- pended particles will be retained either by the plumose hairs fringing the distal flaps of the first maxilliped endopodites as the respiratory water is further withdrawn anteriorly, or by a similar fringe at the base of the third maxillipeds if the water passes out by that route. Experimentally, the flow of respiratory water into the buccal cavity can be stimulated by placing a flake of pablum on the first maxilliped endites. This re- action may be a result of weight on the bristles of the endites, or perhaps the bristles have gus- tative sensory properties. This sorting process, which utilizes the dif- ference in weight between the fine food material and coarse mineral matter as a criterion of selec- tion, is supplemented by the coordinated actions of the mouth parts. These appendages hasten sorting by passing the food accumulated on their various combs and hair-covered surfaces on to the mouth, as well as by aiding to rid the buccal cavity of the larger and heavier material. The mouth parts also appear to be capable of freeing food material from the coarse particles, which increases the efficiency of the sorting process. Food material is passed between the third maxillipeds and into the buccal cavity by the minor chela at a point near the tips of the en- dopodite palps. The tips of the palps lift outward and the endopodite plates swing open and lower slightly to make room for the food behind their broad, curved surfaces. The palps then return to their normal position and the endopodite plates partially flex closed, pressing the material up toward the first maxillipeds. The palps, aided by long hairs projecting from their tips, may also assist in pushing the material upwards. Sorting by the mouth parts begins as the material is moistened by respiratory water. The particular role of the first and second maxillipeds in the feeding process varies with the type of substratum on which the crab feeds. Uca pugi- lator typically inhabits protected sandy tidal areas. When the crab feeds on the sandy sub- strate, relatively little food material is placed in the buccal cavity by the minor chela in pro- portion to the number of sand particles. Also, much of this food material clings to the sand grains, rather than being freely suspended in the interstitial water. Thus the feeding process in- volves more than merely a separation of the heavy mineral fraction from the lighter material, as is achieved by the flotation process. Rather, if the crab is to obtain a very large amount of food, it must be capable of removing from the mineral particles any food material which clings to them. The efficiency of the feeding process depends on the thoroughness and rapidity with which this is achieved. The cleaning of adhering food material from sand grains is achieved in part, in U. pugilator, by modification of the bristles on the basal en- dites of the first maxillipeds. In general, the bristles are considerably enlarged, appearing quite stout and stubby. Toward the more median half of the endite they are broadened and flat- tened, with two edges of the bristles lobed from the base. The lobes curve inwards slightly to make the bristles cupped. Thus, the first max- illipeds of U. pugilator present an irregular sur- face against which sand grains may be scraped and food consequently removed as the second maxillipeds sweep the grains toward the sides of the buccal cavity. The bristles on the upper portion of the endites are set with their broad- ened surfaces perpendicular to the arc circum- scribed by the second maxillipeds, permitting food material to accumulate on the broad, cupped surfaces as it is freed from the sand. The bristles on the lower portion of the endites have their cupped surface turned upwards, which enables them to catch any fine material washed down by the flotation process. In addition the modified bristles also serve to brush fine material from the first maxilliped endopodite combs which underlie them. The cleaning process is also facilitated by similar modifications of the tips of some of the 94 Zoologica: New York Zoological Society [46: 8 hairs on the meropodite of the second maxilli- peds. These hairs are generally referred to as spoon-tipped. A spoon shape is achieved by the tips broadening and becoming shallowly cupped as the lobes near the tip arch inwards. There are five or six large lobes fringing two sides of the hair near the tip, in addition to a wide terminal lobe which is turned down over the tip to com- plete the cup. As the hair narrows proximally, it is fringed by several small lobes which grade into a series of serrate projections that soon be- come a mere fringe continuing along two sides of the shank for approximately half its length. These modifications make the hairs capable of effectively drawing large particles across the first maxilliped endites. Though some spoon-tipped hairs are present on the second maxilliped meropodites in all spe- cies of Uca considered here, it is in U. pugilator that the highest degree of hair-tip modification is found. This is true both in the number of hairs which are spoon-tipped, and in the size of the modified tip. Almost all the medianly-projecting hairs of the meropodite in U. pugilator are mod- ified (Plate I, Fig. 1). The short hairs covering the surface of the segment create a surface of stout, cupped bristles very similar to that de- scribed on the first maxilliped basal endites. This surface extends beyond the inner edge of the segment as the hairs become longer, with the more median hairs of the meropodite pro- jecting well into the buccal cavity. As the hairs increase in length, the depth of the spooned portion and the extent of lateral lobing increases slightly. Proceeding posteriorly along the length of the meropodite, the hairs in each row de- crease in length so that the spoon-tips are ar- ranged in diagonal rows, which places the broad- ened hair-tips closely together to produce a continuous surface. While the spoon-tipped hairs extending from the second maxilliped mer- opodite of U. pugilator vary in number among individuals, it may be generally stated that there are close to one hundred spoon-tipped hairs which project beyond the inner edge of the mer- opodite in the smaller crabs. In the larger in- dividuals over two hundred such hairs may be present on the meropodites. The shorter mod- ified hairs, which do not project beyond the inner edge of the meropodite, range in number from seventy to one hundred in this species. The spoon-tipped hairs on the second maxil- liped of U. pugilator are believed to function primarily in the feeding process. Although some of the sand taken into the buccal cavity will fall to the base of the third maxillipeds because of its weight, as the cavity is flooded with res- piratory water, many of the grains become lodged among the bristles of the first maxilliped endites, particularly the long bristles fringing the coxal endite. It is this sand which is picked up between the modified hair-tips of the mero- podite and swept across the bristled endites of the first maxillipeds, resulting in rapid and efficient removal of adhering food material. The sand grains are usually too large to be carried individually by a spoon-tip; instead, a grain is caught between neighboring spoon-tips, where it is firmly held as the close mat of broad- ened tips exerts pressure from the sides, and the stiff underlying hairs prevent the grain from falling through the mat. Thus, the spoon- tipped hairs, when placed closely together, ap- pear to contribute to the feeding process by providing an enlarged surface of stout hair- tips which can prevent sand grains from being forced up into or through the meropodite hairs when the second maxilliped moves across the bristles of the first maxilliped endites. Such a surface appears necessary in crabs which feed on a sandy habitat, for with loss of sand grains from the surface of the bristled endites only a small percentage of sand would be cleaned and little food material would be recovered. Food freed by the cleaning process is of a very light and filmy texture, which readily ad- heres to the modified bristles of the first maxil- liped endites and the fringes and spoon-tips of the second maxilliped meropodite hairs. Little food accumulates in the spooned portion of the second maxilliped hairs, but it is apparently passed directly to the first maxilliped endites as the meropodite moves over them. In addition to facilitating the cleaning of organic material sand grains, the sweeping action of the second maxillipeds across the buc- cal cavity also serves to carry coarse particles away from the central portion of the cavity. This can be considered as another general role of the second maxillipeds and of the spoon- tipped hairs in feeding. Once the particles have been carried to the sides of the buccal cavity, the meropodites lift away from the first maxil- liped endites, which permits the sand to fall from the appendages as a vibratory action of the second maxillipeds and the flooding of the buccal cavity with respiratory water provides impetus to free the grains from the hair surfaces. The utilization of water to remove coarse par- ticles from the buccal cavity is suggested by the fluid consistency of the ball of sand which accumulates at the base of the third maxillipeds for discard. On the maxilliped meropodite of the marsh fiddler crab, U. pugnax, fewer of the medianly- projecting hairs are spoon-tipped, as compared 1961] Miller: Feeding Mechanism of Fiddler Crabs 95 to the sand-inhabiting species, U. pugilator. Also, in U. pugnax the modified hairs are lim- ited to the upper portion of the meropodite (Plate I, Fig. 2). There are several rows of shorter hairs which do not project beyond the edge of the meropodite and which are well-lobed and broadened, but they are of a soft texture, which makes their effectiveness in manipulation of material questionable. However, there are five or six additional diagonal rows of stout, spoon- tipped hairs which extend beyond the meropo- dite, of a size and shape very similar to the spoon-tips in U. pugilator. The thirty to forty modified hairs in U. pugnax are not as closely grouped as in U. pugilator, somewhat decreasing the effectiveness of the narrow band of spoon- tips in picking up material. The band of spoon- tipped hairs does coincide in position with the upper edges of the first maxilliped endites in U. pugnax. Thus, the second maxillipeds appear to aid the feeding process primarily by picking up coarse particles from those edges of the en- dites and sweeping them away from the central portion of the buccal cavity. It is doubted whether as much food material is scoured from the coarser particles as a consequence of this action as in U. pugilator, for the bristles cover- ing the first maxilliped endites of U. pugnax are not stout, but have a fine, downy texture. The thoroughness with which the crabs are able to sort food from the mineral fraction of the substratum corresponds to the availability of food material. On the sandy beach inhabited by U. pugilator, while some detritus and nema- todes are present in the interstitial water, much bacteria and algae is on the surface of the grains. This crab appears to ingest a minimal amount of sand and the food must therefore be separ- ated from the mineral fraction. For the marsh- inhabiting U. pugnax, on the other hand, food is more readily available because the silt with which it is generally associated is of sufficiently fine texture to be ingested. In U. minax, modification of the second max- illiped hairs occurs to a much lesser extent than that observed in either of the other two species. There are very few short hairs present along the upper surface of the meropodite and they are mostly limited to its edges. The modification observed in many of the medianly-projecting hairs consists of a feathering of two sides of the hair, the increase in surface adapting it to handle fine particle matter. In addition, in the larger crabs, many of the longer medianly-projecting hairs are hooked at their tip. In the more medium-sized crabs (of approximately 13.0 mm. to 19.0 mm. carapace width) these hair tips are microscopically flattened, with delicate lobes fringing the two sides of the flat surface near the tips, which then blend proximally into the ser- rulate projections that feather the shaft. Also, among the medium-sized crabs, several of the longer hairs of the meropodite are spoon-tipped, although in U. minax the spooned portion is much smaller and accordingly more shallow than those previously described (Plate I, Fig. 3). The spoon-tipped hairs extend medianly from the upper portion of the meropodite as three or four diagonal rows, each consisting of five mod- ified hairs. These fifteen to twenty-four spoon- tipped hairs form a narrow band in the same position under the first maxilliped endites as in U. pugnax. However, in U. minax, the spoon- tips are widely spaced and are believed capable of doing little more to supplement the manipula- tion of material than what is typically achieved by the medianly-extending rows of unmodified hairs. The presence of only a few modified hairs in U. minax is significant in view of the greater availability of food in the high marsh. Since U. minax can readily obtain the food associated with the finely sorted muddy substratum or on partially decayed marsh vegetation, there is neither a need for a highly efficient method of sorting food nor a systematic means of ridding inorganic material from the buccal cavity. It is probable that the medianly-projecting hairs contribute to the feeding process merely by dis- persing material across the endites of the first maxillipeds, whereby a larger percentage of the particles within the buccal cavity are exposed to washing and respiratory water and food material is more readily separated for ingestion. In all three species of Uca considered here, the last two segments of the second maxilliped palps are fringed with stout spoon-tipped hairs along their lower edge. The modified hairs of the terminal segment have their cupped surfaces facing downwards, while the spoon-tipped hairs of the propodite have the cupped surfaces turned inwards. So placed, the short, well-lobed hairs make the flexible palps very effective in manip- ulating material within the buccal cavity. In addition to facilitating removal of material from among the mouth parts, these hairs also enable the palps to collect food material which is ac- cumulated on the hairy surfaces of the mouth parts, such as on the first maxilliped endites, or the fringes of plumose hair on the outer edges of the second maxillipeds, where it is deposited after the flotation process. The palps then carry the food to the maxillae, where it is closer to the mouth. The spoon-tipped hairs are conspicuously abundant on the palps of U. pugi- lator, while in the other two species they are 96 Zoologica: New York Zoological Society [46: 8 somewhat fewer in number and less deeply curved. Yet the basic arrangement of these spoon-tipped hairs remains fairly consistent among the three species, indicating the impor- tance of the palps in performing these functions in the feeding process, regardless of the charac- ter of the substratum on which the crab nor- mally feeds. The second and first maxillae work together to accumulate the finer material for ingestion. The upper pairs of endite lobes of the second maxilla appear to serve primarily to retain food material as it is sorted from the coarse particles, for they are abundantly covered with plumose hairs of moderate length. Underlapping these second maxillae endite lobes are several spoon- tipped hairs which project from the anterior edge of each first maxilliped basal endite. It appears that the hairs are modified primarily to function as surfaces against which the plu- mose surfaces of the second maxilla can rub, for the lateral movements made by the maxilla as the scaphognathite moves within the gill cav- ity do move the basal endites against these first maxilliped hairs. With this action, any fine material caught in the plumose hairs of the basal endite of the second maxilla will be carried toward the median edge of the segment, where it is in a position to be passed on toward the mouth. There are also a few hairs which pro- ject anteriorly from the second maxilla, which are similarly hooked. Likewise, these hairs would move fine material to the median edges of the first maxilla basal endites which lie above them. There is little species difference in the form or arrangement of the hair modifications discussed here. Food material is removed from the median edge of the second maxilla basal endite as the segment moves posteriorly and upward, scraping its plumose surface against the anterior edge of the coxal endite of the first maxilliped. This action places the fine material above the first maxillipeds and in a position to be carried to the maxillary combs by respiratory water. In addition, food can be passed from the first maxilla basal endites to the tufted edges of the mandibular palps, as the mandibles move later- ally above the endites. This material also ap- pears to be carried from the palps to the maxil- lary combs as the mandibles are washed with respiratory water. The combs of the second maxillae curve med- ianly from the base of the coxal endites, closely adhering to the bristled, convex surface of the first maxilla coxal endites. The combs appear to serve as surfaces to retain material carried into the upper portions of the buccal cavity by the flotation process. The maxillary comb and brush surface of the coxal endite of the first maxilla appear to perform the same function as that described for the first maxillipeds; how- ever the comb and brush surfaces of the max- illae appear to be capable of handling finer material since the hairs of these surfaces are set closer together. The fine food material and silt on the combs are transferred to the median edges of the first maxilla coxal endites as the endites brush the convex combs beneath them. As food material accumulates near the median edge of the bristled surface, the first maxilla can move forward a short distance, placing the inner edge of the coxal endite, and the food material, in the mouth. Disscussion Certain aspects of the feeding mechanism of Uca help to explain the ecological limitations imposed upon the crabs by their mode of feed- ing. Analysis of the feeding mechanism shows two general processes to be involved: flotation, and coordinated action of the mouth parts. Considering the flotation process, it is instruc- tive to examine the copious use of water from the gill cavities of the crabs in light of water conditions prevailing in their respective habitats. After the flooding of the buccal cavity with res- piratory water, some water is returned to the gill cavities, either via the excurrent openings as the scaphognathite reverses its direction of movement, or by way of incurrent canals, located at the base of the branchiostegite above the coxa of the chelae. The latter route was suggested by Altevogt. However, some water is lost as food is ingested, as well as when material is discarded from the buccal cavity. Additional water is lost by evaporation while the mouth parts are exposed to the air. Thus, to continue feeding, the crab must have access to an external supply of water in order to re- plenish its respiratory water. This factor is im- portant in limiting the areas where the crabs may feed, and is reflected by the moist condition of the material from which they prefer to feed. Of more general significance, the inclusion of the flotation process as an integral part of the feeding mechanism is one major factor prevent- ing Uca from living in a terrestrial habitat. Among the marsh-inhabiting fiddler crabs, the requisite of standing water for the flotation process affects its distribution little, other than limiting the crab to the intertidal areas, for within that area the substratum remains moist and water is available even during low tide in drainage depressions and in the depression sur- rounding the burrow entrance. Uca can readily 1961] Miller: Feeding Mechanism of Fiddler Crabs 97 replenish its respiratory water from such small pools by lowering the thorax into the water, submerging the incurrent canals at the base of the branchiostegite. U. pugnax has been ob- served to stray from its burrow while feeding and to lower its thorax to take up additional water, only to find the moisture insufficient. Immediately the crab returned to its burrow, where it could take up water, and then once again it began feeding. This relationship between the burrow and a source of respiratory water may be a key factor contributing to the high sense of burrow-centered territoriality which has been observed in U. pugnax. That such ter- ritoriality has not been observed with U. minax is in agreement with its preference for feeding in the muddier portions of the marsh, away from the burrowing area. In the sandy areas inhabited by U. pugilator, the need of water for the flotation process has a more pronounced effect on the movements of the crab. As well as being limited to the moist intertidal portion of the protected beach, the crab may also be required to move to the water's edge to feed, should the beach elevation at the burrowing area be too great for water to remain in the burrow during low tide. Both the tidal magnitude, as influenced by the phase of the moon and the season of the year, and the con- tour of the beach, will affect the water level, and thus be factors contributing to vertical move- ments of the crab during feeding. Ecological considerations may be deduced by examining the role of the mouth parts in feed- ing. Generally speaking, the mouthparts man- ipulate material within the buccal cavity by two coordinated actions: that which passes food material toward the mouth and one which re- moves coarse material from the central portion of the buccal cavity. Species differences are observed in the hairs covering the mouth parts in Uca, which reflect the relative importance of the several mouth part actions in the feeding process of each species. Since the majority of these mouth part differences can be correlated with the type of substrate on which the crabs typically live, they can be looked upon as mod- ifications which enable the crabs to feed in that habitat. Of the three fiddler crabs considered here, the mouth part hairs of U. minax appear to be the least modified, it being assumed that the undifferentiated hairs are the more primitive. This species typically lives in the mature Spar- tina marshes, well up the estuary where lower salinities are experienced. Field studies indicate that U. minax prefers to feed in low areas where the mud is very fluid. The crab also feeds ex- tensively on bacterial slime on decaying plant material, which is abundant throughout the marsh. The material carried to the buccal cavity by the minor chela is generally of such a fine texture that little sorting would be required by the mouth parts before ingestion could take place. Respiratory water is still pumped into the buccal cavity, however, where it doubtless serves as a solvent aiding in the dispersal of the fine material to the various mouth parts. Once on these hair-covered surfaces the mouth parts pass the fine silt and food material upwards to the mouth. Should material be rejected from the buccal cavity during feeding in U. minax, rejection oc- curs immediately after the material is placed within the cavity, and all is discarded. Since re- jection takes place without sorting, the palps of the second maxillipeds, rather than respiratory water, would serve to direct the material from the buccal cavity. The palps also aid in the ma- nipulation of particles within the cavity, for which they are particularly adapted due to the arrangement of the spoon-tipped hairs on the terminal segments. A similar modification oc- curs with the longer, medianly-extending hairs of the meropodite, although here it is limited to a minute hooking or flattening of the tip, which is fringed by delicate lobes and setules. These long hairs may increase the efficiency of the sorting process by spreading the material across the first maxilliped endites, where it will be more accessible for washing by respiratory water, as well as providing another surface on which fine material may accumulate. The presence of flat-tipped hairs on the sec- ond maxilliped meropodites in U. minax is felt to be a modification serving to increase the effi- ciency of the sorting process and to enable crabs to invade and survive in intertidal areas having a coarser substrate, where food material is not readily available. A slight advancement in this direction is seen with the modification of a num- ber of hair tips to a deeply spooned shape in U. pugnax, with an accompanying ability of the crab to feed in sandy portions of the marsh. The modified hairs, which may approach thirty in number, extend from the upper portion of the second maxilliped meropodites, where they are effective in removing coarse material from the upper edges of the first maxilliped endites and withdrawing material from the central part of the buccal cavity. The spoon-tipped hairs pres- ent on the second maxilliped palps are similarly arranged in both U. pugnax and U. minax, as is also the luxuriant fringe of plumose hairs on the outer edges of the second maxillipeds and on the maxillae, which basically adapts U. pugnax to feed on a muddy substrate. 98 Zoologica: New York Zoological Society [46: 8 The greatest modification of the hairs of the mouth parts appears in U. pugilator. Here the terminal lobing of the spoon-tipped hairs is more pronounced, making the tips more rounded and deeply cupped, which, coupled with the greater number of spoon-tipped hairs present, greatly increases the efficiency of the mouth parts in handling coarse material. The most striking sur- faces of closely set spoon-tips are those formed by the hairs extending medianly from the mero- podite of the second maxillipeds. The number and size of the spoon-tipped hairs of the second maxilliped palps are also increased in U. pugi- lator, making those two opposing surfaces of spoon-tipped hairs on the dactylopodite and pro- podite more effective. Thus the palps further supplement the action of the meropodite in manipulating material to be discarded from the buccal cavity, as well as carrying accumulations of food to the maxilla for passage to the mouth. This efficiency in handling coarse material within the buccal cavity enables the crab to feed in such habitats as the protected sand beach. In addition to the necessity of selecting against coarse mineral particles, another problem which appears to confront fiddler crabs living in a sandy habitat is the paucity of available food material, for there the organic matter is closely associated with the coarse sand. In such a form, the food material is not as readily available for ingestion as is the organic matter in the marsh, which is associated with fine silt. Thus, a beach- inhabiting fiddler crab must also be capable of efficiently separating from the sand, a large per- centage of the food material which is taken into the buccal cavity. In U. pugilator, the efficiency of the sorting process is increased by a cleaning action which involves the drawing of sand par- ticles over the bristled surface of the first maxilliped basal endites by the second maxilli- peds. The bristles of the basal endites are modi- fied in U. pugilator in a manner similar to that of the tips of many of the hairs on the second maxillipeds. Such enlarged bristles projecting perpendicularly from the endites offer a brush- like surface against which sand grains may be scraped and food material removed as the sec- ond maxillipeds sweep the grains across the endites. The bulk of the freed food is then re- tained on the shallowly-cupped endite hairs, which are placed in such positions as to catch the maximum amount of loose food material which may be washed, toward the base of the buccal cavity by respiratory water. The closely set spoon-tipped hairs of the second maxilliped meropodites facilitate this cleaning process by providing a sufficiently rigid surface to achieve efficient scouring action on the sand grains as they are carried across the first maxillipeds. Distribution of Uca is generally believed to be governed by a complex of physical and biotic factors, operative during the pelagic develop- mental stages of the crab, as well as after it has invaded the intertidal areas. However, in con- sidering the influence of the observed modifica- tions of the mouth part hairs on the distribution of Uca, only the adult crabs and their intertidal habitat need be examined, for it is only in the adult stage that the mouth-part hairs are well developed, and thus fully functional. In U. minax, the condition of the mouth-part hairs indicates that they are not able to sort food mate- rial from coarse inorganic matter with any effi- ciency. This leads to the inference that the crab would be able to survive only in areas where it has access to an abundance of food material on the surface of a silty substratum, such as in the salt marsh. However, U. minax does not inhabit all portions of the tidal marsh, but it is generally restricted to the high regions, where a dense mat of Spartina alterniflora roots and rhizomes im- part great stability to the substrate. In view of this restricted distribution within the marsh, though nutritive conditions appear fairly con- sistent throughout, there are doubtless limiting factors other than an inability of the crab to ob- tain food. It is probable that the stability of the substrate, which differs from that of the lower marsh, is critical for the crabs’ burrowing activities. The more advanced modification of some of the hairs on the mouth appendages in U. pugnax does appear to be an adaptation permitting the crab to be more widely distributed. This is dem- onstrated by the crabs’ presence in the sandy areas of young marshes. There feeding is facili- tated by the band of modified hairs extending from the upper portion of the second maxilliped meropodites, which can manipulate coarse ma- terial in the central region of the buccal cavity. The inability of U. pugnax to survive on pro- tected sandy beaches, however, where surface food material is less available, reflects on the un- modified condition of the first maxilliped basal endites and the wide spacing of the spoon-tipped hairs present on the second maxilliped, which limits the efficiency with which food can be cleaned from the coarse grains. Thus the lack of modified hairs is one factor limiting U. pugnax to marshy areas, where food material can be more readily obtained. The ability of U. pugilator to obtain sufficient food from its sandy habitat is seen to lie prin- cipally in the highly modified condition of the first and second maxillipeds. However, it is diffi- cult to believe that this structural adaptation renders the crab incapable of surviving in 1961] Miller: Feeding Mechanism of Fiddler Crabs 99 marshy areas, particularly when its presence in the sand-fringed young marsh is considered. There U. pugilator inhabits primarily the high sand rim when it is flooded by a spring tide. During the neap tides, when the high sandy areas remain dry, the crab moves into the siltier areas of the marsh where the surface substratum may contain only 40% sand. With the next spring tide, a large portion of the population returns to the sand fringe, demonstrating an apparent preference for the sandier areas. Population pressure may play a role in this latter movement, for when U. pugilator is in the silty marsh areas, it shares the space with U. pugnax. But more generally, the ability of U. pugilator to feed with efficiency in a sandy area is looked to as a basis for this preference. When the crab feeds from a sandy substrate, the inorganic fraction of the material passed into the buccal cavity is of suffi- cient size that it may be readily discarded during the sorting process, and little inorganic material is ingested. This, coupled with the effective cleaning process in U. pugilator, indicates that material of a relatively high total nutritive value is ingested during feeding. In the marsh, how- ever, the efficiency of the feeding process is diminished, since the silt with which the organic matter is associated is of too fine a texture to permit separation of the food material. This is further borne out by examination of the fecal material of Uca, which indicates that these fiddler crabs do not remove fine silt from the organic matter. Therefore, in the marsh the nutritive value of the material ingested by the crab is reduced in comparison to that of sand- inhabiting species. While observations show U. pugilator to be capable of surviving in areas with little sand in the substratum, the mouth part modifications which have aided it in sur- viving in sandy areas have also exposed it to such advantageous nutritive conditions that the sandy habitat has become the more favorable. However, since this factor appears to be more of a preference than an absolute requirement, there are probably other factors, such as the consistency of the substratum for burrowing, which prevent U. pugilator from living in the marsh. Summary 1. A study has been made of the feeding mech- anism of three species of fiddler crabs: Uca pugilator, U. pugnax and U. minax. 2. Fiddler crabs feed by scraping organic mat- ter from the surface of the substratum, with subsequent separation of food, material from the coarser fraction within the buccal cavity. 3. The two basic aspects of the feeding mech- anism are the coordinated actions of the mouth parts and a flotation process which utilizes water from the gill cavities. 4. Species-specific modification of the hairs which cover the mouth parts are described and correlated with the characteristic sub- stratum on which each species feeds. 5. The distribution of the fiddler crabs is dis- cussed, with consideration of ecological adaptations of the feeding mechanism with- in each species. References Altevogt, R. 1957. Untersuchungen zur biologie, okologie und physiologie indischer winkerkrabben. Z. Morph, Okol. Tiere, 46:1-110. Bouyoucos, G. J. 1928. Making mechanical analysis of soils in fif- teen minutes. Soil Sci., 25:473-480. Crane, J. 1941. Eastern Pacific Expeditions of the New York Zoological Society. XXVI. Crabs of the genus Uca from the west coast of Cen- tral America. Zoologica, 26:145-208. 1943. Display, breeding and relationships of the fiddler crabs (Brachyura, genus Uca ) in the northeastern United States. Zoologica, 28:217-223. Gray, E. H. 1942. Ecological and life history aspects of the red-jointed fiddler crab, Uca minax (Le Conte), region of Solomons Island, Mary- land. Publ. No. 51, Chesapeake Biol. Lab., 1-20. Huxley, T. H. 1878. A Manual of the Anatomy of Inverte- brated Animals. New York: D. Appleton & Co., pp. 263-276. Lochhead, J. H. 1950. Callinectes sapides., in Selected Inverte- brate Types, ed. by F. A. Brown, Jr. New York: John Wiley & Sons, pp. 448-450, 458-460. Matthews, L. H. 1930. Notes on the fiddler crab, Uca leptodactyla Rathbun. Ann. and Mag. Nat. Hist., (Ser. 10) 5:659-663. Pearse, A. S. 1912. The habits of fiddler-crabs. Philippine J. Sci., (Sec. D) 7:113-133. 1914. On the habits of Uca pugnax (Smith) and U. pugilator (Bose). Trans. Wise. Acad, of Sci., Arts, and Letters, 17:791-802. Schwartz, B. & S. R. Safir 1915. The natural history and behavior of the fiddler crab. Cold Spring Harbor Monogr. No. 8. 100 Zoologica: New York Zoological Society [46: 8: 1961] SlEBOLD, C. T. VON 1874. Anatomy of the Invertebrata. (Translated by W. I. Burnett). Boston: James Camp- bell. pp. 328-329. Snodgrass, R. E. 1950. Comparative studies on the jaws of man- dibulate arthropods. Smithsonian Misc. Coll., 116:1-85. 1952. A. Textbook of Arthropod Anatomy. Ithaca: Comstock Publishing Associates, pp. 158-179. Teal, J. M. 1958. Distribution of fiddler crabs in Georgia salt marshes. Ecol., 39:185-193. EXPLANATION OF THE PLATE Plate I Fig. 2. Median portion of the left second maxilli- Fig. 1. Median portion of the right second max- Pe(^ meropodite, U. pugnax, dorsal view. illiped meropodite, U. pugilator, dorsal Fig. 3. Median portion of the right second max- view. illiped meropodite, U. minax, dorsal view. MILLER PLATE I FIG. 3 THE FEEDING MECHANISM OF FIDDLER CRABS, WITH ECOLOGICAL CONSIDERATIONS OF FEEDING ADAPTATIONS FIG. 1 FIG. 2 9 Hybridization Experiments in Rhodeine Fishes (Cyprinidae, Teleostei). Intergeneric Hybrids Obtained from Acheilognathus lanceolata X Rhodeus amarus and Rhodeus amarus X Acheilognathus tabira J. J. Duyvene de Wit Zoology Department, University of the Orange Free State, South Africa (Plate I) IN a previous publication1 the successful in- tergeneric hybridization of female Rhodeus ocellatus (Kner) and mal e Acanthorhodeus atremius (Jordan & Thompson), both of Jap- anese origin, was reported. All 15 hybrid speci- mens obtained were males. A number of these were crossed back to the maternal species, R. ocellatus, and the offspring obtained from this combination consisted of interfertile males and females. This paper deals with results obtained by crossing (1) female Acheilognathus lanceo- lata (Temminck & Schlegel) of Japanese origin with male Rhodeus amarus (Bloch) of western Europe, and (2) female Rhodeus amarus with male Acheilognathus tabira (Jordan & Thomp- son) of Japanese origin. Berg (1948-1949) con- sidered R. amarus a subspecies of R. sericeus. Recently, however, Holcik (1959) has revised the taxonomic status of the different forms of European bitterling which had been considered to represent only minor variations within the taxon R. sericeus amarus. As a result of exten- sive comparative investigations he proposed to subdivide the group into three lower taxonomic units, "Rhodeus sericeus amarus amarus (Bloch)” which inhabits the system of the river Elbe in Germany, "Rhodeus sericeus amarus danubicus” which populates the Danube river in Austria, and "Rhodeus sericeus amarus svetovi- dovl” occurring in the Dnepr and Bug rivers in the U.S.S.R. The form that has been used in the present investigation is identical with the form called "R. sericeus amarus amarus” by Holcik. Hybrids between A. lanceolata and R. amarus Because of the incompatibility of their spawn- ing behavior, A. lanceolata and R. amarus are incapable of interbreeding under laboratory conditions with the aid of freshwater mussels. Therefore, artificial insemination was applied in the usual way. Thirty-four eggs of one female A. lanceolata were inseminated with sperm of one male R. amarus. Fertilization was complete. All the larvae hatched and 33 of them reached the free- swimming stage within 24 days. Twenty arbi- trarily selected fry were raised to the adult stage. A representative adult hybrid specimen is illus- trated in Plate I, Fig. 3, together with specimens of the parental species. All hybrids showed a male phenotype and displayed full nuptial colors in the breeding sea- son. Tubercles were abundant on top of the snout. Spawning behavior, however, was poorly developed. The quantity of milt produced by stripping was very small, and from this it is con- cluded that spermatogenesis was impaired. All hybrids were fairly uniform in size. With respect to body size they were, however, not intermedi- ate between the parental species. They attained the body lengths of the maternal species, A. lanceolata, which is larger than R. amarus. The hybrids were more resistant to infectious dis- eases than either parental species. In their gen- eral appearance and liveliness they clearly dem- onstrated the presence of hybrid vigor. Hybrids between R. amarus and A. tabira In R. amarus and A. tabira, the patterns of spawning behavior are very similar. Therefore the interbreeding of both species with the aid of freshwater mussels was attempted under labora- tory conditions. 101 102 Zoologica: New York Zoological Society [46: 9: 1961] One male A. tabira and two female R. amarus were placed together in an aquarium provided with three South African najads, Aspatharia wahlbergi (Krauss). After five weeks, five fry were seen swimming near the surface of the water. Only one specimen reached the adult stage. It is illustrated in Plate I, Fig. 6, together with specimens of the parental species. The hybrid was more or less intermediate be- tween the parental species and showed a male phenotype. In the breeding season it displayed full nuptial colors and spawning behavior. The red pigment, which is present in male R. amarus but lacking in A. tabira , was also absent in the hybrid. Tubercles were abundant on top of the snout, but the quantity of milt produced by means of stripping was extremely small. A taxonomic description of the two hybrid forms recorded here will be published separately. Summary With the aid of artificial insemination, inter- generic hybrids between female Acheilognatus lanceolata of Japanese origin and male Rhodeus amarus from western Europe were obtained. All hybrids were males. They displayed full nuptial colors in the breeding season, but spawning be- havior was poorly developed and milt produc- tion was impaired. The interbreeding of Rhodeus amarus and Acheilognathus tabira from Japan with the aid of mussels was successfully carried out under laboratory conditions. Only one hybrid reached the adult stage, however. It appeared to be a male. In the breeding season it displayed full nuptial colors and spawning behavior, but its production of milt was very limited. Acknowledgements I am greatly indebted to Prof. Tokiharu Abe and Dr. Yoshitsugu Hirosaki for supplying the species of Japanese bitterling, and to Mr. F. G. du Jardin for making photographs of the illus- trated specimens. This investigation was gener- ously supported by the Council for Scientific and Industrial Research of the Union of South Africa. References Berg, L. S. 1948-1949. The fishes of the fresh waters of the USSR and adjacent countries. Keys to the Fauna of the USSR, Akad. Nauk USSR, Vol. 27: 1-466; Vol. 29: 467-925; Vol. 30: 926-1381. [In Russian]. Holcik, J. 1959. The systematic status of the European bitterling, Rhodeus sericeus amarus (Bloch), 1783. Vop. Ikhtiol., Akad. Nauk USSR, No. 13: 39-50. [In Russian]. EXPLANATION OF THE PLATE Plate I Fig. 1. Female Acheilognathus lanceolata. Fig. 2. Male Rhodeus amarus. Fig. 3. Adult intergeneric hybrid between female A. lanceolata and male R. amarus. Stand- ; ard lengths: 74, 56 and 105 mm. Fig. 4. Female Rhodeus amarus. Fig. 5. Male Acheilognathus tabira. Fig. 6. Adult intergeneric hybrid between female R. sericeus amarus and male A. tabira Standard lengths: 55, 62 and 75 mm. 1Zoologica, Vol. 46 (2): 25-26, 1961. DE WIT PLATE I FIG. 6 INTERGENERIC HYBRIDS OBTAINED FROM AC H El LOG NAT HUS LANCEOLATA X RHODEUS AMARUS AND RHODEUS AMARUS X ACH El LOG NATH US TABIRA 10 Some Observations on the Metamorphosis of the Frog Rana curtipes Jerdon Lucy Lobo Biological Laboratory, Fordham University, New York 58, N. Y. (Plate I; Text-figure 1) RANA CURTIPES Jerdon is relatively unknown scientifically, but is found in abundance in the thick forests of Dan- deli, Londa, Castlerock, Supa, Nagargali, Anmode and other neighboring forest areas along the Kali River on the west coast of India. Tadpoles in various stages of development are often seen swarming together in the Kali River (Kali = black), where their jet black color blends with the black waters of the river. The frog is often found in small puddles within the forest and seems to prefer the cool water of shady areas. It apparently estivates during the hot season extending from March to May, when all the pools and puddles dry up. In the rainy season ( i.e from June to September) it is seen in large numbers. Tadpoles collected from various forest areas along the Kali were brought to the Karnatak College laboratory at Dharwar, India, where they were allowed to metamorphose in an aquar- ium well supplied with algae and other food material. It was not possible to undertake a detailed study of the development from fertilization of the egg to the emergence of the tadpoles, since the spawn could not be found; neither could early tadpole stages with external gills be col- lected. Only young tadpoles (Plate I, Fig. 1), jet black in color, are seen swimming in the Kali River in the months of September and October. A young tadpole measures about 2 inches in length. It has a long and coiled gut about 8 inches long. These tadpoles are voracious and are purely vegetarian in diet. Growth takes place by an increase in size of the tadpole and by a relative elongation of the gut. No other morphological changes take place. The large tadpole (Plate I, Fig. 2) varies in length from 3 to 4 inches. It is very stout and its sluggish nature may be attributed to its feeding habits. Large tadpoles are blackish-brown with a large head, and a distinct parotoid-like gland is found behind each eye. Posteriorly the long tail is flattened like a fin. The mouth is wide and bears two horny lips and 4 to 6 rows of labial teeth over the dorsal lip (Text-fig. 1 ) . A row of papillae surrounds the mouth parts. Scattered irregularly along the body and tail are black dots, which are less prominent on the head region. The following table shows the approximate growth of a tadpole from October to January. Month Length of Body Breadth of Body Length of Intestine Remarks October 2" I" 8" Small tadpoles November 3" 1" 14'/2" Large tadpoles December 4" VA" 15" ” ” January A' A" VA" 14" ” ” No structural modifications occur in the tad- pole during these four months, but only an in- crease in size. The tadpole attains its maximum growth about the middle of January, after which it becomes even less active. About the end of January the hind limbs make their appearance. At this stage the tadpole has a very long tail ( 2-3") and the head is about 1 inch long (Plate I, Fig. 3). The gut now shows a reduction in length and measures 14 inches. About the first week of February the fore limbs begin to bud and early in March the hind and fore limbs are fully formed. The tadpole is 103 104 Zoologica: New York Zoological Society [46: 10: 1961] still jet black and has a long tail, as seen in Plate I, Fig. 4. The tail is gradually absorbed and a reduction in size of the gut ensues. By the end of March the last traces of the tail have disap- peared and metamorphosis is complete. The young frog (Plate I, Figs. 5 & 6) has now changed its diet and feeds on insects. The young frog is jet black but it gradually loses the black pigment on the dorsal and ventral surfaces, which take on a brown color as age advances. The prolonged growth and metamorphosis in the forest-dwelling Rana curtipes is peculiar to its kind, and extends over a total period of 9 months, usually from July to March. Acknowledgements I wish to express by appreciation to Dr. J. C. Uttangi, Dept, of Biology, Karnatak Science College, Dharwar, India, for help and facilities rendered to me in making my observations. Thanks are also due to Dr. R. S. Miller, Dept, of Biology, University of Saskatchewan, Saska- toon, Sask., Canada, for the necessary correc- tions made in the manuscript. Bibliography Blair, W. F. & A. P. Blair 1957. Vertebrates of the United States. Mc- Graw-Hill, New York. Boulenger, G. A. 1890. Fauna of British India. (Reptilia and Batrachia). Taylor and Francis, London. 1903. Description of three new frogs from Southern India and Ceylon. Journ. Bom- bay Nat. Hist. Society, XV: 430-431. Butler, A. L. 1903. A list of Batrachians known to inhabit the Malay Peninsula, with some remarks on their habits and distribution. Joum. Bom- bay Nat. Hist. Society, XV: 193-205, 387-402. Ferguson, H. S. 1904. A list of Travancore batrachians. Joum. Bombay Nat. Hist. Society, XV: 499-509. Holmes, S. J. 1926. The Biology of the Frog. The Macmillan Company, New York. Noble, G. K. 1931. The Biology of the Amphibia. McGraw- Hill, New York. Rugh, R. 1951. The Frog. Reproduction and Develop- ment. The Blakiston Company, Philadel- phia. EXPLANATION OF THE PLATE Plate I Fig. 1. Small tadpole of R. curtipes. Fig. 2. Large tadpole of R. curtipes. Fig. 3. Tadpole with hind limbs. Fig. 4. Tadpole with fore and hind limbs. Fig. 5. Young frog with tail. Fig. 6. Young frog with reduced tail. LOBO PLATE I SOME OBSERVATIONS ON THE METAMORPHOSIS OF THE FROG RANA CURTIPES JERDON NEW YORK ZOOLOGICAL SOCIETY GENERAL OFFICE 30 East Fortieth Street, New York 16, N. Y. PUBLICATION OFFICE The Zoological Park, New York 60, N. Y. OFFICERS PRESIDENT VICE-PRESIDENT SECRETARY TREASURER Fairfield Osborn Laurance S. Rockefeller George Wall Merck David H, McAlpin SCIENTIFIC STAFF: John Tee-Van General Director William G. Conway. . Director , Zoological Park Christopher W. Coates. .Director, Aquarium ZOOLOGICAL PARK Joseph A. Davis, Jr.. .Associate Curator, Mammals Grace Davall Assistant Curator, Mammals and Birds William G. Conway . . Curator, Birds Herndon G. Dowling . Curator, Reptiles Charles P. Gandal . . Veterinarian Lee S. Crandall General Curator Emeritus William Beebe Honorary Curator, Birds AQUARIUM James W. Atz Associate Curator Carleton Ray Assistant to the Director Ro*s F. Nigrelli Pathologist & Chair- man of Department of Marine Biochem- istry & Ecology C. M. Breder, Jr Research Associate in Ichthyology Harry A. Charipper. . .Research Associate in Histology Sophie Jakowska Research Associate in Experimental Biology Klaus D. Kallman Research Associate in Genetics Louis Mowbray Research Associate in Field Biology Homer W. Smith Research Associate in Physiology GENERAL William Bridges . . Editor & Curator, Publications Dorothy Reville . . .Editorial Assistant Sam Dunton Photographer Henry M. Lester. .Photographic Consultant DEPARTMENT OF TROPICAL RESEARCH William Beebe Director Emeritus Jocelyn Crane Assistant Director David W. Snow Resident Naturalist John Tee-Van Associate William K. Gregory. . . .Associate AFFILIATE L. Floyd Clarke Director, Jackson Hole Biological Research Station EDITORIAL COMMITTEE Fairfield Osborn, Chairman James W. Atz William G. Conway William Beebe Lee S. Crandall William Bridges Herndon G. Dowling Christopher W. Coates John Tee- Van ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME 46 • PART 3 • NOVEMBER 24, 1961 • NUMBERS 11 TO 15 PUBLISHED BY THE SOCIETY The ZOOLOGICAL PARK, New York Contents PAGE 11. A Study of the Biology and Behavior of the Caterpillars, Pupae and Emerg- ing Butterflies of the Subfamily Heliconiinae in Trinidad, West Indies. Part II. Molting, and the Behavior of Pupae and Emerging Adults. By Anne J. Alexander. Plate I; Text-figures 1-3 105 12. Melanoma, Renal Thyroid Tumor and Reticulo-endothelial Hyperplasia in a Non-hybrid Platyfish. By Pamela A. Mac Intyre & K. France Baker-Cohen. Plates I & II; Text-figure 1 . 125 13. Eastern Pacific Expeditions of the New York Zoological Society. XLV. Non-intertidal Brachygnathous Crabs from the West Coast of Tropical America. Part 2: Brachygnatha Brachyrhyncha. By John S. Garth. Text-figures 1 & 2 } 133 14. Nematodes and Cestodes from the Australian Monitor, Varanus gouldii. By Horace W. Stunkard & Charles P. Gandal. Text-figures 1-6 161 15. Urinary Amino Acids of Non-human Primates. By Jack Fooden. Plates I-III; Text-figures 1-4 167 11 A Study of the Biology and Behavior of the Caterpillars, Pupae and Emerging Butterflies of the Subfamily Heliconiinae in Trinidad, West Indies. Part II. Molting, and the Behavior of Pupae and Emerging Adults1,2 Anne J. Alexander Zoology Department, Rhodes University, Grahamstown, South Africa (Plate I; Text-figures 1-3) [This paper is one of a series emanating from the Tropical Field Station of the New York Zoological Society, at Simla, Arima Valley, Trinidad, West Indies. This station was founded in 1950 by the Zoo- logical Society’s Department of Tropical Research, under the direction of Dr. William Beebe. It com- prises 200 acres in the middle of the Northern Range, which includes large stretches of undisturbed government forest preserves. The laboratory of the Station is intended for research in tropical ecology and in animal behavior. The altitude of the research area is 500 to 1,800 feet, and the annual rainfall is more than 100 inches. [For further ecological details of meteorology and biotic zones, see “Introduction to the Ecology of the Arima Valley, Trinidad, B.W.I.,” William Beebe, Zoologica, 1952, 37 (13): 157-184. [The success of the present study is a large meas- ure of the cooperation of the staff at Simla, especially of Jocelyn Crane and Constance Carter, the former contributing much of her knowledge of the animals, the latter helping with recording of observations]. Contents Page I. Introduction 105 II. Molting Behavior 106 III. Pupation 109 a. Selection of Pupational Site 109 b. Pupational Behavior Ill IV. Pupal Behavior and Adaptations 112 Contribution No. 1010, Department of Tropical Research, New York Zoological Society. 2This study has been aided by a grant from the Na- tional Science Foundation (G6376). Financial support from the Royal Commission for the Exhibition of 1851 and the South African Council for Scientific and Indus- trial Research is also gratefully acknowledged. V. Emergence from the Pupa 115 VI. Phylogenetic Discussion 119 VII. Summary 122 VIII. References 123 I. Introduction A preliminary study has been made of the /% behavior of as many species of larval A ft heliconiines as were available over a period of AVi months spent in Trinidad, W. I. A comparative account of the general activities, such as feeding, resting, locomotion and social and defensive behavior, has already been given (Alexander, 1961). Observations on the late larvae, however, made it clear that some of the significance of their behavior should be sought in the pupa and emergent adult. Hence descrip- tions of some of these aspects of heliconiine be- havior are presented here as Part II of the report, together with observations on molting behavior throughout larval life. The information relates to eleven of the four- teen species of heliconiines found in Trinidad, namely Heliconius doris doris (Linnaeus), Heli- con ins sara thamar Hubner, Heliconius erato hydara Hewitson, Heliconius melpomene eur- yades Riffarth, Heliconius ricini insulana Stichel, Heliconius aliphera aliphera (Godart), Helicon- ius isabella isabella (Cramer), Dryadula phae- tusa phaetusa (Linnaeus) , Dryas iulia iulia (Fab- ricius), Agraulis vanillae vanillae (Linnaeus) and Dione juno juno (Cramer). Descriptions and illustrations of the external characteristics of the larvae referred to here are given by Beebe, Crane & Fleming (1960). 105 106 Zoologica: New York Zoological Society [46: 11 Table I. Molting Behavior in Species of Heliconiine Caterpillars Records of times abstracted from approximately three cases for each species, but confirm general impressions of molts whose details were not recorded. Species ' Position on vine prior to molt Pose prior to shedding skin Time of day when skin normally shed Dione juno Stem Arranged in circle, heads to center, bodies straight Mid-a.m. Agraulis vanillae Midrib/stem — Mid-a.m. to mid-p.m. Dryas iulia Midrib/stem/ tendril “J”/straight Mid-a.m. to early p.m. Dryadula phaetusa Midrib “J”/straight Noon to early p.m. Heliconius Midrib Curved Mid-a.m. aliphera Heliconius Leaf blade Straight p.m. isabella Heliconius melpomene Midrib Straight/head slightly to one side Mid-a.m. to early p.m. Heliconius erato Stem, seldom midrib Straight Mid-a.m. Heliconius Midrib/stem ? Mid-a.m. ricini Heliconius sara Stem, few on leaves Straight Early a.m. to noon II. Molting Behavior As in most butterflies, the Heliconiinae typi- cally have four molts separating five larval stages. Six instars may occur occasionally in some species—//, sara, H. Isabella, H. aliphera and to a lesser extent H. melpomene and D. juno (Beebe, Crane & Fleming, 1960). Normal molting is very similar in all species observed and a generalized description will be given first. During the afternoon before a caterpillar actually sheds its skin, it selects a site and takes up a position ready for molting, both of these actions varying slightly from species to species (Table I). The caterpillar rests with thorax sharply arched into the air and head bowed down. During this period the dorsal skin between the prothorax and head capsule becomes enorm- ously stretched and a pale band appears beneath it— the new head capsule showing through the old skin. It is this which Wachter (1930), in her classical paper on the molting of the silkworm, calls the “triangle stage.” Next day at some time between 7:30 a.m. and 3 p.m. the skin is actually shed, the actual timetendingtodifferfromspeciestospecies (Table I ) . From the initial appearance of the pale triangle until shortly before the shedding of the skin, the animal remains almost motionless. (Text-fig. la) . It is, however, still able to walk and gives a per- fectly recognizable defensive response if dis- turbed. For about half an hour before the skin splits, the larva makes numerous small contortions within the old skin. The skin appears shrivelled and the scoli along thorax and abdomen are moved independantly of each other, due pre- sumably to local contractions inside the larva as the muscle attachments are being freed from the old cuticle. This activity may continue for about ten minutes. The head is then bent grad- ually back so that the old capsule is at right angles to the long axis of the body (Text-fig. lb) . The cuticle between the old head capsule and the ventral part of the first thoracic segment spilts and the legs begin to emerge3. The whole skin is shifted backward as the larva surges for- ward and this puts a strain across the dorsal cuticle behind the old capsule. Thus about forty seconds after the third pair of legs has been freed, this connection between the old head capsule and dorsal prothorax breaks and the head is sud- sWachter (1930) says that in Bombyx mori the bend- ing back of the head occurs because contact between old head capsule and skin is maintained at the mid- dorsal point after being broken ventrally and laterally. This may be true for the Heliconiinae as well. 1961] Alexander: Biology & Behavior of Heliconiinae in Trinidad 107 a bed Text-fig. 1. Diagrammatic representation of four stages during the shedding of the skin of a heliconiid caterpillar, e, The pose adopted during the “triangle stage”; b, head bent sharply back as the skin begins to tear; e, head more extremely bent as the legs emerge; d, head jerked down and the caterpillar emerging from the skin which is gathered into folds as it passes back. Newly exposed cuticle is stippled, old head capsule is black and the “triangle” is marked with an arrow. denly jerked down to its normal position, the old capsule now being stuck loosely onto the mouthparts of the new one. Meanwhile the rest of the old skin is being gradually shifted back by muscular movements which run along the body. Tn some species, H. sara, H. ricini and D. juno, which tend to be gregarious, the newly emerged caterpillar puts its feet onto the surface and begins to crawl forward out of its skin, so that waves of movement are assisted by the walk- ing of the larva. There is no strict distinction be- tween these two methods since the behavior is graded from D. iulia, in which there is almost no sign of walking forward, to D. juno where it is very marked. Partially correlated with the extent of this walking forward is the degree to which the cast skin is gathered together as a heap or left spread. D. iulia and D. phaetusa are alone marked in their habit of producing a closely folded mass of exuvium. The complete shedding of the skin, from the time the head is bent back until the anal prolegs are freed, takes about four minutes. There are, however, slight indications of specific differences in duration (Table IT) although these are not consistent enough to evaluate satisfactorily with- out a larger sample from each species. Once the skin is freed from the last segment, the hind end is lifted into the air. The anal pro- legs contract and expand several times and then the caterpillar begins to rid itself of the old head capsule. This it does by wiping off the capsule against a leaf or stem. As soon as its new capsule is bared, the caterpillar puts its spinneret to the substratum and weaves a small patch of silk just ahead of it. It walks forward a few milli- metres and weaves again. There is a fairly wide variation in the pose adopted during the process of expanding the scoli and spinules of the head and body, but information is too limited to say whether any specific constancy exists among these variations. All species are alike, however, in having head and thorax held away from the substratum. Ex- pansion seems to be effected by periodically rais- ing the internal pressure of the body and blowing out the scoli and spinules. The contraction which produces this pressure increase is a shortening of the body, the legs being kept free of the ground and the head and thorax pressed back into the abdominal region. During this contrac- tion the arch already present in the thorax may be flattened somewhat towards the substratum, as often happens in A. vanillae. The thorax and head may be lifted further and the extent of arching actually increased, an event which is more noticeable in H. melpomene and D. iulia. The principle effect, a longitudinal shortening of the body, is the same in all species and is often the only one visible. During any one contraction a scolus will expand from the base up and the spinules arising from this expanded portion will “click” into position. Normally it takes about 6-11 contractions to expand a scolus and its spinules fully. There is a very slight tendency for scoli and spinules to collapse again once the pres- sure which erected them has subsided. A series of 7 contractions lasts approximately 5 minutes, by the end of which time all scoli and spinules are usually fully expanded. It may be noted that although spine expansion normally begins within a few minutes of the animal’s being freed of its skin, disturbance by another caterpillar or other interference may delay its onset for at least 20 minutes without obvious effect upon the form of the spines. Once the expansion is complete, darkening and hard- ening begin. The hardening and darkening process occurs gradually and there is no marked activity for 20 to 40 minutes. During this time all species rest with thoracic segments arched, legs some- what spread and head slightly bowed. D. iulia and H. aliphera frequently adopt a “J” position as well, the head and thorax between 0° and 90° to the rest of the body in both dorso-ventral and lateral planes. The caterpillar then relaxes so that its legs touch the leaf or stem again, but it may remain motionless for another one to four hours. It then turns slowly round, may rest again but finally walks back to its empty skin and begins to eat it. 108 Zoologica: New York Zoological Society [46: 11 Table II. Examples of Records of Behavior during Shedding of Larval Skin by Species of Heliconiine Caterpillars Species Duration of the shedding of the skin (mins., secs.) Duration of spine expansion (mins., secs.) Number of contractions in spine expansion Dione juno 5,15 ? ? 5,00 ? ? 4,00 5,20 7 23,40 5,00 5 5,00 4, 10 5 6,00 ? ? Agraulis vanillae 4,40 ? ? 4,45 6, 10 6 3,10 5,40 6 5,35 6,20 10 Dry as iulia 3,20 ? ? 2,20 6,20 9 2,30 4,35 8 ? 5,20 9 2,45 8,20 11 Dryadula 3,20 5, 30 5 phaetusa ? 4, 00 + 3 Heliconius ? 10, 15 11 aliphera 2,30 ? ? 2,30 7,45 7 Heliconius 3,40 10, 55 8 melpomene 3,30 5, 10-6, 50 11 3,05 6, 50 5 2,50 4,55 7 Heliconius 5,00 3,25-7,25 ? erato Heliconius 3,05 6, 15 2 ricini Heliconius 3,20 6,00 ? sara 3,30 7,00 2, indistinct 2.10+ 3, 00+ ? In D. iulia and D. phaetusa, but not in other seen to occur naturally among gregarious species species, there is a great deal of arching of both head and hind end during the period between the end of hardening and eating the skin. The legs and anal and last abdominal prolegs are freed and arched up so that the hind end and head almost meet above the caterpillar’s body. If there happens to be the cast skin of another caterpillar immediately in front of a larva which has just finished its hardening and darkening, it will not turn round and go back to eat its own skin but will walk forward and accept the for- eign one. This indicates that the behavioral sequence leading to the eating of exuvia is not strictly stereotyped. After the skin has been eaten there is a further period of rest before the cater- pillar begins to feed on a leaf. If, however, a second skin is offered to a newly molted cater- pillar it will eat this as well, a phenomenon often such as H. sara. Thus eating a single empty skin does not, in itself, provide the whole consum- matory act. D. juno and H. sara, two of the gregarious species, differ from A. vanillae, D. iulia, D. phaetusa, H. aliphera, H. isabella, H. melpom- ene, H. erato and to a lesser extent H. ricini in the appearance of the spine expansion process. Instead, of being a definite series of clear-cut contractions, there is an indistinct period during which the scoli gradually expand. While this could be a genuine effect, it seems more likely that it is due to the observational set-up. The gre- garious species have shorter scoli and less de- veloped spination (Beebe, Crane & Fleming, 1960). Thus they provide less distinct indicators of the process which is being observed than would caterpillars of solitary species. In addition, 1961] Alexander: Biology & Behavior of Heliconiinae in Trinidad 109 Table in. Prepupational Behavior in Species of Heliconiine Caterpillars Data on Dione juno relate to a single group of 13 larvae observed in this study but are confirmed by the record of another group kept by Constance Carter. Dryadula phaetusa record refers to a single animal. Information relating to Heliconius sara supplied by Constance Carter. Approx, duration Locomotion during Normal time of day for: Species of fast before site select, (hours) site select., mins./meter Site select. Hanging up Shedding last larval skin Dione 8 1,44-3,04 8:30 p.m. 11:00 p.m. 1:00-4:00 p.m. juno Agraulis 6-14 2,22-7,59 Indistinct 5:00 p.m. 7:00-11:30 a.m. vanillae Dryadula 4 10,12 3:00 p.m. 5:00 p.m. Before noon phaetusa Dryas iulia 5 1,24-1,58 2:30 p.m. 5:00-8:00 p.m. 8:00 a.m. -noon Heliconius 0 2,05-3,43 2:30 p.m. 3:00-7:00 p.m. ? isabella Heliconius 0 -5,30 2:00 p.m. 3:00-7:00 p.m. 9: 30-noon aliphera Heliconius melpomene 1-8 2,55-3,16 11:00 1:30 a.m.- p.m. 1:30-7:00 p.m. 7:00-11:30 a.m. Heliconius erato 1 3,18-4,25 9:30 12:30 a.m.- p.m. 12:30-3:00 p.m. 8:00-10:30 a.m. Heliconius ricini 14-1 3,20-4,50 10:30 1:30 a.m.- p.m. 1:00-5:00 p.m. 7:30-9:30 a.m. Heliconius ? 3,14- 10:30 a.m. 5:00 p.m. ? sara the caterpillars are together in a group and more difficult to distinguish individually. Moreover they are frequently moving and disturbing each other. III. Pupation a. Selection of Pupational Site As can be seen from Table III, most species stop feeding some hours before they start to find a site suitable for pupation. There are no noticeable behavioral changes during this pe- riod but changes take place in the coloration (Beebe, Crane & Fleming, 1960). The cater- pillar rests either on the leaf it had been eating or on the stem. Once they have begun to search for a pupa- tion site, the caterpillars no longer feed. H. isabella is exceptional for it may sometimes stop and eat again even after it has abandoned a feed- ing position and begun the ambulatory phase. None of the species which pupate on leaves has ever been seen to do so on a leaf which has already been chewed by a caterpillar or one on which another larva is resting. Some indication of the speed with which caterpillars of different species walk is given in Table III. (This has not been corrected for the slightly different temperatures at which the ob- servations were made.) Only D. iulia is excep- tional in walking invariably fast. D. phaetusa is odd in being very slow but as only one individual was tested, no valid conclusions can be drawn from this. The rate of walking varies during the ambulatory phase, starting slowly, increasing, then apparently falling again— an effect which partially accounts for the wide variation in re- sults presented here. Except in the cases of A. vanillae, D. phaetusa and H. erato, the walk is interrupted by few rests. A caterpillar may drop from one part of the vine to another or to the ground, covering a small area of its leaf or stem with silk and then swinging down from this on a safety thread. Sometimes the drop may be as much as a meter. Once a caterpillar has started walking, it may pass numbers of what appear to be suitable pupation sites, a behavioral feature noted in the cecropia silkworm (v.d. Kloot & Williams, 1953). It should be noted, however, that at least some D. iulia and H. melpomene are able to settle at a pupation site after walking only a few inches. The location of such a site differs from species to species, as can be seen in Table IV, yet it is always such that the pupa can hang 110 Zoologica: New York Zoological Society [46: 11 Table IV. Sites Selected for Pupation by Prepupational Heliconiine Caterpillars Having Fairly Unrestricted Choice L=Living, D=Dead in reference to stem and tendril. Species Leaf tip Leaf blade Pet- iole Midrib Leaf margin Stem, L., D. Tendril Dione juno 3L 10L Agraulis vanillae _ 2 6L 1 Dryadula phaetusa _ 1L _ Dryas iulia — 1 3 2 - 2L 5 Heliconius isabella 12 & 12D Heliconius aliphera 4 12 _ Heliconius melpomene _ 10 3 2 Heliconius erato 2 1 1 2L 2 Heliconius ricini 10 3 1L 5 Heliconius sara 2 1 down below it, whether it be beneath a leaf, stem, tendril, flower stalk or some foreign hori- zontal surface. Occasionally a caterpillar may attach against a vertical surface. One feature which normally determines if a site is suitable, is whether it provides enough free space beneath for the pupa. This is estimated by means of a behavior pattern which will be termed “trying for height.” Taking H. mel- pomene as an example, the caterpillar walks along the ventral midrib of a fresh, unchewed leaf as it goes weaving on either side of the mid- rib. Towards the center of the leaf it stops and weaves more actively. Then it arches its head back and releases its foothold so that its head and thorax hang loose beneath the leaf. If it touches a second leaf or some other object be- low, it will usually abandon the position as a possible pupation site. If it does not touch any- thing, it releases the hold of its first and. then second prolegs so that its body hangs further down beneath the leaf. It may swing gently, reaching out around it. If it now touches some object, the caterpillar does not grasp it and crawl onto it as it would if it were trying to move away, but will usually contract up onto the mid- rib again and walk off to try another leaf. A caterpillar usually tries for height three or four times, facing first one way along the midrib and then another. Finally, if the leaf is acceptable, it will rest before beginning to spin its pupal silk pad. H. aliphera and H. isabella differ from the other species in that, when trying for height, the prolegs are never freed from the leaf above. In these two species the pupae do not hang directly but are bent almost at right angles to their sup- port (Plate I). They therefore need less space beneath their pupation site. Other immediate features which render a site desirable for pupation are less clear. Field obser- vations show that all species except H. isabella, and to a more limited extent H. aliphera, choose to hang up on a ridge, a knob or other protuber- ance or irregularity, never on a smooth surface. In the laboratory H. isabella, given the choice of an irregular surface, such as aluminum gauze, or a smooth one such as glass, will hang up on the latter. The same preference is also shown by H. aliphera (Plate I) , which in the field pupated on the ventral midribs of leaves in 12 out of 16 cases. Those four which were not on midribs were, nevertheless, in positions which would have been acceptable to H. isabella. In broader terms a pupation site must, so far as possible, protect the pupa and permit a healthy butterfly to emerge. Protection against climatic factors is unlikely to be important, in that the pupae are in an environment which already sup- ports the caterpillars. Nothing at all is known 1961] Alexander: Biology & Behavior of Heliconiinae in Trinidad 111 of pupal predators, although ants will attack pupae which are injured or unhealthy; they have not been seen to molest them otherwise. Various insect parasites, both dipteran and hymenop- teran, can be bred out of pupae and it has been established that the eggs of one of these are laid by the female in the pupa itself. What does seem to be a significant hazard, however, is that an- other caterpillar, of the same or a different species, will begin to eat away the leaf from around a pupa so that the pupa falls to the ground. This has happened at least four times on the observational vines when there have been plenty of other leaves available for the larvae. Some of the consequences of a pupa falling to the ground will be considered later. Meaningful discussion of the significance of a preferred pupation site, and consequently of the prepupational behavior of the caterpillar, must necessarily be unsatisfactory until more exact information is available about potential pupal predators and parasites. Those species which hatitually move off their leaves to pupate (e. g., D. iulia). risk attack by predators such as ants or mantids (and both have been observed to kill caterpillars searching for pupational sites) during this extended locomotory stage. There may be compensation for such dangers in their subsequent security from being chewed off a leaf. b. Pupational Behavior Once the site has been chosen, the caterpillar weaves generally around it and may tie the selected structure firmly onto the main vine. Then it spins silk more particularly in the region where the pupal pad is to be placed. There are several rest periods during which, if oriented under a leaf or along a stem, the caterpillar faces first one way, then the other. Eventually, after 30 minutes to two hours, spinning of the pad itself begins. The caterpillar first spins on one side of the midrib, stem or tendril, then, bringing its spinneret to the place where the pad is to be, touches the spot, lifts its head and pulls it backwards, touches the spinneret again to the position of the pad and then repeats the pro- cedure, but turning this time to the other side. The process is repeated many times before the spinning becomes concentrated on the pad itself. The drawing back of the head over the area where the pad is being spun is concerned with forming loops of silk which make up much of the pad and has been discussed in Part I of this paper as “yawning.” The pad is cone-shaped with an elliptical base whose long axis is about 1 mm. long. It is about 0.5 mm. tall and its apex is asymmetrically placed, being nearer to the caterpillar as it spins than the mid-point of the base. The color of the silk varies from species to species (Table V). Once the pad is completed, the caterpillar moves forward until the pad is either just before or behind its fourth pair of prolegs. In this posi- tion the animal waits until it has emptied its gut of the last fecal pellet. Then it walks further forward so that its anal prolegs touch the pad. These structures work themselves into the silk, Species Dione juno Agraulis vanillae Dryadula phaetusa Dryas iulia Heliconius isabella Heliconius aliphera Heliconius melpomene Heliconius erato Heliconius ricini Heliconius sara Table V. Colors of Silk which Heliconiine Prepupational Larvae Produced in Spinning Their Pupal Pads White Pale pink Pink Red Indeterminate 13 9 2 21 9 1 1 1 5 1 2 3 2 8 12 2 2 3 112 Zoologica: New York Zoological Society [46: 11 presumably much of the attachment being by way of the crochets. The caterpillar jerks its hind end forward and back, either to make the attachment more secure or to test it. Some species immediately loose their hold on the leaf or stem and hang freely from the pad, e. g. D. iulia and H. melpomene. Others such as A. van- illae may wait for as long as half an hour. When the prolegs release their hold on the substratum, it is the hindermost abdominal pair which let go first, then the third prolegs and finally the second and first. The caterpillar rests in this position until the following morning, when it sheds its skin. If the observer pokes a larva which has just released its foothold and is hanging by its anal prolegs, it can be induced to climb back onto the leaf or stem. If the disturbance is rough enough, the caterpillar will break its attachment to the silk pad. It may then rest a while before reattaching itself once more to the original pad or may spin a second one either beside the first or a short distance off. There is no need for the caterpillar to revert to the previous locomotory stage before it can spin a second pad, in the way that cecropia larvae seem constrained to do if removed from a partly constructed cocoon (v. d. Kloot & Williams, 1953). D. iulia and H. melpomene can be induced by continued teasing to spin a third silk pad, but this is invariably very thin and small. There appear to be differences in the ease with which different species can be teased into abandoning their pads or spinning a second. Thus A. vanillae is markedly reluctant to climb back onto the substratum and will twist and try to bite at the teasing forceps for much longer than would H. melpomene. No standards were established for the teasing, however, and a great deal more experimentation is necessary before evaluating any such specific differences. A peculiar point emerging from these latter observations is that the pattern of hanging free after attachment to the pad changes with repe- tition. On the first occasion it is invariably the hinder of the prolegs which loose their hold first. On the second attempt the prolegs hang free in order from the front to the back. The possible interest attaching to the point is that it suggests that the second hanging-up is not in all ways a simple repetition of the first. IV. Pupal Behavior and Adaptations Of the ten species of heliconiine pupae which have been observed, all show a certain amount of movement. This is distinct from any flexibility which they may have at the point of attachment. This latter characteristic differs from species to Table VI. Behavioral and Pigmentary Characteristics of Heliconiine Pupae Information relating to Heliconius sara supplied by Constance Carter; details of methods of measuring flexibility not strictly comparable with those of other species. Position of H. sara in the series is therefore open to suspicion. Species Degree of flexibility Bend in pupa Odor production Stridulation Gold spots Heliconius isabella 160 Permanent + Heliconius aliphera 160 Permanent + Agraulis vanillae 147 Changing + ( Heliconius sara) (144) ? ? + Dione juno 136 Changing _ + Dryas iulia 126 + + + + + Dryadula phaetusa 125 _ _ _ + + Heliconius ricini 116 _ + + + + Heliconius melpomene 79 + + + + + + Heliconius erato 65 - + + + + 1961] Alexander: Biology & Behavior of Heliconiinae in Trinidad 113 species, presumably depending on how tightly the pupa has attached its cremaster to the silk pad. Flexibility may be expressed by how far the pupa rotates while its support is rotated through 180° in the vertical plane. An absolutely rigid attachment gives the full 180° turn while a fully flexible pupa will not rotate at all. In this way, the ten species available for testing can be ar- ranged in a series (Table VI). Hinton (1955) discusses the selective advan- tage derived by such pensile pupae from their flexibility of attachment, contending that swing- ing hinders predators such as birds from getting an effective peck or that is it useful in shaking off ants. Whether such factors are involved among the heliconiine pupae or not, flexibility certainly adds to the effect of the movement of the pupa itself. When the species of Heliconiinae studied are arranged according to other aspects of pupal be- havior, the series established for flexibility of attachment tends to be repeated. Thus in the two species which are rigidly attached, H. isa- bella and H. aliphera, the posterior segments of the abdomen are bent so that the ventral sur- face lies near or touches the surface of the leaf and the angle between body and leaf is invar- iably less then 35°. The pupae of these species are the least active of all those studied. When touched by caterpillars or the experimenter, they swing from side to side but produce only faintly audible stridulation. Both are light colored, creamy white or translucent (with a suggestion of green in H. aliphera ) . Their mark- ings are faint brownish or black without gold spots or other “attention calling” devices. Neither produces any perceptible odor. Consid- ered as a whole, the behavior and appearance of the pupae of these two species suggest that the evolutionary trend in the pupal stages of H. aliphera and H. Isabella has been towards camouflage and concealment rather than adver- tisement. The next two pupae on the flexibility scale, A. vanillae and D. juno, share with H. aliphera and H. Isabella the characteristic of bending their bodies so that they do not hang vertically down as do the other five species. They do not, however, bend up close to the support nor do they remain in this position until emergence. The major bend in the body of the pupa is more character- istically lateral as opposed to the dorso-ventral one in H. aliphera and H. isabella; further, A. vanillae and D. juno frequently have two distinct bends, one almost at right angles to the other (Text-fig. 2). During the nine or ten days of pupal life the orientation of these bends con- stantly changes; without any apparent environ- mental stimulus a pupa will twist its body into a new position. A few hours before emergence of the adult, they straighten out so that the pupa hangs vertically down. This again is dis- tinct from the behavior of H. aliphera and H. isabella, whose adults emerge from a pupa which is still curved upwards towards the leaf. It may be noticed that, for the butterfly to expand its Text-fig 2 . Diagrammatic representation of the main morphological and some behavioral characteristics of the pupae discussed in the text. Dorsal view of: a, H. erato; b, D. iulia; c, H. aliphera; d, A. vanillae. Gold markings are shown in solid black and the ventral bend in the posterior end of H. aliphera is indicated by stippling. 114 Zoologica: New York Zoological Society [46: 11 wings, a free space is needed beneath the empty pupal case to which it clings. Since the pupae of H. aliphera and H. Isabella remain bent during emergence, less space is required than when the pupae straighten out as do those of D. juno and A. vanillae. It will be remembered in this con- nection that H. aliphera and H. Isabella do not release their prolegs when “trying for height” (see page 110) while D. juno and A. vanillae us- ually do. The pupae of D. juno and A. vanillae make no audible sound when stimulated, nor do they pro- duce any noticeable odor. Their movement is slight, rather slow and limited. They have no bright spots or grotesque frills or spines and their usual colors are neutral straw to brown- gray (Beebe, Crane & Fleming, 1960). ,4. vanillae pupates on fresh vine, either on the stem, ten- drils, flowers or leaves, and almost invariably on its own food-plant. The brownish color combines with the bendings of the body to give the pupa a resemblance to a shrivelled leaf or withered flower of P. foetida. Like A . aliphera and H. isa- bella, the pupae of this species seem to be adapted rather toward concealment than warning or frightening away possible predators. Of a group of 13 healthy D. juno caterpillars given a choice of pupational sites, all hung up on the stem, tendrils or leaves of a fresh vine. Although they resembled A. vanillae in their choice of a pupation site, they remained together as a group and it is hard to see any attempt at concealment in their behavior. The fact that both D. juno and A. vanillae continually change from one twisted position to another during the pupal stage is interesting when compared with the constant bend of the abdomen of H. aliphera and H. isabella. The latter two are flexed directly back, i. e., their bend is a bilaterally symmetrical one. It seems likely that if a pupa is asymmetrically twisted throughout its entire pupal life, the development of a bilaterally symmetrical butterfly would be more difficult. Thus a pupa of D. iulia which had been put onto a flat surface immediately after it had molted, and was still soft, hardened with a permanent twist to one side. When the butterfly emerged, the wing on the concave side was shorter than the other and buckled; the animal was barely able to flutter. Thus it may be advantageous for a pupa to remain bilaterally symmetrical. Since shrivelled leaves are seldom so, A. vanillae (perhaps D. juno as well) has to lose its symmetry to mimic such objects. Possibly the continual changing of the asymmetrical pose is a compensatory measure which ensures that the developing adult is not continuously exposed to a single pattern of asymmetric forces. The pupae of the five remaining species hang vertically and are fairly flexible. With the ex- ception of D. phaetusa (of which only one specimen was tested), they all wiggle suddenly and furiously when stimulated. Again with the exception of D. phaetusa, they produce a clearly audible squeaking. All except D. phaetusa pro- duce a sudden and noticeable smell when dis- turbed. H. melpomene smells most strongly, an odor very like that produced by the adult butter- fly when disturbed. This is unpleasant to some people, not to others — it passes very quickly. The pupae of D. iulia smell of musty honey, those of H. ricini and H. erato have a weaker odor somewhat like a mixture of D. iulia and H. melpomene. There are in addition bizarre anatomical effects (Beebe, Crane & Fleming, 1960). All five species, D. phaetusa included, have showy golden or silver spots on their backs (Text-fig. 2). H. melpomene is chestnut colored with black spines and is fairly obvious against the green of the leaves under which it normally pu- pates. H. ricini and H. erato have elaborate antler-like frontal horns, protuberances and spines on their bodies. All these features, behavioral and structural, seem to advertise the pupae, and are thus the converse of the tendency shown in H. aliphera, H. isabella, D. juno and A. vanillae. However the site in which D. iulia chooses to pupate, up against the bark of a tree or on a shrivelled ten- dril or flower, combines with the mottled appear- ance of the animal to make it almost invisible in its natural position. The gold spots, so obvious in H. melopene and especially D. phaetusa, are in D. iulia very much smaller and the most anterior pair are almost invisible. In D. iulia and probably also A. vanillae, there is a tendency for the shade of the pupa to be influenced by the lighting conditions under which the larva hung up, a tendency well known among lepidopterous larvae and first investigated by Boulton as early as 1887. If D. iulia is kept in the dark during the day on which is it due to hang up, the resultant pupa is much darker than one kept in the light4. In these features therefore the pupa of D. iulia appears to be cryptic. In this case selection may be working in these two directions at once, and they do not necessarily cancel each other out nor are they mutually exclusive. A pupa which is concealed may be discovered acciden- tally by a predator and it would then be advan- tageous for it to have an intimidation display. If an “advertisement policy” like that of H. mel- 4 Although this experiment was carried out in a single room with only a partition separating “light” from “dark” larvae, there is no guarantee that some tempera- ture difference did not exist. 1961] Alexander: Biology & Behavior of Heliconiinae in Trinidad 115 pomene were modified so as to retain only those elements such as noise, movement or odor, which the pupa can control, then a condition like that in D. iulia would result. Perhaps the turning point in such an evolutionary pathway would be that of changing the pupational site from beneath a leaf to a stem, tendril or foreign surface. This change in its turn could have re- sulted from selection pressure of caterpillars eating the sheltering leaves or from the fact that leaves on the particular species of vine are scarce or very flexible— both of the latter tend to be true in the case of P. tuberosa, the vine on which £>. iulia feeds. H. ricini and H. erato are usually cryptically placed, but less so than D. iulia (Table IV). About half of the cases observed were not on living leaves. The explanation suggested, for D. iulia can be put forward for these two species as well, although with perhaps less justification. The case of D. phaetusa is not explicable in such terms. The only feature which could be regarded as intimidating is the spotting. It has no marked movement, sound or smell. On the other hand, the pupa does not appear to be well concealed, being light beige without striking pe- culiarities of any sort. Perhaps some simple adaptive explanation will be revealed by obser- vations in the field. V. Emergence from the Pupa The duration of the pupal stage in the Heli- coniinae is generally 9 or 10 days. There is a suggestion of some specific differences in what time of day emergence takes place; thus H. erato is usually out of its pupal case before 9 a.m., while H. melpomene tends to come out through- out the morning, with a concentration around 10:30 a.m. The pupal case cracks across behind the head, down the center of the thoracic tergites, then across and down either side of the wing covers. The lateral edges of the split behind the head run down onto the ventral surface of the pupal case and along the ventral edges of the wing covers. Thus the thoracic part of the pupal case is divided longitudinally into three sections which can be easily separated from each other and from which the wings and legs of the butterfly can emerge without excess straining. The sep- aration of the three sectors seems to be effected mainly by movements of the bases of the wings, which, expanding and contracting against the loose sections, gradually work them apart. Pre- sumably in its normal hanging position this would be helped by gravity as the body tends to fall out of the pupal case. The legs do, how- ever, grasp feebly at the edges of the case and within a few seconds either by their effort, by waves of movement passing back or by gravity, the swollen abdomen is freed of the case. The butterfly immediately turns around so that the small wing buds, antennae and abdomen hang down behind it. The wings at this time are about one-third of their final length and during the next two to five minutes they expand until they reach full length. When fully expanded, the wings are almost completely flattened. Although they are soft and flaccid to the touch, they do not fall into folds as an equally soft cloth would do. The antennae hang straight back, lying between the wings. The labial palps have moved from their original position, folded down into the thorax, to their final one, bent up beside the compound eyes. At intervals for a period of 15 to 30 minutes the wings flap slightly open and then close again, the intervals between the opening and closing depending to some extent on the species, ex- amples of which are shown in Text-fig. 3. Dur- ing these movements the two wings of each side are held together with the hind wing a little fur- ther forward relative to the front than is nor- mally the case. As the wings open, the head moves slightly forward and the abdomen swings between them. When the wings close the head and abdominal movements are reversed. Exten- sion of the proboscis occurs fairly frequently (Text-fig. 3), but without definite relation to the wing movements. The palps begin to tremble a short time after the wing fanning starts and there is a marked tendency for this to be asso- ciated with wing closing rather than opening. At the end of wing fanning, the wings are closed and then the hind ones open again so that they are well separated from the forewings. The latter open only very slightly. The wings are then held motionless in this position for an interval which varies from less than four to more than 15 minutes; the interval will be referred to as the “relaxation period.” Depend- ing on the species, the palp trembling will or will not continue throughout the relaxation period (Table VII). The end of relaxation is marked by the hind wings closing up against the fore- wings. These are, however, still held slightly apart from each other for a further two to four minutes so that the ending is gradual. A variable time after the start of wing flap- ping, the antennae move forward to the lead- ing edges of the wings and, when the wings next close, fold back so as to lie lateral to the wings instead of between them. Subsequently 116 Zoologica: New York Zoological Society [46: 11 the antennae, together with the head, can be seen to move forward and back as the wings flap. By the beginning of the relaxation period they come to lie together at the leading edge of the wings. During or after relaxation they move anterior to the wings and are held together for a variable time, sometimes opening and closing slightly. In some species, such as A. vanillae, this forward movement of the antennae is very characteristic and coincides with the end of palp trembling, though not usually at the end of relaxation. The first drops of meconium are voided at a iiliiiilllii iii DIONE JUNO • M I* I • M ~l I* I • I 8 A k AGR.AU LIS VANILLAE ^ ^ ^ M • I • I “I • I • 1 • I - 1 • I • * A k DRYADULA PHAETUSA . | ^*V*»'^W^A^/WVV^^*ArAVWVWAAWVVV\^V^WW*ArVvVVV •I *1 ■H,N •1*1’ HELICONIUS MELPOMENE !• M HELICONIUS RICINI h I* 1 • I s I a I • I * 1*1 HELICONIUS SARA i • |'MS!+ I* f’hl'IW N° I' A i 8 a a & a a i n a & A I 1 S S 1 I I I I I t Text-fig. 3. Examples of records of emergence behavior of D. juno, A. vanillae, D. phaetusa, D. iulia, H. isabella, H. melpomene, H. ricini and H. sara. The time scale at the top and bottom of the figure is in minutes and zero time for each record is on the left hand side and is the moment when the pupal case split open. The upper saw-toothed line of each record represents activity of the palps, while a continuous line marks periods when inactivity was observed. A vertical stroke indicates wing opening during the period of wing fanning, wing closing being shown by a solid black circle, while a dotted line shows the relaxation period. The open circle with a cross in it marks the liberation of the first drop of meconium. Below this in each record is a representation of the state of the wings, starting with a pair of wavy lines at the left of each record and representing the crumpled wings at the last point of time when they were mentioned in the laboratory notes as being crumpled prior to the straightening, which is represented by straight lines. The position of the antennal tips is shown relative to the wings as pairs of black dots. 1961] Alexander: Biology & Behavior of Heliconiinae in Trinidad 117 Table VII. Behavior of Emergent Heliconiine Butterflies Number in [ ] after each species denotes number of cases watched in detail; number fol- lowed by an “e” indicates additional cases of experimental animals. (See Table IX). Species Number of flaps and number of minutes of phase Number of flaps per minute Palp movement obvious during flapping Palp movement obvious during relaxation period Minutes before 1st meconium shed. Co-efficient of variation (bracketed) Final palp move- ment and head twisting Dione juno [8] 14/16-18/17 1.0 + 4- 20.7 (.07) + 4 — r Agraulis vanillae [5] [2e] 10/9.0-17/10 1.3 Absent or late 4- 29.6 (.06) 4-4 — b Dryadula phaetusa [1] 32/23 1.4 ++ 32.0 (?) H — 1“ Dryas iulia [6] [9e] 18/13-24/13 1.6 ++ 17.8 (.21) H — 1“+ Heliconius Isabella [3] [le] 26/17-28/16 1.6 4-+ late 4- 32.0 (.03) Heliconius aliphera [2] 21/16-29/19 1.55 21.5 Heliconius melpomene [8] [2e] 19/15-27/17 1.4 +4 — p (on shut) 10.8 (.45) _ Heliconius erato [1] [3e] 22/15 1.4 4- (on shut) 23.1 (.58) _ Heliconius ricini [4] [2e] 21/14-25/15 1.5 4- (on shut) 21.3 (.16) Heliconius sara [2] 22/10 2.1 4-4-4- (on shut) - 13.7 (.10) — intervals of a few minutes until the animal is ready to fly. The first two or three sets of drops contain a fine precipitate but later drops become more transparent and paler in color until after a period varying from 6 to 40 minutes the fluid is altogether clear. Once the relaxation period is over and the wings are once more apposed, only slight move- ments of the antennae take place before the animal flies off. They do, however, gradually move further and further apart until their nor- mal position is achieved. Just before flight a peculiar pattern has been noticed in D. juno, A. vanillae, D. phaetusa and D. iulia. The butterfly twists its head from one side to the other, at the same time extending and waggling its palps, one after the other. Flight takes place after the antennae are well apart, although a disturbance before this time may cause the butterfly to flap its wings. The stimulus for flight often comes from the move- ment of a nearby butterfly or a puff of wind but frequently there is no apparent external stimulus at all. The insect walks up its empty pupal case, flapping its wings, then suddenly takes off, sometimes still dropping meconium as it flies. The specific differences among the 10 species observed are listed in Table VII and some can be seen in Text-fig. 3. During a natural emergence there is very little rearrangement of the legs and body necessary before the butterfly is in the position appropri- ate for wing fanning. The preferred position is one in which the longitudinal axis of the thorax is not quite horizontal, sloping slightly upward at the anterior end. In the majority of species examined, the pupae have projections on the dorsal part of abdominal segments and these are placed so that the newly emerged butterfly can obtain its preferred pose by clinging onto them with the first pair of functional legs and onto the margin of the head case with the others. How- ever, the stance does not invariably occur and it may not be correct to regard such projections as having been evolved primarily for this func- tion. Emergence from the pupal case is almost in- variably prolonged when the animal is not hang- ing in the normal way. If, however, an object is placed near the butterfly’s legs as soon as these are free, it is grasped and the body and wings pulled out quite quickly. If a recently emerged butterfly is not provided with an appropriate resting place, it walks until it finds one and, in H. melpomene at least, is 118 Zoologica: New York Zoological Society [46: 11 capable of searching for at least 27 minutes. Its walking is initially directed toward the lightest part of the horizon but this orientation gradually becomes confused and the butterfly tends to begin circling movements, possibly due to mech- anical difficulties with its wings. During this walking stage the inflation of the wings is not delayed and within five minutes of its exit from the pupal case the animal’s wings are fully expanded and flattened. If the butterfly is allowed to hang up at this stage, it shows no apparent deviation from the normal pattern of wing flapping, palp trembling, relaxation and finally flight. Yet even when it is forced to walk on a flat surface there may be some damage to the wings, either from scraping on the pupal case or injury during walking. There may also be a fair amount of bleeding from both wings, although the effects are not as serious as Brocher (1919) reported for punctures made at an earlier stage in the wings of Agrionidae where full expansion was prevented. In natural condi- tions, after a pupa has fallen to the ground as a result of the chewing activity of some other caterpillar, its newly emerged butterfly may have to walk over twigs, leaves or other debris and will probably damage its wings more severely. If the butterfly is forced to continue walking after the wing expansion is completed, the wings trail limply behind it and during this time there are no signs of any flapping behavior, nor are the palps trembled at all. Meconium will never- theless be dropped at the appropriate time. When eventually the butterfly is allowed to hang up, the wing flapping may be prolonged, or much curtailed; it may be much faster, or just irregular (Table IX). If the butterfly is forced to walk for as long as 10 minutes, its wings, when even- tually it does hang up, fall into folds instead of remaining flattened. Although they appear still to be soft, the animal seems unable to hold them in such a way that they harden as flat sur- faces. Thus, although they are fully expanded, the ability to flatten the wings seems to dis- appear about eight minutes after emergence from the pupal case. Normal wing expansion may sometimes take as long as five minutes, so the flattening process cannot be postponed for more than about three minutes after its nor- mal time. Hardening of the wings must occur very shortly after the loss of flattening ability, for when hanging up is delayed for longer than eight minutes, the wings begin to take on the shape they had when trailing along behind the walking butterfly. If they were twisted over the right side, the hardened wings later show the same twist. Indeed, it is likely that straightening and hardening normally take place concurrently. Even if its wings have hardened in a distorted and useless condition, the butterfly, after a per- iod of some wing flapping and violent palp trembling, relaxes for a period, rests and then finally tries to fly. Although the wings may be so twisted that their pattern is unrecognizable, the butterfly still rests for the normal relaxation period and then attempts to flap the wings prior to take off. Like the wings, the antennae may remain un- straightened unless they are allowed to hang straight back between the wings. The critical period for this event appears to be the same as that for wing flattening and hardening. Wing flapping and palp trembling are defin- itely not concerned with wing expansion. They normally occur during the process of flattening and hardening the wings and are curtailed when enforced walking is continued so long that the wings are never correctly flattened and hardened. It thus seems likely that these behavioral activi- Table VIII. Colors of First Meconium Produced by Emergent Heliconune Butterflies Species Brown Chestnut Dark Light Yellow Dione juno 4 4 Agraulis vanillae — 5 Dryadula phaetusa — — Dryas iulia 2 5 Heliconius isabella — — Heliconius aliphera — — Heliconius melpomene 2 (+ red) 1 (-f red) Heliconius erato — — Heliconius ricini — 4 (+ red) Heliconius sara — — 1 1 6 3 4 3 Gray 15 5 3 1961] Alexander: Biology & Behavior of Heliconiinae in Trinidad 119 ties are concerned in some manner with pro- cesses associated with hardening. In relation to survival the important point is that if within eight minutes of emerging from its pupal case an adult can reach a position suit- able for flattening and hardening of its wings, no harm should result if it had fallen or been chewed from its natural support. It must how- ever be recognized that a pupa lying on the ground may well be more susceptible to attack by ants or small mammals or to destruction by molds. Thus the critical significance of selection of pupation sites, in relation to being chewed free by other larvae, cannot be dismissed yet. VI. Phylogenetic Discussion Similarities and differences between the larval behavior of some of the species of the Heli- coniinae became apparent in Part I of this paper and phylogenetic considerations dependent on them have already been briefly dealt with. Part II provides more material and, in addition, fea- tures of the pupae and behavior patterns of the emerging imago. As in Part I, attempts have been made throughout the description to explain the presence of these in terms of the factors which control their appearance and of their possible selective advantage to the species. It is necessary now to see how far these latter behavioral char- acteristics may be of phylogenetic importance. Considering the phylogenetic age of the molt- ing patterns and the improbability that they are strongly influenced, by the food plant of the caterpillar, it is in fact surprising that there are any specific differences here at all. Yet the position taken up on the vine prior to molting is specifically determined to a large extent and can be partially correlated with feeding and rest- ing positions (Part I). Those species, H. erato, D. iulia and A. vanillae, which tend to rest on the stem, frequently molt there as well— in con- trast to H. aliphera, H. melpomene, H. isabella and to a lesser extent H. ricini. It is noticeable that H. aliphera is more like the other species of Heliconius in molting under the midrib of a leaf than is H. isabella, which molts on the blade of the leaf. This intermediate position of H. ali- phera was evident in the consideration of feed- ing of the larvae and will be remarked later in relation to behavior of pupae and emergent butterflies. Table IX. Examples of Records of Fanning Behavior in Emergent Butterflies under Normal Conditions and when Forced to Walk for Various Lengths of Time after Emerging from Pupal Case Number of wing beats during wing fanning, given in minutes and seconds: Species Normal conditions Butterfly forced to walk for number mins, shown in parentheses Agraulis vanillae 17 in 10 ? in 9, 30 (15) 13 “ 10 11 “ 10 11 “ 9 10 “ 9 Dryas iulia 24 in 13 18 in 10, 30 (6) 21 “ 13 17 “ 10 (5V2) 18 “ 13 15 “ 7 (12) Heliconius isabella 26 in 17 23 in 15 (7) 25 “ 17 28 “ 16 +22 “ + 13 Heliconius melpomene 27 in 17 31 in 22 (6) 22 “ 17 20 “ 10, 50 (214 ) 20 “ 15 19 “ 15 ? “ 15 22 “ 14 Heliconius erato 22 in 15 24 in 13, 10 (12) 12 “ 11 (7+2) Heliconius ricini 22 in 17 25 “ 15 21 in 11, 15 (8) 21 “ 14 16 “ 10, 35 (6+2) 120 Zoologica: New York Zoological Society [46: 11 Gregarious species, D. juno, H. sara and, less markedly, H. ricini, molt together on a stem although in normal resting they are often found on the leaf they had been eating. Likewise the pose adopted prior to molting is related to that usually taken by a resting cater- pillar and is thus species-specific to some extent. There is less variation between the species when it comes to the actual shedding of the skin and subsequent expansion of the scoli and spines. There is but a slight suggestion that A. vanillae waits longer during the triangle stage than do the others and that H. sara, H. ricini, H. erato, H. melpomene, H. aliphera and D. juno are towards the other extreme in a short wait. Once the caterpillar does begin shedding its skin, the process is almost invariably slower in A. vanillae than other species, a feature which might well appear among caterpillars which are relatively unspecialized or unlikely to be interrupted by others during the molting process. Conversely the habit of shedding empty head capsules before the skin is free is one which might be retained in an unspecialized species or acquired as part of the plasticity necessary when gregarious cater- pillars are molting together. It occurs in A. van- illae, D. juno, H. ricini and H. sara. D. iulia and D. phaetusa were seen in Part I to share many behavioral characteristics of feed- ing and resting. Both show two peculiarities in molting behavior— leaving the cast skin in a small heap instead of spreading it out as they emerge from it, and the marked head-to-tail arching after spine expansion. There are quite marked differences in the choice of pupation site and these can be corre- lated to some extent with specific differences in feeding, resting and weaving behavior. In feed- ing both H. aliphera and more noticeably H. isabella orient on the blade of a leaf. Resting always takes place on the blade and molting has not been seen to occur anywhere other than on a leaf blade in these two species. The emphasis on weaving has already been mentioned in H. ali- phera and H. isabella and its possible correlation with larval life spent primarily on a smooth blade. Selection of a pupation site can be seen as part of the same syndrome : fifth instar H. isa- bella caterpillars invariably hang up on the ven- tral surface of a leaf blade if possible while those of H. aliphera show definite indications of the same tendency. H. isabella is in many other char- acteristics more specialized than H. aliphera and these two species are alone among the Trinidad heliconiines in their choice of pupation site. Thus it seems justifiable to assume that the choice of the smooth surface of a leaf blade as a pupation site is a specialization among them. Among the species studied, D. iulia is unique in its rapid and almost invariably long journey to find a pupation site on a foreign plant or sur- face. On the other hand, it is the most catholic in its final choice— a characteristic which is sel- dom associated with specialization. D. phaetusa, which in respect of feeding, resting and molting is very like D. iulia, cannot really be judged from observations on a single individual, but it seems probable that its behavior in selecting a site for pupation is more like that of H. erato than D. iulia. H. melpomene, H. ricini and to a lesser extent H. erato show clear similarities in their choice of pupation site. This is not surprising since they have shared characteristics of feeding and rest- ing. There is little information on H. sara and because I did not see the behavior myself, I feel unable to comment on it. There is no evi- dence to suggest that it should be considered different from H. melpomene in its choice of pupation site. If it were assumed, then, that leaf surfaces were the pupation site among primitive heli- coniines and in particular the midrib of the leaf, the situation has changed in D. iulia, D. phaet- usa, H. erato, A. vanillae and D. juno. The argu- ment relating to scarcity of leaves and their flimsiness and flexibility on P. tuberosa might explain the abandoning of the presumed primi- tive site in the cases of D. iulia , D. phaetusa and H. erato. Much of the manner of eating, position and posture in feeding and resting as well as position during molting has already been related to characteristics of this vine. It might be pos- sible to extend the argument to cover A. vanillae as well, in that its food plant P. foetida, has fairly thin leaves. D. juno, which pupates as a group, could perhaps have been forced into abandoning leaves as pupational sites because of lack of available space. It seems more reason- able, however, to suggest that midrib selection is a specialization adopted within the genus Hel- iconius and that the more basic choice was wider, including stem, petiole, flower stalk and midrib. The tendency to orient with respect to the mid- rib is particularly strong when the latter is as defined as that in a leaf of P. laurifolia. It has already been pointed out that H. melpomene and the solitary H. ricini and H. sara orient to the midrib in feeding, rest along it and molt on it. If H. aliphera were “in the course of” aban- doning a preference for protuberances in favor of flat surfaces as pupation sites, a shift to the midrib or a fine vein would be the last stage be- fore reaching a leaf blade position like that of H. isabella. In other prepupational behavior H. aliphera 1961] Alexander: Biology & Behavior of Heliconiinae in Trinidad 121 and H. Isabella are distinct from H. melpomene, H. erato, H. ricini and H. sara. They eat until the search for a pupation site begins and H. Isa- bella may even stop to eat during the walk. The search begins distinctly later in the day than it does among other Heliconius species. Differ- ences in “trying for height” behavior are almost certainly part of a syndrome of pupal adaptation and adult emergence and in all this H. aliphera and H. isabella are alike and differ from other Heliconius. The significance of the color differ- ences in the silk of the pupal pads is obscure. It is worth noticing, however, that H. aliphera and H. isabella again resemble each other in pro- ducing white silk and in differing from H. mel- pomene, H. erato and H. ricini. The red or pink silk characteristic of these latter species is cer- tainly a specialization in that D. iulia, D. phae- tusa, D. juno, A. vanillae and even H. sara pro- duce white silk like that of H. aliphera and H. isabella. D. iulia, D. phaetusa and especially D. juno select their sites late in the day. A. vanillae, on the other hand, has usually chosen its site and often hung up by mid- or late afternoon, only slightly later than is usual for H. aliphera and H. isabella. It spends an even longer period between its last larval feed and the time when it hangs up, sometimes resting almost motionless from 6:30 a.m. until the afternoon. A long fast also occurs in D. iulia but since fifth instar cater- pillars of this species do not normally feed dur- ing the morning, its relation to the fast in A. vanillae is not clear. Among lepidoptera in gen- eral it seems that pupational behavior only starts some time after the last larval meal— thus Crowell (1943) describes a protracted resting period in Prodenia larvae prior to pupation and the larvae of the cecropia silk moth do not begin their wandering phase until they have emptied their guts of the last larval meal (v.d. Kloot & Williams, 1953). This would support the idea that H. aliphera and H. isabella are specialized in having no rest period at all prior to site se- lection and that A. vanillae is in this respect the most primitive of the species considered here. The time of day at which a pupating cater- pillar sheds its last larval skin is linked at least partially with the time at which it spun its pad and hung up on the day before. Thus H. erato, H. melpomene and H. ricini tend to be earliest. H. aliphera, H. isabella, A. vanillae and D. iulia are somewhat later. D. juno is the latest and here the last larval skins are still being shed during mid-afternoon of the day after hanging up. Among the behavioral characteristics of the pupa itself, those of H. isabella and H. aliphera again are in an association separating them sharply from H. melpomene, H. ricini, H. erato, D. iulia and D. phaetusa, though less strongly from A. vanillae and D. juno. The two species mentioned last are perhaps more closely allied to each other in their pupal than in any other behavior. The behavior described for H. melpomene and D. iulia so far has been fairly distinct. In pupal behavior, however, they are very alike, movements, sounds and odor production being very similar. The odors are perhaps different enough to suggest that this might be due to con- vergence. Yet it has already been pointed out that D. iulia might be losing the gold spots which are suggested as part of the equipment of ad- vertisement. The direction of evolution in such a case can only be estimated in the light of other closely related species. It is thus very unfortunate that information on pupal behavior of D. phae- tusa is so limited as to be valueless in such a con- nection. Since this species so often shows charac- teristics intermediate between those of D. iulia and the Heliconius species, it might well be of importance here. The close relationship of H. ricini and H. erato with each other and with H. melpomene has been continually evident throughout the study and these two species are so alike in pupal be- havior and form that it is possible to confuse them. Information on the emergence behavior of the species is admittedly scanty, yet it is sufficient to divide the butterflies into four groups. H. melpomene, H. erato, H. ricini and H. sara compose the first of these. All of the four show palp movements clearly during the wing fanning phase of emergence but lack them alto- gether during the subsequent relaxation. None of them extend the palps or twist their heads from side to side just prior to flight. The actual fan- ning phase is variable, both in duration and num- ber of wing beats. With the exception of H. sara, however, the range of variation of this group would have to be extended to accommodate any of the other species studied in Trinidad. The rate of wing beating is very close in H. melpomene, H. erato and H. ricini, though H. sara beats very much faster. H. aliphera and H. isabella are closer to the other Heliconius in the emergence behavior of their images than any other aspects so far. Move- ment of the palps during fanning is however less marked and tends to occur later. There is also a slight tendency to include palp movements in the relaxation period. The wing beats of fanning are somewhat quicker in the other Heliconius but the duration of the behavior is the same. 122 Zoologica: New York Zoological Society [46: 11 The six species considered so far show no sign of palp twitching and twisting of the head before flight but all of the remaining species do. Of these D. iulia and D. phaetusa form one group, A. vanillae and D. juno the other. The former are like the four Heliconius species in that palp movement is obvious during fanning and does not extend into the relaxation period. The rate of fanning in D. phaetusa is approximately the same as that in the H. melpomene group, but in D. iulia it is much higher. Both D. juno and A. vanillae show practically no palp movement during fanning but have a very distinct burst of it during relaxation. The wing beats during fanning are much slower than in any other species. Information relating to the period between emergence from the pupal case and shedding of the meconium is surprising. Firstly because the co-efficient of variation varies so much between species, being very high in H. melpomene and H. erato and only an eighth the size in D. juno and A. vanillae. It seems possible that the point at which meconium is shed is changing in some species or that an unsuspected feature, such as the sex of the individual, is coming into the matter. The second point for comment is that the time of shedding meconium differs so between species which would be regarded as closely related on other grounds. Thus D. iulia and D. phaetusa contrast strongly, as do H. aliphera and H. isa- bella. Since the liberation of meconium is not effected by whether or not fanning occurs, it is logical that the point of time at which it occurs bears no relation to duration or rate of fanning. Nor is there any direct correlation between the color of meconium and time of shedding. Consideration of the color of the meconium supports all previous evidence linking H. ali- phera and H. isabella and separating them from the other species of Heliconius. This character also links D. juno and A. vanillae as do other features of their larval and pupal behavior. Meconium of D. iulia and D. phaetusa cannot be distinguished by color. All of the remaining species of Heliconius clearly belong together as a group for, although H. ricini produces none of the gray meconium typical of the others, it is nevertheless closer to H. melpomene than to any other species in respect of color. In conclusion, much of the information given in this paper supports the hypotheses raised tentatively in the discussion of Part I. Relation- ships gauged on general behavior of the cater- pillars corresponds fairly closely with those sug- gested by considering larval molting, and the behavior of pupae and emerging adults. It is interesting, however, that the apparent degree of relationship between species differs at differ- ent stages. Thus the early larval instars of H. aliphera, H. isabella and D. juno are very alike in their feeding patterns. As they grow, similar- ities decrease. Pupal behavior is distinctly dif- ferent and the question arises as to whether the points of agreement could not be due to con- vergence. The behavior of emergent H. aliphera and H. isabella is no closer to that of D. juno than any other species. Conversely, the relation- ship between these two species of Heliconius and the H. melpomene, H. erato, H. ricini, H. sara group is extremely obscure during larval and pupal life and it is only at the emergence of the adult that similarities become apparent. The last mentioned group of Heliconius may indeed warrent Michener’s description of “large and relatively homogeneous” (1942) in adult be- havior but among the larvae there are at least some cases of non-conformity. Thus H. erato, on the basis of late larval, pupal and emergent behavior, is obviously closely related to H . mel- pomene. In its general larval behavior, however, even including molting, this is masked by modi- fications which probably arose in relation to its particular food plant. There are indications that it may originally have lived on a vine other than its present P. tuberosa and it has been suggested that this was somewhat like the thick-veined P. laurifolia on which H. melpomene and H. ricini feed. Thus the behavior of larval lepidoptera can clearly be of taxonomic significance. The point which is perhaps of more importance is that such behavioral features, together with those of pupal and adult stages, are essential to form a picture of evolutionary relationships which is anywhere near complete. Taxonomists once favored the use of characteristics which are as far as possible independent of the environment, preferentially those with no apparent functional significance: in the past features which could be regarded as adaptive have been taxonomically suspect. There is now a wider recognition of the fact that to understand the pathway along which an animal has evolved, its mode of life, in as many aspects as possible, must be considered. While nobody is ever likely to suggest that animals be classified on their behavioral characteristics alone, or on their physiology, biochemistry or ecology, such disciplines are essential in interpreting an- atomical features. VII. Summary 1 . Behavioral observations were made on cater- 1961] Alexander: Biology & Behavior of Heliconiinae in Trinidad 123 pillars of butterflies of the subfamily Heli- coniinae. Some of these observations concern 1 1 of the 14 species present in Trinidad, some con- cern 10 and some 9. 2. A brief description is given of the general molting behavior. This comprises a quiescent period, the movements during the actual shed- ding of the skin, expansion of the scoli, hard- ening and darkening of the cuticle and a further period of rest. The larva then turns to eat its cast skin before resuming normal feeding. Slight systematic differences occurring within the be- havior are described. 3. It is established that caterpillars will eat sev- eral cast skins, one after the other, and that they do not distinguish between those of their own and other species. The age of the exuvium is of no significance but that of a fifth instar animal, i.e., last larval molt, is usually rejected. 4. Behavior leading to the selection of a pupa- tion site is described. This involves a locomotory phase and some estimation of the amount of free space beneath a potential site. Differences in the preferred pupation sites of the species are de- scribed and the significance briefly considered. 5. Behavior leading to the attachment of a late fifth instar larva to its silk pad is described and the results of experimental interference related. The degree of flexibility of attachment, the ex- tent of movement, stridulation and production of odor are compared in the ten species whose pupae were studied. These features are correla- ted with pupation site, form and color of the pupae. 6. Emergence of a butterfly from its pupa is described and specific differences in this be- havior noted. Observations made when recently emerged butterflies were forced to walk for vary- ing periods before wing expansion are described and the implications of the results are discussed in relation to larval behavior and the pupation site. 7. The phylogenetic implications of the sys- tematic differences in molting, late larval, pupal and emergent adult behavior are discussed. Prac- tically all the information suggests the same pic- ture of the relationship of the species concerned. This coincides to a great extent with that drawn from considering general larval behavior. VII. References Alexander, Anne J. 1961. A study of the biology and behavior of the caterpillars, pupae and emerging but- terflies of the Subfamily Heliconiinae in Trinidad, West Indies. Part I. Some aspects of larval behavior. Zoologica, 46 ( 1 ) : 1-24. Beebe, William, Jocelyn Crane & Henry Fleming. 1960. A comparison of eggs, larvae and pupae in fourteen species of heliconiine butter- flies from Trinidad, W.I. Zoologica, 45 (9): 111-154. Brocher, F. 1919. Le mechanisme physiologique de la derniere mue des larves des Agrionides. Ann. Biol. Lacustre Bruxelles. 9: 183-200. Crowell, H. H. 1943. Feeding habits of the southern army worm and the rate of passage of food through its gut. Ann. Ent. Soc. Amer., 36: 243-249. Hinton, H. E. 1955. Protective devices of endopterygote pupae. Trans. Soc. Brit. Ent., 12: 50-92. v. d. Kloot, W. G. & C. M. Williams 1953. Cocoon construction by the Cecropia silk- worm. I. The role of the external environ- ment. Behaviour, 5: 141-156. Michener, C. D. 1942. A generic revision of the Heliconiinae. (Lepidoptera. Nymphalidae). Amer. Mus. Nov., no. 1197, Oct. 1942. Poulton, E. B. 1887. An enquiry into the cause and extent of a special colouration relation between cer- tain exposed lepidopterous pupae and the surfaces which immediately surround them. Phil. Trans. Roy. Soc. (B), 178: 311-441. Wachter, S. 1930. The moulting of the silkworm and a his- tological study of the moulting gland. Ann. ent. Soc. Amer., 23: 381-389. 124 Zoologica: New York Zoological Society [46: 11: 1961] EXPLANATION OF THE PLATE Plate I Pupa of H. aliphera, showing the dorso-ventral bend in the body which causes it to lie almost parallel to the surface under which it is attached, in this case a sheet of glass. ALEXANDER PLATE I FIG. 1 A STUDY OF THE BIOLOGY AND BEHAVIOR OF THE CATERPILLARS, PUPAE AND EMERGING BUTTERFLIES OF THE SUBFAMILY HELICONIINAE IN TRINIDAD, W.l. 12 Melanoma, Renal Thyroid Tumor and Reticulo-endothelial Hyperplasia in a Non-hybrid Platyfish1 Pamela A. Mac Intyre & K. France Baker-Cohen2 Genetics Laboratory of the New York Aquarium, New York Zoological Society (Plates I & II; Text-figure 1) MELANOMAS consistently occur in certain platyfish-swordtail and other interspecific hybrid combinations among fishes of the genus Xiphophorus. Tumors of the thyroid gland develop in certain species and strains of Xiphophorus when they are main- tained in iodine-poor water. Genetic factors are important in the etiology of these atypical growths of pigment and thyroid cells (Gordon, 1958; Atz, 1959; Baker, 1958a; Mac Intyre, 1960). The discovery, in our laboratory, of a non- hybrid platyfish with a melanoma evoked con- siderable interest, and when this specimen was also found to exhibit tumorous thyroidal growths in both the pharyngeal region and kidney, an abnormal pituitary gland and hyperplasia of reticulo-endothelial cells, the value of studying it in detail became even more apparent. During the study large pigmented masses, interpreted as melanin-laden macrophages, were found in the kidney— a phenomenon not known to have been previously described. In these fishes, melanosis necessarily precedes melanoma. One melanotic specimen of the same species was found and is described. Description of Melanomatous Fish with Thyroid Tumor History and Gross Anatomy The specimen lThis work was supported by grants from the National Cancer Institute to the late Dr. Myron Gordon (C-297 ) and to Dr. Sylvia Greenberg (C-4945) and by a re- search fellowship (CF-6184) from the National Cancer Institute to the junior author. It was aided by the labora- tory facilities of the American Museum of Natural His- tory, New York 24, New York. 2Present address: Department of Anatomy, Albert Einstein College of Medicine, Eastchester Road and Morris Park Avenue, Bronx, New York. was an adult male spike-tailed platyfish, Xipho- phorus variatus xiphidium, from the sixth labo- ratory-reared generation of fish collected in 1939 from the Rio Purificacion, Tamaulipas, Mexico (strain 53) . The identification was veri- fied by Dr. Donn E. Rosen of the American Museum of Natural History, The fish measured 28 mm. in standard length and 12 mm. in depth. It carried the dominant sex-linked gene, Sp, for macromelanophore spotting on the body. In this specimen much of the body was intensely black. Externally a swelling, which was caused by the melanoma, was visible on the right side under- neath the dorsal fin (Text-fig. 1 ) . The extent of the swelling was approximately 8 X 6V2 X 2 mm. Its surface was smooth. The thyroid tumor was not visible externally. The specimen was fixed in Bouin’s fluid, de- calcified in formic acid, embedded in paraffin and sectioned at 3, 7 and 10 microns; staining was with Harris’s hematoxylin and eosin, Masson’s trichrome stain or Heidenhain’s iron hematoxylin. Before staining, the melanin in some sections was bleached with hydrogen per- oxide or with potassium permanganate and oxalic acid. Microscopic Appearance of the Melanoma.— The melanoma was essentially similar in micro- scopic appearance to those previously described in hybrids between Xiphophorus maculatus and X. hellerii by Reed & Gordon (1931), Gordon & Smith (1938), and others. It has caused ex- tensive destruction of the body musculature (Plate I, Fig. 2) . The dermis was infiltrated with tumor cells, and it contained melanin-laden macrophages. The epidermis was considerably thickened in some regions and also contained melanin-laden macrophages (Fig. 3). The scales were in abnormal positions. 125 126 Zoologica: New York Zoological Society [46: 12 The predominant type of cell in the melanoma was the lightly pigmented melanocyte. There were scattered melanin-laden cells — macro- phages and melanophores.3 In areas adjacent to muscle cells, the melanocytes were arranged in whorls, and their nuclei were spindle-shaped and small, generally 2Vi X 5 microns (Fig. 4). Elsewhere the nuclei of the melanocytes were round, oval or irregular and generally somewhat larger; the size varied but the majority were under 5 microns in diameter (Fig. 5). The cell outlines of the melanocytes were indistinct. The nuclei contained one or more prominent nu- cleoli. Mitotic figures were rare. A number of nuclei showed median or budlike constrictions, possibly indicating amitosis. Some giant nuclei were observed (Fig. 6). Text-fig. 1. Lateral and dorsal views of a male Xiphophorus variatus xiphidium with a melanoma. Microscopic Appearance of the Thyroid Tumor— The pharyngeal thyroid was markedly hyperplastic and extended into the bases of the gills (Plate II, Figs. 10, 11). This hyperplasia was chiefly made up of small to large follicles, all filled with dense, acidophilic colloid. The col- loid was not vacuolated and varied from a very dense homogeneous condition in the small follic- les to a course granular state in the large ones. Some afollicular cells occupied interstices be- tween the follicles. Thyroid tissue was present in small amounts in the chorioid gland of the eye and was either follicular with granular colloid or afollicular. In the kidneys, thyroid tissue was rep- resented by many large, colloid-filled follicles, similar to those in the pharynx, and by swollen cystic follicles almost or completely devoid of colloid (Figs. 12, 13). No thyroid tissue was seen elsewhere in the body. The renal thyroid 3In lower vertebrates, the melanocyte, which is the melanin-synthesizing cell common to all classes of verte- brates, differentiates into a pigment-effector cell called the melanophore; the melanophore is not a pigment- carrying macrophage (Gordon, 1953). was different from the thyroid tumors in the kidneys of Xiphophorus maculatus described by Baker et al. (1955) in that the normal kidney tissues were less disturbed by the proliferation of thyroid tissue than they were in X. maculatus during the late stages of tumor growth and that large amounts of colloid were present through- out both the pharyngeal and renal thyroid masses. Hyperplasia of the Pituitary Gland— Although the pituitary gland of this fish was not sectioned or stained in a manner suitable for the demon- stration of cell types, it was at once apparent that the gland was abnormal (Plate II, Fig. 14). In the anterior part of the gland, the vertical cross-section was much elongated; the elonga- tion was made up of a homogeneous mass of light-staining basophils. Eosinophils were not delineated by the staining methods employed. Normally, at that cross-sectional level, the pitui- tary is round and chiefly composed of darkly staining basophils; the paler cells appear in small numbers along the ventral border and increase somewhat in the posterior direction, as the neural portion of the hypophysis appears. In the tumorous fish, the anterior basophilic over- growth also continued along the entire length of the pituitary, making the entire organ mis- shapen. Hypertrophy of the paler basophilic elements of the pituitary gland has been found in X. maculatus with goitrous thyroids or with thyroid tissue regeneration following radioiodine treat- ment (Baker-Cohen, unpublished data). In such fish, however, the proliferation of pale basophils did not assume the proportions seen in the tu- morous X. v. xiphidium. Hyperplasia of the Reticulo-endothelial Sys- tem— In the kidney there were numerous areas of cells heavily laden with melanin (Plate II, Figs. 12, 13). These cells were spherical and from 5 to 10 microns in diameter. Their nuclei, which were visible only in bleached sections, were compressed at the periphery of the cells because of the large amount of melanin. The nuclei were oval or irregular in shape, generally 1 to 3 microns long and Vi to 1 Vi microns wide (Figs. 8, 9). In the anterior part of the mela- noma there was a distinct area that consisted of a large mass of melanin-laden cells similar to the ones described above. This appeared to be a distinct growth and not part of the melanoma proper. This mass was ventral to the vertebral column, surrounding the urinary ducts and dorsal to the air bladder, which it compressed (Plate I, Fig. 7). Posteriorly this mass was re- duced to a few cells ventral to the vertebral col- umn (Fig. 1). A few such cells were found in 1961] Mac Intyre & Baker-Cohen: Melanoma in a Non-hybrid Platyfish 127 the liver, surrounding the brain, and in blood vessels within the brain. These cells appear to be macrophages, present in hyperplastic propor- tions (see Discussion). Closely associated with the melanin-laden macrophages in the kidney were many nodular areas containing pale cells with indistinct out- lines and round nuclei (2 Vi to 5 microns in di- ameter). Scattered macrophages were found in many of these nodules. We interpret these pale cells as primitive reticular cells, present in hyper- plastic proportions. Some of these cells were also found associated with the area of melanin-laden cells in the anterior part of the melanoma. Ex- tensive nodules of these cells were found among the pharyngeal thyroid follicles (Fig. 1 1 ) . In the pharyngeal area, macrophages were rarely found in these nodules. Description of Melanotic Fish History and Gross Anatomy.— This specimen also was a laboratory-reared, adult male X. v. xiphidium. It was identified by Dr. Donn E. Rosen. The fish measured 27 mm. in standard length and 11 mm. in depth. It carried the Sp gene; in this specimen the body was almost com- pletely black (blacker than the body of the melanomatous fish). The specimen was fixed in formalin, decalcified in formic acid, embedded in paraffin and sectioned at 7 and 10 microns; staining was with Harris’s hematoxylin and eosin. Microscopic Appearance of the Melanosis.— In the posterior part of the body, melanophores were present not only in the dermis but had spread along the septae of the body musculature. Some lightly pigmented melanocytes had also spread among the muscles, and in the left dorsal side melanocytes had multiplied sufficiently to compress some of the muscles. There was slight destruction of the musculature. The epidermis in this area was thickened. It is not possible to say whether, if the fish had remained alive, this melanosis would, have developed into a mela- noma. Microscopic Appearance of the Thyroid Tis- sue and Pituitary Gland— The pharyngeal thy- roid appeared normal. It consisted of a number of follicles, 20 to 175 microns in diameter, none containing colloid (probably an artifact, result- ing from formalin fixation). The follicular cells were squamous to cuboidal. No thyroid tissue was observed in the kidney. The kidney con- tained many pearl-like basophilic concretions, similar to those described by Berg et al. (1954) in regressing thyroid tumors that had been treated with potassium iodide. The hypophysis of this fish, although not well preserved because of the slow formalin fixation, appeared to be normal. Its structure seemed to be the same as seen in X. maculatus. Macrophage Hyperplasia.— There was a large mass of spherical melanin-laden cells, similar to those described as macrophages in the melano- matous X. v. xiphidium, ventral to the vertebral column, surrounding the urinary ducts and uri- nary bladder. Numerous masses of these cells were present in the kidney, and some were found in the lining of the air bladder. Some primitive reticular cells were also present in the kidney. Discussion Melanoma.— Thousands of wild and labora- tory-reared poeciliid fish, including Xiphophorus variatus xiphidium, have been examined in this laboratory and elsewhere, but melanomas have been found only in interspecific hybrids, with the exception of the Fu ( fuliginosus ) domesti- cated strain of X. maculatus (Kosswig, 1938) and a few specimens of X. montezumae cortezi carrying the genetic factor for spotted caudal peduncle, Sc (unpublished data). The macro- melanophore genes in platyfish require genetic buffering to control their growth, which is upset in hybrid combinations (Gordon, 1958; Atz, 1959). In the melanomatous X. v. xiphidium, which carried the Sp gene for macromelano- phore spotting, the normal genetic constitution of the species was presumably altered, perhaps as a result of mutation or inbreeding. Thyroid Tumor.— The discovery of extensive thyroidal proliferation in the kidneys of the melanomatous X. v. xiphidium adds another species to the growing roster of teleosts in which extra-pharyngeal thyroid tissue has been found. These have been summarized recently by Baker (1959) and Baker-Cohen (1959). In all cases to date, the fish exhibiting displaced thyroid tissue have had some history of a domesticated environment, sometimes with a known iodine deficiency, and all were freshwater species. Both the pharyngeal and renal thyroidal growths in the X. v. xiphidium, like those de- scribed in a cherry barb, Barbus titteya (Baker, 1959), and unlike those reported in common platyfish, X. maculatus (Baker et al., 1955), were extensively colloidal. They thereby more nearly resembled colloid goiter than adenoma, as described in man. Thyroid tumors have rarely been found in this species. However, they may be more fre- quent than previously supposed if they are not visible externally, as was the present one. Thy- roid tumors in poeciliid fish generally produce externally-visible swellings of the gill region or of the body. 128 Zoologica: New York Zoological Society [46: 12 Relation between the Melanoma and Thyroid Tumor— Stolk (1959) recently reported that the incidence and development of melanomas was increased in thiouracil-fed (hypothyroid) platy- fish-swordtail hybrids and reduced in thyroxine- fed (hyperthyroid) ones. Similar results were obtained by Meites (1958) for carcinogen-in- duced skin tumors in Swiss mice; his article gives references to the few and contradictory studies that have been made on the influences of the thyroid gland on tumor growth. Endocrine secretions, particularly certain pi- tuitary hormones, play an important role in pigment cell formation and maturation and in melanin synthesis in normal fish (Pickford, 1957; Chavin, 1959; Kosto et al., 1959). Con- sidering the extreme rarity of melanomas in X. v. xiphidium, it is conceivable that the de- velopment of the melanoma described here might have been influenced, by the abnormal state of the fish’s endocrine system, particularly in view of the unusual state of hyperplasia of the hypophyseal basophils. Hyperplasia of the Reticulo-endothelial , Sys- tem.—Morphologically, the melanoma and melanosis described here are similar to those re- ported in platyfish-swordtail hybrids, with the exception of the large number of heavily mel- anin-laden cells in the kidney and ventral to the vertebral column. Metastases are very rare in fish melanomas. Breider (1938, 1953) described pigment deposits in the kidneys of melanoma- tous platyfish-swordtail hybrids. He considered them to be acellular pigment deposits but did not make any studies of bleached sections; the deposits shown in his figures appear very similar to the masses of melanin-laden cells found in our X. v. xiphidium. Mac Intyre (unpublished data) has found similar, though less extensive, masses in the kidneys of some melanomatous X. maculatus belonging to the Fu strain and in some melanomatous platyfish-swordtail hybrids. We do not believe that these represent me- tastases, but rather macrophages that have both multiplied to hyperplastic proportions and massed in these areas. Cell movement alone could not account for the large quantity of these cells. What happens to melanin after it has been produced by the melanocyte is not fully known (Dalton & Felix, 1953). Much of it is trans- ferred to macrophages. The recent findings of Speece et al. (1959) in human melanomas in- dicate that macrophages not only ingest dead cells but pick up melanin discharged into the intercellular fluid by active tumor cells. Obser- vation of cut pigment cells in amphibians showed that some melanin-containing macrophages re- mained at the area of ingestion, while others migrated and entered the blood stream (Lehman, 1953). Melanin debris may not be immediately ingested by macrophages; it may be removed from tissue spaces by the lymphatics (Smith, 1931) and thus carried to lymph nodes or other areas of macrophage concentration. Ultimately, the melanin may be broken down intracellularly in macrophages, or, as Cowdry (1950) stated, “once formed, melanin is a very stable sub- stance and may be excreted in the urine.” In fish and amphibians, some melanin-containing mac- rophages are sloughed off through the epidermis (Goodrich & Hansen, 1931; Gordon & Lansing, 1943; Niu, 1959). Niu (1959) has suggested that in amphibians the pigment debris is gen- erally carried by macrophages to the liver, where it is eliminated via the biliary system. In teleost fishes, the kidney is the organ rich- est in lymphoid tissue (Bertin, 1958) and the principal hemopoietic organ (Grasse, 1958). In the kidney sinuses of normal poeciliid fish, large numbers of lymphocytes and erythrocytes and small numbers of primitive reticular cells and macrophages are found. Primitive reticular cells can differentiate into lymphocytes and macro- phages (Walvig, 1958). Studies in the eel have shown that injected dye and bacteria are picked up chiefly by lymphoid cells in the kidney (Schmidt, 1959). Bertin (1958) stated that ink or carmine injected into a fish’s body is elimi- nated in the urine, after several hours, across the renal lymphoid tissue. In the two specimens of X. v. xiphidium, his- tological observations indicate that much of the enormous amount of melanin produced by the melanoma and melanosis was picked up by macrophages in the kidney, which were evi- dently not able to eliminate the melanin, by intracellular digestion or by transfer into the urine, at the rate that it was produced by the abnormally large number of melanocytes. In addition to this process, much of the melanin was phagocytized by macrophages in the im- mediate area, some of which remained there and others of which were sloughed off through the epidermis. That these are methods of mel- anin disposal in poeciliid fishes is confirmed by observations we have made in normally pig- mented fish : e.g. small numbers of macrophages containing melanin are customarily found in the kidney and other lymphoid tissues. The nodules of primitive reticular cells found in the kidney and in the pharyngeal thyroid tumor resemble cells that were found by Rasquin & Rosenbloom (1954) in Astyanax raised in the dark and described by them as reactive hyper- plastic reticulo-endothelial cells. Baker (1958b) 1961] Mac Intyre & Baker-Cohen: Melanoma in a Non-hybrid Platyfish 129 found similar masses, associated with degenerat- ing thyroid follicles, in X. maculatus whose renal thyroid tumors had been treated with potassium iodide. Relations between the Reticulo-endothelial Hyperplasia and the Melanoma and Thyroid Tumor.— Several studies have been made in mammals of the relation between the reticulo- endothelial system and neoplastic diseases ( e.g . Pelner, 1957; Old et al., 1960). The latter found a slight increase of RES activity during the growth of certain spontaneous mouse tumors and suggested that this activation might repre- sent response to some unique property of the tumor, stimulation by products of tumor growth, or response to associated necrosis, hem- orrhage, or infection. In the fish with the mela- noma and thyroid tumor, we conclude that there was a marked reactive hyperplasia of the reticulo-endothelial system, particularly of mac- rophages and primitive reticular cells, as a result of the presence of the tumors, for one or more of the above reasons. Reaction to exces- sive melanin production may have been a causa- tive factor in the hyperplasia, or the melanin phagocytosis may have been a secondary phe- nomenon. The cells of this reactive hyperplasia were present in such great quantity that they themselves probably caused considerable dam- age through pressure on the air bladder and kid- ney tissue. Summary Melanomas consistently occur in certain platyfish-swordtail and other interspecific hy- brid combinations of xiphophorin fishes. A melanoma in a non-hybrid xiphophorin, the spike-tailed platyfish, Xiphophorus variatus xiphidium, is described, in which the normal genetic mechanisms buffering pigment cell growth were presumably upset. The melanoma consisted mainly of lightly pigmented melano- cytes and had destroyed much of the body mus- culature on the right side. This fish had an ex- tensive thyroid tumor in the pharyngeal region and in the kidney, consisting of colloid-filled and swollen cystic follicles. In the pituitary gland, a hyperplasia of light-staining basophils was pres- ent. In addition to these abnormal growths, there was a marked reactive hyperplasia of the reticulo-endothelial system, consisting of large masses of melanin-laden macrophages in the kidney and ventral to the vertebral column, and of nodules of primitive reticular cells in the kidney and in the pharyngeal thyroid tumor. The disposal of melanin and the reticulo-endo- thelial system in fish are discussed; it appears that much of the melanin produced by pigment cells is picked up by phagocytic cells in the lymphoid tissue of the kidney., Another specimen of X. v. xiphidium is de- scribed, in which melanosis was present. Large masses of melanin-laden macrophages were also found in its kidney and ventral to the vertebral column. Acknowledgments We wish to acknowledge our gratitude to the late Dr. Myron Gordon for encouraging this research and to Dr. James W. Atz for critical reading of the manuscript. Literature Cited Atz, J. W. 1959. Morphological and genetic studies on the pigmentary patterns of xiphophorin fishes and their hybrids. Ph. D. thesis, New York University, New York, N. Y. Baker, K. F. 1958a. Heterotopic thyroid tissues in fishes. I. The origin and development of heterotopic thyroid tissue in platyfish. Jour. Morph., 103: 91-134. 1958b. Heterotopic thyroid tissues in fishes. II. The effect of iodine and thiourea upon the development of heterotopic thyroid tissue in platyfish. Jour. Exper. Zool. 138: 329- 354. 1959. Heterotopic thyroid tissues in fishes. III. Extrapharyngeal thyroid tissue in Monte- zuma swordtails, a guppy and a cherry barb. Zoologica, 44: 133-139. Baker, K. F., O. Berg, A. Gorbman, R. F. Nigrelli & M. Gordon 1955. Functional thyroid tumors in the kidneys of platyfish. Cancer Res., 15: 118-123. Baker-Cohen, K. F. 1959. Renal and other heterotopic thyroid tissue in fish. In “Comparative Endocrinology.” Ed. by A. Gorbman. John Wiley, New York: 283-304. Berg, O., M. Gordon & A. Gorbman 1954. Comparative effects of thyroidal stimu- lants and inhibitors on normal and tumor- ous thyroids in xiphophorin fishes. Cancer Res., 14: 527-533. Bertin, L. 1958. Appareil circulatoire. In “Traite de Zoo- logie.” Ed. by P. Grasse. Masson et Cie, Paris, 13: 1399-1458. Breider, H. 1938. Die genetischen, histologischen und zyto- logischen Grundlagen der Geschwultsbild- ung nach Kreuzung verschiedener Rassen und Arten lebendgebarender Zahnkarpfen. Zeit. r. Zellf. u. mikro. Anat., 28: 784-828. 130 Zoologica: New York Zoological Society [46: 12 1953. Zur Physiologie der Nigra (N) — und Spotted (Sp)— Faktoren lebendgebarender Zahnkarpfen. Sonderdrunk aus “Strahlen- therapie,” 9 1 : 316-320. Chavin, W. 1959. Pituitary hormones in melanogenesis. In “Pigment Cell Biology.” Ed. by M. Gor- don. Academic Press, New York: 63-84. Cowdry, E. V. 1950. A Textbook of Histology. 4th ed. Lea & Febiger, Philadelphia, 640 pp. (see p. 537). Dalton, A. J., & M. D. Felix 1953. Phase contrast and electron micrography of the Cloudman S91 mouse melanoma. In “Pigment Cell Growth.” Ed. by M. Gordon. Academic Press, New York: 267- 276. Goodrich, H. B., & I. B. Hansen 1931. The postembryonic development of Men- delian characters in the goldfish Carassius auratus. Jour. Exper. Zool., 59: 337-358. Gordon, M. 1953. Preface. In “Pigment Cell Growth.” Ed. by M. Gordon. Academic Press, New York. 1958. A genetic concept for the origin of mel- anomas. Ann. N. Y. Acad. Sci., 71: 1213- 1222. Gordon, M., & W. Lansing 1943. Cutaneous melanophore eruptions in young fishes during stages preceding mel- anotic tumor formation. Jour. Morph. 73: 231-245. Gordon, M., & G. M. Smith 1938. Progressive growth stages of a heritable melanotic neoplastic disease in fishes from the day of birth. Amer. Jour. Cancer, 34: 255-272. Grasse, P. 1958. Addenda— Milieu interieur. In “Traite de Zoologie.” Ed. by P. Grasse. Masson et Cie, Paris, 13: 2626-2630. Kosswig, C. 1938. Uber einen neuen Farbcharakter des Platypoecilus maculatus. Rev. Fac. Sci. Univ. Istanbul, 3: 395-402. Kosto, B„ G. E. Pickford & M. Foster 1959. Further studies of the hormonal induction of melanogenesis in the killifish, Fundulus heteroclitus. Endocrinology, 65: 869-881. Lehman, H. E. 1953. Observations on macrophage behavior in the fin of Xenopus larvae. Biol. Bull., 105 : 490-495. Mac Intyre, P. A. 1960. Tumors of the thyroid gland in teleost fishes. Zoologica, 45: 161-170. Meites, J. 1958. Effects of thyroxine and thiouracil on in- duction of skin tumors in mice by 9, 10- dimethyl- 1,2-benzanthracene and croton oil. Cancer Res., 18: 176-180. Niu, M. C. 1959. Some aspects of the life history of amphib- ian pigment cells. In “Pigment Cell Biology.” Ed. by M. Gordon. Academic Press, New York: 37-50. Old, L. J., D. A. Clarke, B. Benacerraf & M. Goldsmith 1960. The reticuloendothelial system and the neoplastic process. Ann. N. Y. Acad. Sci., 88: 264-280. Pelner, L. 1957. Host-tumor antagonism. IX. The reticulo- endothelial system and neoplasia. Jour. Amer. Geriatrics Soc., 5: 916-931. Pickford, G. E. 1957. The chromatophore hormones of the pitu- itary. In “The Physiology of the Pituitary Gland of Fishes.” By G. E. Pickford & J. W. Atz. N. Y. Zoological Society, New York: 32-59, 281-321. Rasquin, P., & L. Rosenbloom 1954. Endocrine imbalance and tissue hyper- plasia in teleosts maintained in darkness. Bull. Amer. Mus. Nat. Hist., 104: 363-425. Reed, H. D., & M. Gordon 1931. The morphology of melanotic overgrowths in hybrids of Mexican killifishes. Amer. Jour. Cancer, 15: 1524-1537. Schmidt, F. 1959. Vorkommen und Verhalten von Sternzel- len der Leber des Aales und ihre Beziehung zum reticuloendothelialen System. Zeit. f. Zellforschung u. mikro. Anat., 49: 401- 417. Smith, G. M. 1931. The occurrence of melanophores in certain experimental wounds of the goldfish ( Car- assius auratus). Biol. Bull., 61: 73-84. Speece, A. J., J. P. Chang & W. O. Russell 1959. A microspectro - photometric - autoradio- graphic study of tyrosinase activity in human melanoma. In “Pigment Cell Biology.” Ed. by M. Gordon. Academic Press, New York: 371-387. Stolk, A. 1959. Effect of thiouracil and thyroxine on devel- opment and growth of cutaneous melan- oma in killifish hybrids. Nature, 184: 562- 563. Walvig, F. 1958. Blood and parenchymal cells in the spleen of the icefish Chaenocephalus aceratus (Lonnberg). Nytt Mag. Zool., 6: 111-120. 1961] Mac Intyre & Baker-Cohen: Melanoma in a Non-hybrid. Platyfish 131 EXPLANATION OF THE PLATES Plate I Fig. 1. Cross section of a melanomatous male Xiphophorus variatus xiphidium through the posterior part of the melanoma, which consists mainly of lightly-pigmented mel- anocytes. 13 X. Fig. 2. Infiltration of the musculature by melan- ocytes. 140X. Fig. 3. Thickening of the epidermis. 260 X. Fig. 4. Bleached section in an area of muscle in- filtration, showing whorls of melanocytes with small spindle-shaped nuclei. The large dark objects are muscles. 345 X. Fig. 5. Bleached section from a more central part of the tumor. The nuclei are larger and round or irregular in shape. 920 X. Fig. 6. A giant nucleus. 700 X. Fig. 7. Cross section through the anterior part of the melanoma. To the left are melanocytes which have invaded the musculature. A large hyperplastic mass of melanin-laden macrophages occupies part of the region where the air bladder is normally located. 70X. Plate II Fig. 8. Bleached section showing the edge of the melanin mass of Fig. 7. At the right are melanin-laden macrophages which, after bleaching, appear as empty shells. At the left are melanocytes. 450 X. Fig. 9. Bleached melanin-laden macrophages. The small, irregularly-shaped nuclei are com- pressed at the periphery of the cells. 945 X. Fig. 10. Thyroid tumor in the same fish. Pharyngeal region. 70 X. Fig. 11. Detailed view of pharyngeal region. At the right are thyroid follicles. At the left is a mass of hyperplastic primitive reticular cells. 375 X. Fig. 12. Thyroid tumor in the kidney showing large colloid-filled follicles and a large cyst at the upper right. The solid black areas are groups of hyperplastic macrophages, laden with melanin. 70 X. Fig. 13. Detailed view of melanin-laden macro- phages and thyroid tissue in kidney. 375 X. Fig. 14. Hyperplasia of light-staining basophils in the pituitary gland of the same fish. 100 X. MACINTYRE & BAKER-COHEN PLATE I MELANOMA, RENAL THYROID TUMOR AND RETICULO-EN DOTHELIAL HYPERPLASIA IN A NON-HYBRID PLATYFISH MACINTYRE & BAKER-COHEN PLATE II v4 Vt ••• T ... ;;T; V §X MELANOMA. RENAL THYROID TUMOR AND RETICULO- ENDOTHELIAL HYPERPLASIA IN A NON-HYBRID PLATYFISH 13 Eastern Pacific Expeditions of the New York Zoological Society. XLV. Non-intertidal Brachygnathous Crabs from the West Coast of Tropical America. Part 2: Brachygnatha Brachyrhyncha1’2 John S. Garth Allan Hancock Foundation, University of Southern California (Plate I; Text-figures 1 & 2) [This is the forty-fifth of a series of papers deal- ing with the collections of the Eastern Pacific Ex- peditions of the New York Zoological Society made under the direction of William Beebe. The present paper is concerned with specimens taken on the Templeton Crocker Expedition (1936) and the Eastern Pacific “Zaca” Expedition ( 1937-1938). For data on localities, dates, dredges, etc., refer to Zoologica, Vol. XXII, No. 2, pp. 33-46, and Vol. XXIII, No. 14, pp. 287-298.] Contents PAGE Introduction 134 Ecological Considerations 135 Geographical Considerations 135 Systematic Considerations 135 Restriction of Synonymies 136 Measurements 136 Acknowledgment 136 Systematic Discussion 137 Tribe Brachyura Subtribe Brachygnatha Superfamily Brachyrhyncha Family Portunidae Portunus (Port units) xantusii (Stimpson) 137 Portunus ( Portunus ) acuminatus (Stimpson) 137 Portunus (Portunus) asper (A. Milne Edwards) 138 Portunus (Portunus) panamensis (Stimpson) 138 Portunus (Achelous) brevimanus (Faxon) 139 Portunus (Achelous) pichilinquei Rathbun 139 Contribution No. 1011, Department of Tropical Re- search, New York Zoological Society. Contribution No. 243, Allan Hancock Foundation, University of Southern California. Portunus (Achelous) affinis (Faxon) 139 Portunus (Achelous) tuberculatus (Stimpson) 140 Portunus (Achelous) iridescens (Rathbun) 141 Callinectes arcuatus Ordway 141 Callinectes toxotes Ordway 142 Arenaeus mexicanus (Gerstaecker) 142 Cronius ruber (Lamarck) 143 Euphylax dovii Stimpson 144 Euphylax robustus A. Milne Edwards 145 Family Xanthidae Medaeus lobipes Rathbun 145 Medaeus spinulifer (Rathbun)... 146 Xanthodius stimpsoni (A. Milne Edwards) 146 Hexapanopeus costaricensis Garth 146 Hexapanopeus nicaraguensis (Rathbun) 147 Hexapanopeus orcutti Rathbun. . . 147 Hexapanopeus sinaloensis Rathbun 147 Hexapanopeus beebei, new species 148 Panopeus purpureus Lockington. . 149 Panop eus bermudensis Benedict & Rathbun 149 Eurytium tristani Rathbun Eurytium tristani minor (Bott), new combination 149 Micropanope polita Rathbun .... 150 Micropanope xantusii (Stimpson) 150 Micropanope (?) maculatus (Rathbun) 151 Paraxanthias taylori (Stimpson).. 151 Menippe obtusa Stimpson 151 Pilumnus pygmaeus Boone 151 Pilumnus limosus Smith 151 Pilumnus stimpsonii Miers 152 Heteractaea peterseni Garth 152 Quadrella nitida Smith 152 133 134 Zoologica: New York Zoological Society [46: 13 Family Goneplacidae Pseudorhombila xanthiformis Garth 154 Euryplax polita Smith 154 Chasmophora macrophthalma (Rathbun) 154 Speocarcinus granulimanus Rathbun 154 Speocarcinus californiensis (Lockington) 155 Speocarcinus ostrearicola Rathbun 155 Chasmocarcinus latipes Rathbun . . 155 Hexapus williamsi Glassell 156 Family Cymopoliidae Cymopolia lucasii (Rathbun) .... 156 Literature Cited 156 Introduction THE brachygnathous crabs of the families Portunidae, Xanthidae, Goneplacidae and Cymopoliidae constitute the subject matter of the second part of this paper, the Maj- idae and Parthenopidae having been treated in part one. The Pinnotheridae are not included, since they present problems in identification not encountered in the other families, which for the area covered are much better known. The general statements made in the introduc- tion to part one apply equally to part two. Inso- far as these were restricted to the Oxyrhyncha, however, they need supplementation for the sARENA bank ^GORDA BANKS SAN LUCAS B. 'V MAZATLAN BANDERAS BA CHAMELA B ■V, V CLARION ISL. TENACATITA B. MANZANILLO- SIHUATANEJO ACAPULCO DULCE R y PORT ANGELES-’* PORT GUATULCO SANTA CRUZ B. TANGOLA-TANGOLA B. GULF OF FONSECA < CONCHAGUITA ISL .. LA UNION, MEANGUERA ISL .. MONYPENNY PT, POTOSI R.. FARALLONE ISL . ) GULF OF (ALCATRAZ ISL , CEORO ISL., . NEGRITOS ISL . CORINTO SAN JUAN DEL SUR PORT PARKER — £ MURCIELAGO B. POTRERO GRANDE E PORT CULEBRA BRAX I LI TO B. PIEDRA BLANCA B NICOYA BALLENAS B JASPER ISL, PUNTARENAS } \ GULF OF DULCE (AOLyro), EASTERN PACI FIC E X P E D 1 T 1 0 N S N E V/ YORK ZOOLOGICAL SOCIETY S H 0 R E COLLECT! NG STATIONS PARI BA / v-BAHIA ISL- / HONDA 1 COIBA ISL. GORGONA ISL. • & GA L APAGOS IS. ECUADOR PERU Text-fig. 1. Shore collecting stations of the Eastern Pacific Expeditions of the New York Zoological Society. For exact locations of associated dredge stations, refer to Zoologica, vol. XXII, no. 2, and vol. XXIII, no. 14. 1961] Garth: N on-intertidal Brachygnathous Crabs 135 Brachyrhyncha. Of this group the cancroid or cyclometopous crabs were the subject of mon- ographic treatment as recently as Rathbun (1930), while the grapsoid or catometopous crabs have received no comprehensive survey since Rathbun (1918). It is to be expected, therefore, that the largest number of new records will be found among the latter group. Ecological Considerations As in the earlier part of this report, the field notes on color, behavior and food habits pro- vided by Miss Jocelyn Crane have been utilized freely and fully to supplement the routinely sys- tematic portions of this paper. Of the 15 species of Portunidae, color in life is recorded for all but a few; these perhaps were not recognized as different in the field. Color notes on populations of the same species from widely separated locali- ties are included in order to establish a basis for a future consideration of the geographical variation of color and pattern. Notes on behavior are given for Portunus (Achelous) tuberculatus, Callinectes arcuatus and Cronius ruber. The food habits of Arenaeus mexicanus are dis- cussed, and the use of Euphylax dovii and E. robustus as food for the fish, Caranx caninus, is mentioned. Among the Xanthidae, color in life is recorded for but 9 of the 21 species, and usually from but a single locality. Color notes for the Goneplacidae and Cymopoliidae are not available. The species that are duplicated from Crane’s (1947) report on the intertidal forms are few in number. All come from her vertical zones 3 to 7, inclusive. Common to zone 3 (stones near low tide level) are Xanthodius stimpsoni and Pilumnus pygmaeus. Common to zone 4 (tide- pools) are the above two plus Menippe obtusa. Common to zone 5 (PociUopora coral) are Me- daeus spinulifer, Xanthodius stimpsoni and Mi- cropanope xantusii. Common to zones 6 and 7 (mangroves and mudflats, respectively) are Panopeus purpureus and Eurytium tristani. The fact that the PociUopora coral zone is more properly subtidal or adtidal than intertidal, and that mangrove and mudflat biotopes continue below low-tide level, was discussed in the previ- ous section of this report (Garth, 1959). Fur- thermore, since the specimens of Menippe ob- tusa, Panopeus purpureus and Eurytium tristani here reported carry no indication of depth, while Micropanope xantusii and Pilumnus pygmaeus were taken once each in a dead pearl oyster, also without indication of depth, it is possible that all five species were collected intertidally, but that lacking also evidence assigning them to a particular intertidal zone or habitat, they were set aside for later treatment with the non-inter- tidal material. Certainly, they form a marginal group when considered in this category. Geographical Considerations The present collection adds appreciably to our knowledge of distribution within the Panamic faunal province, broadly defined as extending from Lower California and the Gulf of Cali- fornia to southern Ecuador or northern Peru. (See Text-fig. 1). Of species heretofore known to occur in the Gulf of California, the following may be report- ed as having ranges extended southward along the mainland, those marked with an asterisk hav- ing been reported previously from the Galapagos Islands (Garth, 1946) as well: Euryplax polita to southern Mexico, * Micropanope (?) macula- tus, Speocarcinus granulimanus and S. cali- forniensis to Costa Rica, Portunus (Achelous) iridescens and *Micropanope polita to Panama. Of species known to occur in Mexico south of Cape Corrientes, Hexapanopeus orcutti and Pi- lumnus stimpsonii may be reported southward to Costa Rica, while of species known to occur in Costa Rica, Hexapanopeus nicaraguensis and H. costaricensis may be reported southward to El Salvador and Panama, respectively. Of species known to occur in Peru, Speocar- cinus ostrearicola may be reported northward to Nicaragua; known to occur in Colombia, Pseudorhombila xanthiformis may be reported northward to Costa Rica; known from Panama, Pilumnus limosus and Chasmophora macroph- thalma may be reported northward to west Mexico; while Menippe obtusa and Pilumnus pygmaeus, known from Nicaragua and Costa Rica, respectively, may be reported northward to southern Mexico. (See also Table I). Three species, *Medaeus spinulifer, *Micro- panope polita and Cymopolia lucasii, are re- ported for the first time from Clarion Island, Mexico. An important Panama record for Het- eractaea peterseni links the previous records from Colombia and the Gulf of California for that species. The southern record of Arenaeus mexicanus of Callao, Peru, is confirmed. Systematic Considerations The 45 species of Brachyrhyncha compare with the 44 species of Oxyrhyncha previously reported in part one of this paper. The number of species by families is as follows: Portunidae, 15; Xanthidae, 21; Goneplacidae, 8; Cymopoli- idae, 1. While the total number of species for the eastern Pacific is not as readily available for the Brachyrhyncha as for the recently mono- graphed Oxyrhyncha (Garth, 1958), it may be stated that of the Portunidae, largely a tropical 136 Zoologica: New York Zoological Society [46: 13 Table I. Extensions of Range | From To Xanthidae Medaeus spinulifer Mainland Clarion I. Hexapanopeus costaricensis Costa Rica Panama Hexapanopeus nicaraguensis Costa Rica El Salvador Hexapanopeus orcutti NW Mexico Costa Rica Micropanope polita Gulf of California, Galapagos Clarion I., Panama Micropanope ( ?)maculatus Gulf of California, Galapagos Costa Rica Menippe obtusa Nicaragua Mexico Pilumnus pygmaeus Costa Rica Mexico Pilumnus limosus Panama Mexico Pilumnus stimpsonii Mexico Costa Rica Goneplacidae Pseudorhombila xanthiformis Colombia Costa Rica Euryplax polita Gulf of California S Mexico Chasmophora macrophthalma Panama Mexico Speocarcinus granulimanus Gulf of California Costa Rica Speocarcinus calif orniensis Gulf of California Costa Rica Speocarcinus ostrearicola Peru Nicaragua Cymopoliidae Cymopolia lucasii Gulf of California, Galapagos Clarion I. family and exclusively estuarine or pelagic, the “Zaca” obtained a complete representation for the territory covered, lacking only the few en- demics from Chile-Peru, the Galapagos, and the Gulf of California (the latter obtained by the “Zaca” in 1936 and reported by Crane in 1937) to complete the list of species known from the entire eastern Pacific. Among the Xanthidae, a new species of Hexapanopeus is described from Cor into, Nica- ragua, and the megalops of Quadrella nitida is described and figured for the first time. The depth range is increased for several species, among them Pilumnus stimpsonii and P. limo- sus. Panopeus convexus Bott (not A. Milne Ed- wards) is considered a synonym of Eurytium tristani Rathbun, the subspecies minor Bott (1955) as belonging to that species also, hence a new combination. It is among the Goneplacidae, however, that the greatest number of “firsts” has been estab- lished. The first specimens since the types may be reported for Pseudorhombila xanthiformis Garth, Speocarcinus ostrearicola Rathbun and Hexapus williamsi Glassell. Moreover, each was known from but a single specimen, and the op- posite sex of each is now made known, i.e., the male of Pseudorhombila, the female of the other two. Restriction of Synonymies In keeping with the format established in the earlier section of this report, synonymies are restricted to the original description, the first use of the name in its current combination, and the citation placing it in the territory covered, if not included in the above two. Reference is also made to the appropriate monograph of Rathbun, either the cancroid (1930) or the grapsoid (1918) volume, and to all reported occurrences of the species in the eastern trop- ical Pacific since then. Measurements To the total length of the largest and smallest specimen examined in each class, male, female and ovigerous female, as given for the Oxyrhyn- cha, a second measurement, that of total width including spines, if any, has been added. In the Portunidae, where length of lateral spine is fre- quently a diagnostic character, a third measure- ment, width excluding spines, is given. Thus the figures 14.5 X 29.5 (21.5) imply length and breadth with (and without) lateral spines. The figures 29.5 — 21.5 divided, by 2 will give the length of the lateral spine, in this case, 4 mm. Acknowledgment In addition to those to whom gratitude was expressed in the earlier part of this study, the writer wishes to thank Dr. Jens W. Knudsen, Pacific Lutheran College, Tacoma, Washington, an authority on larval development of the Xan- thidae, for the illustration of the magalops of Quadrella nitida that appears as Text-fig. 2. 1961] Garth: N on-intertidal Brachygnathous Crabs 137 Systematic Discussion Tribe Brachyura Subtribe Brachygnatha Superfamily Brachyrhyncha Family Portunidae Portunus (Portunus) xantusii (Stimpson) Achelous xantusii Stimpson, 1860, p. 222. Portunus ( Portunus ) xantusii, Rathbun, 1923, p. 620 (part); 1930, p. 50, pi. 18. Glassell, 1935, p. 105. Not Portunus ( Portunus ) xantusi, Boone, 1930, p. 163, pi. 56, figs. A, B. Range. — From Santa Barbara, California (Glassell), to Cape San Lucas, Lower Califor- nia; Gulf of California at Agua Verde and Con- cepcion bays. (Rathbun, 1930). Material Examined. — San Benito Islands, west coast of Lower California, Mexico, No- vember 9, 1937, Station 178, L-l (night light), 8 males, 2 females. Measurements— Male specimen, 15.6 X 31.2 (22.6) mm., female specimen, 14.5 X 29.5 (21.5) mm. Habitat— Pelagic. Color in Life— Exceedingly variable: dullest specimen grayish speckled with black and white, pepper-and-salt fashion. Tips of legs, including chelae, pinkish; ambulatories banded white and brown. Brightest specimen pinkish pepper-and- salt with entire gastric region raspberry red. Others pinkish pepper-and-salt with anterolat- eral margin and that of front ringed, well inside spines, with black. One with carapace and chelae mottled brown, on the carapace a Y of brown, the prongs extending from the inner margins of the eyes to gastric region, the base along mid- line through cardiac and intestinal regions to posterior edge of carapace. Gastric and frontal regions between forks of Y rosy pink. Chelae with pinkish chestnut band across base and an- other across tips. Ambulatories overcast with pink above and below. (J. Crane, field notes). Remarks.— A megalops possibly of this species was taken at the same time as the adults; it is translucent with large black spots. Portunus (Portunus) acuminatus (Stimpson) Achelous acuminatus Stimpson, 1871, p. 112. Portunus ( Portunus ) acuminatus, Garth, 1940, p. 73, pi. 19, figs. 1-3; 1948, p. 33. Not Rathbun, 1930, p. 56, pi. 19. Range— From Isabel Island, Mexico, to La Libertad, Ecuador. 2-50 fathoms. (Garth, 1948). Material Examined— 88 specimens from 12 stations: Mexico Manzanillo, November 22, 1937, Station 184, D-2, 30 fathoms, 29 males, 26 females (8 ovig- erous). 17 mi. SE X E of Acapulco, November 29, 1937, Station 189, D-l, 20 fathoms, 1 male. 4 mi. SSW of Maldonado Point, November 30, 1937, Station 192, D-l, D-2, 26-33 fathoms, 2 males. Port Guatulco, Station 195, December 4, 1937, D-2, 3 fathoms, 2 young; December 6, 1937, D-ll, 5 fathoms, 1 young; D-12, 6 fath- oms, 1 young; December 7, 1937, D-19, 17 fathoms, 2 young. Tangola-Tangola Bay, Station 196, Decem- ber 9, 1937, D-6, D-7, 7-6 fathoms, 2 young; December 12, 1937, D-14, D-15, 5 fathoms, 3 young; December 13, 1937, D-16, 16 fathoms, 2 young. Nicaragua Corinto, January 7, 1938, Station 200, D-27 to D-30, 3 fathoms, 2 young males. Costa Rica Port Parker, Station 203, January 20, 1938, D-l to D-3, 10-15 fathoms, 2 males, 2 females ( 1 ovigerous) ; January 22, 1938, D-ll, 2-4 fath- oms, 1 male. Port Culebra, January 30, 1938, Station 206, D-2, 14 fathoms, 1 female, 1 young. Cedro Island, Gulf of Nicoya, February 13, 1938, Station 213, D-l to D-10, 4-10 fathoms, 1 young. 14 mi. S X E of Judas Point, March 1, 1938, Station 214, D-2, D-3, 43-50 fathoms, 1 male. Golfito, Gulf of Dulce, March 9, 1938, Sta- tion 218, D-4, D-5, 6 fathoms, 1 young male. Panama Bahia Honda, March 18, 1938, Station 222, D-l to D-3, D-5, 3-11 fathoms, 3 young. Measurements. — Males from 9.3 X 18.5 (14.3) to 16.0 X 41.9 (27.2) mm., females from 8.0 X 16.4 (12.1) to 16.2 X 37.7 (26.8) mm., ovigerous females from 8.0 X 16.4 (12.1) to 15.3 X 36.2 (26.0) mm., young from 5.0 X 9.0 (7.3) mm. Habitat— Shelly mud, shelly sand; gravelly mud, gravelly sand; sandy mud; crushed shell; mangrove leaves; rock; dead coral. Color in Life.— Of an Acapulco, Mexico, spec- imen: Chestnut mottled with darker. Of Man- zanillo, Mexico, specimens: “Plain” and “or- ange branchialed”; eggs raspberry. Gastric spot present or absent as in Portunus (Achelous) af- finis. (J. Crane, field notes.) 138 Zoologica: New York Zoological Society [46: 13 Remarks.— The identification of this and the following two species of Portunus (Portunus) occurring widely throughout the Panamic Prov- ince has been facilitated by a prior study (Garth, 1940, p. 73) based on Hancock Expedition ma- terial in which the true P. (P.) acuminatus (Stimpson) was recognized and a neotype estab- lished. The result was to accord equal and full specific rank to acuminatus (Stimpson), pana- mensis (Stimpson), and asper (A. Milne Ed- wards) [= transversus (Stimpson) ], rather than to consider them members of the so-called "acuminatus-asper-panamensis group” (Rath- bun, 1930, p. 53). The acuminate lateral spine and the slender, almost filiform chelae are diag- nostic, now that ample material is available to show these distinctive features. Portunus (Portunus) asper (A. Milne Edwards) Neptunus asper A. Milne Edwards, 1861, p. 325, pi. 30, figs. 3-3c. Portunus ( Portunus ) asper, Rathbun, 1930, p. 56, pi. 20, figs. 2, 3, pi. 21, pi. 22, figs. 1, 2. Garth, 1948, p. 33; 1957, p. 36, synonymy. Range. — From Mazatlan, Mexico, to Chile. To 16 fathoms. (Garth, 1957). Material Examined— 33 specimens from 10 stations: Mexico 17 mi. SE X E of Acapulco, November 29, 1937, Station 189, D-l to D-3, 13-20 fathoms, 2 males, 1 ovigerous female. Mouth of Dulce River, November 30, 1937, Station 191, D-l, 8 fathoms, 1 male, 1 young. Port Guatulco, December 6, 1937, Station 195, D-ll, 5 fathoms, 1 young; D-12, 6 fath- oms, 1 male. Tangola-Tangola Bay, Station 196, Decem- ber 9, 1937, D-l, D-2, D-5, 5-9 fathoms, 4 young; December 13, 1937, D-16, 16 fathoms, 2 males, 1 female, 1 young. El Salvador Meanguera Island, Gulf of Fonseca, Decem- ber 23, 1937, Station 199, D-l, 16 fathoms, 1 male. Costa Rica Murcielago Bay, January 23, 1938, Station 204, D-l, D-2, D-4, 2-4 fathoms, 2 young. Port Parker, January 30, 1938, Station ?, depth ?, 1 young male. Piedra Blanca Bay, Station 208, February 1, 1938, L-l, surface at light, 1 male, 1 ovigerous female, 7 young; February 5, 1938, D-l to D-3, D-6, D-7, D-9, 3-6 fathoms, 1 male, 2 young. Cedro Island, Gulf of Nicoya, February 21, 1938, Station 213, L-l, surface at light, 1 male. Golfito, Gulf of Dulce, March 7, 1938, Sta- tion ?, depth ?, 1 female without chelipeds. Measurements. — Males from 10.2 X 22.0 (15.9) to 41.9 X 96.5 (69.3) mm., females from 14.7 X 32.8 (23.7) to 39.6 X 85 (65.2) mm., ovigerous females from 38.1 X 85.5 (61.7) mm., young from 5.2 X 9.3 (7.6) mm. Habitat. — Sand, mud, sandy mud, gravelly sand, crushed shell, rocks, algae. Color in Life.—Oi Piedra Blanca, Costa Rica, specimens: 39 mm. ovigerous female and 20 mm. male olive buff above except for white- tipped spines. Carpus, manus and dactyls of ambulatories (but not of swimming legs) lilac. Swimming legs with posterior half of paddle only lilac. Middle of fixed finger with a band of brick red; tips of both dactyls white. Underparts white. Eggs bright orange. 1 1 mm. young lack the violet and are grayer, not buffy, with suggestions of red bar across fixed finger. (J. Crane, field notes). Of Gulf of Fonseca, El Salvador, male: Cara- pace and all legs pale olive brown; a white spot on posterior lateral margin. Movable dactyl dull violet basally, distal part and fixed finger white; dactyls of ambulatories rose red, tips white, dis- tal half of swimmerets violet red. (J. Crane, field notes). Of Acapulco, Mexico, specimens: Carapace and chelipeds pale olive brown, ridges darker. Ambulatories and cheliped wine colored. Manus and dactyls of swimmerets white with two longi- tudinal stripes of dark brown or pale buff. One patch faintly visible on posterolateral region of large male; below this, on margin, a white spot. (J. Crane, field notes). Underside white. Remarks. — The broad anterolateral arc and teeth that show little reduction are characters useful in separating this species from both Por- tunus (P.) acuminatus and P. (P.) panamensis, while the heavier cheliped will serve to separate this species from P. (P.) acuminatus even in the young, where the relative lengths of the lateral spines might not suffice. Portunus (Portunus) panamensis (Stimpson) Achelous panamensis Stimpson, 1871, p. 112. Portunus ( Portunus ) panamensis, Rathbun, 1910, pp. 577, 610; 1930, p. 58, pi. 20, fig. 1, pi. 22. fig. 3, pis. 23, 24. Finnegan, 1931, p. 626, text-fig. 5. Garth, 1948, p. 34. Range— From Panama Bay to Bay of Sechura, Peru (from Angeles and Mulege Bays, Gulf of California, Mexico, only if Rathbun’s synonymy of Amphitrite paucispinis Lockington be accept- ed). To 33 fathoms. (Garth, 1948). Material Examined.— 115 specimens from 6 stations: 1961] Garth: N on-intertidal Brachygnathous Crabs 139 Nicaragua Corinto, January 5, 1938, Station 200, D-12 to D-19, 1-13 fathoms, 2 young. Costa Rica Port Parker, Station 203, January 20, 1938, D-l to D-3, 10-15 fathoms, 29 males, 24 fe- males (10 ovigerous), 1 young; January 22, 1938, B-4, 7 fathoms, 3 males, 3 females; D-6, 1 fathom, 1 young; D-7, 9-5 fathoms, 1 female; D-8, 9 fathoms, 1 female; D-9, 1.5-4 fathoms, 1 male, 1 female; D-ll, 2-4 fathoms, 1 male, 1 ovigerous female; D-12, 2 fathoms, 1 male; D-15, 9-2 fathoms, 2 males, 2 females, 1 young. ?Golfito, Gulf of Dulce, March 9, 1938, Sta- tion 218, D-8, 6 fathoms, 1 male, 1 young. Panama Bahia Honda, March 18, 1938, Station 222, D-l to D-3, D-5, 3 fathoms, 2 males, 3 females, 2 young. Colombia At sea near Gorgona Island, March 27, 1938, from mangrove seeds floating in tide rip, 5 young, questionably of this species. Gorgona Island, March 31, 1938, Station 232, D-l, 2-8 fathoms, 1 male, 25 young. Measurements— Males from 4.4 X 6.7 (6.0) to 11.2 X 22.0 (17.6) mm., females from 4.5 X 7.3 (6.2) to 11.0 X 21.5 (16.8) mm., ovig- erous females from 5.5 to 9.1 (7.8) to 8.6 X 15.6 (12.7) mm., young from 3.0 X 5.0 mm. All but the largest male came from the first Port Parker series. Habitat— Sandy and shelly mud, shelly sand, crushed shell, gravel, rocks, coral, algae, and mangrove leaves. (These all from Station 203). Remarks— A small species, as shown by the size of the ovigerous females, which would be in the size range of young in either the Portunus (Portunus) asper or P. (P.) acuminatus series. The young taken at sea off Gorgona Island, the largest of which is only 3.5 mm. in length by 5.6 mm. in width, may be of two species. The two larger specimens show the alternation of large and small anterolateral teeth expected in P. (P.) panamensis; the three smaller specimens have anterolateral teeth of equal size, as in P. (P.) asper. The species finds its optimum conditions in the shallow bays of Costa Rica and Panama, judging from the tremendous breeding popula- tion found by the “Zaca” at Port Parker, and oc- curs sparingly to the north and south. Portunus ( Achehus ) brevimanus (Faxon) Achelous spinimanus, Faxon, 1895, p. 23. Not Portunus spinimanus Latreille. Achelous brevimanus Faxon, 1895, p. 23. Portunus ( Achelous ) brevimanus, Rathbun, 1898, p. 593 (part: not the Galapagos specimens); 1930, p. 68, pis. 29, 30. IPortunus ( Achelous ) spinimanus, Finnegan, 1931, p. 628. Not Portunus spinimanus Latreille. Range— Revilla Gigedo Islands, Mexico, and Cocos Island, Costa Rica. (Rathbun, 1930). Material Examined. — 2 specimens from 2 Revilla Gigedo Islands stations: Sulphur Bay, Clarion Island, May 11, 1936, from night light, 1 male. 3 mi. off Pyramid Rock, Clarion Island, May 12, 1936, Station 136, D-2, 55 fathoms, 1 female. Measurements. — Male 8.4 X 14.3 (11.3) mm., female 17.6 X 28.0 (23.8) mm. Habitat— Not given. Remarks —Aside from the suggestion by Glas- sell (1934, p. 454) that specimens from Perlas Islands, Panama, and Puntarenas, Costa Rica, attributed to Portunus (Portunus) xantusii (Stimpson) by Boone (1930, p. 163, pi. 56, figs. A, B), might instead represent Faxon’s species, a suggestion questioned by this writer, P. (Achel- ous) brevimanus has not been reported from the Central or South American mainland, unless specimens from Gorgona Island attributed to the Atlantic P. ( Achelous ) spinimanus Latreille by Finnegan (1930) be of this species. Specimens from the Galapagos Islands earlier attributed to P. (A.) brevimanus by Rathbun (1898) were subsequently described by her (1902) as P. (A.) stanfordi. Portunus (Achelous) pkhilinquei Rathbun Portunus ( Achelous ) pichilinquei Rathbun, 1930, p. 78, pi. 37. Crane, 1937, p. 67. Range— From Magdalena Bay, west coast of Lower California, and Cape Tepoca, Gulf of California, to Cape San Lucas. 0.5 to 33 fath- oms. (Crane, 1937). Material Examined.— San Lucas Bay, Lower California, Mexico, November 13, 1937, Sta- tion 135, D-27, 2-6 fathoms, 1 young male. Measurements.— Young male, 6.0 X 9.8 (8.2) mm. Habitat.— Sand bottom. Color in Life.— Mottled olive and grayish and black. Legs grayish banded with black. Under- side pure white. (J. Crane, field notes). Remarks— The single specimen was taken in the same dredge haul with Arenaeus mexicanus. Portunus (Achelous) affinis (Faxon) Achelous affinis Faxon, 1893, p. 155 (part: not the Guaymas, Mexico, specimens); 1895, p. 23, 140 Zoologica: New York Zoological Society [46: 13 Portunus ( Achelous ) affinis, Rathbun, 1898, p. 595; 1930, p. 80, pis. 38, 39. Portunus affinis, Coventry, 1944, p. 538. Range.— From Cape San Lucas, Lower Cali- fornia, Mexico, to Ecuador. (Rathbun, 1930). Material Examined. — 59 specimens from 5 stations: Mexico Tenacatita Bay, November 21, 1937, Station 183, D-2, D-3, 30-40 fathoms, 4 males, 6 fe- males. Manzanillo, November 22, 1937, Station 184, D-2, 30 fathoms, 7 males, 7 females (5 ovig- erous) . Port Guatulco, Station 195, December 3-5, 1937, light, 1 male; December 6, 1937, D-13, 8 fathoms, 2 young; D-16, 10 fathoms, 1 male, 7 young; December 7, 1937, D-17, 6 fathoms, 1 young; D-19, 17 fathoms, 1 young male; D-20, 23 fathoms, 2 males; D-21, 18 fathoms, 1 young male. Tangola-Tangola Bay, Station 196, Decem- ber 9, 1937, D-l, D-2, D-5, 5 fathoms, 3 young; D-6, D-7, 7-6 fathoms, 1 young; December 13, 1937, D-16, 16 fathoms, 1 young; D-17, 23 fathoms, 13 young. Colombia Gorgona Island, March 31, 1938, Station 232, D-l, 2-8 fathoms, 1 young. Measurements. — Males from 8.0 X 12.5 (11.2) to 25.4 X 44.6 (38.4) mm., females from 14.8 X 25.1 (22.0) to 21.1 X 36.8 (30.9) mm., ovigerous females from 16.0 X 27.3 (24.0) to 21.1 X 36.8 (30.9) mm., young from 4.0 X 5.9 mm. Habitat. — Sand, mud, sandy mud; gravelly sand, gravelly mud; crushed shell. Color in Life.—Ot Tenacatita Bay specimens: Apricot buff, striations brown, chelipeds and ambulatories, especially inner sides of manus and dactylus, streaked with violet. A constant white spot in middle of posterolateral margin. Of Manzanillo specimens : White spot on gas- tric region and spot above base of swimming legs on abdomen may be present, absent, or faint. Posterolateral spot, however, constant. Eggs raspberry. ( J. Crane, field notes) . Remarks.— This species is at once separated from the Portunus (Portunus) species of the Panamic Province with which it ranges coex- tensively by its short lateral spine and spinulous merus of the fourth ambulatory (or natatory) leg. In the latter respect it resembles P. (P.) xantusii of southern California — west coast of Lower California. Portunus (Achelous) tubereulatus (Stimpson) Achelous tubereulatus Stimpson, 1860, p. 223. Portunus ( Achelous ) tubereulatus, Rathbun, 1898, p. 596; 1930, p. 90, pi. 44. Finnegan, 1931, p. 629. Crane, 1937, p. 68. Garth, 1946, p. 421, pi. 71, fig. 2; 1948, p. 34. Range.— From Cape San Lucas, Lower Cali- fornia, Mexico, to off Ecuador. Galapagos Is- lands. 3-70 fathoms. (Garth, 1948). Material Examined— 200 specimens from 6 stations: Mexico Chamela Bay, November 17, 1937, Station 182, D-4, 16 fathoms, 1 female. Port Guatulco, Station 195, December 4, 1937, D-3, 3.5 fathoms, 1 female; December 5, 1937, D-8, D-9, 6-7 fathoms, 4 males, 6 fe- males; December 7, 1937, D-16, 10 fathoms, 2 males, 11 young; D-17, 6 fathoms, 1 male, 2 fe- males, 4 young; D-18, 6 fathoms, 8 males, 6 females. Tangola-Tangola Bay, Station 196, December 9, 1937, D-l, D-2, D-5, 5-9 fathoms, 1 female, 3 young; D-6, D-7, 6-7 fathoms, 2 males, 2 fe- males (1 ovigerous), 26 young; D-8, 9 fathoms, 4 males, 2 females, 43 young; December 12, 1937, D-9 to D-12, 4.5 to 7.5 fathoms, 3 young; D-13, 10 fathoms, 1 male, 4 young; D-14, D-15, 5 fathoms, 3 males, 3 females (2 ovigerous), 27 young; December 13, 1937, D-16, 16 fathoms, 1 young. Costa Rica Port Parker, Arriba rocks, January 16-17, 1938, 2 males; Station 203, January 22, 1938, D-ll, 2-4 fathoms, 1 female. Piedra Blanca Bay, February 5, 1938, D-l to D-3, D-6, D-7, D-9, 3-6 fathoms, 1 male, 1 young. Colombia Gorgona Island, March 31, 1938, Station 232, D-l, 2-8 fathoms, 1 male, 1 female, 22 young. Measurements. — Males from 6.0 X 11.6 (8.6) to 11.3 X 26.4 (18.3) mm., females from 6.9 X 14.4 (10.5) to 10.2 X 22.9 (16.0) mm., ovigerous females from 7.8 X 15.1 (11.2) to 9.0 X 18.5 (13.7) mm., young from 4.6 X 7.3 (6.3) mm. Habitat. — Predominantly sand, often with crushed shell or algae; occasionally rock; rarely mud. Color in Life. — Of Chamela Bay, Mexico, female: General color light brown marbled with black; base of lateral spines tinged with crimson. Legs barred cream and brown. Underparts white (J. Crane, field notes). 1961] Garth: N on-intertidal Brachygnathous Crabs 141 Behavior. — Kept alive in aquarium in 2 Vi inches of water on native sand (coarse sandy bottom with tiny shells). Much more responsive to light than Cycloes and more nervous and active; continually changing position, burrowing quickly in sand, hind end first. Digs with am- bulatories, kicking sand out forward to chelipeds. Sinks all of self except front and eyes. (J. Crane, field notes). Remarks. — Of small size but distinctively or- namented with tubercles, and bearing a spine at the posterolateral angles of the carapace, as well as a long, straight lateral spine, Portunus ( Ache - lous) tuberculatus is easily segregated from the several other species of Portunus with which it customarily occurs, often in the same dredge hauls. Portunus (Achelous) iridescens (Rathbun) Neptunus ( Hellenus ) iridescens Rathbun, 1893, p. 240. Portunus ( Achelous ) iridescens, Rathbun, 1930, p. 93, pi. 46. Crane, 1937, p. 66. Portunus ( Achelous ) spinicarpus, Finnegan, 1931, p. 628. Not Achelous spinicarpus Stimpson. Range— From off Santa Margarita Island, west coast of Lower California, and from off Diggs Point to off La Paz Bay, Gulf of Califor- nia, Mexico. 18-112 fathoms. (Rathbun, 1930). Gorgona Island. (Finnegan). Material Examined.— 31 specimens from 2 stations: Costa Rica 14 mi. S X E of Judas Point, March 1, 1938, Station 214, D-2, D-3, 50 fathoms, 13 males, 13 females (6 ovigerous). Panama Gulf of Chiriqui, March 13, 1938, Station 221, D-3 to D-5, 35-40 fathoms, 2 males, 3 females. Measurements.— Males from 14.4 X 29.3 (21.5) to 24.7 X 48.0 (36.5) mm., females from 16.2 X 31.0 (24.9) to 25.3 X 48.5 (37.4), ovigerous females from 17.0 X 36.3 (25.6) to 21.1 X 42.7 (31.3) mm. Habitat.— Mud; sandy mud. Color in Life— Not recorded. Remarks— Distinguished from all other Paci- fic Portunidae by the long inner carpal spine, a character shared with the Atlantic Portunus {Achelous) spinicarpus (Stimpson). The Costa Rican and Panamanian localities above would represent the first records for the species from the Bay of Panama, were it not for the previous report of the “St. George” from Gorgona Island, Finnegan (1931) attributing it to the Atlantic species. Callinectes arcuatus Qrdway Callinectes arcuatus Ordway, 1863, p. 578. Rathbun, 1930, p. 121, pi. 52. Garth, 1948, p. 35; 1957, p. 36, synonymy. Holthuis, 1954, p. 27. Bott, 1955, p. 56. Range— From Anaheim Slough, California, to ? south Chile. (Garth, 1957). Material Examined.— 100 specimens from 13 stations : Mexico Chamela Bay, North Lagoon, November 17, 1937, 1 male, 3 females (2 ovigerous). Acapulco Beach, November 26-28, 1937, 1 male. Honduras Cutuco and Potosi Lights, Gulf of Fonseca, December 20, 1937, 4 males, 6 females, 3 young. El Salvador La Union, Gulf of Fonseca, December 27, 1937, Station 199, D-8 to D-16, 5-6 fathoms, 1 female. Nicaragua Corinto, December 29, 1937, Station 200, D-7, 2 fathoms, 1 young; January 7, 1938, D-20 to D-26, 1. 5-6.5 fathoms, 2 young. Castenones Lagoon and mid-harbor, January 6, 1938, 4 males, 4 females. Costa Rica Port Parker, January 13, 1938, shore, 1 male, 3 females (1 ovigerous), 4 young; January 22, 1938, Station 203, D-9, 1.5-4 fathoms, coral, 1 young. Port Culebra, January 26, 1938, 8 males, 3 females, 2 young; January 30, 1938, 1 young male. Piedra Blanca, February 6, 1938, 2 males, 2 females. Cedro Island, Gulf of Nicoya, February 12, 1938, 1 male. Golfito, Gulf of Dulce, March 5, 1938, 1 male, 1 female, 4 young; March 9, 1938, Station 218, D-4, D-5, 6 fathoms, 3 young; D-8, 6 fathoms, 2 young; same date, mudflats, 1 young. Panama Bahia Honda, March 16, 1938, 3 males, poor condition; March 19, 1938, tidepool, 1 male, lacks chelipeds. Bella Vista, Panama City, 1944, 5 males. Canal Zone Balboa, April, 1938, night light, 3 males, 2 females, 8 young; 1940, 3 males, 5 young. 142 Zoologica: New York Zoological Society [46: 13 Ecuador Puerto Bolivar, April, 1944, 1 male. Measurements.— Males from 17.0 X 39.6 (29.2) to 42.8 X 98 (75.2) mm., females from 16.0 X 35.2 (26.4) to 51.8 X 102.8 (85.7), ovigerous females from 24.7 X 52 (42) to 51.8 X 102.8 (85.7) mm., young from 5.7 X 12.9 (10.5) mm. Habitat— Mud, shells, mangrove leaves; fre- quently taken in mudflats or shallow lagoons, at Gulf of Dulce “mostly salt, slightly brackish.” Came often to shipside light at night. One coral station. Color in Life.—Ot Port Parker male: Carapace dull olive gray-green. Chelipeds olive green dor- sally, whitish ventrally, washed with bluish- violet and chelae tipped with pale yellow-brown. Legs turquoise washed with olive; hairs straw gold; swimming legs olive green with suggestion of turquoise, paddles washed with black; hairs straw; tubercles at leg joints golden orange; eyes straw with brownish streaks; underparts pure white. (J. Crane, field notes). Of Chamela Bay females : Carapace in general blue with olive pile: central portion (without pile) blue-violet; anterolateral margins deep purplish-vinaceous. Base of merus of cheliped olive, inner margin of manus blue-violet, rest of cheliped purplish. Spines of cheliped and tips of anterolateral spines white. Chelae varied, fixed finger usually tipped with white; both fingers barred with purple. Ambulatories Italian blue, hairs olive; swimming legs same with tubercles at joints and all margins narrowly violet; swim- ming feet sometimes turquoise green. Abdomen violet, joints white; plastron white; under sides of legs, however, colored like upper sides. (J. Crane, field notes). Behavior.— Of Chamela Bay females: 8 large females [were seen] near the mouth of the lagoon, all with buffy eggs. [Each was] swim- ming singly, at least 25 feet away from the near- est other one. [There were] no males in the vicinity, nor were there [any] small females. ( J. Crane, field notes) . Remarks— In view of the overlapping ranges of this and the following species, it was con- sidered more than likely that some of the young listed above would prove to be Callinectes tox- otes, to which only two large males are referred. Specimens from Acapulco Beach, Mexico, Gol- fito, Costa Rica (Sta. 218, D-8), and Puerto Bolivar, Ecuador, all of which showed blunted or rounded frontal teeth, were compared with specimens of like size in the Hancock collections determined by M. J. Rathbun as C. toxotes. Not only did the “Zaca” young fail to show the nar- row intramedial area, but they proved dissimilar from C. toxotes in other characters as well. It was concluded, therefore, that all young Callin- ectes obtained by the “Zaca” were C. arcuatus. (See also Remarks under the following C. tox- otes) . Callinectes toxotes Ordway Callinectes toxotes Ordway, 1863, p. 576. Rathbun, 1930, p. 127, pi. 54. Garth, 1948, p. 35; 1957, p. 37, synonymy. Holthuis, 1954, p. 27. Bott, 1955, p. 56. Range— From Cape San Lucas, Lower Cali- fornia, Mexico, to mouth of River Tumbes, Peru. Juan Fernandez Island, Chile. (Garth, 1957). Material Examined— 2 specimens from as many stations: Costa Rica Piedra Blanca, February 6, 1938, 1 large male. Golfito, Gulf of Dulce, March 6, 1938, 1 large male. Measurements.— The two males measured 46.5 X 89.2 (76.4) and 52.8 X 101.3 (87.5) mm., respectively. Habitat.— Not stipulated, but presumably as in the preceding species. Color in Life.— Not noted. Remarks— A more granulate species than Cal- linectes arcuatus, C. toxotes is further charac- terized by having the frontal teeth rounded, the middle pair equally advanced with the outer pair in the young. The intramedial area, that portion of the gastric region behind the posterior of the gastric carinae, is as long as its posterior width. These characters are all apparent in young from Costa Rica from among Hancock collections de- termined for the writer prior to 1935 by the late Mary J. Rathbun. The absence of young, and of mature females, from the “Zaca” series would indicate that C. toxotes is much less abundant, and that it may have narrower eco- logical tolerances than C. arcuatus. Arenaeus mexicanus (Gerstaecker) Euctenota mexicana Gerstaecker, 1857, p. 131, pi. 5, figs. 3, 4. Arenaeus mexicanus, Faxon, 1895, p. 22. Rathbun, 1930, p. 137, pi. 58, fig. 1, pi. 61. Garth, 1948, p. 35. Holthuis, 1954, p. 28. Range— From Ballenas Bay, Lower California, and Carmen Island, Gulf of California, Mexico, to Ancon, Peru. (Garth, 1948). Material Examined.— 101 specimens from 18 localities : Mexico San Lucas Bay, Lower California, November 1961] Garth: N on-intertidal Brachygnathous Crabs 143 13, 1937, Station 135, D-27, 2-6 fathoms, 1 male, 3 young. Chamela Bay, lagoon shore, November 17, 1937, 2 males. Passavera Island, November 19, 1937, 1 male, 3 females (1 ovigerous). Acapulco beach, November 26-28, 1937, 1 male. Port Guatulco, December 3-5, Station 195, L-l to L-3 (light), 2 females. Tangola-Tangola Bay, Station 196, December 9, 1937, D-l, D-2, D-5, 5-9 fathoms, 9 young; December 12, 1937, D-9 to D-12, 7.5-4 fathoms, 8 young. Nicaragua Corinto, Station 200, January 5, 1938, D-12 to D-19, 3-13 fathoms, 1 male, 6 young; January 7, 1938, D-20 to D-26, 1.5-6.5 fathoms, 12 young. San Juan del Sur, January 9, 1938, 1 carapace. Costa Rica Potrero Grande Bay, January 20, 1938, 1 male, 2 ovigerous females. Murcielago Bay, January 23, 1938, Station 204, D-l, D-2, D-4, 2-4 fathoms, 9 young. Piedra Blanca Bay, February 5, 1938, Station 208, D-l to D-3, D-6, D-7, D-9, 3-6 fathoms, 1 young; February 6, 1938, 1 male. Cedro Island, Gulf of Nicoya, February 12, 1938, 1 male, 3 females, 3 young. Ballenas Bay, Gulf of Nicoya, [February 25, 1938], 1 male, found dead in mangroves. Uvita Bay, March 3, 1938, seine, 6 males, 2 females. Panama Isla Parida, Gulf of Chiriqui, March 12, 1938, 1 male. Bahia Honda, March 14, 1938, 1 male. Pacheca Island, Pearl Islands, July 4, 1933, tidepools, “Antares,” 1 male. Colombia Gorgona Island, March 28, 1938, 2 males; March 31, 1938, Station 232, D-l, 2-8 fathoms, 7 young. Peru Immediately S of Callao, 1941, 9 males, caught by natives, gift of Mrs. Sherman P. Haight. Measurements. — Males from 8.1 X 19.1 (13.8) to 35.0 X 80.3 (60.1) mm., females from 12.4 X 31.0 (21.0) to 29.0 X 65.5 (47.4) mm., ovigerous females from 17.6 X 39.0 (30.0) to 29.0 X 65.5 (47.4) mm., young from 4.0 X 7.5 (6.0) mm. Habitat — Sand, rarely with mangrove leaves, rocks, or algae. Food.— Four stomachs: amphipods (2), sand, algae, and iridescent, fine nacre shell (very thin, inner layers only apparently) (2). (J. Crane, of 15 to 33 mm. specimens seined at Piedra Blanca) . Color in Life— Of San Lucas Bay specimens: All mottled, gray and black, spotted with white. Underside of legs and carapace, except abdomen, speckled with black. Chelipeds and legs grayish spotted with black. Two conspicuous black spots on carapace, one on each side of mid-gastric region. (J. Crane, field notes). Of Chamela Bay specimens: Olive-tinged pepper-and-salt with bright white spot in middle of posterior gastric region and another on intes- tinal region. (J. Crane, field notes). Of Passavera, Chamela Bay, specimens: Olive spotted finely with white. Eggs bright orange. Distal segments of legs pale gray. (J. Crane, field notes) . Remarks.— The Haight specimens, included here for convenience although not of “Zaca” col- lecting, confirm the southern limit of range for the species, Callao, Peru, being just a few miles south of Ancon, where a single specimen was obtained by R. E. Coker (Rathbun, 1930). Cronins ruber (Lamarck) Portunus ruber Lamarck, 1818, p. 260. Cronius ruber, Stimpson, 1860, p. 225. Rathbun, 1930, p. 139, pis. 62, 63. Finnegan, 1931, p. 630. Garth, 1946, p. 422, pi. 72, figs. 3, 4; 1948, p. 36. Holthuis, 1954, p. 28, text-fig. 10. Range— From Point San Bartolome, Lower California, Mexico, to Paita, Peru. Galapagos Islands. 4-20 fathoms. Occurs also in the Atlan- tic. (Garth, 1948). Material Examined— 48 specimens from 12 stations or localities: Mexico Chamela Bay, November 17, 1937, Station 182, D-l, 8 fathoms, 1 young. Manzanillo, November 22, 1937, Station 184, D-l, 25 fathoms, 1 young. Port Guatulco, Station 195, December 5, 1937, D-5, D-7, 2-4.5 fathoms, 2 males, 1 fe- male, 3 young; D-8, D-9, 6-7 fathoms, 1 female, 7 young; D-8, D-9, 6-7 fathoms, 1 female; December 6, 1937, D-l 4, 4 fathoms, 1 female; December 7, 1937, D-l 8, 6 fathoms, 1 young. Tangola-Tangola Bay, December 9, 1937, Station 196, D-6, D-7, 7-6 fathoms, 1 male, 1 young; December 12, 1937, D-l 4, D-l 5, 5 fathoms, 2 young. 144 Zoologica: New York Zoological Society [46: 13 Honduras Cutuco and Potosi Lights, Gulf of Fonseca, December 20, 1937, 1 male. Gulf of Fonseca, date?, fumarole, 1 young, 1 ovigerous female. Costa Rica Port Parker, Arriba rocks, January 15-18, 1938, 1 male, 1 female. Port Culebra, January 30, 1938, 1 male, broken. Piedra Blanca, February 2, 1938, 1 female; same locality, February 5, 1938, Station 208, D-l, D-2, B-3, B-6, B-7, D-9, 5 fathoms, 6 young. Panama Bahia Honda, Station 222, March 18, 1938, D-l, D-2, D-3, D-5, 3-11 fathoms, 2 young. Colombia Gorgona Island, March 30, 1938, 1 male; from coral, 1 ovigerous female; same date?, 2 males. Gorgonilla Island, April 2, 1938, 1 male, 1 female, 3 young. Measurements. — Males from 10.9 X 16.8 ( 14.9) mm. to 23.9 X 39.3 (35.0) mm., females from 9.0 X 13.0 (11.9) mm., ovigerous females from 15.3 X 24.3 (21.9) mm. to 44.2 X 68.0 (59.7) mm. The largest specimen, a male from Port Culebra, is in damaged condition. It meas- ures aproximately 47 X 72 mm. in length and breadth. Habitat. — Off Mexico and Costa Rica, from sand bottom with algae, rocks, or crushed shell; coral. Off Panama, from mud bottom with rocks, dead coral, shell, and leaves. “Under stone com- pletely out of water and in upper tidal zone; alive and all right.” Color in Life. — Of Chamela Bay young fe- male: Dark brownish-black streaked with gray. Paddle legs chestnut brown. (J. Crane, field notes). Of Piedra Blanca female: Carapace olive brown speckled finely with cream. Transverse ridges blue-black. A prominent oval cream spot on middle of posterolateral margin. Anterolat- eral spines violet tipped with reddish-brown. Chelipeds like carapace, both as to background, ridges, and spines, above, but ridges definitely dark blue or green. Lower (outer) half of cheli- peds creamy white. Chelae purple, tips buffy white, a greenish spot in middle of movable dac- tyls above. Ambulatories mottled dark green and white. Dactyls reddish-brown. Swimming feet pumpkin orange. Underside of carapace and maxillipeds and anterior edge of sternum orange streaked with white. Rest of sternum, abdomen, and under side of merus of ambulatories white. A purple line down middle of abdomen. Carpus to dactylus of ambulatories like upper side. (J. Crane, field notes). Of Gulf of Fonseca ovigerous female: Dark purplish-black; pile dark buff; ridges and cara- pace and legs purplish-red and purplish-blue. Same color on abdominal crests. Swimmerets rusty orange. Chelae dark purple. Under side white with buffy pile, except carpus, manus, and dactylus of legs, which are like upper parts of same. Eggs buffy orange. ( J. Crane, field notes) . Behavior.— Threatens. (J. Crane, field note). Remarks. — The young of this species were frequently included in mixed lots of Portunus species, which they resemble greatly. The alter- nation of large and small anterolateral teeth and narrow carapace even suggests the subgenus Achelous. The presence of four spines on the manus, however, serves at once to distinguish them from all other eastern Pacific Portunidae. Euphylax dovii Stimpson Euphylax dovii Stimpson, 1860, p. 226, pi. 5, figs. 5, 5a. Rathbun, 1930, p. 147, pi. 65. Boone, 1930, p. 190, pi. 65. Garth, 1946, p. 423, pi. 72, figs. 1, 2. Coventry, 1944, p. 539. Euphylax dowi, Garth, 1957, p. 38. Range.— West coast of Mexico? Panama to Talcahuano, Chile. Galapagos Islands. Material Examined. — Identifiable material from 3 stations, as follows: Panama Bahia Honda, March 15, 1938, fragments, food of Caranx caninus Gunther. * Hannibal Bank, March 20, 1938, Station 224, D-2, D-3, 35 fathoms, 3 chelae. 22 mi. ESE of Jicaron Island, March 20, 1938, Station 226, L-l (night light), 1 female. Measurements. — Female specimen, length 23.4 mm., width 37.8 mm. Habitat.— Pelagic. Frequently comes to light at night. Color in Life— Carapace and merus of all legs deep purple; other segments of legs wine red. Underside of carapace, meri of legs, and maxil- lipeds blue; sternum white; abdomen brownish; undersides of rest of ambulatories wine red. (J. Crane, field notes) . Remarks. — The three chelae from Hannibal Bank were taken in a dredge from a bottom of either rocks, mud, and dead coral or sand, shells, *The fish Caranx caninus Gunther is considered by some authors to be a synonym of the Atlantic C. hippos (Linnaeus). 1961] Garth: N on-intertidal Brachygnathous Crabs 145 and algae. A specimen was seen by John Tee- Van swimming at the surface in daylight, on the same day, above Hannibal Bank. (J. Crane, field notes). This corresponds with the experi- ence of the “Velero III,” which encountered the crabs in numbers at Cocos Island, Costa Rica, (Garth, 1946), and of the “Askoy,” which en- countered them at Malpelo Island, Colombia. The observations of Dr. R. C. Murphy are re- corded in Garth (1948, p. 9). Euphylax robustus A. Milne Edwards Euphylax robustus A. Milne Edwards, 1874, p. 249; 1879, p. 205, pi. 37. Rathbun, 1930, p. 148, pis. 66, 67. Coventry, 1944, p. 540. Garth, 1948, p. 37. Range— From Isabel Island, Mexico (Coven- try), to Octavia Bay, Colombia (Garth). Material Examined— 9 specimens from 5 sta- tions : Mexico 17 mi. SE X E of Acapulco, November 29, 1937, Station 189, D-l, 20 fathoms, 1 female. Tangola-Tangola Bay, Station 196, Decem- ber 12, 1937, D-9 to D-12, 7.5-4 fathoms, 1 young; December 13, 1937, D-17, 23 fathoms, 2 young. Costa Rica Port Culebra, January 30, 1938, Station 206, D-3, 14 fathoms, 2 young females. Panama Parida Island, Gulf of Chiriqui, March 12, 1938, 1 damaged specimen, food of Caranx caninus. Bahia Honda, March 18, 1938, Station 222, D-5, 11 fathoms, 2 young. Colombia At sea near Gorgona Island, March 27, 1938, from mangrove seeds floating in tide rip, 1 young. Measurements.— Large female, length 60 mm., width 96 mm., exorbital width 84.2 mm., frontal width 15 mm., cheliped 130 mm., chela 74 mm., dactyl 44.2 mm., height of palm 30 mm., first walking leg 116.5 mm. Young from 3.5 mm. length. Habitat— Sand, mud, sandy mud; with shell, leaves, or algae. Color in Life.— Carapace and legs above gray blue-green. Chelipeds gray blue-green except olive brown manus and dactyls. Ridges and dactyls tinged with pink. Tubercles on chelipeds white. Eyestalks bright violet. Swimmerets pale horn. Abdomen barred with violet and white. Legs barred with violet and white below. Eggs pale salmon. ( J. Crane, field notes) . Remarks.— While credit for the rediscovery of A. Milne Edwards’s lost species rightfully be- longs to the Fifth George Vanderbilt Expedition (Coventry, 1944), “Zaca” scientists may consid- er as one of their more significant contributions the finding of a specimen of the opposite sex and of a size comparable to the 56 X 90 mm. holotype of Euphylax robustus. This 60 X 96 mm. female, dredged on sandy mud bottom near Acapulco, Mexico, by its detailed resemblance to A. Milne Edwards’s unique male, upholds the writer’s con- viction, based upon the examination of imma- ture specimens only (Garth, 1948, p. 37), that “[£.] robustus is a valid species and not con- specific with [E.] dovii, as suggested by Rath- bun (1930, p. 148).” That specimens of this large and distinctive species have not escaped the eyes of discriminating collectors, but have merely failed to be reported in the literature, is attested by a pair of comparable size from Peru sent the writer by Dr. Albert Panning of the Hamburg Museum. A redescription based on this new material, together with photographs, will appear in a subsequent monograph. Family Xanthidae Medaeus iobipes Rathbun Medaeus Iobipes Rathbun, 1898, p. 583, pi. 44, fig. 1; 1930, p. 275, text-fig. 44, pi. 114. Crane, 1937, p. 70. Garth, 1946, p. 442, pi. 77, fig. 2; 1948, p. 39. Range.— From Santa Inez Bay, Gulf of Cali- fornia, Mexico, to Guayabo Chiquito, Panama. Galapagos Islands. 5.5-150 fathoms. (Garth, 1948). Material Examined.— 35 specimens from 4 stations; Mexico Manzanillo, November 22, 1937, Station 184, D-l, 25 fathoms, 1 young male; D-2, 30 fath- oms, 15 males, 13 females (1 ovigerous). Costa Rica Port Parker, January 20, 1938, Station 203, D-l to D-3, 10-15 fathoms, 2 males, 1 female. Port Culebra, January 30, 1938, Station 206, D-l, D-3, 14 fathoms, 1 male. Golfito, Gulf of Dulce, March 9, 1938, Sta- tion 218, D-4 to D-7, 4-6 fathoms, 1 female, 1 young. Measurements. — Males from 4.9 X 6.9 to 18.6 X 28.1 mm., females from 5.7 X 8.1 to 16.4 X 24.2 mm., ovigerous female 11.6 X 17.1 mm., young from 3.0 X 4.0 mm. 146 Zoologica: New York Zoological Society [46: 13 Habitat.— Sand, gravelly sand; sandy mud and crushed shell; mangrove leaves, mud, and shell. Color in Life.—Ot Manzanillo specimens: At least half had orange carapace of varying de- grees of brightness with dark brown median longitudinal band and same brown on anterola- teral angles. In the rest the orange was replaced by light brown or white. Chelipeds orange or light brown externally, white internally and on distal part of manus. Underparts white sprin- kled posteriorly with brown. (J. Crane, field notes). Breeding— Mexico in late November. Remarks— The Port Culebra male, a young specimen, is granulate to the point of spinulosity. Medaeus spinulifer (Rathbun) Pilumnus spinulifer Rathbun, 1898, p. 585, pi. 42, figs. 6-8. Finnegan, 1931, p. 643. Medaeus spinulifer, Rathbun, 1930, 276, text-fig. 45. Garth, 1946, p. 443, pi. 75, figs. 5, 6; 1948, p. 40. Crane, 1947, p. 75. Range.— From Cape San Lucas, Lower Cali- fornia, Mexico, to Utria Bay, Colombia. Gala- pagos Islands. Shore to 73 fathoms. (Garth, 1948). Material Examined. — 3 specimens from 2 stations: Mexico 3 mi. off Pyramid Rock, Clarion Island, May 12, 1936, Station 163, D-3, D-4, 50 fathoms, 1 female. Manzanillo, November 22, 1937, Station 184, D-2, 30 fathoms, 1 male, 1 young. Measurements.— Male specimen 9.0 X 13.3 mm., female specimen 7.0 X 10.1 mm., young specimen 2.6 X 3.3 mm. Habitat— Gravelly sand. Color in Life.— Hot noted. Remarks.— The Manzanillo specimens were sorted out from among a large number of Med- aeus lobipes Rathbun. The species is now re- corded from the Revilla Gigedo Islands. Xanthodius stimpsoni (A. Milne Edwards) Xantho stimpsoni A. Milne Edwards, 1879, p. 252, pi. 46, figs. 2-2b. Finnegan, 1931, p. 631. Buiten- dijk, 1950, p. 277. Xanthodius stimpsoni, Rathbun, 1930, p. 315, pi. 143, figs. 5-7. Crane, 1947, p. 77. Garth, 1948, p. 41. Daira ecuadorensis Rathbun, 1935, p. 49. Range.— From Cape San Lucas, Lower Cali- fornia, Mexico, to Santa Elena Bay, Ecuador. 7-27 meters. (Garth). Material Examined. — 15 specimens from 3 stations: Mexico Banderas Bay, November 16, 1937, from oys- ter-bearing rocks, 3 young. Port Guatulco, Station 195, December 4, 1937, D-3, 3.5 fathoms, 1 male, 1 young; D-4, 4.5 fathoms, 1 male, 2 ovigerous females, 1 young; December 5, 1937, D-5, 2 fathoms, 2 males, 2 females (1 ovigerous), 1 young. Nicaragua Corinto, January 5, 1938, Station 200, D-15, 1 fathom, 1 young. Measurements— Males from 4.9 X 7.2 to 6.2 X 9.3 mm., females from 5.0 X 7.2 to 6.0 X 9.0 mm., ovigerous females from 5.0 X 7.3 to 6.0 X 9.0 mm., young from 2.8 X 3.7 mm. Habitat.— Sand, with algae or crushed shell; mangrove leaves. Color in Life.—Ot Port Guatulco specimens: Carapace white speckled with violet; chelipeds bright orange; legs dark brown except last (white). (J. Crane, field notes). Marked with shades of white and dark red. (Idem.). Breeding.— Mexico in early December. Remarks. — Since depth is not mentioned in Rathbun (1930), the “Askoy” records of 6-10 feet and 7-27 meters and the “Zaca” records of 1-4.5 fathoms aid materially in defining the bathymetric range. Hexapanope us costarkensis Garth Hexapanopeus costaricensis Garth, 1940, p. 79, pi. 21, figs. 1-4. Range.— From Port Parker and Puerto Cule- bra, Costa Rica. 3-10 fathoms. (Garth). Material Examined. — 14 specimens from 2 stations: Costa Rica Port Parker, January 20, 1938, Station 203, D-l to D-3, 10-15 fathoms, 5 males, 3 females (1 ovigerous); January 22, 1938, D-7, 5-9 fath- oms, 1 male, 1 female; D-8, 9 fathoms, 1 male. Panama Bahia Honda, March 18, 1938, Station 222, D-l to D-3, D-5, 3-11 fathoms, 2 males; D-3, 8 fathoms, 1 male. Measurements.— Males from 4.1 X 5.3 to 6.0 X 7.7 mm., non-ovigerous females from 4.0 X 5.5 to 4.6 X 6.0 mm., ovigerous female 3.5 X 4.9 mm. Habitat. — Sandy mud, crushed shell; shelly sand, algae; shelly mud; dead coral. Color in Life.— Not noted. Breeding— Costa Rica in late January. 1961] 147 Garth: N on-intertidal Brachygnathous Crabs Remarks— The “Zaca” records confirm those of the “Velero III” from Port Parker, the type locality, and extend the range of the species south from Costa Rica to northern Panama. Hexapanopeus nicaraguensis (Rathbun) Lophopanopeus nicaraguensis Rathbun, 1904b, p. 162. Hexapanopeus nicaraguensis, Rathbun, 1930, p. 395, text-fig. 61. Range.— Known only from the type locality, Realejo [Corinto], Nicaragua. Material Examined.— A specimens from 2 sta- tions : El Salvador La Libertad, December 16, 1937, Station 198, D-l, 13 fathoms, 1 male. Nicaragua Corinto, Station 200, December 29, 1937, D-l, D-3, D-8, 2-6.6 fathoms, 1 male, 1 ovig- erous female; January 7, 1938, D-27 to D-30, 3 fathoms, 1 male. Measurements— Males from 4.7 X 6.9 to 7.3 X 10.6 mm., ovigerous female 5.4 X 7.7 mm. Habitat— Mud, mangrove leaves. Color in Life— Not noted. Breeding.— Nicaragua in late December. Remarks— The above specimens are smaller than would be suggested by the unique male holotype, an 8.7 X 13 mm. specimen. Their presence at the original locality is confirmed, Realejo being the classic locality, Corinto its modern counterpart. A fine specimen from La Libertad, which is widest opposite the last mar- ginal tooth, extends the known range a full de- gree of latitude north to El Salvador. Hexapanopeus orcutti Rathbun Hexapanopeus orcutti Rathbun, 1930, p. 397, pi. 170, figs. 3, 4. Range.— Known only from the type locality, near Modesto, Sinaloa, Mexico. Material Examined. — 32 specimens from 6 stations comprising 8 separate localities: Mexico Banderas Bay, November 16, 1937, from oys- ter-bearing rocks, 1 male, 5 young. Chamela Bay, November 17, 1937, Station 182, D-3, 15 fathoms, 1 ovigerous female. Port Guatulco, Station 195, December 4, 1937, D-l, 2.5 fathoms, 1 male; D-2, 3 fathoms, 1 male, 1 young; December 6, 1937, D-10, 4 fathoms, 1 ovigerous female. Nicaragua Monypenny Point, Gulf of Fonseca, De- cember 24, 1937, Station 199, D-2, 5 fathoms, 1 male; D-5, D-6, 4-7 fathoms, 2 females (1 ovigerous). El Salvador La Union, Gulf of Fonseca, December 27, 1937, Station 199, D-8, 6 fathoms, 6 males, 4 females (2 ovigerous); D-17, D-21, 3-4 fath- oms, 2 males, 1 female. Costa Rica Port Parker, January 22, 1938, Station 203, D-4, 7 fathoms, 1 female, 1 young; D-10, 2.5-6 fathoms, 1 male, 4 females (2 ovigerous). Golfito, Gulf of Dulce, March 9, 1938, D-4 to D-7, 4-6 fathoms, 1 male, 1 female. Measurements.— Males from 3.0 X 4.0 to 6.0 X 8.3 mm., females from 3.0 X 3.9 to 3.7 X 5.1 mm., ovigerous female 3.7 X 5.0 mm., young from 2.2 X 2.8 mm., Gulf of Fonseca speci- mens: males from 4.1 X 5.2 to 9.5 X 12.6 mm., females from 4.0 X 5.3 to 7.0 X 9.2 mm., ovige- rous females 4.6 X 6.0 to 5.5 X 7.2 mm., young not present. Color in Life. — Of Chamela Bay, Mexico, specimens: Black all over except ambulatories, which are barred with black and cream. Chelae tipped with cream; abdomen pale buffy. (J. Crane, field notes). Of Port Guatulco, Mexico, specimens: Gray marked with white. (J. Crane, field notes). Habitat— Sand, algae; gravelly sand, crushed shell, dead coral; mud, mangrove leaves; gravel, algae. Breeding— Mexico in mid-November and ear- ly December; Nicaragua in late December. Remarks— As will be noted under Measure- ments above, the Gulf of Fonseca specimens (Monypenny Point, La Union) represent a giant race as compared to Mexican and Costa Rican specimens. They appear also to have longer legs and somewhat different chelae, and should perhaps on this account be segre- gated from the above series. It is clear, however, that the range of the species should be extended from northwest Mexico all the way to the Gulf of Dulce, Costa Rica, the depth to 15 fathoms. The Port Guatulco male has two minor chelae, possibly the result of regeneration. Hexapanopeus sinaloensis Rathbun Hexapanopeus sinaloensis Rathbun, 1930, p. 398, pi. 170, figs. 1, 2. Garth, 1948, p. 41. Hexapanopeus setipalpus Finnegan, 193 1, p. 641. Range.— From Boca Tecapan, Sinaloa, Mex- 148 Zoologica: New York Zoological Society [46: 13 ico, to Malaga Bay, Colombia, 4-9 meters. (Garth, 1948). Material Examined.. — 31 specimens from 6 stations: Nicaragua Meanguera Island, Gulf of Fonseca, Decem- ber 23, 1937, Station 199, B-l, 16 fathoms, 2 males. Corinto, Station 200, December 29, 1937, B-l, D-3, D-8, 2-6.6 fathoms, 1 male, 1 female; January 7, 1938, D-26, 2.5 fathoms, 1 ovigerous female; D-27 to D-30, 3 fathoms, 1 male. Costa Rica Port Parker, January 22, 1938, Station 203, D-5, 3 fathoms, 1 male; D-13, 7-9 fathoms, 1 ovigerous female. Cedro Island, Gulf of Nicoya, February 13, 1938, Station 213, D-l to D-10, 4-10 fathoms, 7 males, 6 females (2 ovigerous). Golfito, Gulf of Dulce, March 9, 1938, Sta- tion 218, D-4 to D-7, 4-6 fathoms, 3 males, 3 females (2 ovigerous) ; D-8, 6 fathoms, 2 males, 2 young. Panama Bahia Honda, March 18, 1938, Station 222, D-3, 8 fathoms, 1 male. Measurements— Males from 3.2 X 4.4 to 5.7 X 8.7 mm., females from 3.0 X 4.0 to 4.2 X 5.9 mm., ovigerous females same; young from 2.9 X 3.8 mm. Habitat. — Mangrove leaves; shells and dead coral or algae; mud, sand, and crushed shell. Color in Life.— Not noted. Breeding.— Nicaragua in early January; Costa Rica from late January to early March. Remarks.— Previously recorded from Taboga Island, Panama, as Hexapanopeus setipalpus Finnegan. The records for Nicaragua and Costa Rica are new. Hexapanopeus beebei, new species (Plate I) Type.— Male holotype, A.H.F. No. 377, and male and female, paratypes, N.Y.Z.S. No. 37,718, from Corinto, Nicaragua, December 29, 1937, “Zaca” Station 200, D-l, 3, 8, 2-6.6 fathoms. For additional paratypes see Supple- mentary Material below. Measurements. — Male holotype, length 5.4 mm., width 7.2 mm., of fronto-orbit 5.4 mm., of front 2.6 mm., length of cheliped 8.3 mm., of chela (lower margin) 6.1 mm., of dactyl 3.5 mm., height of palm 3.5 mm. Female paratype, length 5.2 mm., width 7.1 mm. Diagnosis— Fifth lateral tooth minute. Supra- orbital lobe advanced to level of outer orbital tooth. Fingers white, the whitened portion not continuing backward and upward on palm. Ma- jor dactyl strongly curved in male, a denticle but no large tooth at base. Male first pleopod with a reflected medial spine, an opposing lance- olate lobe, and a truncated or collared hood with three terminal setae. Description— C arapace nearly flat, almost de- void of pubescence, smooth anteriorly but gran- ulate posteriorly, clearly divided by furrows into regions each surmounted by one or more rows of granules, disposed as follows: two pro- togastric, of which the posterior is the more oblique, one epibranchial, in line with the last tooth, one metagastric, interrupted at the mid- dle, one cardiac, also medially divided, and one metabranchial; in addition to these an hepatic, in line with the anterior protogastric, and a post- frontal, in advance of the anterior protogastric. Front not exceptionally narrow, a little less than one-third carapace width, produced, thin-edged, lateral margins oblique, anterior margins some- what oblique, straight or slightly sinuous, frontal lobes separated by a shallow but distinct median V, a suggestion of an outer lobe. Inner supra- orbital border elevated, swollen, orbital emar- ginations V-shaped, the included lobe equally advanced inwardly with the outer orbital tooth. Second anterolateral tooth low, rounded, and separated from the first or exorbital by a shallow sinus, their combined width greater than that of the third tooth; third tooth with a short, straight, oblique anterior margin and a long, arcuate posterior margin, tip rectangular; fourth tooth spiniform, narrower at base than third and more projecting, directed obliquely for- ward, anterior margin transverse or slightly con- cave, posterior margin straight or nearly so; fifth tooth minute, separated from fourth by a closed fissure, and appearing as a notch on the posterior margin of the fourth. Merus of outer maxilliped subrectangular, broader than long, anterolateral angle slightly produced and rounded, anterointernal angle shallowly notched at insertion of palpus. Chelipeds markedly unequal in the male; merus with a superior distal tubercle; carpus with a marginal carina incompletely outlining a rhomb enclosing a rectangular sulcus in its outer portion, a scattering of tubercles above; manus of major chela smooth, inflated, two parallel ridges above. Fingers white, the white “color” not continued appreciably on palm, major dac- tylus strongly curved, almost forming a quarter- circle, a granulate ridge above, a small denticle basally in place of the customary larger tooth; 1961] Garth: N on-intertidal Brachygnathous Crabs 149 this with a small denticle distally defining a gape into which a somewhat larger tooth of the pollex fits incompletely; pollex not deflexed. Minor manus slenderer than major, fingers elon- gate, ridged and compressed, meeting without a gape but with crossed tips; pollex deflexed. Male abdomen with sides of third segment rounded, sides of fused segments 3-5 concave, narrowest at base of segment 6, segment 7 tri- angular with a blunt tip. Male first pleopod with a long, backward-pointing medial spine, an equally long and oppositely directed lanceolate lobe, and a rimmed hood bearing three terminal setae. Female noticeably more convex than male, ridges more prominent, interspaces more felted. Chelipeds less robust, carpus and minor manus more nodose, palms more granulate. Fingers compressed, those of major manus meeting with- out a gape but possessing larger teeth than those of minor manus. Supplementary Material.— In addition to spec- imens mentioned under Type above, the follow- ing specimens, also from “Zaca” Station 200: 0-7, 2 fathoms, 1 female, paratype; D-27 to 30, 3 fathoms, mangrove leaves, 1 male, 1 ovigerous female, paratypes. The male has two minor che- lipeds, the result of regeneration of a major cheliped. Remarks. — The proposed new species is most closely allied to HeXapanopeus caribbaeus (Stimpson), from which it differs in having the fingers white, the whitened portion not running backward and upward on the palm, the front not especially narrowed, and the fingers of the major chela of the male conspicuously gaping. I take pleasure in naming this diminutive panopeid for Dr. William Beebe, director emeri- tus of the Department of Tropical Reseach, New York Zoological Society, whose “Book of Bays” (1942) so delightfully describes the ex- pedition on which it was collected. Panopeus purpureas Lockington Panopeus purpureus Lockington, 1877b, p. 101. Rathbun, 1930, p. 344, pi. 158, fig. 1, pi. 159. Crane, 1947, p. 79. Range— From Magdalena Bay, Lower Cali- fornia, and Guaymas, Sonora, Mexico, to mouth of Rio Tumbes, Peru. (Rathbun). Material Examined— Puntarenas, Costa Rica, February 22, 1938, 1 male. Measurements. — Male specimen length 11.5 mm., width 16.5 mm. Habitat. — In stony mud on edges of man- grove swamps and open mudflats. (Crane). Remarks.— It was undoubtedly due to an over- sight that the above specimen was not reported by Crane (1947), who listed the species from Culebra, Ballenas, and Golfito, Costa Rica, and Puerto Bolivar, Ecuador. Panopeus bermudensis Benedict & Rathbun Panopeus bermudensis Benedict & Rathbun, 1891, p. 376, pi. 20, fig. 2; pi. 24, figs. 14, 15. Rathbun, 1930, p. 360, pi. 165, text-fig. 56. Range.— Eastern Pacific from Magdalena Bay, Lower California, Mexico, to Matapalo (near Capon), Peru. Western Atlantic from Florida and the Bahamas to Brazil. Bermuda. (Rath- bun). Material Examined.— Chamela Bay, Mexico, November 17, 1937, Station 182, D-l, 8 fath- oms, 17 males, 6 ovigerous females. Measurements.— Males from 4.3 X 5.7 to 6.3 X 8.5 mm., ovigerous females from 3.9 X 5.3 to 4.9 X 6.6 mm. Habitat.— Sand and algae. Fine bits of sea- weed were entwined among specimens. Remarks. — When considered independently, the Chamela Bay specimens could scarcely be reconciled with specimens from Bermuda to which the name bermudensis was originally ap- plied. It is when considered in the context of a larger representation of this geographically vari- able complex from many west coast localities that their relationship to a similar array from numerous Caribbean localities is appreciated. While as a result of studies now in progress it may be decided to segregate the Pacific material from the Atlantic as a distinct species, to do so now on the basis of a single lot of specimens from an isolated locality would be premature. The Chamela Bay specimens serve to link the present representation from Lower California and the Gulf of California with that from Cen- tral and northern South America. The small size of the females, all of which are ovigerous, is noteworthy. Eurytium fristani Rathbun Eurytium trisf&ni minor (Bott), n. comb. Eurytium tristani Rathbun, 1906, p. 100; 1910, pp. 543, 585, pi. 47, fig. 1; 1930, p. 425, pi. 176, fig. 3; pi. 177, fig. 3. Crane, 1947, p. 80. Panopeus convexus minor Bott, 1955, p. 57, pi. 6, figs. 9a, 9b. (Not P. convexus A. Milne Edwards, 1880). Range. — From El Triunfo, El Salvador, to Salto (near Capon), Peru. Material Examined— Puntarenas, Costa Rica, February 22, 1938, 26 males, 29 females (1 ovigerous), 1 young. Measurements— Males from 3.5 X 4.5 to 12.0 X 17.8 mm., females from 3.7 X 5.1 to 9.9 X 150 Zoologica: New York Zoological Society [46: 13 15.1 mm., ovigerous female 6.1 X 8.7 mm., young 3.2 X 4.3 mm. Habitat— Not stated, but most certainly from mud flats exposed at low tide. Color in Life.— Not noted. Remarks. — Of the two pairs of specimens from Puntarenas sent to Frankfurt, Germany, for comparison with the types of Panopeus con- vexus minor Bott, Dr. Richard Bott writes as follows: “They agree with the type in all char- acteristics mentioned by me; the first pleopods are completely identical.” It was noted that, while in the published figure of the male holo- type the left cheliped is larger than the right, among “Zaca” specimens the right cheliped is larger than the left in most instances. Dr. Bott affirms that this is also true of the remaining male from the type series of P. convexus minor. The decision to transfer Bott’s subspecies to Rathbun’s species of another genus was the result of examining a male specimen from El Triunfo reported as Panopeus convexus convexus (Bott, 1955, p. 57) and loaned the writer by Dr. Bott for use in connection with studies on the family Xanthidae. This proved to be none other than Eurytium tristani Rathbun. It therefore follows that the somewhat smaller specimens reported by him as Panopeus convexus minor should be called Eurytium tristani minor (Bott) instead. The small size of the “Zaca” specimens, and in particular, of the ovigerous female, support their continued distinction. Micropanope polita Rathbun Micropanope polita Rathbun, 1893, p. 238; 1930, p. 440, text-fig. 40, pi. 180, figs. 3, 4, synonymy. Crane, 1937, p. 71. Garth, 1946, p. 459, pi. 77, fig. 4. Range.— From Magdalena Bay, Lower Cali- fornia, and Santa Inez Bay, Gulf of California, Mexico, to Cocos Island, Costa Rica. Galapagos Islands. 3-150 fathoms. (Garth). Material Examined. — 35 specimens from 3 stations: Mexico SE of Cedros Island, Lower California, No- vember 10, 1937, Station 126, D-19, 25 fath- oms, 1 male, 1 female. From holes in rocks. 3 mi. off Pyramid Rock, Clarion Island, May 12, 1936, Station 163, D-2, 55 fathoms, 30 specimens. Panama Hannibal Bank, March 20, 1938, Station 224, D-l to D-3, 35-40 fathoms, 1 male, 2 females. Measurements.— Males from 4.0 X 6.0 to 5.4 X 8.6 mm., females from 2.9 X 4.2 to 4.8 X 7.2 mm. The larger specimens are from the more northerly localities. Habitat. — Rocks, mud, dead coral; sand, shells, algae. Color in Life. — Of Cedros Island, Mexico, specimens: Male tan and cream mottled. Front and manus rosy red. Dactyls dark brown tipped with white. Ambulatories banded tan and cream tinged with pink. Female carapace entirely crim- son, but brightest on front. Chelipeds coral red. Chelae dark brown, tipped with white. Under- parts pinkish. ( J. Crane, field notes) . Remarks— Food preferences and breeding in the southern part of the Gulf of California are discussed by Crane (1937). The records from the Revilla Gigedo Islands, of which Clarion is the most remote, and from the mainland of Panama are new. Micropanope xantusii (Stimpson) Xanthodes xantusii Stimpson, 1871, p. 105. Micropanope xantusii, Rathbun, 1930, p. 438, pi. 179, figs. 1-4. Crane, 1937, p. 72; 1947, p. 80. Garth, 1946, p. 457, pi. 77, fig. 6; 1948, p. 42. Xanthias serrulata Finnegan, 193 1, p. 634, text-fig. 6. Range— From Arena Bank, Gulf of Califor- nia, Mexico, to Santa Elena Bay, Ecuador. Gal- apagos Islands. (Garth, 1948). Occasionally to 40 fathoms. Material Examined.— 20 specimens from 3 collections made at 2 localities: Mexico Sulphur Bay, Clarion Island, May 15, 1936, coral, 1 young male. Port Guatulco, December 4, 1937, in dead pearl oyster, 1 male; December 6, 1937, Station 195, D-15, 1.5 fathoms, 8 males, 10 females (4 ovigerous). Measurements.— Males from 3.0 X 4.2 to 7.0 X 9.8 mm., females from 4.0 X 5.6 to 7.25 X 11.0 mm., ovigerous females from 4.5 X 6.4 to 6.3 X 9.2 mm. Habitat.— Co ral obtained by diving. Color in Life.— Variable, but majority dark red mottled with lighter and darker. Sulci on major cheliped of adult males may be almost lacking. (Crane, 1947). Breeding.— Mexico in early December. Remarks.— According to Crane (1947), who reported the species from the intertidal of Mex- ico (Clarion Island, Sihuatenejo, Acapulco) and Costa Rica (Port Parker, Culebra, Jasper Island) , “always found in Pocillopora coral ex- cept for 3 young found at Port Parker among algae-covered stones.” 1961] Garth: N on-intertidal Brachygnathous Crabs 151 Micropanope (?) maculatus (Rathbun) Lophopanopeus maculatus Rathbun, 1898, p. 588, pi. 40, figs. 10, 11; 1930, p. 330, text-fig. 51. Garth, 1946, p. 453, pi. 78, figs. 3, 4. Micropanope (?) maculatus, Menzies, 1948, p. 24. Range— From Magdalena Bay, Lower Cali- fornia, and southern part, Gulf of California, Mexico. Galapagos Islands. 2-70 fathoms. (Garth). Material Examined— Port Parker, Costa Rica, January 22, 1938, Station 203, D-9, 1.5-4 fath- oms, 1 male, 1 female. Measurements. — Male specimen 4.7 X 6.5 mm., female specimen 4.1 X 5.7 mm. Habitat.— Coral bottom. Color in Life.— Not noted. Remarks.— The Port Parker specimens are in good condition and show the distinctive char- acters well. Micropanope (?) maculatus is now recorded from the Central American mainland. The exclusion of the species from Lophopano- peus is the result of Menzies’ revision of that genus. Its referral to Micropanope in this in- stance is tentative and without prejudice to sys- tematic studies by the WTiter now in progress. Paraxanthias taylori (Stimpson) Xanthodes taylori Stimpson, 1860, p. 208, pi. 3, fig. 5. Paraxanthias taylori, Odhner, 1925, p. 85. Rathbun, 1930, p. 466, pi. 188, pi. 189, fig. 1, synonymy. Range. — From Monterey Bay, California, to Magdalena Bay, Lower California, Mexico. Shore to 55 fathoms. (Rathbun, 1930). Material Examined. — SE of Cedros Island, Lower California, Mexico, November 10, 1937, Station 126, D-19, 25 fathoms, 1 male. Measurements.— Male specimen, length 6.8 mm., width 9.8 mm. Habitat— Rocks, algae. From hole in rocks. (J. Crane, field notes). Color in Life. — Pinkish tan; chelae brown tipped with white. (J. Crane, field notes). Remarks.— A southern California— west coast of Lower California warm-temperate species. Menippe obtusa Stimpson Menippe obtusa Stimpson, 1859, p. 53. Rathbun, 1930, p. 478, pi. 197, pi. 198, figs. 1, 2. Sivertsen, 1933, p. 16. Garth, 1946, p. 470, pi. 82, figs. 3, 4; 1948, p. 45. Crane, 1947, p. 80. Range— From Corinto, Nicaragua, to La Plata Island, Ecuador. Galapagos Islands. Shore to 6.5 meters. (Garth, 1948). Material Examined.— Passavera Island, Cham- ela Bay, Mexico, November 19, 1937, 1 female. Measurements. — Female specimen, length 12.8 mm., width 18.3 mm. Habitat— Not given. Color in Life.— Brown to apricot orange. (J. Crane, of Corinto, Nicaragua, specimens). Remarks.— For a discussion of the interesting habits of this and the related Menippe frontalis A. Milne Edwards, including use of the stridu- lating ridges of the inner surface of the manus, the reader is referred to Crane (1947, p. 80). Chamela Bay, Mexico, represents a northward extension of range from Corinto, Nicaragua, an earlier reported “Zaca” record. Pilumnus pygmaeus Boone Pilumnus pygmaeus Boone, 1927, p. 221, text-fig. 81. Rathbun, 1930, p. 515, pi. 207, figs. 4, 5. Garth, 1946, p. 472, pi. 80, fig. 4; 1948, p. 48. Crane, 1947, p. 81. Range.— From Port Parker, Costa Rica, to Utria Bay, Colombia. Galapagos Islands. Shore to 13 fathoms. Material Examined.— 2 specimens from 2 sta- tions: Mexico Port Guatulco, December 4, 1937, in dead pearl oyster, 1 female. Costa Rica Port Parker, [date?], 1 male. Measurements.— Male specimen 2.5 X 3.3 mm., female specimen 2.7 X 3.6 mm. Habitat— From weed in tidepools and on un- dersides of overgrown rocks close to low tide. (Crane, 1947, of Costa Rican specimens). Breeding.— Eggs in January. (Crane, 1947). Remarks— The known range of the species is extended northward from Port Parker, Costa Rica, to Port Guatulco, Mexico. Pilumnus limosus Smith Pilumnus limosus Smith, 1869, p. 285. Rathbun, 1930, p. 518, pi. 208, pi. 209, figs. 1-3. Range.— From Panama to Paita, Peru. Shore. (Rathbun, 1930). Material Examined.— 2 specimens from 2 sta- tions: Mexico 4 mi. SSW of Maldonado Point, November 30, 1937, Station 192, D-l, 26 fathoms, 1 female. Tangola-Tangola Bay, December 13, 1937, Station 196, D-l 8, 30 fathoms, 1 male. Measurements. — Male specimen 5.0 X 6.8 mm., female specimen 7.4 X 9.9 mm. Habitat.— Mud. 152 Zoologica: New York Zoological Society [46: 13 Color in Life. — Brownish. (J. Crane, field notes). Remarks— A distinctive species, with a char- acteristic pattern of pubescence that obscures the scattered granules of the carapace. Pilumnus limosus is now recorded from Mexico, a north- ward extension from Panama, and to a depth of 30 fathoms. Pilumnus stimpsonii Miers Pilumnus tnarginatus Stimpson, 1871, p. 109. Not P. marginatus Stimpson, 1858. Pilumnus stimpsonii Miers, 1886, p. 147 (name sub- stituted for P. marginatus, preoccupied). Rath- bun, 1930, p. 524, pi. 205, figs. 5, 6. Range.— Cape San Lucas, Lower California, and Manzanillo, Colima, Mexico. (Rathbun, 1930). Material Examined.— Pori Parker, Costa Rica, January 22, 1938, Station 203, D-10, 2.5-6 fath- oms, 1 male, 6 females (3 ovigerous). Measurements— Male 2.9 X 3.4 mm., females from 3.2 X 4.0 to 4.0 X 4.9 mm., ovigerous fe- males from 3.2 X 4.0 to 3.8 X 4.5 mm. Habitat— Rocks. Color in Life.— Not recorded. Breeding.— Costa Rica in late January. Mexico in mid-July. (Rathbun, 1930). Remarks.— Since Stimpson’s type is non-ex- tant and the material reported by Rathbun (1930) consists in its entirety of 3 specimens collected by Orcutt in Mexico, the “Zaca” series from Port Parker represents more than double the number of specimens existing in collections and reported upon to date. Orcutt’s material also included an ovigerous female. The range of Pil- umnus stimpsonii is extended southward to Costa Rica, the depth to 6 fathoms. Heteractaea peterseni Garth Heteractaea peterseni Garth, 1940, p. 81, pi. 22, figs. 1-5. Range.— From San Francisco Island, Gulf of California, Mexico, to Octavia Bay, Colombia. 35-44 fathoms. (Garth). Material Examined.— Hannibal Bank, Pana- ma, March 20, 1938, Station 224, B-l to D-3, 40-35 fathoms, 1 female, 2 young. Measurements.— Female specimen 5.5 X 8.0 mm., young 3.9 X 5.7 and 4.4 X 6.3 mm. Habitat.— Rocks, dead coral, mud; sand, shells, algae. The fact that the holotype, from Octavia Bay, Colombia, was cracked from rock makes the former habitat the more likely. The chan- neled meri of the walking legs are believed to represent an adaptation for breathing from a rock-bound enclosure. Unfortunately, materials from two dissimilar bottom types at Station 224 were combined. Color in Life.— See Garth (1940, p. 83) . Remarks.— The findings by the “Zaca” of Het- eractaea peterseni off Panama bridges the gap between the two “Velero III” stations in Colom- bia and the Gulf of California. The uniformity of depth for the three stations is perhaps more than coincidental: Octavia Bay, Colombia, 35-40 fathoms; San Francisco Island, Gulf of Cali- fornia, 43-44 fathoms; and Hannibal Bank, Panama, 35-40 fathoms. Quadrella nitida Smith (Text-fig. 2) Quadrella nitida Smith, 1869, p. 288. Rathbun, 1930, p. 561, pi. 229, synonymy. Crane, 1937, p. 74. Garth, 1946, p. 494, pi. 80, fig. 6. Range— From off Cape San Lucas, Lower California, and Arena Bank, Gulf of California, Mexico, to Pacheca Island, Panama. Galapagos Islands. 6-75 fathoms. Material Examined. — 15 specimens from a single station: Mazanillo, Mexico, November 22, 1937, Station 184, D-l, 25 fathoms, 1 male, 3 females, 2 young, 1 megalops; D-2, 30 fath- oms, 3 males, 3 females, 2 young. Measurements— Males from 5.0 X 5.5 to 6.9 X 8.0 mm., females from 4.9 X 5.4 to 7.5 X 8.6 mm., young from 3.3 X 3.4 mm. Habitat— Gravelly sand. Color in Life.— Two middle-sized specimens not pure white or with bold orange chelipeds like a northern form [cf. Crane, 1937, p. 74]. Two larger specimens all carapace mouse gray with posterolateral margins curdled white. Cheli- peds light brown with a band of dark gray polka- dotted with white at distal end of each segment. Rest of segment dotted with fine darker brown. Ambulatories banded gray and white. Smaller specimen white with chestnut front and chestnut chelipeds. Largest specimen curdled white; cheli- peds translucent violet gray or pale rosy. Defi- nitely matching lavender shade of some of the gorgonids. (J. Crane, field notes). Remarks— Although the only reference to the presence of gorgonian corals at this station is to be found in the color description above, it is a safe presumption that the specimens were taken on Muricea miser Verrill or a related species. The symbiotic relationship between these crabs and the corals is discussed by Crane (1937). The megalops of the species has not previously been described. Description of the Megalops. — Carapace smooth and bare, slightly broader than long, sides convex, hepatic margins also slightly con- 1961] Garth: N on-intertidal Brachygnathous Crabs 153 Text-fig. 2. Quadrella nitida Smith, megalops. A, dorsal view; B, third maxilliped; C, second maxilliped; D, left third walking leg; E, left fourth walking leg; F, left chela. J. W. Knudsen, del. (B and C, same scale; D and E, same scale). vex, posterior margin concave. Front wide, ros- trum deflexed, bluntly triangular, straight-sided, broadly separated from lateral frontal spines; these spines pointing upward and not extending as far forward as rostrum. Transorbital breadth greater than breadth of carapace. Second maxilliped with a two-segmented exo- podite bearing 5 terminal setae; endopodite with four segments bearing 2, 0, 8, and 10 or more setae from merus to dactylus. Third maxilliped with a two-segmented exopodite bearing 5 or 6 plumose setae terminally; endopodite with isch- ium and merus broadened; ischium subrectangu- lar, with 16 or more setae on inner margin; merus subquadrate, with 8 or more setae on inner margin and 2 at anteroexternal angle; carpus rounded distally, a few setae on outer margin; propodus with 6 or more setae arranged in tufts, one on inner, 2 on outer margin; dac- tylus with 1 2 or more setae on inner, 4 on outer margin, with a terminal brush of a dozen or more longer hairs. Cheliped without an ischial spine; merus elongate, cylindrical, broadening distally; carpus ovate, three setae on outer distal margin; manus elongate, compressed, broadening distally, lower margin sinuous; dactylus with a row of 5 setae, closing against propodus without a gape, their tips crossing. Walking legs slender, cylindrical; meri long and of a length equal to their respective propodi; carpi only half as long as meri and of a length equal to their dactyli; legs finely and sparingly setose. Dactyli adapted to clinging to gorgoniace- ous corals, having a curved nail and superior bristle terminally, followed by a succession of teeth diminishing regularly in size from tip to base along lower margin, accompanied by stout bristles. Abdomen with six segments and a telson; sixth segment and telson smaller than preceding segments; telson arcuate. 154 Zoologica: New York Zoological Society [46: 13 Family Goneplacidae Pseudorhombila xanthiformis Garth Pseudorhombila xanthiformis Garth, 1940, p. 86, pi. 24, figs. 1-5. Range— Known only from the type locality, Gorgona Island, Colombia. 40-60 fathoms. Material Examined— 14 mi. S X E of Judas Point, Costa Rica, March 1, 1938, Station 214, D-l, D-3, 42-50 fathoms, 3 males, 1 female. Off Ballenas Bay, Gulf of Nicoya, Costa Rica, Feb- ruary 25, 1938, Station 213, D-12, 35 fathoms, right and left chelae, of large size; D-16, 45 fathoms, carapace fragment. Measurements— Males 12.6 X 17.3 and 15.6 X 22.9 mm., female 8.9 X 12.7 mm., young male 6.4 X 8.9 mm. Habitat— Mud and shell. Color in Life.— Not recorded. Remark.— Known only from the type female, a 15 X 22 mm. specimen, Pseudorhombila xan- thiformis is now represented by males of good size, and from a locality midway between the type locality, Gorgona Island, Colombia, and the type locality of Oediplax granulata Rathbun. When males of the latter species become avail- able, it should be possible to elucidate the rela- tionship between the two species, and to tell whether they should belong in the same or in different genera. (See discussion, Garth, 1940, p. 88). Description of the Male Abdomen— Segments 3-5 fused, suture lines clearly visible. Abdomen widest at base of segment 3, narrowest at middle of segment 6, sides of segments 4-6 concave. Segment 7 broader than long, sides straight and convergent, tip rounded. Euryplax polita Smith Euryplax politus Smith, 1870, p. 163. Euryplax polita , Rathbun, 1918, p. 36. Range. — Apart from a listing by Glassell (1934) as occurring in the Gulf of California, but without specific locality, the species is known only from Panama. Material Examined. — Tangola-Tangola Bay, Mexico, Station 196, December 9, 1937, D-6, D-8, 9-7 fathoms, 2 females, 2 young; December 12, 1937, D-13, 10 fathoms, 1 male, 1 young; December 13, 1937, D-16, 16 fathoms, 1 young female; D-17, 23 fathoms, 1 female. Measurements. — Male 8.7 X 14.8 mm., females from 4.4 X 5.9 to 5.8 X 8.8 mm., young from 3.2 X 4.4 mm. Habitat.— Sand, gravelly sand, crushed shell, and mud. Color in Life— Not noted. Remarks— The. male above is larger than the male holotype, a 6.9 X 11.2 mm. specimen. Euryplax polita is now known from a definite west Mexican locality. Included with the adult specimens enumerated above are a total of 11 megalopa possibly referable to this species. Chasmophora macrophthalma (Rathbun) Eucratopsis macrophthalma Rathbun, 1898, p. 601, pi. 43, figs. 3, 4. Chasmophora macrophthalma, Rathbun, 1914, p. 119, text-fig. 2; 1918, p. 37, text-fig. 12. Range.— Known only from the type locality. Bay of Panama. 51.5 fathoms. (Rathbun, 1898). Material Examined.— 32 specimens from 3 stations : Mexico 4 mi. SSW of Maldonado Point, November 30, 1937, Station 192, D-l, 26 fathoms, 4 males, 4 females (3 ovigerous). Santa Cruz Bay, December 7, 1937, Station 195, D-21, 18 fathoms, 2 females. Tangola-Tangola Bay, December 13, 1937, Station 196, D-17, 23 fathoms, 12 males, 10 females (1 ovigerous). Measurements.— Males from 2.1 X 3.1 to 4.5 X 8.8 mm., females from 2.7 X 4.6 to 4.4 X 8.3 mm., ovigerous females from 3.2 X 5.9 to 4.4 X 8.3 mm. Habitat— Mud. Color in Life.— Not noted. Remarks.— The larger males have the cheli- peds tremendously developed for their rather small size. The smaller females (3.0 X 5.1 and 3.0 X 5.3) show spinulous anterolateral teeth. Chasmophora macrophthalma may now be re- ported from west Mexico, a northward range extension from Panama. Speocctrcinus granulimanus Rathbun Speocarcinus granulimanus Rathbun, 1893, p. 242; 1918, p. 40, text-fig. 15, pi. 9. Range.— Off Cedros Island, Lower California, and from off Consag Rock to off Point Fermin, Gulf of California, Mexico. 23-33 fathoms. (Rathbun, 1893). Material Examined— 6 specimens from 4 sta- tions: El Salvador La Libertad, December 16, 1937, Station 198, D-l or D-2, 13-14 fathoms, 1 female. Meanguera Island, Gulf of Fonseca, Station 199, December 23, 1937, D-l, 16 fathoms, 2 young; December 27, 1937, D-ll, D-12, 5 fath- oms, 1 male. 1961] Garth: N on-intertidal Brachygnathous Crabs 155 Costa Rica Cedro Island, Gulf of Nicoya, February 13, 1938, Station 213, D-l to D-10, 4-10 fathoms, 1 young male. Golfito, Gulf of Dulce, March 9, 1938, Sta- tion 218, D-8, 6 fathoms, 1 female. Measurements. — Male 11.3 X 14.3 mm., females from 6.0 X 7.9 to 7.3 X 9.6 mm., young male 3.6 X 4.3, young female 4.6 X 6.0 mm. Habitat— Mud; sand, mud, and crushed shell; mangrove leaves, mud, and shells. Color in Life.— Not noted. Remarks.— As characteristic of mud bottoms as the following Speocarcinus calif orniensis, and occurring in the same dredge haul with it at Golfito, S. granulimanus may now be recorded from El Salvador and Costa Rica, a southward range extension from the Gulf of California. Speocarcinus californiensis (Lockington) Eucrate? californiensis Lockington, 1877a, p. 33. Speocarcinus californiensis, Rathbun, 1904a, p. 190, pi. 9, fig. 1; 1918, p. 42, text-fig. 16, pi. 10, figs. 2, 3. Schmitt, 1921, p. 249, text-fig. 148. Range.— San Pedro to San Diego, California. (Rathbun, 1918). Also Gulf of California, with- out specific locality. (Glassell, 1934). Material Examined. — 12 specimens from 3 stations: Costa Rica Port Parker, January 20, 1938, Station 203, D-l to D-3, 10-15 fathoms, 1 male. Port Culebra, January 30, 1938, Station 206, D-2, D-3, 14 fathoms, 1 male, 4 females (1 ovigerous) . Golfito, Gulf of Dulce, March 9, 1938, Sta- tion 218, D-8, 6 fathoms, 4 males, 1 female, 1 young. Measurements.— Males from 3.9 X 5.1 to 5.8 X 7.6 mm., females from 4.1 X 5.3 to 5.6 X 7.2 mm., the latter ovigerous, young from 2.3 X 2.8 mm. Habitat. — Sandy mud, crushed shell; man- grove leaves, mud, and shell. Color in Life.— Not noted. Breeding— Costa Rica in late January. Remarks.— Like the preceding Speocarcinus granulimanus, S. californiensis may now be re- corded from Costa Rica, although not from El Salvador, a range extension southward from the Gulf of California, locality uncertain. Speocarcinus ostrearicola Rathbun Speocarcinus ostrearicola Rathbun, 1910, p. 545, pi. 48, fig. 2; 1918, p. 41, pi. 10, fig. 1. Range.— Known only from the type locality, Matapalo (near Capon), Peru, where it occurs in oyster beds. (Rathbun, 1910). Material Examined.— 95 specimens from four localities: Nicaragua Monypenny Point, Gulf of Fonseca, Decem- ber 24, 1937, Station 199, D-2, D-6, 4-5 fathoms, 2 males. El Salvador La Union, Gulf of Fonseca, December 27, 1937, Station 199, D-14, 5 fathoms, 1 male, 1 female. Panama Bahia Flonda, March 16, 1938, no station number, depth not given, 14 males. Bella Vista, Panama City, date not given, 75 males, 4 females. Measurements— Males from 7.3 X 10.0 to 16.0 X 24.6 mm., females from 6.8 X 9.3 to 9.2 X 13.0 mm. Habitat— Mud. Color in Life.— Not noted. Remarks.— In view of the exceeding abund- ance of this species, both at Bahia Honda and Bella Vista, Panama (where males predominated over females, however), it is difficult to see how it has remained a rarity in collections. Known only from Peru, and from the unique type speci- men, at least insofar as recorded in the literature, Speocarcinus ostrearicola may now be said to occur off Central America as far north as Nica- ragua. Specimens show great difference as to the amount of pubescence on the carapace, some being entirely bare, others almost “furry.” The interspace between the first and second antero- lateral teeth is more U-shaped than V-shaped. Chasmocarcinus latipes Rathbun Chasmocarcinus latipes Rathbun, 1898, p. 602, pi. 43, fig. 5; 1918, p. 57, text-figs. 25, 26. Crane, 1937, p. 75. Garth, 1948, p. 52. Chasmocarcinus ferrugineus Glassell, 1936, p. 216. Crane, 1937, p. 75, pi. 7, fig. 24. Range.— From Cedros Island, Lower Califor- nia, and Arena Bank, Gulf of California, Mexico, to off Esmeraldas, Ecuador. 20-51 fathoms. (Garth, 1948). Material Examined.— 22 specimens from 5 stations : El Salvador La Libertad, December 16, 1937, Station 198, D-l or D-2, 13-14 fathoms, 1 male, 1 female. Meanguera Island, Gulf of Fonseca, Decem- ber 23, 1937, Station 199, D-l, 16 fathoms, 3 males, 5 females. 156 Zoologica: New York Zoological Society [46: 13 Costa Rica Off Ballenas Bay, February 25, 1938, Station 213, D-12, B-16, 35-45 fathoms, 1 male, 1 fe- male. 14 mi. S X E of Judas Point, March 1, 1938, Station 214, D-2, D-3, 43-50 fathoms, 5 males, 4 females. Panama Gulf of Chiriqui, March 13, 1938, Station 221, D-4, 38 fathoms, 1 female. Measurements— Males from 4.0 X 5.6 to 9.1 X 12.2 mm., females from 4.2 X 5.8 to 8.6 X 12.0 mm. A larger female, crushed, could not be measured. Habitat— Mud, sand, mud and crushed shell; sandy mud. Color in Life— Carapace and chelipeds gray- ish-white; center of carapace pink; basal seg- ments of all specimens brown. Pubescence brown. (Crane, 1937, of Gulf of California spe- cimens). Remarks— Specimens of the reddish-brown color associated with Chasmocarcinus ferrugin- eus Glassell occur with normally colored indiv- iduals at Judas Point, thus supporting the syn- onymy given above. The ruddy color is due to a reddish mud. It is noteworthy that all Chasmo- carcinus collected by the “Zaca” were C. latipes, and that none were C. longipes Garth (1940), which has as its range the Panama Bight, from Secas Islands, Panama, to La Plata Island, Ecuador. The equal and similar chelae of the large male from Judas Point serve to distinguish C. latipes from C. longipes, the chelae of which are unequal and dissimilar. Hexapus williamsi Glassell Hexapus williamsi Glassell, 1938, p. 445, pi. 35, figs. 1-4. Range— Known only from the type locality, San Jose, Guatemala. Material Examined.— 1 mi. W of Champerico, Guatemala, December 15, 1937, Station 197, D-2, 14 fathoms, 1 ovigerous female. Measurements— Female specimen, length 9.7 mm., width 15.5 mm. Habitat— Mud. Color in Life.— Not noted. Breeding.— Guatemala in mid-December. Remarks. — Named for M. Woodbridge Williams and described under the title “New and obscure Decapod Crustacea from the west American coasts,” the species has not been taken again since its discovery in April, 1937, by the “Stranger” of Capt. Fred E. Lewis. The “Zaca” specimen is therefore the first since the type, and the only female, the holotype being a male. A slight extension of range can be reported, since Champerico is about 75 miles north and west of San Jose, Guatemala. Family Cymopoliidae Cymopolia lucasii (Rathbun) Palicus lucasii Rathbun, 1898, p. 600, pi. 43, fig. 2. Cymopolia lucasii, Rathbun, 1918, p. 193, text-fig. 119, pi. 44, figs. 1, 2; Crane, 1937, p. 76. Garth, 1946, p. 500, pi. 87, fig. 1. Range.— From Arena Bank, Gulf of Califor- nia, to Cape San Lucas, Lower California, Mexico. Galapagos Islands. 5-60 fathoms. (Garth, 1946). Material Examined. — 2 specimens from 1 station: Mexico Sulphur Bay, Clarion Island, May 11, 1936, Station 163, D-l, 20 fathoms, 1 male, crushed. 3 mi. off Pyramid Rock, Clarion Island, May 12, 1936, Station 163, D-2, 55 fathoms, 1 male. Measurements.— Males 5.2 X 6.0 and 11.9 X 13.4 mm. Habitat— At Arena and Gorda Banks on muddy and sandy bottoms. (Crane, 1937). Remarks.— Cymopolia lucasii is now recorded from Clarion Island, an intermediate locality between Cape San Lucas and the Galapagos Islands. Literature Cited Beebe, W. 1942. Book of Bays. Pp. (xviii) 302. Harcourt, Brace and Co., New York. Benedict, J. E., & Mary J. Rathbun 1891. The genus Panopeus. Proc. U. S. Nat. Mus., vol. 14, pp. 355-385, pis. 19-24. Boone, Lee 1927. The littoral crustacean fauna of the Gala- pagos Islands. Part I: Brachyura. Zoo- logica, vol. 8, no. 4, pp. 127-288, text-figs. 34-102. 1930. Scientific results of the cruises of the yachts “Eagle” and “Ara,” 1921-1928, William K. Vanderbilt, commanding. Crustacea: Stomatopoda and Brachyura. Bull. Vanderbilt Mar. Mus., vol. 2, pp. 1-228, pis. 1-74. Bott, R. 1955. Dekapoden (Crustacea) aus El Salvador. 2. Litorale Dekapoden, ausser Uca. Senck- enbergiana, Biologica, vol. 36, pp. 45-70, pis. 3-8, text-figs. 1-7. 1961] Garth: N on-intertidal Brachygnathous Crabs 157 Buitendijk, Alida M. 1950. Note on a collection of Decapoda Brachyura from the coasts of Mexico, in- cluding the description of a new genus and species. Zool. Mededel. Rijksmus. Natuur. Hist. Leiden, vol. 30, no. 17, pp. 269-282, pi. 10, text-fig. 1. Coventry, G. A. 1944. Results of the Fifth George Vanderbilt Expedition (1941). (Bahamas, Caribbean Sea, Panama, Galapagos Archipelago and Mexican Pacific Islands). The Crustacea. Monog. 6, Acad. Nat. Sci. Philadelphia, pp. 531-544. Crane, Jocelyn 1937. The Templeton Crocker Expedition. III. Brachygnathous crabs from the Gulf of California and the west coast of Lower California. Zoologica, vol. 22, pp. 47-78, pis. 1-8. 1947. Intertidal brachygnathous crabs from the west coast of tropical America with spe- cial reference to ecology. Zoologica, vol. 32, pp. 69-95, text-figs. 1-3. Faxon, W. 1893. Reports on the dredging operations off the west coast of Central America to the Galapagos, to the west coast of Mexico, and in the Gulf of California ... by the U. S. Fish Commission steamer “Alba- tross,” during 1891 . . . VI. Preliminary descriptions of new species of Crustacea. Bull. Mus. Compar. Zool. Harvard, vol. 24, pp. 149-220. 1895. Reports on an exploration off the west coasts of Mexico, Central and South America, and off the Galapagos Islands ... by the U. S. Fish Commission steamer “Albatross,” during 1891 . . . XV. The stalk-eyed Crustacea. Mem. Mus. Com- par. Zool. Harvard, vol. 18, pp. 1-292, pis. A-K, 1-56. Finnegan, Susan 1931. Report on the Brachyura collected in Cen- tral America, the Gorgona and Galapagos Islands, by Dr. Crossland on the “St. George” Expedition to the Pacific, 1924- 25. Jour. Linn. Soc. London, Zool., vol. 37, pp. 607-673, text-figs. 1-6. Garth, J. S. 1940. Some new species of brachyuran crabs from Mexico and the Central and South American mainland. Allan Hancock Paci- fic Exped., vol. 5, no. 3, pp. 53-127, pis. 11-26. 1946. Littoral brachyuran fauna of the Gala- pagos Archipelago. Allan Hancock Pacific Exped., vol. 5, no. 10, pp. (iv) 341-601, pis. 49-87, text-fig. 1. 1948. The Brachyura of the “Askoy” Expedi- tion with remarks on carcinological col- lecting in the Panama Bight. Bull. Amer. Mus. Nat. Hist., vol. 92, art. 1, pp. 1-66, pis. 1-8, text-figs. 1-5. 1957. Reports of the Lund University Chile Ex- pedition 1948-1949. No. 29. The Crus- tacea Decapoda Brachyura of Chile. O Lunds Univ. Arsskr., n.s., Avd. 2, vol. 53, no. 7, pp. 1-127, pis. 1-4, text-figs. 1-11. 1958. Brachyura of the Pacific coast of Amer- ica. Oxyrhyncha. Allan Hancock Pacific Exped., vol. 21, pp. (vi) 859, pis. A-Z, Z1-Z4, 1-55, text-figs. 1-10. In 2 parts. 1959. Eastern Pacific Expeditions of the New York Zoological Society. XLIV. Non- intertidal brachygnathous crabs from the west coast of tropical America. Part 1: Brachygnatha Oxyrhyncha. Zoologica, vol. 44, pt. 3, pp. 105-126, pi. 1, text-figs. 1, 2. Gerstaecker, C. E. A. 1857. Carcinologische Beitrage. Arch. f. Natur- gesch., vol. 22, pt. 1, pp. 101-162, pis. 4-6. Glassell, S. A. 1934. Some corrections needed in recent car- cinological literature. Trans. San Diego Soc. Nat. Hist., vol. 7, pp. 453-454. 1935. New or little known crabs from the Pacific coast of northern Mexico. Trans. San Diego Soc. Nat. Hist., vol. 8, pp. 91-106, pis. 9-16. 1936. The Templeton Crocker Expedition. I. Six new brachyuran crabs from the Gulf of California. Zoologica, vol. 21, pp. 213- 218. 1938. New and obscure decapod Crustacea from the west American coasts. Trans. San Diego Soc. Nat. Hist., vol. 8, pp. 411-453, pis. 27-36. Holthuis, L. B. 1954. On a collection of decapod Crustacea from the republic of El Salvador (Central America). Zool. Verhandel. Rijksmus. Natuur. Hist. Leiden, no. 23, pp. 1-43, pis. 1-2, text-figs. 1-15. Lamarck, J. B. P. A. de M. de 1818. Histoire naturelle des animaux sans verte- bres. 1st ed., vol. 5, pp. 612. Paris. Lockington, W. N. 1877a. Remarks on the Crustacea of the Pacific coast, with descriptions of some new spe- cies. Proc. Calif. Acad. Sci., vol. 7, pp. 28-36. 1877b. Remarks on the Crustacea of the west coast of North America, with a catalogue of the species in the museum of the Cali- fornia Academy of Sciences. Proc. Calif. Acad. Sci., vol. 7, pp. 94-108. 158 Zoologica: New York Zoological Society [46: 13 Menzies, R. J. 1948. A revision of the brachyuran genus Lophopanopeus. Allan Hancock Found. Pubs., Occas. Paper No. 4, pp. 1-45, pis. 1-6. Miers, E. J. 1886. Report on the Brachyura collected by H. M. S. Challenger during the years 1873-1876. Rpt. Zool. Challenger Exped., vol. 17, pp. (1) 362, pis. 1-29. London, Edinburgh & Dublin. Milne Edwards, A. 1861. Etudes zoologiques sur les crustaces re- cents de la famille des portuniens. Arch. Mus. Hist. Nat. Paris, vol. 10, pp. 309-428, pis. 28-38. 1874. [No title.] In Fisher, P., L. de Folin, and L. Perier, Les fonds de la mer. Paris, vol. 2. 1873-1880. fitudes sur les xiphosures et les crus- taces de la region mexicaine. Mission Scientifique au Mexique et dans l’Amer- ique centrale, pt. 5, pp. 368, pis. 1-61. Paris. Odhner, T. 1925. Monographierte Gattungen der Krabben- familie Xanthidae. Goteborg’s K. Vet. Handl., Fjarde Foljden, vol. 29, no. 1, pp. 1-92, pis. 1-5, text-figs. 1-7. Ordway, A. 1863. Monograph of the genus Callinectes. Bos- ton Jour. Nat. Hist., vol. 7, no. 4, pp. 567-583. Rathbun, Mary J. 1893. Scientific results of explorations by the U. S. Fish Commission steamer Albatross. XXIV. Descriptions of new genera and species of crabs from the west coast of North America and the Sandwich Islands. Proc. U. S. Natl. Mus., vol. 16, pp. 223- 260. 1898. The Brachyura collected by the U. S. Fish Commission steamer Albatross on the voy- age from Norfolk, Virginia, to San Fran- cisco, California, 1887-1888. Proc. U. S. Natl. Mus., vol. 21, pp. 567-616, pis. 41-44. 1902. Papers from the Hopkins-Stanford Expedi- tion, 1898-1899. VIII. Brachyura and Macrura. Proc. Washington Acad. Sci., vol. 4, pp. 275-292, pi. 12. 1904a. Decapod crustaceans of the northwest coast of North America. Harriman Alaska Expedition, vol. 10, Crustaceans, pp. 210, pis. 1-10, text-figs. 1-95. Washington. 1904b. Descriptions of three new species of American crabs. Proc. Biol. Soc. Washing- ton, vol. 17, pp. 161-162. 1906. Descriptions of three new mangrove crabs from Costa Rica. Proc. Biol. Soc. Wash- ington, vol. 19, pp. 99-100. 1910. The stalk-eyed Crustacea of Peru and the adjacent coast. Proc. U. S. Natl. Mus., vol. 38, pp. 531-620, pis. 36-56. 1914. New genera and species of American brachyrhynchous crabs. Proc. U. S. Natl. Mus., vol. 47, pp. 117-129, pis. 1-10. 1918. The grapsoid crabs of America. Bull. U. S. Natl. Mus., no. 97, pp. 1-461, pis. 1-161, text-figs. 1-172. 1923. Scientific results of the expedition to the Gulf of California ... by the U. S. Fish- eries steamship “Albatross,” in 1911 . . . XIII. The brachyuran crabs collected by the U. S. Fisheries steamer “Albatross” in 1911, chiefly on the west coast of Mexico. Bull. Amer. Mus. Nat. Hist., vol. 48, pp. 619-637, pis. 26-36, text-figs. 1-7. 1930. The cancroid crabs of America of the fam- ilies Euryalidae, Portunidae, Atelecycli- dae, Cancridae, and Xanthidae. Bull. U. S. Natl. Mus., no. 152, pp. 1-609, pis. 1-230, text-figs. 1-85. 1935. Preliminary descriptions of six new species of crabs from the Pacific coast of America. Proc. Biol. Soc. Washington, vol. 48, pp. 49-52. Schmitt, W. L. 1921. The marine decapod Crustacea of Cali- fornia. Univ. California Pubs. Zool., vol. 23, pp. 470, pis. 1-50, text-figs. 1-164. Sivertsen, E. 1933. The Norwegian Zoological Expedition to the Galapagos Islands 1925, conducted by Alf Wollebaek. VII. Littoral Crustacea Decapoda from the Galapagos Islands. Oslo Zool. Meddel. no. 38, pp. 1-23, pis. 1-4. Smith, S. I. 1869. Notes on new or little-known species of American cancroid Crustacea. Proc. Bos- ton Soc. Nat. Hist., vol. 12, pp. 274-289. 1870. Notes on American Crustacea. No. 1. Ocy- podoidea. Trans. Connecticut Acad. Arts Sci., vol. 2, pp. 113-176. Stimpson, W. 1858. Prodromus descriptionis animalium ever- tebratorum quae in Expeditione ad Ocean- urn Pacificum Septentrionalem, a Republi- ca Federata missa, Cadwaladaro Ringgold et Johanne Rodgers ducibus, observavit et descripsit. Pars. IV. Crustacea Cancroidea et Corystoidea. Proc. Acad. Nat. Sci. Philadelphia, vol. 10, pp. 3 1-40. 1859. Notes on North American Crustacea, No. 1. Ann. Lyceum Nat. Hist. New York, vol. 7, pp. 49-93, pi. 1. 1860. Notes on North American Crustacea, in the Museum of the Smithsonian Institu- tion. No. II. Ann. Lyceum Nat. Hist. New York, vol. 7, pp. 176-246, pis. 2, 5. 1871. Notes on North American Crustacea, in the Museum of the Smithsonian Institu- tion. No. III. Ann. Lyceum Nat. Hist. New York, vol. 10, pp. 92-136. 1961] Garth: N on-intertidal Brachygnathous Crabs 159 EXPLANATION OF THE PLATE Plate I Hexapanopeus beebei, new species Fig. 1. Male holotype, dorsal view. Fig. 2. Male holotype, ventral view. Fig. 3. Female paratype, dorsal view. Fig. 4. Female paratype, ventral view. (Right third walking leg of female paratype missing) GARTH PLATE 1 NON-INTERTIDAL BRACHYGNATHOUS CRABS FROM THE WEST COAST OF TROPICAL AMERICA 14 Nematodes and Cestodes from the Australian Monitor, Varanus gouldii Horace W. Stunkard1 & Charles P. Gandal2 (Text-figures 1-6) THE parasites reported here were taken from the digestive tract of a lizard, Var- anus gouldii (Gray) , which died Novem- ber 10, 1959, in the New York Zoological Park. The animal, a mature male, was 1.58 meters long and weighed 7.8 kilograms. It was taken near Karumba, on the lower Norman River, North Queensland, June 21, 1959. When captured, it was eating a salt-water crocodile, Crocodylus porosus. Varanus gouldii is a terrestrial species which inhabits the arid interior areas of Austra- lia. Autopsy revealed several hundred nematodes with their anterior ends deeply embedded in the mucosa of the stomach. The stomach wall was greatly thickened and dark purple in color. The initial portion of the intestine, just beyond the stomach, contained about a dozen tapeworms with their scolices deeply embedded in the mu- cosa. In this area there were also several yellow- ish scars in the mucosa, presumably from the detached tapeworms. There was general peritoni- tis throughout the abdominal cavity. The liver was extremely swollen, with fibrinous deposits over the external surface. It contained numerous deep necrotic abscesses, which were light tan in color in contrast to the normal, red-colored tis- sue. Beginning about 10 cm. from the anus, the intestine was much distended for some 5 cm. and, on section, this region showed numerous scars in the mucosa with necrosis and sloughing. Tanqua tiara (von Linstow, 1879) Blanchard, 1904 The nematodes are identified as Tanqua tiara, a species described by von Linstow (1879) as 1 Research Associate, The American Museum of Natural History, Central Park West at 79th Street, New York 24, New York. 2 Veterinarian, The New York Zoological Park, New York 60, N. Y. Ascaris tiara from Varanus ornatus in Natal, South Africa. Linstow ( 1 904a) erected the genus Ctenoceph- alus to contain A. tiara, and in the species he included worms from Varanus bengalensis taken in India and from Varanus salvator, a widely dis- tributed lizard found in southeast Asia, Ceylon and the islands of the East Indies. Blanchard (1904) noted that Ctenocephalus von Linstow was a homonym of Ctenocephalus Kolenati, 1857, a genus erected to contain the siphonap- terous species, hyanae, canis, felis, etc., and he renamed the genus Tanqua, with tiara as type species. Meanwhile, Stiles had informed von Linstow that the name Ctenocephalus was preoccupied, whereupon von Linstow ( 1904b) substituted the name Tetradenos which, published later, became a synonym of Tanqua. The nomenclatorial com- bination, Tanqua tiara, was made by Stiles & Hassall (1905). Parona (1898) reported Ascaris tiara from Varanus salvator taken in Sumatra and from Varanus gouldii taken in Australia or New Guinea— the locality was not definite. Lei- per (1909) identified worms from Varanus nilo- ticus taken on the White Nile as T. tiara. Baylis (1916) suggested that the specimens described by von Linstow were from Varanus albigularis, rather than V. ornatus. He emended and ampli- fied the earlier descriptions of T. tiara; to this species he assigned specimens from V. niloticus taken at Accra on the African Gold Coast, from Varanus exanthematicus taken in northern Ni- geria, from T ropidonotus quincunciatus ( = T. asperrimus ) taken in Ceylon, and others from an unnamed species of Varanus taken in Zanzi- bar. Baylis transferred Heterakis anomala von Lin- stow, 1904, from the stomach of Tropidonotus piscator taken in Ceylon, to Tanqua as T. ano- 161 162 Zoologica: New York Zoological Society [46: 14 mala (von Linstow) and described as T. dia- dema n. sp., specimens from the South American freshwater snake, Helicops angulatus. Tanqua diadema was distinguished by the shorter length of the esophagus and the more anterior location of the vulva. Tanqua anomala is a large species and like T. diadema, has an anterior vulva. The specimens from V. gouldii agree so com- pletely with Baylis’s definition of T. tiara that their allocation is readily made. Tanqua tiara is one of the gnathostomes, spi- rurid nematodes, in which typically the life-cycle involves three hosts, a crustacean, a fish or am- phibian and the final vertebrate host, although in certain species with hard-shelled, embryon- ated eggs, a terrestrial arthropod may serve as the single intermediate host. In T. tiara the eggs are thin-shelled. According to Monnig (1947), juvenile worms migrate in the abdominal organs, particularly the livers of their final hosts. In pass- ing through the liver, they destroy tissue and leave characteristic yellow mosaic markings on the surface and burrows filled with necrotic ma- terial in the parenchyma. They wander through other organs, including the diaphragm. Adult worms penetrate the wall of the stomach, pro- ducing cavities filled with sanguino-purulent fluid and a marked gastritis. Bothridium parvum (Johnston, 1913) The cestodes from V. gouldii belong to the genus Bothridium de Blainville, 1824, whose members are parasitic primarily in pythons and boas. The type species is B. pithonis de Blainville, 1824, from the intestine of Python molurus, al- though de Blainville (1828) stated that the para- site occurred also in other pythons. Joyeux & Baer (1927) gave a historical review of the genus and the species from snakes which had been assigned to it. They suppressed Pro- dicoelia Leblond, 1836, and Solenophorus Crep- lin, 1839, as synonyms of Bothridium. Prodicoe- lia was based on P. ditrema Leblond, 1836, from the South American anaconda, listed as Boa scytale ( =Eunectes murinus Wagl.). Joyeux & Baer recognized B. ditremum (Leblond, 1836) and two species from African and Asian pythons, together with a smaller variety of the Asian species. The African species was identified as B. ovatum (syn. Solenophorus ovatus Diesing, 1850) from Constrictor hieroglyphicus (= Py- thon sebae Kuhl. ) . The larger Asian specimens were assigned to B. pithonis de Blainville, 1824, and the smaller worms from Python reticulatus Gray were described as Bothridium pithonis var. parvum nov. var. Joyeux & Houdemer (1928) listed both B. pithonis and B. pithonis var. minor Joyeux & Baer, 1926, from the intestine of Python reticulatus. In addition to pythons and boas, B. pithonis was reported from the king co- bra of Malaya, Naja hannah, by Loewen (1945) and Yeh (1956). Species of Bothridium have been described also from varanid lizards. Valenciennes (1850) recorded an unnamed species from Varanus nilo- ticus. Johnston (1913) described specimens from Varanus varius taken in North Queensland as Bothridium pythonis var. parva var. nov. and later (1916) he listed these worms as Bothridium parvum. Notified of the homonomy of B. pith- onis var. parvum Joyeux & Baer, 1927, and B. parvum (Johnston, 1913), Joyeux & Baer (1928) proposed the name B. pithonis var. minor to re- place B. pithonis var. parvum. They followed de Blainville in writing the specific name pithonis, although they stated that correct transliteration would be spelled, pythonis. Certain authors in- cluding Monticelli & Crety ( 1891 ) and Johnston (1913) have used the correct orthography. Bay- lis (1935) noted that the name given by Joyeux & Baer (1927) was B. pithonis var. parvum, not B. pithonis var. minor Joyeux & Baer, 1926, as stated by Joyeux & Houdemer (1928). Baylis observed that the specific name, minor, is avail- able to distinguish between the specimens studied by Joyeux & Baer and those of Johnston. Joyeux & Baer (1936) admitted that the specimens named B. pithonis var. minor are identical with those named earlier B. pithonis var. parvum; that the latter name disappears as a homonym and the correct designation is B. pithonis de Blain- ville, 1824; var. minor Joyeux & Baer, 1928. The descriptions of B. parvum and B. pithonis var. minor are so brief and inadequate that satis- factory comparisons between them and the pres- ent specimens can not be made. Johnston (1913) noted differences in the form of the bothria as a result of muscular contractions and that the relative thickness of the ring of longitudinal mus- cles varies inversely with the elongation of the proglottid, so certain measurements that ap- pear precise may have little value. The form of the ovary, especially its anteroposterior exten- sion, may change as the proglottid is relaxed or compressed by contraction of the longitudinal or transverse muscles. Since the descriptions of the species named above are so inadequate, it ap- pears worth while to present a more complete account of the worms from V. gouldii. They agree most closely with B. parvum and, although somewhat larger, are tentatively assigned to that species. It is noteworthy that V. varius, host of B. par- vum, is an arboreal species of rain-forests where- as V. gouldii lives in very dry interior areas and does not climb trees. The differences in locality 1961] Stunkard & Gandal: Parasites from Varanus gouldii 163 and ecology, together with difficulties the para- site would encounter in transferring from one area to the other, suggest distinct species of Bothridium, but clear-cut specific differences are not apparent. Furthermore, two possibilities ex- ist to account for the same species in different hosts and climatic regions. The species of Var- anus are carnivorous and an intermediate host containing the larvae may migrate seasonally or otherwise between desert and rain-forest re- gions. Mertens (1942) reported that monitors feed on other lizards, on birds and many types of animals; also that a specimen of V. salvator 1.65 m. long was found in the stomach of Python reticulatus, 3 m. long. Thus the parasites of each species are introduced into the digestive tracts of their predators. Another possibility presup- poses that a common ancestor of the two var- anids may have harbored the cestode which has persisted in the existing species without specific change. The problems of specific determination and intraspecific variation in parasitic flatworms were discussed by Stunkard (1957). The specimens from V. gouldii measure as much as 75 mm. in length and 2.5 mm. in great- est width, with scolices so deeply embedded in the intestinal wall that they could be removed only when released by dissection of the sur- rounding tissue. When pulled free, the bothria were filled with tissue of the host and bundles of fibers protruded from the anterior openings (Text-fig. 1). The scolex measures 3. 0-3.4 mm. in the dorsoventral axis; the bothria are 2. 1-2.4 mm. long and 1.45-1.75 mm. in width. The stro- bila is craspedote and a stained, mounted speci- men 68 mm. long comprises 244 proglottids. The original terminal proglottids had been lost. There is virtually no unsegmented neck region; the proglottids increase in size and after one-fifth of the total length, the sides are almost parallel. The reproductive organs mature early and by pro- glottid 150, one-fourth of the length from the scolex, proglottids measure 1.80 mm. in width, 0.20-0.27 mm. in length, and eggs are beginning to appear in the uterus. Further back in the stro- bila the proglottids become more elongate and the one before the last (Text-fig. 5) is 1.85 mm. long and 1.30 mm. wide. The reproductive or- gans have virtually disappeared and it is hardly more than an egg sac. The excretory vessels and nerve trunks lie in the medullary parenchyma. The two excretory ducts on either side extend the length of the strobila; the dorsal ones are smaller and slightly lateral to the larger ventral vessels which are connected by transverse ducts near the posterior end of each proglottid. The chief longitudinal nerve trunks are situated just lateral to the ven- tral excretory ducts and ventral to the dorsal excretory vessels. The testes are arranged in a single, staggered layer, 150-200 follicles in the medullary portion of each proglottid (Text-figs. 3, 6). They extend throughout the length of the proglottid, except in the median field, and in some proglottids there are about as many follicles on the lateral as on the medial sides of the excretory ducts. The fol- licles are oval, 0.032-0.055 mm. in length and 0.015-0.035 mm. in width, with their length in the transverse axis of the proglottid. Vasa effer- entia fuse to form larger ducts which unite and open into a long, much coiled, sperm-filled vas deferens, 0.010-0.015 mm. in diameter (Text-fig. 2) , which functions as a seminal vesicle. The vas deferens opens into a structure with thick mus- cular walls and lined by ciliated epithelium, formerly known as a seminal vesicle or an “Esch- trichtscher Korper,” but Fuhrmann (1931) ob- served that it seldom contains sperm and actu- ally functions in expulsion of spermatozoa. He called it a “Propulsionsblase,” literally a propul- sive bulb, and in the present worms (Text-figs. 2, 4), it is 0.10-0.15 mm. long and 0.08-0.11 mm. in diameter. The structure is continuous with and followed by the cirrus-sac, 0.15-0.20 mm. in diameter, which contains many cells that appear to be secretory, with ducts to the base of the male papilla. The common genital pore is located medially, on the ventral surface, about one-fourth of the length of the proglottid from its anterior margin and at the level of the pos- terior edge of the velum of the preceding pro- glottid. There is a transversely oval genital atrium, lined with cuticula; the opening of the vagina is immediately behind or somewhat lateral to the cirrus-sac. The vagina extends posteriad from the genital pore, below the uterus and either right or left of the uterine pore; it passes dorsal to the ovary where the posterior end expands to form a diag- onally oriented seminal receptacle 0.08-0.12 mm. in length and 0.04-0.06 mm. in diameter. The ovary is reniform to bilobed, situated on the ventral side, near the posterior end of the pro- glottid. It is 0.30-0.45 mm. wide and 0.05-0.08 mm. in anteroposterior measurement. The lat- eral portions extend dorsally, forming a depres- sion which is occupied by the seminal receptacle and the anterior portion of the shell-gland. The oviduct arises at the median posterior face of the ovary, curves laterally anteriorly and dorsally, where it receives a duct from the seminal recep- tacle; it then turns posteriorly and ventrally, re- ceives the common vitelline duct and enters the shell-gland where it coils about and expands to form the ootype. The shell-gland in one series of 164 Zoologica: New York Zoological Society [46: 14 Text-fig. 1. Scolex and early proglottids; scolex turned 90 degrees in mounting so the dorsoventral aspect appears in lateral view; fibers from intestinal wall of host in the bothria. Text-fig. 2. Transverse section through the vas deferens and propulsive bulb, showing cortical and medullary parenchyma, longi- tudinal and transverse muscles and testicular follicles. Text-fig. 3. Frontal section, showing vitelline follicles, longitudinal muscles, dorsal excretory vessels, testes, vas deferens and uterine coils above the ovary. Text-fig. 4. Transverse section through the junction of vas deferens and propulsive bulb. Text-fig. 5. Gravid proglottid, one before the last in a strobila, with accumulation of eggs in the uterus. Text-fig. 6. Transverse section, showing vitellaria, longitudinal, transverse and dorsoventral muscles, testes, excretory ducts, uterus and male papilla. Abbreviations used in Figures: de— dorsal excretory duct; Im— longi- tudinal muscles; pb— propulsive bulb; tm— transverse muscles; ts— testis follicle; ut— uterus; vd —vas deferens; ve— ventral excretory duct; vt— vitelline follicle. 1961] Stunkard & Gandal: Parasites from Varanus gouldii 165 cross sections is 0.024 ram. wide and 0.010 mm. in dorso-ventral extent; frontal sections give an anteroposterior range of 0.010-0.016 mm. From the shell-gland the uterus emerges on the dorsal side of the ovary; it coils ventrally, then dor- sally and anteriorly, passing forward on the dorsal side of the proglottid. As it becomes filled with eggs it forms loops (Text-figs. 3, 6), which overlie the ovary and cirrus-sac; the terminal re- current loop is surrounded by a heavy sphincter and opens at the uterine pore, situated near the middle of the ventral surface of the proglottid. The vitelline follicles lie in the cortical area (Text-figs. 3, 6) and like the testes are interrup- ted in the median field. They are oval, 0.016- 0.035 mm. in length and 0.010-0.022 mm. in width. On either side, ducts from the follicles unite to form larger ducts which pass mediad in the posterior part of the proglottid and join to form the common vitelline duct which opens into the oviduct just before it enters the shell-gland. As the proglottids become gravid, the reproduc- tive organs regress and the uterus becomes a large, egg-filled sac (Text-fig. 5). The eggs are thin-shelled, 0.075-0.080 mm. in length and 0.045-0.050 m. in width. The life-cycle is not known for any species of Bothridium but certain significant data are available. Solomon (1932) obtained develop- ment of B. pythonis to the procercoid stage in Cyclops viridis. Meggitt (1931) reported plero- cercoid larvae from the mesentery, lung and aorta of Naja naja in Burma. Baylis (1933) found encysted plerocercoid larvae, believed to be those of a species of Bothridium , in the liver of a snake ( Bungaris fasciatus) in Java. Baylis ( 1935) reported similar larvae from cysts on the external wall of the intestine and mesentery of an Australian water-rat ( Hydromys chrysogaster) taken at Cromarty, North Queensland. On mor- phological grounds he assigned the plerocercoids to the genus Bothridium, but stated that specific identification is at present impossible. From pres- ent data it is probable that crustaceans are the first intermediate hosts and harbor procercoid stages, that fishes serve as second hosts and har- bor the plerocercoids which migrate and encyst in the visceral organs of reptiles and small mam- mals that ingest infected fishes, and that the stro- bilate stage develops in the final predator. Summary Nematodes and cestodes are reported from the digestive tract of a monitor lizard, Varanus gouldii, which was captured in North Queens- land, Australia, and died in the New York Zoolo- gical Park. The nematodes are identified as Tan- qua tiara (von Linstow, 1879) Blanchard, 1904, and the cestodes are referred tentatively to Both- ridium parvum (Johnston, 1913) Johnston, 1916, although the description of that species and of Bothridium pithonis de Blainville, 1924; var. minor Joyeux & Baer, 1928, are so incom- plete that final determination is equivocal. NOTE: Specimens of Tanqaa tiara and Both- ridium parvum are in the collection of the American Museum of Natural History, New York. Literature Cited Baylis, H .A. 1916. The nematode genus Tanqua R. Blanch- ard. Ann. Mag. Nat. Hist., (8) 17: 223- 232. 1933. On some parasitic worms from Java, with remarks on the acanthocephalan genus Pallisentis. Ann. Mag. Nat. Hist., (10) 12: 443-449. 1935. The plerocercoid larva of Bothridium (Cestoda). Ann. Mag. Nat. Hist., (10) 16: 482-485. Blanchard, R. 1904. Tanqua n. g., remplagant Ctenocephalus von Linstow. Arch. Parasitol., 8: 478. Blainville, H. de 1828. Article “Vers.” Dictionnaire des Sciences Naturelles: p. 609. Cited after Joyeux & Baer, 1927. Fuhrmann, O. 1931. Dritte Klasse des Cladus Platheminthes. Cestodea. In Handbuch der Zoologie, Kiikenthal u. Krumbach, 2: 141-416. Ber- lin u. Leipzig. Johnston, T. H. 1913. Report on the Cestoda and Acanthoce- phala of North Queensland, with plates xv to xvii. Australian Institute of Trop. Medi- cine, Report for the year 1911: 75-96. 1916. A census of the endoparasites recorded as occurring in Queensland, arranged under their hosts. Proc. Roy. Soc. Queensland, 28: 31-79. Joyeux, Ch., & J. G. Baer 1927. Recherches sur quelques especes du genre Bothridium de Blainville, 1824 (Diphyllo- bothriidae). Ann. Parasitol., 5: 127-139. 1928. Rectification of nomenclature. Ann. Para- sitol., 6: 144. 1936. Notices helminthologiques. Bull. Soc. Zool. France, 60: 482-501. Joyeux, Ch., & E. Houdemer 1928. Recherches sur la fauna helminthologique de 1’Indochine (Cestodes et trematodes). Ann. Parasitol., 6: 27-58. Leiper, R. T. 1909. Helminthes contained in Dr. C. M. Wen- yon’s collection from the Sudan. Third 166 Zoologica: New York Zoological Society [46: 14: 1961] Report, Wellcome Research Labs, at the Gordon Memorial College, Khartoum. Dept. Educat. Sudan Gov’t. Egypt, pp. 187-199. Linstow, O. F. B. von 1879. Helminthologische Untersuchungen. Jahr- esb. Ver. Vaterl. Naturk., Wurtemberg, 35: 313-342. 1904a. Nematoda in the collection of the Colom- bo Museum. Spolia Zeylanica, 1: 91-104. 1904b. Beobachtungen an Nematoden und Ces- toden. Arch. Naturg., 70 (7): 379-383. Loewen, S. L. 1945. A new host record for the cestode Both- ridium pithonis de Blainville, 1828. Trans. Kansas Acad. Sci., 48: 107-108. Meggitt, F. J. 1931. On cestodes collected in Burma. Part II. Parasitol., 23: 250-263. Mertens, R. 1942. Die Familie der Warane (Varanidae). Abhandl. Senckenberg. Naturf. Gesellsch. Frankfurt a. Main. pp. 4-74; 238-391. Monnig, H. O. 1947. Veterinary Helminthology and Entomol- ogy. 3rd Edit. London. Monticelli, F. S., & C. Crety 1891. Richerche intomo alia sottofamiglia So- lenophorinae Monti. Crety. Mem. reale Accad. Sci. Torino, 41: 381-402. Parona, C. 1898. Elminti raccolti dal Dott. Elio Modigliani alle isole Mentawei, Engano e Sumatra. Ann. Mus. Civ. Storia Nat., Genova, 19: 102-124. Solomon, S. G. 1932. On the experimental development of Bothridium (— Solenophorus) pythonis de Blainville, 1824, in Cyclops viridis Jurine, 1820. Jour. Helminth., 10: 67-74. Stiles, C. W. & A. Hassall 1905. The determination of generic types, and a list of roundworm genera, with their origi- nal and type species. Bull. No. 79, Bureau Animal Industry, U.S. Dept. Agric., 150 pp. Stunkard, H. W. 1957. Intraspecific variation in parasitic flat- worms. Syst. Zool., 6: 7-18. Valenciennes, A. 1850. Note sur un helminthe rendu par un varan du Nil (Lacerta nilotica Linne; Varanus niloticus Dumeril). Compt. Rendu Soc. Biol. Paris, 1: 184-185. Yeh, L. S. 1956. On some helminths from a king cobra in Malaya, including Occipitodontus edesoni n. sp., and Ophiotaenia kuantanensis. Jour. Helminth., 30: 211-216. 15 Urinary Amino Acids of Non-human Primates Jack Fooden Department of Zoology, The University of Chicago (Plates I-III; Text-figures 1-4) I. Introduction SMALL but measurable quantities of free amino acids are regularly excreted in the urine of all mammals. Individual patterns of urinary amino acid excretion appear to be genetically determined and related to phylogen- etic position. The present paper is a comparative study of the urinary amino acid excretion pat- terns of non-human primates. The urinary amino acid excretion of humans has been the subject of normative and clinical studies andof genetic research. Berry (1953), Stein (1953) and Eades & Pollock (1954) determined the amino acid concentrations of urine speci- mens from normal human individuals. Smith (1958) reported on the association of unusual urinary amino acid patterns with various human diseases. The extensive literature on the inheri- tance of human amino acid excretion patterns has been reviewed by Harris (1959). Only two previously published papers pertain to the urinary amino acid excretion of non- human primates. Datta & Harris (1953) in- cluded one rhesus monkey in their survey of mammalian urinary amino acid patterns. Gartler, Firschein & Dobzhansky (1956) reported on the amino acid excretion of 48 apes— 2 gibbons, 3 orang-utans, 37 chimpanzees and 6 gorillas. II. Materials and Methods The general procedure followed in the present study is as follows. Specimens of urine were col- lected from 112 of approximately 200 primates kept at the Chicago Zoological Park at Brook- field, Illinois. A sample of each specimen was desalted and subjected to two-way paper chrom- atography. The amino acids separated on each urinary chromatogram were converted into vis- ible spots by reaction with a color reagent. The quantity of amino acid represented by each spot was measured by comparing its optical density with the optical density of spots produced by known quantities of the corresponding pure amino acid. As shown in Table I, the 112 animals studied belong to 7 of the 10 living families of non- human primates. Sixteen belong to 3 families of the Suborder Prosimii; 96 belong to 4 families of the Suborder Anthropoidea. The species and sex of the individuals studied are given in Table II. The classification used throughout is that of Simpson (1945), except for the generic sub- division of the Callithricidae, which follows Hershkovitz (1958). All but two of the urine specimens were ob- tained by confining animals individually in a metabolism cage. Specimens from 106 of the 112 animals were collected over a 24-hour per- iod of confinement. Specimens from the other Table I. Numbers of Genera of Primates Studied Compared with Numbers of Living Genera Family Living Genera Genera Studied Individuals Studied Suborder Prosimii Tupaiidae . 6 1 7 Lemuridae . 6 1 2 Indriidae . 3 Daubentoniidae . . 1 Lorisidae . 6 3 7 Tarsiidae 1 Total . 23 ~ 5 16 Suborder Anthropoidea Cebidae 12 6 27 Callithricidae . . . . 3 2 8 Cercopithecidae . 16 9 51 Pongidae . 5 4 10 Total . 36 21 96 167 168 Zoologica: New York Zoological Society [46: 15 Table II. Species and Sex of Animals Studied Species j Males Females Total Common Name Family Tupaiidae Urogale everetti 2 5 7 Tree shrew Family Lemuridae Lemur fulvus 1 2 Lemur Family Lorisidae Nycticebus coucang 2 2 4 Slow loris Perodicticus potto 1 0 1 Potto Galago crassicaudatus 1 1 2 Galago Family Cebidae Cacajao rubicundus 0 2 2 Uakari Pithecia monacha 0 1 1 Said Cebus albifrons 4 1 5 Capuchin Cebus apella 3 1 4 Cebus capucinus 1 1 2 Cebus nigrivittatus 1 0 1 Saimiri sciureus 2 1 3 Squirrel monkey A teles belzebuth 0 3 3 Spider monkey Ateles geoffroyi 1 1 2 Lagothrix cana 0 1 1 Woolly monkey Lagothrix infumata 1 0 1 Lagothrix poppigii 21 0 2 Family Callithricidae Callithrix jacchus 0 4 4 Marmoset Saguinus leucopus 0 1 1 Tamarin Saguinus nigricollis 1 1 2 Saguinus oedipus 1 0 1 Family Cercopithecidae Macaca irus 0 3 3 Macaque Macaca maura 1 1 2 Macaca mulatto 0 2 2 Macaca nemestrina 0 3 3 Macaca nemestrina X M. silenus 1 0 1 Macaca radiata 2 4 6 Macaca silenus 1 0 1 Cercocebus torquatus 1 1 2 Mangabey Papio cynocephalus 7 1 8 Baboon Comopithecus hamadryas 1 0 1 Hamadryas baboon Mandrillus leucophaeus 0 2 2 Drill Mandrillus sphinx 1 1 2 Mandrill Cercopithecus aethiops 1 2 3 Guenon Cercopithecus diana 0 1 1 Cercopithecus I’hoesti 0 2 2 Cercopithecus mitis 0 1 1 Cercopithecus neglectus 1 2 3 Cercopithecus talapoin 1 0 1 Erythrocebus patas 1 0 1 Patas monkey Presbytis entellus 2 1 3 Langur Presbytis obscurus 1 1 2 Colobus polykomos 1 0 1 Guereza Family Pongidae Hylobates lar 1 2 3 Gibbon Pongo pygmaeus 1 1 2 Orang-utan Pan troglodytes 1 3 4 Chimpanzee Gorilla gorilla 1 0 1 Gorilla Total 51 61 112 ]Onc, unavailable for re-examination, may be L. injumata. 1961] Fooden: Urinary Amino Acids of Non-human Primates 169 6 individuals— 1 orang-utan, 4 chimpanzees and 1 gorilla— were collected from a single urinary discharge. To assess individual constancy of amino acid excretion, a second urine specimen was obtained from 12 of the 106 animals origin- ally represented by 24-hour specimens. Eleven of the 12 subsequent specimens were 24-hour col- lections; the other was from a single urination. A third 24-hour specimen was obtained from 1 animal. Following the collection of each specimen its volume was measured, and after thorough mix- ing a sample was taken for amino acid determin- ation. Because it was necessary to store the samples for a time between collection and analy- sis, isopropyl alcohol was added routinely as a preservative. The amount added was such that the original volume of urine equaled 90% of the final volume. The samples were stored at 4°C. for an average of approximately 8 months. The effect of storage on the amino acids of hu- man urine has been studied by Stein (1953). He detected no change resulting from storage ex- cept for an increase in glutamic acid concentra- tion. Stein attributed this increase to the pres- ence in urine of a labile conjugate of glutamic acid which during storage is converted to the free amino acid. Before chromatography each of the stored samples was desalted in a Research Specialties Co. electrolytic desalter in order to improve chromatographic resolution (Smith, 1958). A measured quantity, ranging from 20 to 200 /xl., of each desalted sample was applied as a 7-10 mm. spot centered 2.5 cm. from one corner of a 23 X 28 cm. sheet of Whatman No. 1 filter paper. Five or six sheets spotted with urine were mounted on a chromatography rack together with 5 or 6 sheets spotted with various concen- trations of a known mixture of commercially- obtained amino acids. In each instance 1 1 sheets, including unknowns and standards, were chro- matographed simultaneously on one rack. Two chromatograms were made of each urine sample, usually on separate racks. The first chromatographic solvent, which was run the length of the sheet, was a mixture (30:10:10:1) of 2,6-lutidine (Eastman Prac- tical, 95 % ) , absolute ethanol, distilled water and diethylamine. The second solvent, run at right angles to the first, was a mixture (100:20: 0.6) of liquefied phenol (Mallinckrodt Gilt Label, 88%), distilled water and ammonia. As recommended by Block, Durrum & Zweig (1958), a beaker containing 1% sodium cyan- ide in water was placed in the tank during phenol development. Following phenol development the sheets were allowed to air-dry thoroughly. Color was developed by dipping the dried sheets in a 0.2% solution of ninhydrin in acetone and subsequently heating them in a water-saturated atmosphere for 15 minutes at 76 °C. Results of the technique are illustrated by PI. I, Fig. 1, which shows the typical resolution of a known mixture of amino acids. Within 15 hours after ninhydrin treatment, maximum optical densities of the developed spots were determined by means of a Welch Densichron transmission densitometer. Quanti- tative estimates of urinary amino acid concen- trations were derived by interpolation from loga- rithmically plotted curves for standards devel- oped on the same rack as the urine samples. Concentration values reported in this paper are the average of duplicate determinations. The distribution of differences between pairs of de- terminations is shown in Table III. Table III. Distribution of Differences between Duplicate Urinary Amino Acid Determinatons Concentration Log Difference between Duplicate Determinations1 CL, c n < B 03 5 s Cumulative frequency in percent 0> .3 o O a 1 3 5 < <0.1 44 61 57 47 62 50 60 55 66 68 57 <0.2 70 90 82 76 80 67 69 75 93 78 78 <0.3 85 97 89 84 85 77 77 84 97 85 86 <0.4 94 100 94 93 88 84 85 89 99 88 91 <0.5 95 100 94 94 91 86 89 92 100 92 93 <1.0 100 100 100 100 99 97 94 98 100 97 99 <1.5 100 100 100 100 100 100 100 100 100 100 100 Concentration unit — 10-5 mmoles/ml. 170 Zoologica: New York Zoological Society [46: 15 III. Results Concentrations shown in Table IV are ex- A total of 10 urinary amino acids were chro- Pressed logarithmically based on a concentra- matographically identifiable in the specimens u°n umt °f 1 O'5 mmoles of ammo acid per ml. studied. Traces of unidentifiable ninhydrin-posi- of urine‘ In the interpretation of differences m tive substances were detected in some specimens, concentration, a difference of 1 .0 between con- The clearly identified amino acids are: glutamic centration logs shown for two specimens obvi- acid, aspartic acid, serine, glycine, glutamine, ously represents a tenfold deference in anthme- lysine, taurine, threonine, alanine and hydroxy- tically expressed concentration. As indicated in proline. Table IV gives urinary concentrations a footnote to Table IV, concentration logs of of these 10 amino acids in specimens from the zero are assigned to amino acids absent from a 112 individuals included in the study. The body particular specimen or present in concentrations weight of each animal and the volume of its 24- insufficient to yield a detectable chromatographic hour specimen are also given. spot with the volume of urine applied. Table IV. Amino Acid Concentration of Urine Specimens from 112 Non-human Primates Individual Species X 310 0.9 1.9 0.6 2.8 1.3 0.3 0.6 1.0 1.2 0.5 116 Mandrillus leucophaeus .... 2 14.5 310 1.8 2.5 0.7 2.4 0.5 0.4 2.1 0.8 1.5 0 166 M. leucophaeus 2 2.3 275 1.4 1.7 1.1 1.9 2.8 0.9 1.0 0.7 1.0 1.3 110 M. sphinx 2 4.5 521 1.2 2.5 1.1 2.7 2.2 0 0 0.7 1.6 0 111 M. sphinx 8 4.5 n.d.2 1.2 2.0 1.5 2.4 2.3 0 0 0 1.4 1.4 78 Cercopithecus aethiops 2 3.4 59 0.5 1.4 0 2.9 0.8 0.8 0.8 0.3 1.1 0 79 C. aethiops 2 3.9 250 1.0 1.7 0.2 2.4 0.7 0.8 0 0 1.0 0 86 C. aethiops 8 7.9 250 0.5 2.1 0 2.8 1.2 1.1 0 1.0 1.0 0 172 Zoologica: New York Zoological Society [46: 15 Table IV. Amino Acid Concentration of Urine Specimens from 112 Non-human Primates ( cont .) _2 Log of concentration in 10-5 mmoles/ml.1 Individual Species Sex Body wt. (kg Vol. of 24-hr Specimen (rr Aspartic Glutamic Serine Glycine Glutamine Lysine Taurine Threonine Alanine Hydroxy- proline 87 C. diana, 25 Sep. 58 -$ 2.7 193 1.4 3.0 0 2.8 2.2 0.5 0 1.3 2.3 0 130 (Indiv. No. 87), 21 Oct. 58. • $ 2.7 202 2.0 2.8 1.3 2.9 2.8 0 0 1.6 2.3 1.1 73 C. Vhoesti • 9 2.3 280 1.1 1.2 0 2.7 1.1 1.1 0 0.5 1.2 0 74 C. Vhoesti •2 4.3 190 1.2 1.5 0.3 2.8 1.7 0.4 0 0 1.5 0 81 C. mitis, 23 Sept. 58 -2 4.3 400 0.7 1.1 0 3.0 1.2 0 0.5 0 1.1 0.6 131 (Indiv. No. 81), 21 Oct. 58. . •2 n.d.2 250 1.1 1.2 0 3.5 2.2 0.7 2.1 0.8 1.5 0.8 80 C. neglectus ■8 6.6 290 1.0 2.6 0 2.7 1.2 0.9 0.6 0.4 1.9 0 82 C. neglectus .2 3.9 275 1.3 1.9 0 2.5 0 0 0 0 2.0 0 85 C. neglectus, 25 Sep. 58 -2 4.3 175 0.4 2.7 0 2.5 1.8 1.1 0 1.5 2.2 1.4 164 (Indiv. No. 85), 31 Jan. 59. . • 2 5.4 s.u.6 0.8 2.4 0.9 2.0 0 1.0 1.7 1.1 1.3 0.5 107 C. talapoin, 6 Oct. 58 ■ 8 1.4 76 1.1 1.4 0 2.2 1.0 0.3 0.7 0.3 1.3 0 127 (Indiv. No. 107), 20 Oct. 58. ■ 8 n.d.2 92 1.1 1.5 0 2.3 1.3 0 0 0 1.3 0 120 Erythrocebus patas ■ 8 13.6 86 1.4 1.5 0.9 2.0 0.6 1.3 1.6 1.2 0.8 0 113 Presbytis entellus ■ 8 11.7 188 0.7 1.8 2.3 2.8 2.4 1.4 2.0 1.9 1.2 0.5 114 P. entellus • 2 10.4 468 0.9 1.5 1.0 2.5 0 0 2.1 0.7 1.0 0.6 115 P. entellus ■8 15.4 (650)3 0.6 1.8 2.0 3.1 2.0 1.2 1.9 1.5 1.3 0.5 89 P. obscurus • 2 4.8 134 1.2 1.4 2.0 2.6 2.7 1.1 2.2 2.1 1.4 1.2 90 P. obscurus ■8 7.2 200 1.5 2.5 0.6 2.7 2.7 0.9 2.1 1.4 1.7 0 54 Colobus polykomos, 12 Sep. 58 ■8 13.6 425 0.4 1.3 0.3 2.4 0.3 0 0 0 0.9 0 126 (Indiv. No. 54), 20 Oct. 58 . . 8 n.d.2 283 0.7 1.6 0.6 1.9 1.1 0.9 0.8 0.9 0.7 0.7 150 Hylobates lar 2 4.5 50 1.0 1.5 1.4 2.0 1.2 1.4 1.3 0.8 1.6 0 151 H.lar 8 6.4 172 0.9 1.0 1.2 2.3 1.3 0.9 0 0 1.3 0 152 H. lar 2 5.0 153 0.8 0.9 1.1 1.7 1.2 0.6 1.0 0.6 0.9 0 123 Pongo pvgmaeus 2 15.9 n.d.2 1.1 1.6 0.4 2.4 0.9 0.3 1.2 0.5 1.5 0.9 153 P. pvgmaeus 8 24.9 s.u.6 0.5 0.4 0.6 1.5 0.1 0.6 1.2 0.3 0.7 0 159 Pan troglodytes 2 12.7 s.u.6 0.5 0.5 0.3 1.3 0.9 0 2.0 0.1 0.1 0 160 P. troglodytes 8 14.5 s.u.6 0.3 0.6 0.3 1.4 1.3 0.1 1.5 0.8 -0.1 0.6 161 P. troglodytes 2 15.9 s.u.6 0.5 0.8 0.4 1.6 0 0.2 2.1 0.2 0.7 0.1 162 P. troglodytes 2 10.4 s.u.6 1.2 1.3 1.0 1.5 1.3 0.9 2.2 0.9 0.9 0 154 Gorilla gorilla 7 8 n.d.2 s.u.6 1.0 1.6 0.9 2.2 1.4 0.8 1.0 0.8 1.0 0 ]Zero values are assigned to amino acids whose concentration was insufficient to yield a detectable chroma- tographic spot with the volume of urine applied. 2Not determined. Approximate volume. 4Specimen from this species is composite collection of 24-hour output of two female uakaris. 5Cebid individual fed meatless diet. 6Specimen represents single urinary discharge. Specimen from this animal siphoned off previously scrubbed floor of exhibition cage. IV. Discussion A. Urinary Output and Body Weight In Text-fig. 1 generic mean logs of 24-hour urinary output in ml. (U24hr.) are plotted against corresponding mean logs of body weight in grams (B). Superimposed on these points is the line U 24 hr. = 0.1536 B0-82, which is equivalent to Uj hr. = 0.0064 B0-82, the equation found by Adolph (1943; 1949) to express the constant relationship between urinary output and body weight in mammals of diverse orders. Also shown on the graph are values from Adolph for man and for three non-primate genera. The primate data apparently conform to the general mammalian relationship. This casts doubt on the assumption of Gartler, Firschein & Dobzhansky (1956) that the low urinary creatinine concentration of gibbons, orang- utans, chimpanzees and gorillas is compensated for by a high daily volume of urine per unit of body weight. B. Successive Specimens from Individuals The amino acid concentrations of urine speci- 1961] Fooden: Urinary Amino Acids of Non-human Primates 173 mens from the 1 2 individuals sampled more than once are compared graphically in Text-fig. 2. These graphs show the successive specimens of each individual to be generally similar in urinary amino acid pattern. Taurine concentrations, however, do vary considerably in many cases, for 5 of the 12 re-sampled individuals taurine concentration log differences between successive specimens are 1.0 or greater. The general con- stancy of amino acid pattern in successive speci- mens from these non-human primates is in ac- cord with the constant urinary amino acid pat- terns reported for human individuals by Berry (1953) and Harris (1953). Plates I and II, Figs. 2-4, are photographs of chromatograms comparing the successive specimens of 3 re- sampled individuals. C. Effect of Diet The results of previous studies (Sutton, 1951; Gartler, Firschein & Dobzhansky, 1956; Smith, 1958) indicate that the urinary amino acid ex- cretion of humans is generally unaffected by normal variations in diet. In the course of the present study an opportunity to observe the effect of certain dietary factors on the urinary amino acid excretion of non-human primates was afforded by the housing and feeding arrange- ment at the Brookfield Zoo. Prosimians and most South American monkeys are kept in the Small Mammal House, and old world monkeys are kept in the Primate Building. Primates in the Small Mammal House are fed fruit, vegetables, bread and raw ground horsemeat fortified with a vitamin-mineral supplement. The basic diet in the Primate Building includes only fruit, vege- tables and bread. Meat is not given in the Pri- mate Building, although animals of the follow- ing species do receive a raw egg daily: Cebus albifrons, Cebus apella, Macaca nemestrina X M. silenus, Macaca silenus, Cercocebus torqua- tus, Comopithecus hamadryas, Mandrillus leu- cophaeus, Mandrillus sphinx, Cercopithecus diana, Erythrocebus patas. The phylogenetic separation of housing facilities outlined above is not maintained completely. Of the South American monkeys, 4 of 12 capuchins (Cebus) and 3 of 5 spider monkeys (A teles) are kept in the Primate Building. The 7 cebids housed in the Primate Building receive a diet identical with that of the old world monkeys; these animals are individually identified in Table IV. Text-fig. 1. Urinary output in relation to body weight. Closed circles=means for primate genera studied (number of specimens in parentheses); open circles=values for man and non-primates, from Adolph (1943). Superimposed straight line also from Adolph. 174 Zoologica: New York Zoological Society [46: 15 In Table V the capuchins and spider monkeys fed meat and those not fed meat are compared with respect to their frequency distributions of urinary amino acid concentrations. The two groups differ consistently only in distribution of threonine concentrations. The feeding of forti- fied meat thus appears to have no general effect on the urinary amino acid excretion patterns of these cebids. A similar comparison of urinary amino acid concentrations of species fed eggs and of those not fed eggs reveals that this dietary factor also is without apparent effect on urinary amino acid excretion. D. Phylogenetic Comparisons Examination of Table IV reveals no consistent differences in concentration between species within a given genus. Accordingly, for purposes of phylogenetic analysis, attention is centered on comparison of the urinary amino acid patterns of genera and families. Generic means and standard errors of urinary amino acid concentrations are represented graphically in Text-fig. 3. These means are based on data in Table IV for the 106 animals repre- sented by 24-hour specimens. As there appear to be no sexual differences in urinary amino acid concentrations, data for both males and females are included in the computation of each generic mean. Data for the 6 animals represented only by specimens passed at a single urination are not included; the concentration level of these speci- mens may differ from that of the 24-hour speci- mens. Individuals from which two or more specimens were collected are represented by average concentrations. Inspection of Text-fig. 3 reveals that the 10 amino acids identified differ considerably in inter-generic uniformity of concentration. As- partic acid, glutamic acid, glycine and alanine maintain relatively constant concentrations in the urine of the 24 genera studied. Inter-generic variations of mean concentration of these 4 amino acids are small relative to intra-generic variations, represented by the standard errors. The other 6 amino acids show a relatively greater degree of inter-generic variation of mean con- centration. For the 4 amino acids of relatively constant concentration in primate urine, over-all concen- Text-fig. 2. Comparison of amino acid concentrations in successive specimens from 12 non-human primate individuals. 1961] Fooden: Urinary Amino Acids of Non-human Primates 175 Table V. Effect of Diet on Urinary Amino Acid Excretion of 17 Capuchins and Spider Monkeys Diet Meat Meatless Meat Meatless Meat Meatless Meat Meatless Meat Meatless Meat Meatless Meat Meatless Meat Meatless Meat Meatless Meat Meatless Log of amino acid concentration in 10-5 mmoles/ml. 0.5 1.0 1.5 2.0 2.5 0.5 1.0 1.5 2.0 2.5 <0.5 to to to to to <0.5 to to to to to 0.9 1.4 1.9 2.4 2.9 0.9 1.4 1.9 2.4 2.9 1 1 6 Cebus Aspartic Acid Ateles 1 1 2 1 1 1 1 1 . 1 3 Glutamic Acid 4 2 . 2 2 1 1 1 1 1 3 Serine 3 1 1 1 2 1 1 1 1 . • 1 Glycine 2 5 1 1 1 1 1 . 2 1 1 3 Glutamine 3 1 1 1 2 1 1 2 . 1 5 2 Lysine 1 1 1 1 2 . 3 . • 1 1 2 Taurine 1 3 1 1 1 1 2 1 - . 1 2 6 Threonine 1 1 2 1 1 • i 1 1 2 5 Alanine 1 1 1 1 3 . 2 1 1 4 Hydroxyproline 3 2 1 2 1 2 1 tration means may be taken as approximations of values for the order as a whole. These over-all means are given in Table VI. Glycine, with an average concentration of 2.40 pmoles per ml., is the most prominent amino acid in primate urine. Glutamic acid has the next highest average concentration, 0.63 /xmoles per ml., about one- fourth the mean glycine concentration. It should be noted, however, that, because of the previ- ously mentioned effect of storage, all of the glutamic acid measured may not originally have been present as the free amino acid. Urinary alanine and aspartic acid are much lower in mean concentration. The average alanine con- centration is about one-tenth that of glycine; the average aspartic acid concentration, about one- twentieth. The urinary glutamic acid and aspartic acid concentrations reported above for non-human primates are much higher than human urinary concentrations of these two amino acids. This agrees with one of the principal findings of Gart- ler, Firschein & Dobzhansky (1956) , who found glutamic acid and aspartic acid concentrations in ape urine to be significantly higher than in human urine. The high urinary glycine concen- trations found in the present study also agree with the results of Gartler, Firschein & Dob- zhansky; they do not agree with the apparent absence of glycine from the urine of the single rhesus monkey studied by Datta & Harris (1953). The 6 amino acids showing relatively great inter-generic variation of concentration— name- ly, serine, glutamine, lysine, taurine, threonine 176 Zoologica: New York Zoological Society [46: 15 I 1 1 1 _ o ^,2 £ E o 0 o 0 © o 0 0 0 © 0 0 © 0 ' o © © © * 0 0 © 0e © © 0 0 • 0 o o O O 0 O o O • 0 0 © o © O O © © © O o oooo OOOO o o ooo o • ooo o 0 o O0 O 0 0 0 0 o 0 0 0© 0 © © 0© o o o o © o 0 © © o 0 0 o 0e 0 © 0 © 0 © 0 0 0 O © 0 0 O • 0 0 © ©Q O © O 0 © 0 0 © o o © © O 0 • 0 0 © o© 0 0 © 0 0 © 0 © O 0 © 0 © o © 0 o o o© o o o O o o 0 o 00 e © © 0 © 0 0 © 0© © © © 0 © o O 0 O o O O o o • o o o OO O o o O o o o o O o o o o o © o 0 0 0O 0 © © 0 • o o o O o o o o o o Urogate (7) Lemur (2) Nycticebus (4) Perodicticus (I) Galago (2) Cacajao (1) Pithecia (I) Cebus (12) Saimiri (3) A teles (5) Lagothrix (4) Callithrix (4) Saguinus (4) Macaca(!8) Gercocebus (2) Papio (8) Comopithecus (1) Mandrillus (4) Cercopithecus (II) Erythrocebus (I) Presbytis (5) Colobus (I) Hylobates (3) Pongo (I) © o G o o o O Log of concentration in I0*5 mmoles/ml.: o.O 0.1250.25 0.5 1.0 1.5 2.0 30 Text-fig. 3. Comparison, by genera, of mean amino acid con- centrations of urine specimens from 106 non-human primates. Area of each circle proportional to generic mean; bar within circle represents diameter of standard error of mean; horizontal lines separate data for each family. and hydroxyproline— are those among which evi- dence is to be sought concerning phylogenetic trends in primate urinary amino acid excretion. Generic frequency distributions of concentra- tions of these 6 amino acids are compared graphically in Text-fig. 4, from which are omitted genera represented by only a single urine speci- men. Generic excretion tendencies indicated by the frequency distributions are presented in Table VII. Phylogenetically these may be sum- marized as follows : At the subordinal level, prosimian genera all tend to be low in urinary glutamine. At the fam- ily level, tupaiids, as represented by the single species available for study, appear to be high in urinary serine, lysine and taurine. Cebids as a family are uniquely high in urinary hydroxy- proline; among cebid genera, Cebus, represented by 12 individuals of 4 species, clearly tends to be high in all 6 generically variable urinary amino acids. The two callithricid genera studied tend slightly to be high in urinary taurine. Cer- copithecids generally are low in the generically variable urinary amino acids other than gluta- mine; most extreme in this respect is the genus Cercopithecus. Papio and Presbytis, however, are exceptional; Papio is high in urinary taurine, and Presbytis is high in serine, lysine, taurine, and threonine. 1961] Fooden: Urinary Amino Acids of Non-human Primates 111 Table VI. Mean Concentration of Aspartic Acid, Glutamic Acid, Glycine and Alanine in Urine Specimens from 106 Non-human Primates Mean and Mean Standard error of Concentration Amino Acid Concentration Logs1 in p,moles/ml. Aspartic acid 1.05 ± 0.04 0.11 Glutamic acid 1.80 ± 0.04 0.63 Glycine 2.38 ± 0.04 2.40 Alanine 1.41 ± 0.04 0.26 1Unit of concentration = 10-5 mmoles/ml. The hydroxyproline content of cebid urine is distinctive. Twenty-two of 26 cebid specimens have hydroxyproline concentration logs of 1.5 or greater; only 2 of 79 non-cebid specimens have hydroxyproline concentrations this high. The mean urinary hydroxyproline concentration log for cebids is 1.83 ± 0.11 (0.66 /xmoles per ml. ) ; the corresponding mean for non-cebids is 0.27 ± 0.08 (0.02 mmoles per ml.) . The hydroxy- proline spot characteristic of cebid urinary chromatograms is illustrated in PL III, Fig. 5, which shows photographs of one chromatogram from each of the 6 cebid genera studied. These chromatograms also illustrate the dietary inde- pendence of cebid urinary hydroxyproline excre- tion. Although only the A teles chromatogram is from an individual not fed meat, its hydroxy- proline spot is not different from the others. V. Summary Urinary amino acid excretion patterns have been studied chromatographically in specimens from 112 primate individuals representing 7 of 10 non-hominid families. Analysis of successive specimens from 12 animals indicates that uri- nary amino acid patterns of primate individuals Urogale Lemur Nycticebus Galago Cebus Saimiri Afeles Logothrix Callithrlx Saguinus Macaco Cercocebus Papio Mandril lus Cercopithecus Presbytis Hylobates Text-fig. 4. Distributions of urin- ary serine, glutamine, lysine, taur- ine, threonine and hydroxyproline concentrations in 16 non-human primate genera. 178 Zoologica: New York Zoological Society [46: 15 Table VII. Urinary Amino Acid Excretion Tendencies of Primate Genera Suborder Amino Acid Concentration Level1 of Genus Family I Hydroxy- Genus Serine Glutamine Lysine Taurine Threonine proline Prosimii Tupaiidae Urogale + 0 + + (+) 0 Lemuridae Lemur 0 (0) (0) 0 Lorisidae Nycticebus 0 (0) (0) (+) 0 Galago (+) (0) + + (0) Anthropoidea Cebidae Cebus 4“ + + + + + Saimiri ( + ) (+) + (+) + A teles 0 + (0) + La got hr ix (+) (0) (+) + Callithricidae Callithrix (0) + (+) 0 Saguinus 0 0 + 0 Cercopithecidae Macaca 0 0 0 (0) 0 Cercocebus (+) (0) (0) Papio 0 (+) 0 + 0 0 Mandrillus (+) + 0 (0) (0) 0 Cercopithecus 0 (+) 0 0 0 0 Presbytis + + (+) + + 0 Pongidae Hylobates (+) (+) (0) 0 0 H — high; O — low; ( ) = weak tendency. remain relatively constant through time. Dietary variations appear in general to be without effect on urinary amino acid excretion. Daily urinary output bears a constant exponential relationship to body weight. The primate genera studied tend to be uni- form in urinary concentration of glycine, glu- tamic acid, alanine and aspartic acid. Average concentrations of these 4 amino acids, expressed in /imoles per ml. of urine, are as follows: gly- cine, 2.40; glutamic acid 0.63; alanine, 0.26; aspartic acid, 0.11. Inter-generic variations of concentration are relatively great for gluta- mine, serine, lysine, taurine, threonine and hy- droxyproline. Prosimians in general are low in urinary glutamine. Tupaiids, represented in the study by one species, appear to be high in urinary serine, lysine and taurine. Cebids are uniquely high in urinary hydroxyproline; the genus Cebus also is high in serine, glutamine, lysine, taurine and threonine. Cercopithecids tend to be low in inter-generically variable uri- inary amino acids other than glutamine; Papio, however, is high in taurine, and Presbytis is high in serine, taurine and threonine. Acknowledgments During the course of this investigation the author was a National Science Foundation pre- doctoral fellow in the Department of Zoology, University of Chicago. The research was con- ducted under the supervision of Dr. H. H. Strandskov, University of Chicago, to whom the author wishes to express grateful appreciation. Sincere thanks also are extended to Dr. H. H. Strain, Argonne National Laboratory, for advice concerning chromatographic technique, and to Mr. Philip Hershkovitz, Chicago Natural His- tory Museum, for taxonomic assistance. Special thanks are due Mr. Robert Bean, Director of the Chicago Zoological Park at Brookfield, Illi- nois, and his staff. Without the generous coop- eration of officials and staff of the Brookfield Zoo, the present study would not have been pos- sible. Literature Cited Adolph, Edward F. 1943. Physiological regulations. Lancaster: Jacques Cattell Press. 502 pp. 1949. Quantitative relations in the physiological constitutions of mammals. Science, 109: 579-585. 1961] Fooden: Urinary Amino Acids of Non-human Primates 179 Berry, H. K. 1953. Variations in urinary excretion patterns in a Texas population. Am. I. Phys. An- throp., 11:559-575. Block, Richard J., Emmett L. Durrum & Gunter Zweig 1958. A manual of paper chromatography and paper electrophoresis. Second edition. New York: Academic Press. 710 pp. Datta, S. P., & H. Harris 1953. Urinary amino-acid patterns of some mammals. Ann. Eugen., 18:107-116. Eades, Charles H., Jr., & Robert L. Pollack 1954. Urinary excretion of fourteen amino acids by normal and cancer subjects. J. Nat. Cancer Inst., 15:421-427. Gartler, Stanley M., I. Lester Firschein & Theodosius Dobzhansky 1956. A chromatographic investigation of uri- nary amino-acids in the great apes. Am. J. Phys. Anthrop., 14:41-57. Harris, H. 1953. Family studies in the urinary excretion of /3-aminoisobutyric acid. Ann. Eugen., 18:43-49. 1959. Human biochemical genetics. Cambridge: University Press. 310 pp. Hershkovitz, Philip 1958. Type localities and nomenclature of some American primates, with remarks on sec- ondary homonyms. Proc. Biol. Soc. Wash., 71:53-56. Simpson, George Gaylord 1945. The principles of classification and a clas- sification of mammals. Bull. Amer. Mus. Nat. Hist., 85:1-350. Smith, Ivor 1958. Chromatographic techniques, clinical and biochemical applications. New York: In- terscience Publishers. 309 pp. Stein, W. H. 1953. A chromatographic investigation of the amino acid constituents of normal urine. J. Biol. Chem., 201:45-58. Sutton, Harry Eldon 1951. A further study of urinary excretion pat- terns in relation to diet. Univ. Texas Publ. No. 5109:173-180. 180 Zoologica: New York Zoological Society [46: 15: 1961] Fig. 1. Fig. 2. Fig. 3. Plate I Typical chromatographic resolution of known mixture of amino acids. Chromatographic comparison of amino acids in successive urine specimens col- lected on indicated dates from male Pero- dicticus potto. Plate II Chromatographic comparison of amino acids in successive urine specimens col- lected on indicated dates from female Macaca irus. Chromatographic comparison of amino acids in successive urine specimens col- lected on indicated dates from female Cercopithecus neglectus. Plate III Fig. 5. Urinary chromatograms representing each of six cebid genera studied. Arrows indicate hydroxyproline spots. EXPLANATION OF THE PLATES Fig. 4. FOODEN PLATE I SEP. 8 15- 10- 5- cm- Aia Gly Tau Ser GluA cm- • • • i • • • • • cm 5 I 50A3C, Ala Gly Tau - Ser GluA K\ T — r • « » » r r i cm 5 1C FIG. 2 URINARY AMINO ACIDS OF NON-HUMAN PRIMATES FOODEN PLATE II • SEP. 16 f,~TT ^ .OCX 8 15” ASa 15- #AI. 80- %6ly 10" ^j^Gly GluA ^pGiuA 5‘ AspA AspA 11 *~- ■ — moot 22 15 10 5- Alo $ GluA Asp A cm-! '^7~ ■ ' cm o cm| ■ *» ■* < 5 i n, ft* a** cm- r. i — n — r ■ ■ cm jT>A'3r FIG. 3 FIG. 4 Ala URINARY AMINO ACIDS OF NON-HUMAN PRIMATES FOODEN PLATE III O J CACAJAO 4 15* 10* ’j 5*i PITHECIA i cl- CEBUS # * I5*1 10* 5- % 15* 10* 5- * % ■m cm^ 4 — • • t • A <*» V -7 AIMIR! « r • • ,4b L/JSOTHRIX m io-j 5i cm- r » • l • r • • i ■ i iiit t • i cm 5 . I ^ cm 5 15* ,0; ! •> 5J ioA*6 o !•<>* crrn O — i • ^ » t r r l • cm 5 I FIG. 5 URINARY AMINO ACIDS OF NON-HUMAN PRIMATES NEW YORK ZOOLOGICAL SOCIETY GENERAL OFFICE 30 East Fortieth Street, New York 16, N. Y. PUBLICATION OFFICE The Zoological Park, New York 60, N. Y. OFFICERS PRESIDENT VICE-PRESIDENT SECRETARY TREASURER Fairfield Osborn Laurance S. Rockefeller George Wall Merck David H. McAlpin SCIENTIFIC STAFF: John Tee-Van ...... General Director William G. Conway. . Director, Zoological Park Christopher W. Coates . .Director, Aquarium ZOOLOGICAL PARK Joseph A. Davis, Jr. . . Associate Curator, Mammals Grace Davall Assistant Curator, Mammals and Birds William G. Conway . . Curator, Birds Herndon G. Dowling . Curator, Reptiles Charles P. Gandal. . . Veterinarian Lee S. Crandall General Curator, Emeritus William Beebe ..... Honorary Curator, Birds AQUARIUM James W. Atz ....... Curator Carleton Ray Associate Curator Ross F. Nigrelli ...... Pathologist & Chair- man of Department of Marine Biochem- istry & Ecology C. M. Breder, Jr. .... . Research Associate in Ichthyology Harry A. Charipper . . . Research Associate in Histology Sophie Jakowska ..... Research Associate in Experimental Biology Klaus D. Kallman . , . , Research Associate in Genetics Louis Mowbray Research Associate in Field Biology GENERAL William Bridges . . Editor & Curator, Publications Dorothy Reville . . Editorial Assistant Sam Dunton Photographer Henry M. Lester. .Photographic Consultant DEPARTMENT OF TROPICAL RESEARCH William Beebe Director Emeritus Jocelyn Crane Assistant Director David W. Snow ....... Resident Naturalist John Tee-Van ........ Associate William K. Gregory .... Associate AFFILIATE L. Floyd Clarke Director, Jackson Hole Biological Research Station EDITORIAL COMMITTEE Fairfield Osborn, Chairman James W. Atz William G. Conway William Beebe Lee S. Crandall William Bridges Herndon G. Dowling Christopher W. Coates John Tee- Van ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME 46 • PART 4 • DECEMBER 27, 1961 • NUMBER 16 PUBLISHED BY THE SOCIETY Tie ZOOLOGICAL PARK, New York Contents PAGE 16. The Role of the Thyroid in the Development of the Platyfish. By K. France Baker-Cohen, Plates I-IX; Text-figures 1-4.. 181 Index to Volume 46. 223 16 The Role of the Thyroid in the Development of Platyfish1 K. France Baker-Cohen2 Genetics Laboratory of the New York Aquarium, New York Zoological Society (Plates I-IX; Text-figures 1-4) THE role of the thyroid in the physiology of fishes has been a subect of controversy for years (Lynn & Wachowski, 1951; Hoar, 1957; Pickford, 1957). Although a con- siderable body of evidence has accumulated in support of the importance of the thyroid in the growth and maturation of teleosts, there also have been many conflicting reports. Until re- cently, work in this field was hampered by the impossibility of thyroidal extirpation in most teleosts, owing to the diffuse nature of the tissue and its intimate relation to major blood vessels. Observations were limited to the study of effects of antithyroid drugs (“chemical thyroidectomy”) and of thyroid preparations that were adminis- tered to fish with autogenous thyroids (Table 12). The results of such experiments were al- ways subject to criticism, because of the possi- bility of toxic effects exerted independently of the effect on thyroid hormone production, and because non-physiological hormonal excesses may have led to abnormalities. Recently, radioactive iodine (I131) has been employed as a thyroidectomizing agent in tele- osts, with varying effectiveness (Table 11). However, the effects of replacement therapy in athyroid fish have not been reported. With proper controls, the use of I131 may permit analysis of thyroidal function in fish in a manner more nearly comparable to that made possible in higher animals by thyroidectomy. This paper presents observations made, through the use of radioiodine treatment, on thyroidal function in the growth and sexual de- velopment of platyfish, Xiphophorus maculatus. iThis work was supported by a research grant (C-297) to Dr. Myron Gordon, of the New York Zoological Society, from the National Cancer Institute, U. S. Public Health Service, and by a research fellowship (CF-6184) to the author from the National Cancer Institute. 2Present address: Department of Anatomy, Albert Einstein College of Medicine, Bronx 61, New York. The work presented here is an outgrowth of a series of experiments in which radiothyroidec- tomy of very young platyfish was attempted in an effort to obtain evidence relating to the origin of heterotopic thyroid tissue in this species (Baker, 1958a). Many of the radioiodine- treated fish and their controls described here were included in data reported in that publica- tion, but only a few incomplete observations on their anatomy and pathology were given at that time. Materials and Methods 1. Radioiodine Treatment of Young Platyfish. —The origins of the strains of platyfish used in these experiments have been described in Baker et al. (1955). Most of the experiments were made on fish of the BH strain, which was very susceptible to pharyngeal goiter and to thyroid tumors in the kidneys and other organs (Baker, 1958a). a. General Procedures : (Table 1) Groups of platyfish, ranging in age from 17 to 70 days, were treated with radioiodine (I131) by im- mersion in 200 ml. of aquarium water contain- ing 4.5-5.02 me of carrier-free I131. The size of the treated groups ranged from 10 to 24 fish. Their standard length (tip of nose to end of caudal peduncle) was no more than 6 to 8 mm. and thus no deleterious overcrowding existed on a short-term basis. Young fish to be treated were carefully selected by eye, so that in treated and control groups the sizes of the fish were as evenly matched as possible. Exact measurements of length or weight were not made. The exposure time was varied between 24 and 72 hours, but in all but two of the 13 experiments, the expo- sure time was 48 hours or more. Experiment 1 was a preliminary test of the efficiency of the treatment. These fish were of the 20th inbred generation of the 30 strain, which was highly resistant to goiters of all types. 181 182 Zoologica: New York Zoological Society [46: 16 Table 1. Experiments in which Radiothyroidectomy of Young Platyfish Was Attempted by Immersion of Fish in Water Containing Large Amounts of Radioiodine Experi- ment Number1 I131 cone. — mc./200 ml. Hours Exposed Age (days) at Exposure Initial Fish Exposed Exposed Fish Harvested Initial Controls Controls Harvested2 9 $ 9 $ 1. 5.5 48 60 12 0 8 2. 4.5 24 50 17 7 1 3. 5.0 48 63 21 4 3 4. 4.6 48 36 12 3 5 13 1 1 5. 5.0 48 37 12 7 5 11 3 8 6. 5.0 60 17 14 5 3 12 5 7 7. 5.01 72 66 12 2 2 11 1 1 8. 5.0 25.5 52 10 2 3 9 4 2 9. 4.82 60 58 13 0 0 8 0 0 I. 5.02 55.75 42 12 0 5 5 1 4 II. 4.99 52 56 12 0 6 6 2 3 III. 4.99 52 35 13 7 6 6 3 3 IV. 5.02 50 42 24 8 8 8 3 3 Total: 184 100 89 55 Experiment 1 was performed on strain 30 fish, experiments 2 and 3 on strain Fu fish, and the remaining ex- periments on strain BH fish. No controls were kept in experiments 1 to 3. discrepancies between initial controls and harvested controls in experiments 4 to 9 do not necessarily show final death rates, as controls were harvested in small groups at varying ages, for comparison with I131-treated fish, and unused controls sometimes were returned to the breeding stock. See Table 4 for correct death rates among control groups. Four of the fish in this experiment were sacri- ficed immediately post-exposure to determine radioiodine uptake per fish; the remainder were sacrificed at intervals of 6, 17 and 24 days post- exposure to determine the condition of the thy- roid tissue. In all subsequent experiments, fish were not killed for examination until at least 40 days post-exposure. Radioiodine uptake per fish was determined by counting aliquots of a IN NaOH hydrolysate of the head, and aliquots of the radioactive water; per cent, uptake was obtained by the ratio of total counts in each. Total uptake was then obtained by multiplying per cent, uptake by the total original activity known to be in the water. No specific controls were maintained in ex- periments 1 to 3; comparisons made were be- tween treated fish and age-matched members of the same strain. In all other experiments, a closely similar number of broodmate controls was maintained. The controls were subjected to a “sham exposure,” i.e., they were confined with- out feeding in a container in 200 ml. of water for the same period of time as the treated fish. The trip to the treatment site (Department of Zoology, Columbia University) was also taken by the controls, so that any effects of various shocks, such as cold, concussion, sudden dark- ness or light, would be the same. In experiments 1 to 9, the treated fish, after exposure, were washed in radioiodine-free aqua- rium water and transferred to 4-gallon stock tanks, containing gravel, growing plants and snails, and were fed and maintained in the man- ner used throughout the Genetics Laboratory (Gordon, 1950). Control groups were main- tained in adjacent tanks under similar conditions. b. Replacement Therapy: In experiments I to IV, the radioiodine-treated fish were subsequent- ly divided into two or three groups. One of these received no further treatment, while the others received potassium iodide or were fed desiccated thyroid. Non-treated broodmates were kept in equal initial number to each subdivision of the radioiodine-treated group. Further treatments of the I131-exposed fish were begun 5-9 days after removal from the radioisotope. Fish that died during this interval are not included in fig- ures comparing the variously treated groups, but only in total mortality data for radioiodine treatment. In experiment I, the three groups of fish (con- trol, I131-treated, and I131-treated subsequently given potassium iodide) were kept in stock tanks, and one group was given extra KI by the addition of 10 ml. of a 4 mg. /ml. stock solution to its 4-gallon tank every two weeks. Therefore, until the tank water was changed, the concen- tration of remaining KI was unknown. In experi- ments II to IV, each subgroup was maintained in a specified volume of water (4-7 fish in two 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 183 liters, or 8 fish in three liters) in all-glass aquaria, without gravel, plants or snails. These tanks were cleaned and the water changed weekly, and a known concentration of KI, where used, was employed (1 mg./l. final volume). Thyroid was administered by feeding Vt. tablet of desiccated thyroid3 weekly at the start of the experiments; other food was withheld on that day. This was increased to a whole tablet weekly as the fish grew (1 to 3 months after the start of thyroid- feeding, depending on size and number of fish in the group). On all other days, the fish were fed, together with the rest of the fish in the Laboratory, on dried shrimp or liver-pablum (Gordon, 1950). General observations on growth, body shape, coloration, secondary sex- ual development and “health and temperament” were noted during the weekly aquarium clean- ing, when each group was transferred to a small observation vessel while its tank was washed and refilled. Whenever cleaning was done, the water was replaced from the Laboratory breeding tanks in such manner that each subgroup in an experiment received the same mixture of various tank waters. Of 100 radioiodine-treated fish examined in 3Burroughs-Wellcome & Co., “Tabloid” thyroid, U.S.P. Each tablet contained 0.065 gm. desiccated thy- roid. These tablets were tested for metamorphic activity on tadpoles (three years after the fish experiments were completed). In three tests, 14-2 tablets, fed to 4 or 5 three-centimeter Rana pipiens tadpoles, produced tail resorptive processes, hind and fore limb and foot devel- opment, modification of mouth parts, lung breathing, pigmentary differentiation, raised eyes and lack of fur- ther growth. Controls were unchanged within the period (6-11 days) of treatment. these experiments, 87 were radioautographed, using tracer doses of I131 (Baker, 1958a), and all fish were serially sectioned for complete ex- amination of internal organs. 2. Radioiodine Treatment of Adult Platyfish. — (Table 2). In these experiments, radioiodine was given by intraperitoneal injection to fully mature fish of both sexes. Since male playfish are smaller than females, and thus more difficult to inject, fewer males were used. Four groups, totalling 46 fish, were injected; each fish received 48.2 to 100 pc. of I131 in 0.012 to 0.05 ml. of distilled water. The first group was made up of 5 wild-type females of the Fu strain, the mem- bers of which were highly susceptible to thyroid tumors. One fish of this group gave birth to a few young 19 days after the injection; two of these young were radioautographed with the mother 22 days after birth and are described with the group. Groups 2 and 3 included females and males of the 163 strain, which was at least moderately susceptible to thyroid tumors. Group 4 consisted of 12 females of the 30 strain; these fish all died within 5 months post-injection and are included only in the mortality figures. After injection, these fish and their broodmate controls were maintained under ordinary stock conditions. Twelve of the 18 radioiodine-treated fish that were harvested were also radioauto- graphed, and all were serially sectioned for com- plete examination of internal organs. 3. Radiophosphorus Treatment of Young Platyfish— As radiation controls for the X131- treated young fish, 16 platyfish of the BH strain were placed in 200 ml. volumes of aquarium water containing 1.7 me. of P32 for 48.5 to 50.75 Table 2. Experiments in which Radiothyroidectomy of Adult Platyfish Was Attempted by Intraperitoneal Injection of Radioactive Iodine (Starred (*) fish were radioautographed) Group Pedigret Number of Injected Fish pc. 1131 Injected Age (months) when Injected Post- injection Period (months) Number of Controls 9 | $ $ $ 1. 507 Fu Fish started 5 0 Various 8.4 0 0 Fish harvested 1* 0 80 1.4 0 0 2* 0 50 6.5 0 0 2. 163* Fish started 11 96.4 8.7-9.5 10 6 48.2 9.0-9.8 No record Fish harvested i* . 1.3 0 2* 1* 3.5 2* 1* 1 . 19.5 1 3. 163 * Fish started 12 0 50 8.8-11.5 12 0 Fish harvested 5* 0 8.5 0 0 5 0 18.75 3 0 184 Zoologica: New York Zoological Society [46: 16 hours. These were groups of 4-7 fish, aged 30 to 54 days. The amount of P32 used was calcu- lated to give the same roentgen dose of whole- body irradiation within the 200 ml. volume in 48 hours (720 r.) as would 5.0 me. of I131 (Glasser et al., 1952). 4 Fourteen untreated broodmate controls were kept. These fish were maintained in stock tanks after treatment and, when 4.7 to 7 months old, were fixed in formalin and serially sectioned for examination. 4. Analysis of Pituitary Development— A ratio of pituitary to body size was obtained in several experiments, using counts of consecutive 10/a sections of the pituitary as a measure of hypo- physeal size, and standard length of the fish, taken from the specimen itself or its photograph, as the measure of body size. The average ratios obtained by this method were often inconsistent between experiments, when variously treated groups were compared. A better measure of hypophyseal development was obtained by mak- ing a series of outline drawings of each section of the pituitary by means of uniformly magnified projections. The tracings from each fish were then cut out and weighed on an analytical bal- ance. These weights were then averaged within each treatment group and divided by the aver- age standard length within the same group. The ratios obtained by this method proved to be sat- isfactorily consistent between experiments with reference to the relations between the variously treated groups. Volume or weight measurements were not taken on any of the fish before section- ing, so that only length was available as a meas- ure of body size. Objection to a ratio of volume: length, as involving two different types of meas- urements, may be countered, in these experi- ments, with the following observations : Had the proportions of all of the fish been identical, the use of this ratio would have been justified with- out doubt, but if not, it might have led to mis- interpretation. However, the differences in pro- portion observed among the treated groups of fish usually would have enhanced the differences found by using the pituitary-volume-to-body- length ratio. For example, radioiodine-treated fish tended to be dumpy and pot-bellied in shape, which would have increased their weight or vol- ume in proportion to their length, but if such an increased volume had been used to calculate the ratio, the difference between these fish and the controls would have become greater than 4I wish to thank Dr. Edith Quimby, Department of Radiology, Columbia University College of Physicians and Surgeons, for performing the initial calculation of the P32 activity required and for references to the rele- vant published material. indicated at present. This would also hold true for Kl-treated, radioiodine-exposed fish, for the same reason. On the other hand, thyroid-fed fish, if slenderer than their controls, would have had their ratios shifted slightly towards those of the controls. The interpretation of the results ob- tained, in any case, could not have been affected seriously. 5. Histological Methods— All fish were serial- ly sectioned at 10/a and stained with haematoxy- lin-eosin or occasionally with Masson’s tri- chrome. Fixation was routinely in Bouin’s fluid with formic acid as a decalcifier, with the excep- tion of the P32-treated fish which were fixed in 10% formalin. Radioautography was carried out as described in Baker, 1958a. 6. Gonopodial Development.— In experiments I to IV, the anal fins of all fish were removed, after fixation of the whole fish in Bouin’s fluid, dehydrated in alcohols and xylol, and whole- mounted on slides in Permount, unstained, for examination. This was also done with the anal fins of some animals from experiments 2-8 and with those of all P32-treated fish. Results 1. The Efficacy of Radiothyroidectomy in Platyfish. a. Criteria for Thyroidectomy: Thyroidectomy was judged to be complete when I131 radioauto- graphs of serial sections through the thyroid and kidney regions of the fish were blank for thyroid tissue. Radioautographic spotting usu- ally was found also over the auditory region and portions of the pharynx and intestines. In the latter two locations, the spots were produced by I131 that was bound to food particles in the lumen of the alimentary tract or adhering to its lining (PI. II, Fig. 1). These non-thyroidal sources of spotting were readily differentiated by their loca- tion from spots produced by thyroid tissue, even when crude contact radioautography was em- ployed. Location of spots within a section was often aided by the occurrence of a “shadow picture” of the whole section, sometimes includ- ing definition of organ areas contrasting with open areas within the section. In addition to the radioautography, all serial sections were searched histologically in a meth- odical fashion for thyroid tissue in the pharyn- geal and renal areas. In some cases, when radio- autography was not feasible, as in case of pres- ervation after death, histological search was accepted as sole criterion for thyroidectomy. This was used in relatively few cases, however, and was related to the total morphological pic- ture of the individual before any acceptance of 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 185 total thyroidectomy was made. In many in- stances, especially in younger fish, radioauto- graphic speckling was noted, but no thyroid tissue could be found histologically. The prob- able presence of undestroyed thyroid tissue in these cases was revealed by later study of fish from the same group; in these older fish, visible thyroid tissue was often found. Thus, histological criteria for total thyroidectomy were not found to be completely reliable on a short-term basis. b. Radioiodine Treatment by Immersion: On the whole, this method was very successful. In 92 fish examined 40 or more days after treat- ment, the frequency of complete radiothyroid- ectomy, by all criteria, was 68.5%. With the omission of two less successful experiments (Nos. 2 and 5), however, this incidence was 86% (62 of 72 fish harvested). The results appeared to be affected both by exposure time and by the strain to which the fish belonged, when dose was nearly constant. For example, no fish were completely thyroidec- tomized in experiment 2, when strain Fu fish were exposed for 24 hours to 4.5 me. I131, but all fish were thyroidectomized in experiment 8, when strain BH fish were exposed for 25.5 hours to 5.0 me. I131. On the other hand, in experiment 3, most strain Fu fish were thyroidectomized by 48 hours exposure to 5.0 me. I131. The BH strain of platyfish was particularly highly sensitive to low iodine concentration in its environment, as judged by the frequency of thyroid tumors and heterotopic thyroid tissue (Baker, 1958a, b) , and the fish probably concentrated iodide in the thy- roid tissue, when supplied in excess, to a greater degree than platyfish of other strains. The poor results of experiment 5 perhaps were related to an addition of NaHCOs to the radio- iodine solution. This was the only apparent dif- ference between this and later experiments using fish of the BH strain of similar age as well as closely similar amounts of radioiodine and ex- posure times. Although none of these fish was completely thyroidectomized, as judged by spotting on radioautographs, the amount of thy- roid tissue was severely reduced, and none could be found histologically in many cases. Among the group of 29 fish accepted as in- completely radiothyroidectomized, thyroid fol- licles were unquestionably seen in 18 animals (Table 3). In the remaining 11 fish, radioauto- graphic spotting, together with visualization of a few doubtfully identified thyroid follicles in some fish, comprised the evidence for incom- plete thyroidectomy. These doubtful cases were largely confined to animals that were examined within three months after I131 treatment and were no more than 4.5 months old. Among fish Table 3. Comparison of Criteria for Completeness of Thyroidectomy in Platyfish Treated with Radioiodine by Immersion1 Experiment Number2 Total No. Fish Harvested Days Post-treatment when Examined Number of Incomplete Thyroidectomies, Judged from: Radioautograph Spots, No Follicles Seen Follicles Seen Histologically 2 8 41 0 3 150 0 2 176 0 1 539 0 2 3 7 47 1 1 171 0 1 5 12 79 2 2 95 5 0 168 1 1 6 8 60 1 1 277, 289 0 1 I 5 159 1 1 IV 16 140 0 23 Total of incomplete thyroidectomies: 11 18 Fraction of incomplete thyroidectomies: 29 out of 92 I131-treated fish harvested in all experiments, except No. 1. t-All fish radioautographed 40 or more days after treatment. The number of incomplete thyroidectomies that could be shown by radioautographic criteria exceeded by a large increment the number that could be demonstrated by histological examination alone. 2In experiments 4, 7, 8, II and III all fish were totally thyroidectomized. Therefore these experiments are not listed here. 3One of these had only a single thyroid follicle, but in the kidney. 186 Zoologica: New York Zoological Society [46: 16 more than 6 months old there rarely was any question about the presence of undestroyed thy- roid tissue. Not only did the radioautographic spotting per follicle become unusually intense, but the amount of thyroid tissue often became considerable (PI. II, Figs. 3, 4). Whenever vis- ualized and in whatever quantity, thyroid tissue in the non-thyroidectomized fish was hyper- trophied and the colloid was granular and small in amount. The two oldest fish were 19.6 months of age, and in them the thyroid tissue had as- sumed almost goitrous proportions, although each follicle still remained distinct and slightly separated from others and did not form a con- tinuous mass of follicular and afollicular tissue as seen in most goiters among platyfish (PI. II, Fig. 4). In the pilot experiment (No. 1), 12 young fish of a strain insensitive to low concentrations of environmental iodide were used. Four of these fish were killed immediately after exposure and the radioiodine uptake per fish was determined. This proved to be 20-50 /xc. Two fish were ex- amined on the 6th day after treatment, two on the 17th day and four on the 24th day. On the 6th day, the number and distribution of thyroid follicles appeared to be normal. The morphology of the follicles was normal, but the colloid often stained blue with haematoxylin-eosin. Normally, thyroid colloid stained bright red with this stain combination. The blood vessels in the thyroid area were normal, and no other anatomical changes were noted. By the 17 th day, the radio- autographs were “spotty” and the number of thyroid follicles was severely reduced. The cells of these follicles were heightened, often puffy, with very indefinite boundaries and pale cyto- plasm. The colloid was pale pink and coarsely granular. No follicles had normally dense colloid or colloid stained blue. The aorta and walls of the gill chamber in the thyroid area were some- what thickened. No other pathological effects were noted. On the 24th day post-treatment, one fish was completely lacking thyroid by all cri- teria, one had a single spot on the autograph but no visible thyroid follicles, and two retained a few identifiable follicles. The latter were often very indistinct and had pale, enlarged cells and pale-pink granular colloid. In one fish, the thy- roid may have begun to recover as there were two or more follicles with dense vacuolated colloid and flat cells, as well as several pathologi- cal follicles. In all of these fish, the aorta was very shrunken, its walls were thickened, and it was surrounded by a thick layer of gelatinous- appearing connective tissue which attached it to the walls of the anterior pericardial space (PI. II, Fig. 6). One fish had some enlarged tubules containing concretions in the kidneys (the kid- neys of 2 of the 4 fish were not sectioned) . The positive results obtained in this trial, with fish of a goiter-resistant strain, provided a basis for repetition of the method, using the goiter-prone strains, which might be expected to take up radioiodine more avidly, or release it more slow- ly, and therefore be more effectively treated. c. Radioiodine Treatment by Injection: (Table 2) . This method was only moderately successful in producing thyroidectomized fish, although it produced severely hypothyroid animals in all cases. The chief difficulties encountered were the necessity for keeping the injected volume very small, especially for male fish, which are smaller than females, and the tendency for post-injec- tion leakage to occur. Since relatively few fish were treated by this means, and because of the lesser importance of these experiments to the analysis of the chief problem in hand at the time, these difficulties were never fully dealt with. Nevertheless it is likely that successful adminis- tration of the dosages intended for the fish would not have brought about much different results since various amounts of I131 in various volumes were injected, with fairly similar effects. Here again, strain differences were noted. The three female Fu fish of Group 1 were quite successfully thyroidectomized by injections of 50-80 juc. of I131, whereas in Group 2, none of the 163 female fish was completely thyroidecto- mized by injections of 96.4 fxc. of I131. Fish of the 163 strain produced renal thyroid tumors, but with a much lower frequency than did the Fu strain. From this it might be assumed that the thyroids of 163 fish had a lower avidity for, or retention of, iodine than those of Fu fish. Among the 18 fish harvested (Table 2), total histological thyroidectomy was obtained in 9 animals, but two of these had died in I131 tracer solution before fixation, so that the preservation was poor. Radioautographic total thyroidectomy was obtained in only one instance. Two fish were examined 38-41 days post-injection. From Group 1 (strain Fu) , one fish was lacking thyroid histologically, although it retained a faint radio- autographic speckling in the thyroid area. From Group 2 (strain 163), one fish showed thyroid tissue histologically, although extremely re- duced. The thyroid cells were high and the col- loid was pale and granular. All other fish were examined 3.5 or more months after the injec- tions. 2. The Effects of Radioiodine Treatment on Young Platyfish. a. Mortality: The death rate among treated fish in these experiments was very high. Of 172 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 187 DAYS AFTER RADIOIODINE TREATMENT Text-fig. 1. The distribution in time of deaths among radioiodine-treated platyfish. The percentages are based on corrected totals for remaining fish, after subtraction of those killed for examination in each time period. Therefore, the figures do not add up to 100%. The analysis is based on an initial 113 fish, from those experiments in which all deaths were recorded with respect to the time periods (experi- ments 2, 5, 8, 9, and I-IV). Of these fish, 59 were killed after the 30th day post-treatment, and 54 died “naturally.” radioiodine-treated fish, in all experiments ex- cept No. 1, 89 (52%) died during maintenance periods that ranged from 128 to 290 days. These deaths were largely concentrated in the periods between 0 to 2 weeks post-exposure, and be- tween 3 to 5 months post-exposure (Text-fig. 1). A more accurate picture of survivorship, uncom- plicated by early sacrifices and lack of control data, is obtained by the analysis of experiments 7-9 and I-IV only, in which all fish were main- tained untouched until the conclusion of the experiment. Survivorship to the 109th and 130th day post-exposure for these experiments is shown in Table 4. (The 109-day survivorship is pre- sented for comparison with mortality data on P32-treated fish, most of which were kept for only 109 days post-treatment. See below, part 4) . Comparison of the 109-day with the 130-day survivorship further illustrates the heavy loss of treated fish between three and five months after treatment. The death rate among the untreated controls in these experiments was almost negli- gible. b. Growth: Among the most prominent ef- fects of radioiodine treatment was a striking reduction in the growth of the treated fish in comparison with untreated broodmates. Ex- amples of differences in average standard length between treated and control fish, at various ages, are given in Table 5. These differences may be also seen in PI. I, Fig. 2. There was little or no difference between the 2.5-month-old treated and control fish of experiment 6, but in older fish differences of more than 10% were found in 5 out of 6 groups. It must be borne in mind that mortality among the treated fish was invariably highest among the smaller fish, so that the final survivors were a selected group of the largest size. Among the controls, deaths were also most frequent among the smallest fish, but very few of any size died. No measurements of growth rate were made, only the size attained at fixation being recorded. Radioiodine-treated fish were frequently ob- served to develop a hunched, “cretinous” body shape, with a pot belly. This may be seen in PI. I, Figs. 1 and 3, in both living and fixed speci- mens. The pot-bellied appearance was found to be related to the condition of the liver and the abdominal tissues (see below). Treated fish also were noticeably darker in color than the controls in most cases, but this difference was not further investigated. c. General Pathology: The ventral aorta and the bases of the afferent branchial arteries, the bulbus arteriosus and other arteries passing through the thyroid area were usually shrunken in radioiodine-treated fish. The walls of these 188 Zoologica: New York Zoological Society [46: 16 Table 4. Survival Rate in Platyfish Treated by Immersion in Containing Large Amounts of Radioiodine1 Water Experiment Fish Treated Survivors Controls Kept Control Survivors A. Survivorship to the 109th day after treatment: 7 12 9 (11) Incomplete record 8 10 10 9 9 9 13 8 8 8 I 8 3 5 5 II 5 3 6 5 III 7 7 6 6 IV 8 6 8 7 — — — — Total: 63 46 42 40 Percent. Survival: B. Survivorship to the 7 130th day after treatment: 12 732 4 (11) 95 Incomplete record 8 10 10 9 9 9 13 2 8 8 1 8 3 5 5 II 5 2 6 5 III 7 5 6 6 IV 8 6 8 6 — — — — Total: 63 32 42 39 Percent. Survival: 50 93 1These figures do not include fish that were additionally treated with potassium iodide or fed thyroid material. '-Percent, survival would be 70 if two fish in experiment 7, whose deaths were inexactly recorded, had died before the 109th day. vessels, particularly the aorta, were thickened, and connective tissue “adhesions” between the aorta and the pericardial lining frequently were present and developed to a considerable extent (PI. II, Fig. 5) . In some fish, this material formed a thick, structureless, gelatinous-appearing mass around the aorta (PI. II, Fig. 6). The heart was never visibly affected. The area of the thyroid often was filled with fibrous stroma, and in some fish the caudal portion of the area, above the aorta, disappeared through the collapse of the walls of the gill chamber into the vacant space. Most I131-treated fish suffered a loss of lym- phoid tissue in the kidneys to some extent. These are the chief blood-forming organs in teleosts, with the spleen second in importance. In most of the fish, this loss was not severe, but definitely was noticeable upon histological examination. In general the spleen seemed to be normal. No lymphoid cell counts or blood counts were made on these fish. In a few treated fish, almost com- plete loss of lymphoid elements was found. This was first apparent from the empty appearance of the kidneys (PI. Ill, Fig. 2), and was accom- panied by a severely shrunken spleen (PI. Ill, Fig. 7) and an absence of thymus tissue. The thymus was found to persist as a pair of large lymphoid organs in all fish that were totally radiothyroidectomized, regardless of sex, with the exception of those specimens just described. The thymus glands of normal platyfish are very large and solid structures in late embryonic and immature post-natal stages. As the fish mature, these glands often develop a few large internal cysts, sometimes filled with an eosinophilic col- loidal material. In some instances, these cysts closely resembled thyroid follicles, but they were not found to produce I131 radioautographs. They evidently correspond to the cysts and follicular or alveolar structures described in the thymus of Necturus by James ( 1939) . As the male gonad becomes fully mature and full of sperm, the thymus glands shrink and may disappear en- tirely. In female fish, however, large thymus glands were found at all ages up to 20 months, but they more often were cystic in females 9 or more months of age. Pregnancy had no notice- able effect on the thymus. In fish that were not totally radiothyroidectomized, the thymus glands changed in the normal, characteristic way, as sexual maturation took place. Although the glomeruli and the bulk of the kidney tubules were not visibly affected, the lat- ter structures in the kidneys of I131-treated fish Table 5. Difference in Growth between Radioiodine- treated Fish and Their Broodmate Controls1 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 189 a ° 2 9 9 ° ci c$ a> 'S.B’o a ^ a a •• 42 •a 43 ^ tfi «4o a S ' | SP O 0 CQ 1 § !u .s 43 a o ^ C3 -4 a> fl ts cd s-h ■g “ fcM Hh £■0 a s s & P qj *7 ^ ^ H 5 d o B O O ON ON O i-h ON Tt fS ON Tf fsj T-H fO VO C'l m cn V© 00 cn »-h rt ri on no no no’ no* were in part pathological. Many tubules were swollen and their lumens contained unstained masses of material best described as “concre- tions” (PI. Ill, Figs. 3, 4). Tubules also were seen to be degenerating, and sometimes small wormlike structures which were very darkly basophilic were seen (PL III, Fig. 6). These are believed to be regenerating tubules. The liver of radioiodine-treated fish was usu- ally the most noticeably and severely affected internal organ. The pot-bellied appearance of these fish resulted from an enlargement of the liver, which was often extreme, coupled with greatly increased abdominal “vacuolation.” In these enlarged livers, “vacuolation” was found to pervade the entire organ to a remarkable ex- tent (PI. IV, Figs. 2, 4) . The term “vacuolation” will be used throughout this paper to describe the condition of the liver and abdominal tissue in many radioiodine-treated fish. Although it is believed that this condition probably represented fatty deposits, the presence of fat could not be demonstrated because of the routine processing of all material in fat solvents, such as xylol. No frozen sections stained for fat were prepared, as the liver condition was not analyzed at the time when fresh experimental material was available. Vacuolated livers were found in a large pro- portion of the treated fish; severity ranged from small vacuolated areas in the antero-central por- tion of the liver to complete change of the entire organ. Severity was found to be correlated with age (or to time elapsed after treatment) in a remarkably linear fashion (Text-fig. 2). Vacuo- lated livers were found in incompletely thyroid- ectomized fish with nearly as high a frequency as in totally thyroidectomized fish (Table 6), but never was found among the controls, even in the presence of thyroid hypertrophy.6 Included among the fish with vacuolated livers are several specimens whose livers contained large amounts of ceroid (PI. IV, Fig. 3). The ceroid found in the I131-treated platyfish occurred in globules in the liver cells themselves, and also in large masses of large, dark-nucleated cells, which were solidly filled with it. This ob- servation concurs with the distribution in experi- mentally cirrhotic rats reported by Endicott & Lillie ( 1944) , who described ceroid as occurring largely in large round phagocytes which formed broad sheets and trabeculae. In the I131-treated 5 Vacuolated livers were found in many platyfish that had thyroid tumors in their kidneys, although not in all specimens. Most of these fish were relatively old animals. 190 Zoologica: New York Zoological Society [46: 16 Text-fig. 2. Increase with time in severity of liver vacuolation (see text) among radioiodine-treated platy- fish. The fish were classified into four groups of increasing severity from their histological descriptions, and then subgrouped by age and by the period elapsed after treatment. When the two most severely affected groups were compared with the two least severely affected, a straight line relationship to age emerged, and a similar, though less regular, relationship to post-treatment time appeared. These curves show that long- continued hypothyroidism had a progressive deleterious effect on liver function. Pathological effects of radiation, on the other hand, show recovery in time, if the animal is able to survive. Solid circles and continuous line represent age, and open circles with dashed line represent time post-treatment, both on the same time scale. fish, broad masses of ceroid-filled cells also were found near hepatic blood vessels and among the pancreatic tissue outside of the liver. In some fish, vacuolated areas and extensive ceroid de- posits were found together in the liver.6 In the case of the radioiodine-treated platyfish, it was supposed that the conditions of vacuolated liver and “ceroid liver” were related, and, if the as- sumption is made that vacuolation does repre- sent fat, available information appears to sup- port this view (see Discussion). In normal fish, ceroid also was seen in small aggregations of macrophages next to blood vessels, but never was seen in liver cells. 6A re-examination of the livers of 8 thiourea-treated platyfish and their 9 controls (originally described in Baker, 1958b) was made. Ceroid-filled macrophages were found massed in the livers of all treated fish and to an extreme degree in 7 of them. Ceroid deposits, be- sides those in macrophages, were found in the liver cells themselves. Prominent vacuolization, coupled with ceroid deposition, occurred in 4 fish, and in two of these cases there was severe cellular degeneration in the central parts of the liver. In the controls, ceroid was absent from liver cells and scant in macrophages, while vacuolated livers did not occur. In fish with extensive vacuolation in the ab- domen, the pancreatic tissue was severely com- pressed and may best be described as wispy (PI. IV, Fig. 6). In most cases, however, the amount of pancreatic tissue appeared to be rela- tively normal. d. Gonadal Development: No completely radiothyroidectomized fish ever attained sexual maturity, even at an advanced age (there were three possible exceptions, which will be described below, but these fish were not conclusively athyroid) . Gonadal development in treated fish and controls is shown in Tables 7 and 8. From the youngest fish sampled, including incomplete- ly thyroidectomized ones, it was apparent that gonadal development was retarded in all I131- treated fish (PI. V, Figs. 1, 2). Among com- pletely athyroid animals, this retardation fre- quently was extreme. As shown in Table 8, many males had developed no primary spermatogonial cysts (Wolf, 1931) by the age of 6 months or more (PI. V, Fig. 4), when controls were pro- ducing young, and many treated females had developed only a few early basophilic oocytes (PL V, Fig. 6). Although some thyroidecto- 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 191 Table 6. Incidence of Vacuolated Livers (see text) in Radioiodine-treated Platyfish and Their Broodmate Controls Number of Fish Number with Vacuolated Liver Percent, with Vacuolated Liver Fish 5 or less months old: Complete thyroidectomy 7 3 43 Incomplete thyroidectomy 8 5 63 Doubtful 9 7 78 Total I131-treated 24 15 63 Controls 11 0 0 Fish 5.1 to 7 months old: Complete thyroidectomy 24 173 71 Incomplete thyroidectomy 5 3 60 Total I131-treated 29 20 69 I131 + KI treatment 9 6 67 I131 + thyroid feeding 15 0 0 Controls 30 0 0 Fish more than 7 months old: Complete thyroidectomy 9 8 89 Incomplete thyroidectomy 3 2 67 Total I131-treated 12 10 83 Controls 13 0 0 Grand total solely I131-treated: 65 45 69 Grand total controls 54 0 0 10ne additional fish with other abnormalities of the liver not included. Table 7. Gonadal Development m Radioiodine-treated Platyfish and Their Broodmate Controls Experiment Age (months) Radioiodine-treated Controls Total Mature Immature Total Mature Immature A. Total Thyroidectomy 3 6.8 2 0 2 . . 8.4 1 0 1 4 6.0 4 0 4 2 2 0 6 7.5 2 0 2 10.2 1 0 1 7 7 0 7 6.5 4 0 4 2 2 0 8 9.7 5 0 5 6 5 1 I 6.8 2 0 2 5 5 0 II 6.3 2 0 2 5 3 2 III 6.3 6 0 6 6 4 2 IV 6.1 5 0 5 6 1 5 Total: 34 0 34 39 29 10 Percent. Mature: 0% 74% B. Incomplete Thyroidectomy 2 6.6 2 2 0 7.5 1 1 0 3 7.8 1 1 0 . # 5 5.7 1 1 0 6.8 2 2 0 1 1 0 6 9.8 1 1 0 • Total: 8 8 0 1 1 0 Percent. Mature: 100% 100% 192 Zoologica: New York Zoological Society [46: 16 Table 8. Degree of Sexual Development among Totally Radiothyroidectomized Platyfish more than 5 Months of Age1 Females with: Males with: Spermatogonial Spermatogonial Experiment 5 oocytes > 5 oocytes Cysts Absent Cysts Present 3 2 1 4 1 0 2 i 6 . . 3 0 7 3 0 1 0 8 0 2 1 2 I 2 0 II 0 2 m 2 2 0 2 IV 0 2 3 0 Total: 8 7 12 7 1None of these fish was mature: females showed no signs of yolk deposition, while males possessed no germ cells in late stages of meiosis or in sperm formation. Some of the males in which spermatogonial nests or cysts were found exhibited a few cysts with cells in early spermatogonial division. Radioiodine-treated fish that were treated secondarily with potassium iodide or were fed thyroid have been excluded. mized fish were maintained until 8 to 1 0 months of age, gonadal development proceeded no fur- ther than the enlargement of a few oocytes, without yolk deposition, in females (PI. VI, Fig. 1 ) , and the beginning of germinal cyst develop- ment, with some early meiotic cysts, in males (PI. VI, Fig. 2). On the other hand, all fish with incompletely destroyed thyroid tissue were sexually mature by the age of 5 months, al- though no females were found to be gravid (note, however, that only in experiment 2 were treated females and males, that were maturing normally, living together) . Development of the gonopodium (the struc- turally specialized anal fin found in the males of this live-bearing species of fish, which is used for internal fertilization of the female) in thy- roidectomized male fish was inhibited along with the testes, the anal fin remaining juvenile in form and undifferentiated. In one case, where an extremely infantile testis was present, there was elongation of the anal fin, but no segmental differentiation of the rays or of the specialized structures at the tip of the fin had occurred (Text-fig. 4; PI. VII, Fig. 4). This type of anal fin development was described in gonadless hy- brids between Xiphophorus helleri and X. macu- latus by Rosen (1960). Two of the exceptional I131-treated fish men- tioned above were males that developed large mature testes, with spermatophore production, in the apparent absence of thyroid tissue (PI. VI, Figs. 3, 4). One, aged 6 months (experiment 5), unfortunately died and was not radioauto- graphed, but the sections were remarkably good and were stained in a most striking way for thyroid delineation (Masson’s trichrome). The other had been treated with potassium iodide (see below, experiment I) and was radioauto- graphed, but KI is known to interfere with the uptake of further (radioactive) iodide (Baker, 1958b). Both of these fish had anal fins that were elongated but completely undifferentiated (PI. VII, Fig. 4). Both also had among the most extremely vacuolated livers seen (PI. VI, Fig. 4) , with large aggregations of vacuoles compressing the pancreatic tissue in the abdomen, and with concretions in the kidneys. The Kl-treated fish had developed the bright red color characteristic of most males of the BH strain (females and young were more yellowish or brownish) . The third possibly exceptional fish was a female, also from experiment 5, aged 7 months, with maturely yolked oocytes in the ovary. Al- though there were a few spots on the radioauto- graphs, thyroid tissue was not positively identified histologically. This fish also had some vacuolization in the liver, extremely large ab- dominal aggregations of vacuoles, and concre- tions and degenerating tubules in the kidneys. Anomalous development of the testes was found in 4 fish of the radioiodine-treated groups. Two of these were Kl-treated fish of experiment I (see below), the others fish of experiments 5 and II, treated only with radioiodine. Three of these presented the same picture: an anterior testicular mass suspended solitarily from the dorsal peritoneum and unconnected by any dis- cernible duct to the cloacal area or to a slightly posterior second testicular mass. The latter con- tinued with a normal spermatic duct, opening close to the anus. In one case, both masses ap- 1961] Baker-Cohen: Role of Thyroid in Development of Platyfsh 193 peared to be bipartite. In the 4th fish (experi- ment 5), both testes appeared normal, but no spermatic duct could be found. None of these testes were near maturity. Similar anatomical peculiarities were not found among the controls. e. Pituitary Development: No cytological studies were made on the pituitaries of I131- treated fish because the fish were serially sec- tioned with the utmost rapidity in order to make radioautographs, and they therefore were cut relatively thickly (10/x). Staining also was rou- tine; special methods were not used. The only cytological distinction that could be made was between a darkly basophilic cell type in the most anterior portion, a paler basophilic type in the mid-region, and the neural hypophysis. Eosino- phils were not clearly recognizable because of the thickness of the sections and irregularities in staining. A volumetric study of the hypophysis was at- tempted, however, and this produced rather clear-cut results (Table 9). It may be seen that the radioiodine-treated fish had significantly smaller pituitaries per unit of body length than did their controls. The pituitary: body size ratio remained constant at a low level as the En- treated fish grew older (in Table 9, experiments 5 and 8 may be compared), while the pituitary size of the controls increased in relation to their body length. It had been believed that the hy- pophysis of severely hypothyroid or completely athyroid fish might have been enlarged, owing to stimulation of thyrotrophs by the lack of thy- roid hormone, but this was not the case. Instead, the size and appearance of the pituitaries sug- gested exhaustion, and the absence of any cellu- lar increase in response to the environment. The hypophyses of the control fish became larger per unit of body length as the fish grew older and at the same time that thyroid hyper- trophy appeared in many of them. Both features presumably resulted from continued exposure to a low-iodine environment. The large relative size of the glandular portion of the hypophysis in the controls may be seen in Text-figure 3. This was particularly marked in female fish; which is in accordance with the observed higher incidence of goiter in females of the BH strain (Baker, 1958a) . The sex ratios among the treated and control fish were sufficiently alike to preclude any strong bias on a sex basis. Among the control fish, several had developed a marked overgrowth of the paler, intermediate basophiles, so that these overlapped the darker anterior baso- philes ventrally, causing the appearance of a sharp line of demarcation in the anterior region and a distortion of the usually round cross-sec- tion of the hypophysis at that point (Text-fig. 3; PI. VI, Figs. 6,1). This appearance of the hypo- physis was usually found in the presence of thyroid hypertrophy. f. Behavior and Fragility: Although these ob- servations are somewhat subjective and were not measured in a controlled manner, it is believed worth recording that the radioiodine-treated fish were very sluggish. They never were seen darting about the tank, and when caught for tank chang- ing, were listless in avoiding the net. From gen- eral observation and two specific incidents, it was concluded also that they were easily killed by the shocks of handling and other disturb- ances. The first incident was an attempt to run two treated and two control fish in a Warburg apparatus. This also entailed transportation in a small container by ’bus over a considerable distance. The two treated fish died within an hour of being individually set up in Warburg vessels, although the rate of shaking used was very low. The two controls, on the other hand, remained alive and healthy for two days under identical conditions and were removed appar- ently none the worse for the experience. The second incident was the subjection of a treated fish to photography, together with a pair of normal broodmates (PI. I, Fig. 1). This fish was manipulated for some time to obtain two Table 9. Pituitary versus Body Size in Radioiodine-treated Platyfish, Their Controls, and in Radioiodine-treated Fish Fed Thyroid1 Experiment Age (months) Solely I131-treated I131, plus 1 Thyroid-feeding Control Fish 5 4.4 16.6 (3/2) 27.9 (0/5) II 6.3 14.5 (0/2) 19.9 (0/4) 29.4 (2/2)2 in 6.3 13.1 (2/2) 11.7(3/3) 24.5 (3/3) 8 9.6 16.9 (2/3) — 35.8 (2/4) ’Pituitary size was measured volumetrically and body size was measured as standard length. A ratio of the averages of the two parameters is tabulated here, for each group of fish. The number of males and females in each group is given in parentheses, as females/males. 2One fish of unknown sex not included. Its gonad may have been torn out when the anal fin was removed for mounting. 194 Zoologica: New York Zoological Society [46: 16 A ooOOOOOOO QQQ A‘ QQQQQQQQOOOOo 000OO0 »o00OOOO0O OOOOOQQ® QQOQOQ®« oo oooooOOOOOO D OOG iQQQQ® ®(£QQ£)Q®®<© 0)0 OO Text-fio. 3. Projection drawings of serial sections of the hypophyses of representative strain BH platyfish from different experimental groups, all to the same magnification. The hypophyses were cut in cross-section, and are drawn with the ventral aspect upper- most. A comparison of over-all size and neural versus glandular development may be made. A, Radioiodine-treated immature female; B, Thyroid-fed radioiodine-treated mature female; C, Control (pregnant) female; D, Control mature male. It may be seen that the neural hypophysis (solid black areas) in the thyroid-fed fish appears to be nearly as large as that in the controls, but the glandular portion is much smaller. The glandular hypophysis in the normal females usually was much larger than in the normal males, although the male represented here had a slightly larger than average glandular portion; this may be related to the higher incidence of goiter in females of the strain. The hypo- physes shown here were selected both for their representation of their respective groups, and for the most complete and undamaged series of sections. 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 195 photographic exposures. The animal was found to be dying the following morning and had to be fixed without autography to prevent its loss for examination. 3. Replacement Therapy Experiments with Young Radioiodine-treated Platyfish. a. Thyroid-feeding Experiments: In these ex- periments, 18 radioiodine-treated fish were fed on desiccated thyroid once a week. The effects of this treatment were striking (Table 10) . Mor- tality was reduced to a figure insignificantly dif- ferent from that of the untreated controls, and the growth in body length equaled that of the controls, being significantly greater than that of the fish treated only with I131. Body shape was normal, with a tendency towards a more stream- lined appearance than shown by the untreated controls. This is enhanced in the photographs by the greater length of the caudal fin (PI. I, Fig. 2) but the growth of the fins was not noted in the living fish and consequently the tails were not saved or measured. Examination of the pho- tographs taken of each of these groups of fish suggests that the pectoral fins were also elon- gated, but these are often twisted from fixation and one cannot be sure. No sign of exophthal- mia was found in the thyroid-fed fish. These fish produced blank or nearly blank radioautographs, and the most meticulous his- tological search revealed no thyroid tissue. This also was true of the radioiodine-treated controls. Most of these thyroid-fed fish showed a mod- erately to severely shrunken ventral aorta, with thickening of its walls and fibrous adhesions to the pericardium. In other respects they differed but slightly from the untreated controls. Very little lymphoid loss was seen in the kidneys or elsewhere, no degenerating tubules or concre- tions were found in the kidneys, and no vacuoli- zation of the liver or abdominal tissue was present. The sexual development in the thyroid-fed fish was most striking. Not only did fish of both sexes reach maturity by the end of the experi- ments, in equal or greater proportion to the untreated controls, but external male secondary sexual characters were often apparent consider- ably earlier than in the untreated controls (Table 10). Gonopodial structure was perfectly normal in the thyroid-fed males that matured before termination of the experiments (Text-fig. 4; PI. VII, Fig. 2) . At the close of the final experiment (IV), one of the thyroid-fed females was found to be gravid (in these small treatment groups, the sexes were not separated) (PI. VIII, Figs. 3,4). Table 10. Replacement Therapy of Radioiodine-treated Platyfish. A Summary of the Results of 4 Experiments in which Radioiodine-treated Fish Were Subsequently Treated with Potassium Iodide or Were Fed Thyroid Tablets1 Radioiodine-treated Fish Untreated Controls Not Treated Kl-treated Thyroid-fed Fish at beginning of treatments 25 12 18 25 Survivors at end of experiments2 11 9 15 22 Percent. Survival 44 75 83 88 Average age (days) first male seen 114 159 Males with sperm 0 18 6 7 Males with mature gonopodium 0 0 6 7 Females with yolked oocytes 0 0 3 5 Pregnant females 0 0 1 0 Percent, histologically mature fish 0 113 60 55 Average standard length (mm.)4 17.7 20.1 20.5 21.8 22.0 • 24.5 1Each experiment included radioiodine-treated controls and non-treated controls, and the fish in each experi- ment were broodmates. 2Age of these fish at the end of the experiments was 6.1 to 6.8 months. sThis exceptional fish had a mature testis, but the anal fin was undifferentiated (see text). 4Two sets of comparative figures are given because all the measurements for the four groups could not be aver- aged and then compared, since the measurements were taken on photographs which varied slightly in magnifica- tion. Instead, means were taken for each of the groups within each experiment and these means were averaged for the groups shared by two or more experiments. For example: only experiments I and IV included Kl-treated fish, but experiment I did not include thyroid-fed fish. Therefore the mean lengths of the Kl-treated, the solely I131-treated, and the control fish were averaged for these two experiments, but the thyroid-fed fish were omitted. This is represented in the second fine of the comparison given for average standard length. The same method was used with experiments II, III and IV, excluding the Kl-treated fish, and this comprises line 1. By this pro- cedure, each experiment received equal weight, but the varying numbers of differently magnified fish in each treatment group in the experiments pooled was not allowed to distort the result 196 Zoologica: New York Zoological Society [46: 16 Text-fig. 4. The structure of the normal gonopodium of the male platyfish. In mature thyroid-fed, radio- iodine-treated fish, in P32-treated fish, and in controls, no deviation from the arrangement and numbers of the elements shown here was found (compare with photographs of gonopodia of variously treated fish in PL VII). Drawing by Dr. Donn E. Rosen. Among the immature fish in these experi- ments, the thyroid-fed group included 4 fish with extremely undeveloped gonads, that is, males with no spermatogonial cysts and females with less than 5 differentiated ova. One such fish was found in the control group. These were among the smallest fish in each group. All other immatures in the thyroid-fed and control groups appeared to be progressing normally towards eventual maturity. The pituitary glands of some of the thyroid- fed fish were included in volumetric analyses made on radioiodine-treated and untreated fish (Table 9). In two experiments which included thyroid-fed fish, the hypophyses of these proved to be as small, relative to the fish’s body length, as those of I131-alone-treated fish, and much smaller than those of the untreated controls. There was a morphological difference, however, from the radioiodine-treated controls: the neu- rohypophysis, although not measured other than by eye, appeared to be as large in the thyroid-fed fish as in their untreated controls, but the gland- ular portion was very much smaller (Text-fig. 3 ) . In the solely I131-treated fish the whole pitui- tary seemed to be smaller, just as the whole fish was smaller. Accuracy in projection and draw- ing was not good enough to make a sufficiently defined separation of the neural and glandular elements of the hypophyses to quantitatively verify this distinction. It seems reasonable to suppose that the thyroid-fed fish were receiving sufficient thyroid hormone to cause the thyro- trophic cells in their pituitaries to be reduced to a minimum in numbers and/or size, and activity, and thus for the glandular volume of the hypo- physis to be considerably reduced. The thyroid-fed fish were healthy and vigor- ous and as lively, or more so, than the untreated controls. b. Potassium Iodide Treatment: The develop- ment of these fish was not noticeably improved over that of the fish treated only with I131. There was, however, a definite improvement in the sur- vival rate (Table 10), and they appeared to be healthier and livelier than the sluggish I131- treated controls. Retardation of growth was not improved over the I131-treated controls, and none became sexually mature, with the excep- tion of one anomalous male (described above). Internal pathology was similar to that of the treated controls; vacuolated liver with its accom- panying abdominal rotundity was found with almost equal frequency. Gonadal development was improved slightly, as judged by counts made of oocytes in females (average of 45+ in 4 Kl- treated with a range of 20 to 84+ ; 14 in each of 2 radioiodine-treated controls) and by the number of males which had developed sperma- togonial cysts (3 of 5 Kl-treated; 1 of 4 radio- iodine-treated controls) in experiments I and IV. No evidence for the existence of thyroid tissue was found in 8 of 9 Kl-treated, and in 5 of 6 radioiodine - treated controls. Because radio- autography of Kl-treated fish had been found to be unsuccessful, producing totally blank film strips (Baker, 1958b), only three of the present Kl-treated fish were radioautographed. Some 1961] Baker-Colien: Role of Thyroid in Development of Platyfish 197 slight spotting was produced on these radioauto- graphs, but only in the pharyngeal lining and gut areas. The remaining fish were examined section by section with particular care, and in all except one case no thyroid tissue was found in the pharyngeal, renal, splenic and other areas. The exceptions were a male Kl-treated fish with a few regenerating pharyngeal follicles and a testis with the bare beginnings of spermatogonial cysts (experiment I), and a radio iodine-treated female with a single thyroid follicle in the kid- ney (PI. II, Fig. 2) and an immature ovary containing 14 oocytes (experiment IV). The amount of thyroid tissue present in these tv/o fish was apparently insufficient to improve their gonadal and visceral condition appreciably over that of other members of their groups. The latter exceptional specimen was the only instance of finding renal thyroid tissue after the destruction of the pharyngeal thyroid at an early age. This represents an incidence of one out of 79 radioautographed fish, with or without re- generating pharyngeal thyroid tissue (experi- ment 1 excluded). With the exclusion of fish given replacement therapy, the incidence is one out of 42 radioautographed fish, or 2.4%; the incidence of renal thyroid tissue among un- treated controls was 37 out of 50, or 74% (experiments 4-IV, in which sibling controls were kept). These results corroborate the ones previously presented by Baker (1958a) in sup- port of an hypothesis that renal thyroid tissue in the platyfish arises as the result of cell migra- tion from the pharyngeal area, rather than endo- genously in other parts of the body. 4. Radiation Control Experiments with Young Platyfish. Sixteen strain BH platyfish, in two groups, were treated with the amount of phosphorus-32 calculated to deliver whole-body irradiation equivalent to the average dose of radioiodine used. The conditions of exposure were similar. Nine of the 16 treated fish survived until ter- mination of the experiments, 109-178 days after treatment, and all of the 14 broodmate controls survived. One group of 5 fish was exposed to a P32 solution whose pH was not quite physio- logical (6.6), which may help to explain the high death rate seen among these fish, although the deaths did not occur until a month after treatment. As the single surviving P32-treated fish in this group was a male, three female con- trols were discarded. The over-all survivorship of the P32-treated fish was 56%, but that of the 109-day, pH 7.5 group was 73% (8 out of 11 fish). This is not different from the 109-day survivorship shown for I131-treated fish. Because of the small num- bers of fish, no pattern of death distribution was recognizable, but it seems significant that there was no evidence for a bimodal mortality distri- bution such as that found for radioiodine treat- ment. The average length attained by the P32-treated fish did not differ from that of their controls (PI. I, Fig. 4), viz., 20.4 mm. for the P32-treated (range 13.0 to 25.0 mm.), 21.1 mm. for the controls (range 18.5 to 24.5 mm.). Means were not corrected for numbers of fish of each sex, as these were as close to equality as possible in odd-numbered groups, being 5 males, 4 females among the treated fish and 6 males, 5 females among the controls. Upon fixation, at ages be- tween 4.7 and 7 months, all except one fish in the treated and one in the control group appeared to be fully mature by external criteria. More- over, the P32-treated fish did not differ from the controls in body shape, coloration or general health and activity. All of the fish were studied histologically. No abnormalities were found in the viscera of the treated fish, i.e., liver, kidney, spleen and thymus were normal. Goitrous thyroids were found in 3 out of 9 of the treated specimens, and in 5 out of 11 of the controls. Renal thyroid tissue occurred in at least 6 of the treated fish and in 7 of the controls. It should be noted that the kidneys were not carefully searched; follicles were discovered in spot checks of the kidneys, implying that they were quite numerous in the specimens in which they were seen. The only difference found between the P32- treated fish and the controls was in gonadal de- velopment, primarily among males. Five of the 6 control males had fully mature testes, full of spermatophores, as well as completely differ- entiated gonopodia. In the testes of the 6th, spermatogenesis had begun but no spermato- phores had yet appeared; the gonopodium was partially differentiated and was at a normal in- complete stage (Grobstein, 1953). Of the 5 treated males, 4 had fully normal mature gono- podia (PI. VII, Fig. 3), but none had any testi- cular germ cells. Only the ducts were present, but these were complex in structure (PI. VII, Fig. 5). One treated male had juvenile, simple ducts and an undifferentiated anal fin. Among the 5 control females, all had fully mature ovar- ies and two were gravid (PI. VIII, Fig. 1); all 4 P32-treated females also appeared to have com- pletely mature normal ovaries containing fully yolked ova and young yolkless oocytes in all stages of maturation (PI. VIII, Fig. 5). No counts of ova were made, but from the general microscopic appearance of the ovaries, there 198 Zoologica: New York Zoological Society [46: 16 appeared to be no fewer in the P32-treated female fish. From these data it is concluded that P32 ir- radiation, equivalent to that received in the I131 treatment, produced none of the pathologi- cal or developmental effects noted after radio- iodine treatment. The only effects were a very specific and distinctive action on the male gonad, which did not affect secondary sexual characters, and a similar total mortality rate to radioiodine treatment, up to 109 days post-treatment. The specific causes of mortality resulting from the two treatments is probably not similar, when the lower mortality rate of radioiodine- treated fish fed thyroid or Kl-treated is taken into considera- tion. 5. The Effects of Radioiodine Treatment in Adult Platyfish. a. Survival: Forty-six fish, all under one year of age, were injected with I131. Ten of these died within two days after the injection; this is at- tributed to shock or infection and is not consid- ered in relation to radiation damage or to thyroid condition. Two additional fish may also have died this way, but they were not recorded and are not considered here. Eleven more fish died within 8.5 months after injection; that is, Vs of the 34 remaining fish. Unfortunately, data on the mortality of the controls was poorly kept, but in Group 3, 3 out of 12 controls were re- corded as dying within 8.5 months after the injection of their siblings. This suggests a some- what higher death rate among the I131-treated fish, but is by no means conclusive. Six out of 23 radioiodine-injected individuals (omitting those sacrificed when younger and those dying from the immediate effects of injection survived to an age of 27 months or more. At death, histol- ogically visible thyroid was present in some of these old fish and absent in others. b. Gonadal and General Pathological Effects: It is doubtful whether radiothyroidectomy (or near-thyroidectomy) had any profound effect on the gonads of these mature fish, at least up to 6.5 months after treatment. In Group 1, two of the three fish had fully normal mature ovaries. The third fish had given birth 22 days before she was radioautographed and contained no fully yolked ova; normally these should have been present soon after parturition (Tavolga, 1949). Thus a slight delay may be noticeable in this instance. In Group 2 (females and males kept together) young were born in the tank up to three months after treatment. Three females, fixed within 3.5 months of the injection, had normal ovaries and one was gravid; the single male sacrificed within this period had normal testes. Among the female fish examined at longer in- tervals after injection (no males survived) , there appeared to be some effect on the ovaries. These fish include 10 females from Group 3 and one from Group 2, sacrificed 8.5 to 19.5 months post-injection. The youngest 5 were noted to be “thin” when radioautographed. Only one of these had any functional-looking ova; this fish had a single basophilic, pre-yolk-deposition oocyte, and primitive gonial cells appeared to be proliferating from the lining of the ovarian cavity, in small balls resembling primary sper- matogonial cysts in young males (PI. VIII, Figs. 7, 8). Although the future function of these cells is unknown, the condition is suggestive of Essenberg’s (1926) description of the develop- ment of testes from the epithelium of the ovarian cavity in sex-reversing females of Xiphophorus helleri. Essenberg’s suggestion of a relation be- tween depressed metabolism and sex reversal is also of interest in the case of the hypothyroid platyfish. In the other 4 fish, only fibrous atretic follicles were present. The 5 older fish from Group 3, which were 27.5 to 30 months old, in- cluded two with yolked oocytes, but only one of these had ova in earlier stages of development. The other three fish had ovaries with atretic follicles only (PI. VIII, Fig. 6). Although the three sibling controls for these last 5 fish also had atretic follicles in their ovaries, they had many ova in various early stages of maturation as well. The single female from Group 2 was 28-29 months old and had an almost completely degenerate ovary with practically no visible germ cells. This fish exhibited several other path- ological conditions (see below). Vacuolated liver in varying degrees was found in 4 of the 18 I131-injected fish, and ceroid-filled liver cells and/or ceroid-filled macrophages in the liver were seen in three additional fish. In one fish that died while in radioiodine tracer, the liver was decomposed too badly for its con- dition to be judged. Excessive peritoneal vacuo- lation was seen in 9 of the 16 fish that were freshly preserved. Concretions were seen in the kidney tubules of 5 fish, and degenerate tubules in two, one being the extreme case described below. Lym- phoid tissue in the kidneys appeared normal in all but the latter, and likewise the thymus glands appeared normal. The pituitary glands of 11 of the 16 well- preserved fish showed a prominent overgrowth of the pale intermediate basophiles (PI. VI, Fig. 6). This was found in all fish of the Fu strain. It also was found in all strain 163 fish fixed more than 8.5 months after injection, but not in strain 163 fish 38 days to 8.5 months post-injection. 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 199 This was a markedly higher incidence than in the young fish radiothyroidectomized by immer- sion, and is believed to be related to the in- complete degree of thyroidectomy in the injected mature fish, rather than to their age. c. Anomalies Noted: From Group 2, in one male and one female, aged 12-13 months, a large mass of pathological tissue was found in the abdomen among the pancreatic and fatty tissue. In the male, this mass was attached to the nor- mal testis (PI. IX, Fig. 1). These masses were encapsulated and contained collections of whorled fibrous nodes. These nodes resembled, in part, the fibrous “nests” seen in various or- gans of other fish, but did not occur elsewhere in the bodies of these two specimens. A similar anomaly occurred in one 27.5-30-month-old female from Group 3. In Group 1, a 15-month-old female had “nests” of fibrous tissue scattered throughout its body, including the dorsal pericardium and transverse pharyngeal muscles, the kidneys, liver, spleen and abdominal fatty tissue. None occurred in the ovary, heart and thymus areas. These “nests” of cells have been described in Kl-treated platyfish with thyroid tumors in their kidneys by Baker (1958b) and in Astyanax by Rasquin & Rosenbloom (1954). They probably represent a reaction to injection or other bodily insult. It is unlikely that they are directly re- lated to the I131 injection or to processes of thy- roid degeneration, as the fish in which they were seen had been treated 3.5 months or more earlier. The oldest fish in Group 2 (a 28-29-month- old female) was an extreme specimen. Its path- ology included an almost total lack of lymphoid tissue (no thymus, shriveled kidneys and spleen) and an extremely highly vacuolated liver and abdomen, with excessive deposits of pigment in the spleen and kidneys, large necrotic cysts in the liver and a pituitary overgrowth of such pro- portions that it might well be classed as a tumor (PI. IX, Figs. 3-7). There was no sign of hyper- trophied thyroid, as might have been expected from the condition of other organs; instead the thyroid area was filled with fibrous tissue and masses of material which suggested thyroid, but which had flat margins and was stained a pale bluish color. The exact nature of this material is unknown, but the presence of thyroid cells is most unlikely. This fish appears to be the most extreme case of pathological hypothyroidism met with in this study. This fish also contained some of the “nests” of cells mentioned above, especially in subcapsular edematous spaces in the shriveled kidneys and in the necrotic cysts of the liver. The kidneys were deficient in glome- ruli; many tubules seemed excessively small, others contained large concretions and had clear, “blown-up” cells that were unquestionably de- generate.7 None of these anomalies was found among the controls, but because of the small number and the relatively advanced ages of these fish, the peculiarities cannot definitely be related to hypothyroidism, except to surmise that such ani- mals are very likely more prone to infection than normal similarly aged fish. A study of general senile degeneration has not been made on nor- mal untreated Xiphophorus maculatus, but it is suspected that some of the effects noted here may be the result of a slightly premature appear- ance of such degenerative processes. Other ef- fects appear to be definitely related to thyroid function, irrespective of age, e. g., vacuolation of the liver. In addition to the pathological anatomy, it should be noted that renal thyroid tissue was seen, without use of radioautographs, in all of the 6 control fish and in the spleens of three of them, but such tissue was not found or radio- autographically suggested in any of the I131- injected platyfish. d. Young Born after Z131 Injection of Mother: One female of Group 1 gave birth to young 19 days following the injection of approximately 80 /ic. of I131. Two of these young were radio- autographed with the mother when they were 22 days old. One was found to be totally thy- roidectomized, both by autographic and histo- logical criteria. This was a female; the other was a male that had many thyroid follicles. These follicles, as is typical in very young fish, were made up of high basophilic cells. Both young seemed normal, except that the thyroidectom- ized female showed some vacuolization of the liver, and an aorta that seemed shrunken. Both young had undifferentiated gonads, although they definitely were sexable, and the female ap- peared to have large, clear oogonia at the cran- ial end of the oviduct. The testes of the young male fish were almost as well developed as those of some of the totally radiothyroidectomized fish 6-9 months old (PI. VII, Figs. 6, 7). Since no normal young of equal age were sectioned, gonadal development was not controlled in these young fish. According to the accepted gestation period of platyfish, i.e., 22 days (Tavolga, 1949), the mother fish was injected when the young were 7The kidneys of this specimen were in an even more pathological condition than those of fish with the most extreme renal thyroid tumors described by Baker (1958b). 200 Zoologica: New York Zoological Society [46: 16 only three days in development. Since the thy- roid gland does not appear until about the 8th day (Tavolga, 1949), enough radioiodine must have been retained by the yolk or the maternal tissues to affect the young thyroid when it first began to concentrate iodine, 5 days later. An- other possibility is that the treatment slowed down the development of the young, so that they were bom belatedly and were actually at the thyroid development stage when the mother was injected. It is known that most non-thyroid tissue and the yolk of platyfish ova do not re- tain iodine in protein-bound form (Baker, 1958a) , but it is quite possible that iodide may be retained in such tissues for considerable periods. This instance suggests that sufficient I131 could be retained in the body of an adult fish to thy- roidectomize both the adult, and later, young fish developing in its body. There was no pos- sibility of any appreciable amount of I131 re- maining in the aquarium by the time these young were born because of the number of times that the water had been changed since the time of injection. Thus, it seems probable that embryonic platyfish can be thyroidectom- ized in ovarum and still develop to a normal birth. Discussion 1. Radiothyroidectomy Table 11 summarizes published attempts, to date, to extirpate the thyroid of teleosts using radioiodine. These few attempts have met with mixed success. It is possible that the differences between the activities of the thyroid found in various species of fish are responsible for the variety of results obtained (as an example of such differences, see the three species of Fundulus compared by Gorbman & Berg, 1955). It seems likely that less active thy- roid glands, which both concentrate and release iodine and thyroxine or its relatives less rapidly, will receive more irradiation from a relatively prolonged dose of I131 than will very active thy- roids. It is believed that the differences in the efficiency of radiothyroidectomy in the various strains of platyfish described here were at least partly due to genetic differences in iodine re- quirements and metabolism. Those strains that were more goiter-prone (or had less actively secreting thyroids?) seemed to be more readily radiothyroidectomized. Differences in the radio- resistance of thyroid cells might also be a factor among different species or populations of tele- osts. The platyfish discussed here were probably Table 11. Summary of Reports of the Use of Radioactive Iodine in Thyroidectomy of Teleost Fish Author and Date Species and Stage used Degree of Thyroi- dectomy Obtained 1 Observations Made LaRoche & Leblond, 1954 Atlantic salmon ( Salmo salar), parr Complete No effect on growth; reduction of skin pigmentation Baker et al., 1955 Platyfish ( Xiphophorus maculatus), young Some complete No effect on growth and sexual maturation1 Arvy et al., 1956 Rainbow trout ( Salmo gairdneri), parr Incomplete No effect on O2 consumption Fromm & Reineke, 1956 Rainbow trout, finger- lings Complete No effect on O2 consumption Fontaine, Y.-A., 1957 Eel ( Anguilla anguilla ), silver adult females Complete Decreased hypophyseal thyro- tropin; no size difference in hypophysis; no TSH in plasma Fontaine, M. et al., 1957 Eel, adult silver Presumed initially complete Fall in O2 consumption after 48 hours; later recovery pre- sumed correlated with thyroid regeneration Olivereau, 1957 Eel, small males; large females Incomplete Thyroid destruction and regen- eration studied Baker, 1958a Platyfish, young Many complete Some pathological effects not- ed; effect on heterotopic thy- roid development studied Harris, 1959 Killifish ( Fundulus heteroclitus) , adult males Mostly incomplete No effect on salinity tolerance nor on growth or gonads ’These early published conclusions were mistaken and were corrected in the oral presentation of the material (see also Pickford, 1957). Complete results are embodied in the present paper. 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 201 more readily thyroidectomized by I131 than some of the fish used by other workers because they had been bred in an especially iodine-poor environment and because their iodine-to-thyrox- ine turnover was relatively slow. Despite the ad- ministration of thyrotropin prior to each of several I131 injections, Harris (1959) was unable to obtain complete thyroidectomy in most male Fundulus. Perhaps a more protracted period of exposure to I131 would be required if the turn- over rate of iodine in these fish were relatively high within each post-injection period. 2. Thyroid and Development— Various as- pects of the function of the thyroid in fish have been thoroughly reviewed recently by Lynn & Wachowski (1951), Hoar (1957), Pickford (1957) and Leloup & Fontaine (1960). Papers that present evidence for thyroidal involvement in development and sexual function of teleosts, based on observations of non-experimentally treated fish, are included in these reviews. Some of these papers are omitted from discussion here. A synopsis of exclusively experimental observa- tions on the effects of thyroid preparations and antithyroid compounds in the development of teleost fish is presented in Table 12. Many of the entries in this table have been more extensively summarized by Pickford (1957). Therefore the substance of the table will not be discussed here, except in a general way. Among the more striking features to be noted in Table 12 is the heterogeneity of the results obtained with thyroid preparations and the rel- ative homogeneity of those obtained with anti- thyroid drugs. One reason for this is probably the uniformity of dosage employed in the latter groups and the chemical uniformity of the drugs employed. Nearly all investigators agree that treat- ment with thiourea or thiouracil retards both growth and sexual differentiation in young fish, and brings about regression of gonads and loss of weight in adults. The effects of radiothyroidec- tomy of young platyfish observed by the author agree with this consensus, except that gonadal effects on adults were doubtful. LaRoche & Leblond (1954) reported that radiothyroidec- tomy had no effect on growth of salmon parr, but since histological criteria alone were used to determine the extent of thyroidectomy, it is suspected that the fish may not have been totally thyroidectomized. In some of the platyfish de- scribed here, it was found that a barely visible amount of thyroid tissue was sometimes suffi- cient to support sexual maturation and growth to normal adult size, and this barely visible tissue was found only with the aid of radioauto- graphy. The results of treatment with thyroid deriva- tives run the gamut from retardation to enhance- ment of almost every feature of development. This variety in response is probably the result of the different methods of administration, types of preparations and dosages that were employed. In all instances, the fish undoubtedly were re- ceiving abnormally high amounts of thyroid hormones, especially as their own thyroids were also present, even though their activity may have been inhibited. The occurrence of exophthalmia, abnormal gonopodial development, vascular ab- normalitites and deformations of the head cer- tainly cannot be considered as physiological, and it is hard to assess such peculiarities in terms of the normal function of the thyroid hormone. The most commonly described findings seem to be: slenderization in body shape; increased fin, skin and connective tissue development; silvering (increased guanine deposition) and reduced pigmentation in salmonoids; and exo- phthalmos. Yet even among these, there are reports describing the exact opposite, or the absence, of these features— sometimes in the same species. The effects on sexual development and over-all body growth are even more conflicting. Some of this disagreement may be related to the age of the fish tested, as well as to species differences in the coordination of thyroid func- tion with developmental stages. Observations reported here suggest that reproductive proces- ses in fully adult platyfish are relatively inde- pendent of thyroid function, just as growth, once completed, no longer can be generally in- fluenced. Histological evidence suggested a sharp falling off of thyroid activity with age, after full attainment of mature body size, in this species. Among young poeciliid fishes (including the guppy, platyfish and mosquito fish, Gambusia) , increased growth rate, early male secondary sex differentiation, abnormal gonopodial develop- ment, increased fin length, and exophthalmia have usually been produced by treatment with thyroid derivatives. The results of thyroid-feed- ing of athyroid platyfish, reported here, concur with increased growth (increased to normal, that is), slightly accelerated male secondary sexual development, and some elongation of the caudal fin. They disagree in that gonopodial differen- tiation was completely normal in all cases and that no exophthalmia occurred. In comparison with controls, the proportion of mature fish was slightly higher at the conclusion of the ex- periments and only among the thyroid-fed, athyroid fish was a pregnancy found. It is believed that the normal development of these fish took place because they were given very little thy- roid material. For example, the amount of dried 202 Zoologica: New York Zoological Society [46: 16 *-» G O 8 s U t-4 G O to V I G .2 G .2 2 4> o to w to #G (h «D *c/s u Hi CO o §8 o a o 2 <-* &a • o 5i U G £2 'w' CCS e-8 M a 9 CJO .s *2 J8 Ut G to O •*-» G a #o g u 1> a < a a> ■8 O to M ja I * oG to o 1-4 o-. S o ■4-» i & o o G fr’s u W) G § § G O | 8 1 'Si o t> O G G to ,t> O to to O G O to to 04 o X Z e*S z 8 P4 6 4> ^ a o § ■g 3 i’s W>§ O S„ a a* .2 .2 G G to to 2 0) a « a ^ a g S?" S' O to o G o 8 G o> “ 8- O CM c* CM ■S Os O ^ Jfc* o ^ G l-i • >» 4-H 0) ON a> s - to Q> «8 § & U CCS *g to to »-i 8 nJ a :s as w VO rf Os o w c3 to s CO •*t Os t! to 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 203 >> 73 I I I Q) CO •s *3 *8 a *5 a 2 a -B a .S g.9 8 8 -•s s §>§ 3 «a ” S 5 2 ^ SU a'o °°uS 2 ^tT 3 » «« £3 'S O S 0 o> O ^ 1 till o a •*-* 0X3 2'O'y go c/3 +3 d O 2> a o Ou 73 d d d OD OX) .9 CO •S .a a> 8- X3 73 a jl> 00 pL| 0L, a !S .X H cj , w m m 3 rt E Q w fl ra ►- CO co co d Z co - a w : ^ ,2 w) a o .2 ■ s <* 0© o *n ON ~ rQ C r ^ co ^ 3 (SI O *8 * H M £ £ 0-5 •i S * d hJ £ * S .9 O a •« o s: Ps •a V. C> •a 9 «« OX) S| § | sa *d _T C S -d a O .52 -i* o »/-> ON »n On o &oc 5 rt ffi j O d> CO CO -J* o o -5: a •H *55 a •c g 5 o OX) d S3 O c § a -a S a to fc d Oh O o e ^ *d d C/3 O o £ OX) .9 (g s § | I *1 d .H co £ O -g .- -O Oh * 5 « II J o VN On O & d o *n On *5b d o ffl *8 d CQ d O S3 204 Zoologica: New York Zoological Society [46: 16 T3 1) DO 55 fl Oh CO a S 00 8 a 00 o 43 e- o a .2 "9 S3 2 o Pi -s 5 X co O 2 13 .*=1 s « .-a ^ 2 x> c £ id a fe a J3 2 J3 O a c O CH O 3 h >.0 D-M B P\ •B O 2 03 'w' X « 8 2 ^ *1 ’’O a cd fc cd a- «s 00 !3 33 O h-> ^ £ § 8 < 1 m H 'O — T3 J2 « « .« « o ’C T3 O i* ■a *0 ^ S' o .. o.g >^5 53 ^ '" .a & .9 x o !h Jr «o cd g O hSh a ^ 03 ta C4 Pd a> #co a. Oh T— 1 C cd cd o o O >—5 »-) a EC -*-* HH , 43*' o O h -o >» Oh Oh Oh Oh CO a +£ cd q ^ CO 44 03 O T3 >>9 Oh — Oh — 33 33 O < 1 2 o 3 O o o O 03 O a o :P s >/1 ON £ £ W3 O &* X) 6 v o ll c i> 03 4= « o a .2 o> o «sa e> a ~ o - -o w „ § 'D 2 g § is <4-4 C/3 o > (H cd g X! £ E oo a o us & fi* c ^ a J3 t> o .£?« — • s p OX) o i ^ fr o Si o •o a T3 P p & •2 £ 03 o •d . > r, ?o S ►. LO • " X O 03 O ^ gj ■O fl'O si « p°- ■3 >-> p T3 5 U O 00 P ’2 =*- O .2 n .a-S S CO 4-* *— < U . . Q o o L t> a o •p CO T3 g CL> CD vh o d 52b o ■a o _ p O 2 ■a ^s“ 8 ^ -S •s' .p L ’S x — eg o T3 o r u ^ H .EcP, X G T3 03 P G O ^ tT 0) *S~) Os OX) a G o ;>v eg 2 *3 O a (a r** «n 0N IZN ON V=1 D V“4 X T3 *d fi °3 M o cd S3 u 2 c_, p § o o K=»J -1 sr =^c o £ o «o X u O O "G G 73 03 B 3 :s "3 »a s o G 73 03 W) G 1 G O >* _G •g G o > HH T3 T3 O o I-, 2 «) g^g Sw - u 3 CL K O w a ° &G ° 2 o g a CD S P SobS1*1- o P S P P E^ ^ 1 p -o 2 G its <4-1 lllao eS «-( .2 a u. O O -K O > ° S l-H O >n_4 Y o ox) -G x r. hJ .2, G. H H - •c ^ a o O G .•g-g-s O 05 b ° ^ ° »- *o « Hen D, o Cl, CL G o o\ « CL §■ K B. Effects of treatment with antithyroid drugs: Goldsmith et al, 1944 Platyfish X Swordtail hybrids, 0.033% thiourea by immersion Retardation of growth and failure of sex differen- young tiation Nigrelli et al., 1946 Guppy, young 0.03% thiourea by immersion Inhibition of sex differentiation Zaks & Zamkova, 1947 Loach, embryos Thiourea 0.033% by immersion No effect up to 26th day of development 206 Zoologica: New York Zoological Society [46: 16 1 £ f 55 •o a 3 Cl, CO o u O I D 3 « S; | « 35 g*r SCO H co .£p OJ3 O g ■£ ^ O - cij - O 35 3l*g B ’§ .« ^ o vS «0 O S' «s -5 no " O O 4> © d E #2 06 3 ion im a o S2 >v *« « ^ 1 S Ip a> a .§ s a o S'- 3 §1' 3 « o *g 3 Si § a ^ 3 SS o d .o' £ m o o a 60 ? -g :&.fH «a 2 O < s> ,3 <0 N 0 £ 1 8j CO 3 CO Q GS G\ <3\ „ ^ «o ©\ l! - 2 b04 2 c§ W PH >3 CO 5 o Q* & 33 O «8 Oh g 5 5 n os o .a C3 o ir> $ co © d na o cd Ih § 3 >% JO co o o ©a § o p»> b2 | .9 3 g* S 3 O 00 o >» £ O' o 00 o Ps o. Oh 3 o «o ov 3 s- Cu o 33 > o £2 -D a « 3 O t- •H a £ o tp m i 5 * .aja C O t> -^(S % rfX '5*' b ^ S3 ° .E.^3 s-s s Ip hh 333 O MO 3'S s H ? T3 3 •o cd ■§8 §1 fe. a co wo ©s B 0 ja 1 6 CO OX) a as o CC cd N Author and Date Species and Stage Treated Method of Treatment Observations and Conclusions 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 207 a 5 6 •2 o G *0 .a too O Id 4) o Oh O «5 HH * O H-. !» u X Oh^ 2 a:: 4) O 153 P ^ 4> a o a g* c o .2 .ts C/5 x §* toO X a> iT .3 2 § x& I’S O c/5 N cd OOg a T3 o O 4> 4> -r- O -O ^ •O H C u o 2 & s E a M - § 55 o H-H (U . > q T3 S? o £ & I O C/5 4) cd o HI V £3 643 4) CO C/5 < 8 to CO flj S o b z, 3 O 3 £ a 4) > 'Sb X O 1,3 ® 2 3 h e ° 9 2 3 ® E H .o' $ (N © •O a G rH O • S o .5 o S O •» o x 2 Cd - Z 2 ^ cd O £ ° a »o O - 3 :3 - S g § 6 cd z p c4 G >> .2 x G ** •§ g u »o © E o S * S Q> a E x © o o >» x S’ S3 3 s 9 a «o O X O o o D a a >% X) G G3 G o S- o .e ss § «o «+-» I -a 11 CO ’e M E ^7- ■«< Vh 2 § «a * H S' «o o o G 0 1 C >,.2 C a o E 73 < n E 3 X O o\ 25 •=> o 8 S’ fc 2 S 4) X) O 4> 4» O W N to ; 2 5> 00 a 3 o >. ?C, Oh Oh a I Vh CQ 8 •3 E 42 4> O 4» so® II I o cS ~ s ■oj 2 “•SO S > O X E 4> 3^. Oh O. 3 o •D c x X* I! Ill o^lSo E C/5 ,»d* »o C\ a o to .E PQ in ON cd O ffi *cd Q NO ON *0 ON H #3 •S § 50 •c .-S fc o no ON «-T cd h X E «o NO ON t: o G s <*3 cd E G O X N© «o ON 1-4 X E & *o ON G toO tD oo «o On 4> Oh •fr *Im cd X U £ o ON NO ON *-T On o «o 2 ^ *Sb^4 si O NO ON G E a o S S On «*=H G .§ I ‘g4 208 Zoologica: New York Zoological Society [46: 16 thyroid initially given (as food), if dissolved throughout the water, would have been approx- imately 0.0016%; this eventually was doubled but not until the fish had grown considerably larger. This was one-sixth of the concentration used by Hopper (1952) for young guppies (0.01%), and when increased was still only one-third of his dosage. Without chemical as- says, it is hard to compare doses of dried thy- roid with doses of pure thyroxine, but probably many of the latter were excessive. For example, the 5-10 mg./liter dose of Buser & Bougis (1951) seems to be an extremely high concen- tration of thyroxine, since it is almost as high as the dried thyroid concentration first cited, and desiccated thyroid certainly is composed largely of materials other than thyroxine itself. That this dosage produced exophthalmia seems not surprising. In studies on the effects of antithyroid drugs on fish, darkening has been noted: e.g., as a result of the increase in pigmentation (Sembrat, 1954), or a decrease in guanine desposition in the scales, which then revealed the underlying pigment (Dales & Hoar, 1954). Although dark- ening was noted among the radiothyroidectom- ized fish studied here, the effect was not ana- lyzed. 3. Thyroid Activity and Liver Function— Ac- cording to Spellberg (1954), fatty liver in man has been associated with hyperthyroidism, as well as with “anterior pituitary hyperfunction” —among numerous other causes. The first ob- servation differs from experimental results in other mammals. Chaikoff et al, (1948) found fatty liver with fibrosis in thyroidectomized dogs after 193 to 799 days, and McClosky et al. (1947) noted fatty livers, as well as iron-pig- ment deposits in the spleens and livers of cats fed thiouracil for 7-8 months. Thirty days of thy- roid-feeding reduced fat in the livers of normally fed rats, and protected rats on lipotrophic diets from the development of fatty liver (Leathern & Howell, 1950). These authors also noted that thiouracil, fed for 35 days, caused an increase in liver weight and cell size, without change of fat content. Similar observations were made on rats treated with thiourea up to 60 days by May et al. (1946); they found no amelioration of thiourea effects by simultaneous thyroid treat- ment. Leblond & Hoff (1944) also found that thiouracil-feeding of rats for 21 days led to an in- crease in liver weight, but that a decrease oc- curred in thyroidectomized animals. No chem- ical or histological analyses were presented. Sellers & Wen You (1951) found that thyroid hormone influenced the site of fat deposition in the liver of rats. In thyroid deficiency, fat tended to be deposited intracellularly throughout the liver, but in the normal, and especially in the hyperthyroid animal, fat accumulated in the central part of the liver lobule and extracellular fatty cysts were formed. Sellers & Wen You suggested that the protection afforded by pro- pylthiouracil against experimental cirrhosis might be due in part to the reduction of thyroid hormone level, which shifted the pattern of fat deposition from extracellular to intracellular. Finally, Goldberg et al. (1950) found no per- sistent pathological changes in non-cervical tis- sues, e.g., the liver and kidneys, of radiothyroid- ectomized rats up to 8 months after treatment. In fish, only a few observations appear to have been made on the liver in relation to thy- roid function, and these are not altogether in agreement. Hatey (1950) reported that thiourea treatment of carp ( Cyprinus carpio) by immer- sion led to a decrease in liver size and liver gly- cogen within 3.5-11.5 days. Chambers (1953) also found a depletion of liver glycogen in Fun - dulus heteroclitus, treated with thiourea by in- jection for 6 weeks, but this was accompanied by an increase in liver size. Liver cell size was increased and fat content decreased. Bickford (1952) noted a decrease in liver size of hy- pophysectomized male Fundulus heteroclitus treated with thyroxine, but the difference from the controls was not considered significant. The observations of Chambers and Bickford agree with the results obtained for mammals with re- spect to organ and cell size, but both Chambers and Hatey found glycogen level to be affected in the opposite direction than that seen in rats (Leathern & Howell, 1950; May et al., 1946). On the other hand, Fontaine et al. (1953) ob- served a decrease in liver glycogen after long- term, high dosage thyroxine treatment, at high temperatures, of eels ( Anguilla anguilla) and rainbow trout ( Salmo gairdneri), an observa- tion which agrees with the mammalian results. On the basis of these results, these authors hy- pothesized that the drop in liver glycogen seen in smolts might be at least partly due to thyroid hyperfunction. Liver size, histology, and fat and glycogen reserves of teleosts differ in the two sexes, and also vary during different stages of the breed- ing cycle (Olivereau & Leloup, 1950; Immers, 1953; Clavert & Zahnd, 1956a, b; 1957; Zahnd, 1959). In female poeciliids the liver/body size ratio was larger than that of males, but it de- creased during the development of the egg yolk, and fat reserves were depleted (Clavert & Zahnd, 1956a). In inactive females and in males, fat reserves were large. Female hormones given to 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 209 such fish of either sex induced liver changes parallel to those seen in females during vitello- genesis (Clavert & Zahnd, 1957). Tavolga (1949) found that certain female sex hormones when fed, exerted an opposite effect on the livers of platyfish. Liver cells became vacuolated and fatty changes took place— apparently in both sexes. Egami (1955) reviewed the Japanese literature reporting sex and cyclic differences in the liver of fishes. He found in Oryzias that the liver is larger in females and has larger cells. Implantation of estrone pellets into sexually active males led to liver changes in the female direction, but in inactive males there was no effect. Implantation of testosterone pellets or ovariectomy of females caused liver changes in a male direction, but more slowly than the changes in treated males. In sticklebacks ( Gas - terosteus aculeatus) and minnows ( Phoxinus laevis) , glycogen storage was greater in females, but it was largely located in the ovaries (Immers, 1953). Its changes were not regarded as responsible for the fluctuations of liver weight occurring during the sexual cycles. Oguro (1956) found the liver of female sticklebacks to be larger than that of males; there was more fat in the male liver, its cells were larger, its cells contained multiple nucleoli, in contrast to single nucleoli in female liver cells, and the mitochondrial morphology differed in the two sexes. Estrogen injected into males caused liver changes in the female direction. Although the vacuolation of the liver in the radiothyroidectomized platyfish here described could have been due alternatively to accumula- tion of water, glycogen or fat— all of which could have presented the same picture after the solvent treatments used in preparation of slides —the histological picture was not like any of the illustrations offered by the above authors. Sexual inactivity, although a cause of fat or gly- cogen deposition in the liver, did not lead, in the reports cited above, to histological appear- ances even remotely approaching the extreme degree of vacuolization seen in thyroidectom- ized platyfish. Moreover, normally maturing, incompletely thyroidectomized platyfish devel- oped as severely vacuolated livers as immature, totally thyroidectomized specimens. Therefore it may be concluded that the role of sex hor- mones in this instance is of negligible impor- tance, in contrast to that of thyroid hormone. The vacuolated livers in the thyroidectomized and hypothyroid platyfish, if they do represent fatty livers, developed under conditions which accord with the results presented by Chaikoff et al., and the absence of vacuolated liver in thyroid-fed thyroidectomized platyfish is in line with the findings of Leathern & Howell. These mammalian results, and others summarized above, suggested that the presence or absence of thyroid hormone might have a different action on liver physiology than treatment with anti- thyroid drugs. Only the cats of McClosky et al. exhibited fatty liver after chemical thyroidec- tomy. This experiment was, however, the only one in which either mammals or fish were given antithyroid drugs for more than 60 days. Chambers observed ceroid in the livers of all Fund ulus that had been thiourea-treated for 6 weeks, and the present author found ceroid in the livers of platyfish thiourea-treated for 10 weeks. Furthermore, vacuolated livers with cer- oid were found by the author in platyfish treated with thiourea for 1 8 weeks. If ceroid is a product related to neutral fat,8 its over-accumulation might precede and herald the onset of fatty liver. The development of fatty liver might re- quire more prolonged antithyroid treatment than most of these experiments encompassed. The data reported here on thyroidectomized platyfish, as well as on thiourea-treated platy- fish, show that vacuolated liver developed grad- ually over a period of many months (with one exception, vacuolated liver was not found in radiothyroidectomized fish examined prior to 60 days post-treatment). Extensive ceroid deposits found in several specimens were tabulated as equivalent to vacuolated liver. The observations of Sellers & Wen You may reconcile the above-cited association between fatty liver and hyperthyroidism in man with the experimental results on other mammals if, in the human cases, the excess fat were present extra- cellularly. In the thyroidectomized and thiourea- treated platyfish the vacuolation was certainly intracellular, so that no conflict appears here. The failure of Goldberg et al. to find any patho- logical changes in long-term-thyroidless rats alone appears unreconcilable with other obser- vations. Only genetic differences come readily to mind in explanation of this inconsistency. 8Ceroid is a clear yellow pigmented material which first was described in experimentally-produced liver cir- rhosis of rats (Lillie et al., 1941; Edwards & White, 1941). Its histological appearance and histochemical characteristics were described by Endicott & Lillie (1944); among these are insolubility in alcohol and ordinary fat solvents, and non-stainability by haematoxy- lin-eosin or Masson’s trichrome (used exclusively in the experiments reported here). The identification of the material seen in the livers of the author’s radioiodine- treated fish was first suggested by Chambers’ (1953) description of ceroid in the livers of Fundulus treated with thiourea. Endicott & Lillie suggested that ceroid might either be a derivative of neutral fat or a separate product of abnormal liver metabolism. Ham (1952) de- scribed ceroid as an oxidation product of unsaturated fats. 210 Zoologica: New York Zoological Society [46: 16 It seems clear, at any rate, that in platyfish the thyroid hormone is deeply involved in liver metabolism, and that lack of it leads to excessive accumulation of ceroid, and probably of neutral fat.9 A.Thyroid Function andPituitary Development. —Volumetric studies on the pituitaries of thy- roidectomized and normal platyfish suggested that the glandular portion of the hypophysis was sharply reduced in thyroid-fed, athyroid fish, whereas the neural portion developed normally. In solely radiothyroidectomized fish, the entire pituitary appeared to be subnormally developed relative to the size of the animal. This was some- what surprising, as lack of thyroid hormone might be expected to lead to increased thyrotro- phic function on the part of the pituitary, and pos- sibly thereby to an increase in over-all pituitary size. Control platyfish with hypertrophied thyroids were often found to have large ventral baso- philic overgrowths of their pituitaries. Enlarged glandular portions of the pituitary also were found in goitrous teleosts by Stolk (1956a,b). Increases in cell number or size among pituitary basophils have been described in antithyroid drug-treated teleosts by Leloup & Olivereau (1950), Scott (1953) and Sokol (1957), and an increase in the number of eosinophils by Vivien & Gaiser (1952). Without specifying cell type, Honraa & Murakawa (1955) stated that cells of the transitional lobe of thiourea-treated sal- mon larvae were enlarged. However, none of the latter authors mentioned a gross increase in the size of the whole pituitary. Atz (1953) and Barrington & Matty (1955) described degranu- lation of thyrotrophs in the pituitaries of anti- thyroid drug-treated fish, but did not mention increase in cell number or cell size. The excessively small pituitaries of the radio- thyroidectomized platyfish may have been the result of a general exhaustion of the several pituitary cell types in a futile effort to respond to a variety of physiological deficiencies triggered by long-term athyroidism. Severinghaus (1942), as quoted in a review by Maqsood (1952, page 304), stated that early thyroidectomy in mam- mals may inactivate the pituitary to the extent that a virtual hypophysectomy is produced. 9Increased deposition ot iron-containing pigment in liver and spleen was among the effects produced in mice that were fed thiouracil on a long-term basis by Dalton et al. (1946). This was noted also by McClosky et al. (see above) in their thiouracil-fed cats. Although this feature was not systematically studied, it is the author’s impression that excessive brown or black pig- ment deposits were present in the spleens, though not the livers, of thyroidectomized platyfish, and that such deposits also were more frequent in the kidneys, as com- pared with controls (see PI. DC, Figs. 4, 5). Maqsood cited the findings of several other workers that support this statement. The small size of the glandular portions of the pituitaries of thyroid-fed, athyroid fish, on the other hand, was probably due to inactivity, at least on the part of thyrotrophs, for their function was no longer demanded for normal physiological func- tion of the whole animal, owing to the illusory production of an excess of product by the target organ. 5. Thyroid Function and Behavior. An in- crease in sustained activity after thyroxine treat- ment and a decrease in activity following thiourea treatment were observed in goldfish and salmonoids by Hoar et al. ( 1952) and Hoar et al. (1955). The studies of Woodhead (1959) have shown the importance of thyroid gland activity in the long migrations made by sexually immature, as well as mature, cod ( Gadus callar- ias) . Many years ago, Marine & Lenhart ( 1910) noted the sluggish behavior of hypothyroid (goitrous) trout, and they also noted that such trout died more readily than normal fish when taken from the water. Other examples of slug- gish behavior after antithyroid drug treatment may be found in Table 12. Although the be- havioral observations made on radiothyroidec- tomized platyfish in this paper were casual, it seems appropriate to note that the activity pat- tern was in accordance with detailed studies such as those cited, and that greater fragility, as noted by Marine & Lenhart, was found. 6. Radiation Ejects.— Swelling and hyaliniza- tion of blood vessel walls in the thyroid area fol- lowing moderate and high dosages of radio- iodine was noted in adult male rats by Goldberg et al. (1950). Their description applies also to the effects seen on the aorta of radioiodine- treated platyfish. Similar in both rats and fish was the fibrous tissue replacement of thyroid tissue, sometimes in whorled form. The process of thyroid destruction in rats and fish followed parallel courses which were characterized by the swelling of the epithelial cells of the folli- cles, the loss of dense colloid, the disorganiza- tion of follicle structure, and the final disappear- ance of all thyroid tissue, leaving the stromal framework to be infiltrated by fibrous elements. Most work on the effects of whole-body irra- diation on fish has been done with either embryos or mature animals, rather than on young, grow- ing fish equivalent in development to the platy- fish studied here. Bonham et al. (1948) studied the effects of x-rays on fingerling Chinook salmon ( Onchorhynchus tschawytscha) . After 12 weeks, growth in length was affected only in fish that had received 1,000 r. or more. The cumulative death rate in fish receiving more 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 211 than 250 r. was significantly higher than in con- trols. In fish receiving 500 r. or more, the renal hematopoietic cells were reduced in the first two weeks, with recovery by the 4th or 5th week in survivors of 500-1,250 r. The gonads were not studied. Welander et al. (1948) found some- what greater sensitivity in embryonic and larval Chinook salmon than in the older fish. The lowest dose affecting growth in length was 500 r., and the lowest dose affecting hematopoietic tissue was 250 r. At 500 r., neither mortality nor weight was significantly different from the con- trols. Pigmentary development was affected only in fish receiving 1 ,000 r. or more. This also was true of other indices of development, such as eye size and gill growth. Recovery of hemato- poietic tissue occurred in fish receiving 1,000 r. or less. Little or no damage to renal tubules and glomeruli was found in the latter groups. Splenic damage paralleled that of the renal hem- atopoietic tissue in dose effect and recovery. The gonads were retarded by all doses (250 r. and up), but nevertheless, in those receiving 250- 1 ,000 r., they were progressing to definitive sex cell stages, 93 days after irradiation. Other or- gans examined were not affected by doses of 1,000 r. or less. In mammals, including man, the hematopoie- tic tissue is especially sensitive to radiation (Tullis, 1959)., At LDso or higher doses (gamma radiation from atomic explosions), hemato- poietic tissue was destroyed very rapidly in the lymph nodes and bone marrow, less rapidly in other lymphoid tissue. However, Tullis noted that the reticular (stem) cells are among the most radioresistant in the body; recovery of hematopoietic tissues began within about 4 weeks in survivors. The liver and kidneys were not found to be importantly affected. (Tullis men- tioned that “Reversible accumulations of sudan- ophil fat in the liver have been noted in many animal species,” but gave no specific references on this point). Ovaries were noted to be more radioresistant than testes; primordial ova are re- ferred to as being among the most radioresistant of cells. Interstitial cells of the testis also were not noticeably affected. Metcalf el al. (1954) found that in rats given 550 r. whole-body x-radiation, lymphoid degen- eration in lymph nodes, spleen and bone marrow occurred, but returned to normal in 17-40 days. Spermatogonia, after destruction, reappeared in 20-40 days. There was no injury to the liver or kidneys. Lushbaugh (1957), in a review, states that the normal liver is resistant to radiation up to 3,400 r. in mouse, rat and man, and that 20,000 r. did not affect liver regeneration in rats. The pancreas is also radioresistant, requiring more than 2,500 r. for visible effects. The effects of I131 treatment of young and adult platyfish were similar in some respects to those reviewed above, resulting from other forms of irradiation. In particular, the hemato- poietic reduction found was characteristic of radiation damage in other animals. However, the persistence of the condition, sometimes acutely, in survivors for 4-8 months after irra- diation apparently is anomalous. Moreover, hematopoietic tissue in platyfish similarly treated with P32 in equivalent r. dosage (720 r.) was normal 3-4 months after exposure. These ob- servations indicate that athyroidism or extreme hypothyroidism may prevent normal repair of hematopoietic damage resulting from radiation. This is borne out by the improved condition of this tissue in the thyroid-fed, radiothyroidec- tomized platyfish. The effect of the thyroid here, as in other instances, may not be tissue specific, but merely a reflection of a general positive effect on cell growth. Although fatty liver apparently may occur transitorily in various species after radiation, this does not seem to be a commonly noted effect. When the P32-treated platyfish, whose livers were quite normal, are considered in re- lation to the I131-treated fish, the evidence favors the interpretation that the high incidence of severe liver vacuolation found was related en- tirely to thyroid insufficiency and was independ- ent of radiation. Again, the livers of the thyroid- fed, thyroidless fish were completely normal, which further substantiates this assumption. Growth retardation in young salmon, although a definite effect of radiation, was not found to be significant at 720 r., the dosage level received by young platyfish treated with I131 or P32. There was no effect on the body length attained by P32- treated platyfish, nor were their body propor- tions noticeably different from those of the con- trols. Moreover, the thyroid-fed athyroid platy- fish were normal in body length and shape; in fact, a slight growth enhancement appeared in these fish, in the form of elongation of the caudal fin. The retardation of growth found in the solely radiothyroidectomized fish therefore may be attributed to the lack of thyroid hor- mone. No sign of increased pigmentation, similar to the roentgen - pigmentation of goldfish (Ellinger, 1940), was seen in P32-treated platy- fish, nor were any pigmentary differences from controls noted in thyroid-fed I131-treated fish. A generalized darkening was noted in many of the exclusively I131-treated platyfish, but this was of the normal speckled pattern found in 212 Zoologica: New York Zoological Society [46: 16 all fish of the strain employed and was not es- pecially prominent on the head. This probably was a minor effect of the lack of thyroid hor- mone (see above). The effects of treatment with I131 or P32 on the gonads of young platyfish were dissimilar, although both were adverse. Among solely radio- iodine-treated fish, only one developed a gonad of mature appearance. In this male specimen, however, the gonopodium failed to differentiate. A single similar male occurred among I131- treated, Kl-treated fish. Apart from these ex- ceptions, the failure of germinal and duct ele- ments alike to develop beyond juvenile stages in thyroidectomized fish of both sexes stands in sharp contrast to the normal sexual differen- tiation seen in most thyroid-fed, athyroid fish. From these observations, together with those made on P32-treated platyfish (discussed fully below), it may be concluded that the thyroid plays a powerful role in normal sexual matura- tion and that radiation effects were negligible. The complete destruction of the germinal cells in the testes of P32-treated males, together with normal gonadal duct development, con- firms the observations of Vivien (1953a, b) on P32 treatment of immature and adult male gup- pies (Lebistes reticulatus) and swordtails ( Xiph - ophorus helleri). That normal male secondary sexual differentiation was found also agrees with Vivien’s results, and there seems no question that the sex cells themselves have no endocrine role in the development of these male poeciliids. With respect to females, P32 treatment of young platyfish had no readily discernible effect on ovarian development. Fertility was not tested, since the only males made available to these females were their sterile brothers just described. The ineffectiveness of P32 treatment of female platyfish is in disagreement with Vivien (1953b), who found an equal amount of germ cell destruction in both sexes. This difference is particularly surprising when doses are com- pared. The platyfish were treated with 1 .7 me. P32 for 48 hours, to give a total dose of 720 er. Vivien treated his fish, maximally, with 200 jac/liter of P32, which was left to decay through- out the treatment of 130 days, fish removed to clean water after 8 days being no less affected. Using formulas given by Glasser et al. (1952), we calculate Vivien’s procedure to give a total dose of 175 er. (or less, if all the P32 had not decayed in 130 days). This is much smaller than the dose given to the platyfish, yet it had a much more profound effect in the females. Two possible, alternative explanations might be, the chronicity of Vivien’s treatment, or species differences in the fish. The latter is not likely. since Vivien used fish of two different genera with similar results, and one of his species is congeneric with the platyfish. The first alterna- tive also seems unlikely in view of certain ob- servations made on mammals, which are briefly summarized below. Long-term gamma irradiation of mice at var- ious dose rates led to observations showing cer- tain similarities, but also profound differences from both Vivien’s results and our own. Der- inger et al. (1954) found that male mice given 1,100 r. at rates of 4.4 r. per 8-or-24-hour day bred normally, while those sterilized by higher total doses (1,760 r.) at the rate of 8.8 r. per 24- hour day recovered their fertility. The fertility of female mice given less total irradiation than males (770 or 880 r.) at the equal rates of 4.4 or 8.8 r. per 24-hour day, or 8.8 r. per 8-hour day, was reduced and some became totally sterile. Unlike males, females sterilized by higher total doses did not recover their fertility. Thus, the reproductive capacity in female mice was reduced or destroyed by doses of radiation which either did not affect male mice, or from which males recovered. Eschenbrenner & Miller (1954) published similar findings on mice and empha- sized that inhibition of spermatogenesis was cor- related with dose rate, not total dose. They found that interstitial tissue was unaffected up to 5,000 r. total dose. In guinea pigs, the male germ cells were similar in radiosensitivity to those of mice, but the ovaries were highly radio- resistant. The latter was also true of rabbits. The observations on the radioresistance of mammalian testicular interstitial cells agree with all data on P32-treated fish, and also with Vivien’s (1952) earlier observations on the effects of x-rays on the testes of swordtails. How- ever, the results with mice are dissimilar to those of the author for platyfish. In the mouse, at a given dose rate, the ovaries of the females were more radiosensitive than the testes of males, but in the platyfish the females were more re- sistant to radiation damage. Platyfish thus seem to be more similar to the guinea pig or rabbit. In Vivien’s experiments, the initial dose rate would have been 8.4 r. per 24-hour day, which was a rate that sterilized female mice, but was a very much lower rate than that given the platyfish (360 r. per 24-hour day). Yet the ovaries of the platyfish suffered no visible de- generation. Thus, the longer term of exposure of Vivien’s fish to P32 irradiation, although the most likely factor in causing the discrepancy in results on female fish, might be ruled out on the basis of dose-rate considerations. The observations of Samakhvalova (1935) 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 213 and Solberg (1938) on the guppy and the medaka ( Oryzias latipes), respectively, revealed a high degree of radioresistance in the gonads of both sexes. Doses of 1,500 r. to male guppies and 1,980 r. to rnedakas of both sexes, in the form of x-rays, caused degenerative changes at first, but recovery followed in 1-4 months, coupled with the return of fertility in all. Observations such as those reviewed indicate that the effect of P32 may not be qualitatively the same as that of radiation in other forms, as far as the gonads are concerned. Vivien (1953b, c) advanced the hypothesis that the action of P32 on the gonads commences with the diencephalon, where it interferes with neuro- hypophyseal control, and that it is not rever- sible owing to neural destruction. A possible objection to this theory might be the radio- resistance of nervous tissue (Tuliis, 1959). Summary 1. Total radiothyroidectomy of platyfish, Xiphophorus maculatus, 0.5 to 2.0 months old, by means of immersion in solutions of 5.0 mc./200 ml. of I131 for 48 hours was successful, although the death rate after such treatment was high. 2. Gonadal development of completely radio- thyroidectomized fish was severely retarded and growth in length also was significantly impeded. Vacuolated (=fatty?) liver appeared in both completely and partially thyroidectomized fish and it increased in severity in a progressive man- ner as time after treatment lengthened. The hematopoietic tissue in the kidneys was reduced, and the blood vessels in the thyroid area showed radiation damage. Pituitary exhaustion appeared to occur. 3. Feeding of these athyroid fish with mam- malian thyroid restored them to a normal con- dition with respect to death rate, growth, sexual development, liver histology and, in part, hema- topoietic development. The radiation damage to blood vessels persisted. The time of maturation of males was slightly advanced, and the gono- podia and testes were normal. One female be- came gravid. These fish were even more normal than their controls in one respect, viz., their pituitaries showed none of the hypertrophy seen in controls with incipient goiter. Elongation of the caudal fin was noted. No exophthalmia oc- curred. 4. Treatment of these athyroid fish with po- tassium iodine improved the mortality rate but did not improve growth or reduce the develop- ment of vacuolated liver. Gonadal differentia- tion was slightly less retarded, but no fish ap- proached maturity at the time that the controls reproduced. 5. Fish treated with equivalent, non-thyroid- ectomizing, irradiation by means of P32 were completely normal in growth, secondary sexual development, body shape and visceral histology. Male gonads were completely lacking germ cells, but female gonads were unaffected. The death rate was as high as that shown by En- treated fish, but appeared to be differently dis- tributed in time from the I131 death rate. 6. The injection of adult platyfish with I131 led to severe hypothyroidism, but whether it had any effect on the gonads was doubtful. Vacuolated or ceroid-filled livers and enlarged pituitaries were commonly produced. 7. It is concluded that the thyroid plays a significant role in growth, sexual maturation and liver metabolism of the teleost, Xiphophorus maculatus. The thyroid also promotes the re- covery of hematopoietic tissue after irradiation, and strengthens resistance to shock and disease. Acknowledgements The author is deeply indebted to the late Dr. Myron Gordon, without whose encourage- ment and generous provision of fish and the facilities of the Genetics Laboratory, this, as well as most of the earlier works of the author, would not have been accomplished. This paper is dedicated to him. Much appreciation is expressed to Professors Aubrey Gorbman and L. G. Barth, Department of Zoology, Columbia University, who together provided all the necessary facilities and mate- rials for the radioisotope work, and to Drs. Grace E. Pickford, Bingham Oceanographic Laboratory, Yale University, and James W. Atz, New York Aquarium, New York Zoological Society, for critical reading of the manuscript. Appreciation is also expressed to Dr. Pickford for valuable bibliographic aid. Thanks are due Dr. Olga Berg, University of Rochester, for her help with the early work using radioiodine, to Dr. Donn E. Rosen, Department of Fishes, American Museum of Natural History, for help with the gonopodial data and for providing the drawing of the gonopodium published here, and to Dr. Helen Deane, Department of Anatomy, Albert Einstein College of Medicine, for helpful discussion. Literature Cited Albert, A. A. 1945. The experimental production of exoph- thalmos in Fundulus by means of anterior pituitary extracts. Endocrinol., 37: 389- 406. 214 Zoologica: New York Zoological Society [46: 16 Arvy, L., M. Fontaine & M. Gabe 1956. Fonction thyroi'dienne et complexe hypo- thalamo - hypophysaire chez la Truite. Compt. rend. Soc. Biol., 150: 625-627. Atz, E. H. 1953. Experimental differentiation of basophil cell types in the transitional lobe of the pituitary of a teleost fish, Astyanax mexi- canus. Bull. Bingham Oceanograph. Coll., 14: 94-116. Baker, K. F. 1958a. Heterotopic thyroid tissues in fishes. I. The origin and development of heterotopic thyroid tissue in platyfish. J. Morph., 103: 91-134. 1958b. Heterotopic thyroid tissues in fishes. II. The effect of iodine and thiourea upon the development of heterotopic thyroid tissue in platyfish. J. Exper. Zool., 138: 329-354. Baker, K. F., O. Berg, A. Gorbman & M. Gordon 1955. Observations on radiothyroidectomy of juvenile platyfish. Anat. Rec., 122: 453- 454. Baker, K. F., O. Berg, A. Gorbman, R. F. Nigrelli & M. Gordon 1955. Functional thyroid tumors in the kidneys of platyfish. Cancer Res., 15: 118-123. Barrington, E. J. W. 1954. Hormones and control of zoological func- tion. Nature, 174: 720-722. Barrington, E. J. W. & A. J. Matty 1952. Influence of thiourea on reproduction in the minnow. Nature, 170: 105-106. 1955. Identification of thyrotrophin - secreting cells in the pituitary gland of the minnow ( Phoxinus phoxinus). Quart. J. Micr. Sci., 96: 193-201. Baumann, G. & C. Pfister 1936. La thyroxine provoque chez de jeunes alevins de truite, une atrophie et une dis- parition des arcs aortiques, analogues a celles des embryons d’amniotes. Compt. rend. Soc. Biol., 122: 1156-1158. Bjorkland, R. G. 1958. The biological function of the thyroid and the effect of length of day on the growth and maturation of the goldfish Carassius auratus (Linnaeus). Dissert. Abstracts, 19: 1423. Blacher, L. J. 1927. The role of the hypophysis and of the thyroid gland in the cutaneous pigmentary function of amphibians and fishes. Trans. Lab. exp. Biol. Zoopark, Moscow, 3: 37- 81. (Russian, with English summary) Bonham, K., L. R. Donaldson, R. F. Foster, A. D. Welander & A. H. Seymour 1948. The effect of x-ray on mortality, weight, length, and counts of erythrocytes and hematopoietic cells in fingerling chinook salmon, Onchorhynchus tschawytscha Walbaum. Growth, 12: 107-121. Buser, J. & P. Bourns 1951. Allongement des nageoires pectorals pro- voque par la thyroxine chez Gambusia affinis Grd. Arch. zool. expt’l. et gen., 88 : 116-122. Chakoff, I. L., T. Gillman, C. Entenman, J. F. Rinehart & F. L. Reichert 1948. Cirrhosis and other hepatic lesions pro- duced in dogs by thyroidectomy and by combined hypophysectomy and thyroid- ectomy. J. Exper. Med., 88: 1-14. Chambers, H. A. 1953. Toxic effects of thiourea on the liver of the adult male killifish, Fundulus heter- oclitus (Linn.). Bull. Bingham Oceano- graph. Coll., 14: 69-93. Clavert, J. & J. P. Zahnd 1956. Modifications hepatiques survenant pen- dant la vitellogenese chez deux especes de poissons ovovivipares ( Xiphophorus hel- leri et Lebistes reticulatus) . Compt. rend. Soc. Biol., 150: 1261-1262. 1956. Modifications hepatiques durant la gesta- tion chez deux poissons ovovivipares {Xiphophorus helleri et Lebistes reticula- tus). Compt. rend. Soc. Biol., 150: 2257- 2259. 1957. Realisation experimentale, chez Xipho- phorus helleri, de certaines modifications hepatiques qui s’observent au cours de la vitellogenese de la phase de grand accro- isement des follicules. Compt. rend. Soc. Biol., 151: 1234-1236. Dales, S. & W. S. Hoar 1954. Effects of thyroxine and thiourea on the early development of chum salmon ( On- chorhynchus keta). Canad. J. Zool., 32: 244-251. Dalton, A. J., H. P. Morris & C. S. Dubnik 1946. Changes in the thyroid and other organs in mice receiving thiouracil. Fed. Proc., 5: 219. Deringer, M. K., W. E. Heston & E. Lorenz 1954. Effects of long-continued total-body gam- ma irradiation on mice, guinea pigs, and rabbits. IV. Actions on the breeding be- havior of mice. In: R. E. Zirkle, ed., “Bi- ological Effects of External X and Gam- ma Radiations,” McGraw-Hill Book Co., New York. pp. 149-168. 1961] Baker-Cohen: Role of Thyroid in Development of Platyftsh 215 Edwards, J. E. & J. White 1941. Pathological changes, with special refer- ence to pigmentation and classification of hepatic tumors, in rats fed p-dimethyla- minoazobenzene (butter yellow). J. Nat. Cancer Inst., 2: 157-183. Egami, N. 1955. Effect of estrogen and androgen on the weight and structure of the liver of the fish, Oryzias latipes. Annot. Zool. Japon., 28: 79-85. 1957. Inhibitory effect of thiourea on the devel- opment of male characteristics in females of the fish, Oryzias latipes, kept in water containing testosterone proprionate. An- not. Zool. Japon., 30: 26-30. Ellinger, F. 1940. Roentgen-pigmentation in the goldfish. Proc. Soc. Exp. Biol. & Med., 45: 148-150. Endicott, K. M. & R. D. Lillie 1944. Ceroid, the pigment of dietary cirrhosis of rats. Its characteristics and its differ- entiation from hemofuscin. Amer. J. Path., 20: 149-153. Eschenbrenner, A. B. & E. Miller 1954. Effects of long-continued total-body gam- ma irradiation on mice, guinea pigs and rabbits. V. Pathological observations. In: R. E. Zirkle, ed., “Biological Effects of External X and Gamma Radiations,” Mc- Graw-Hill Book Co., New York. pp. 169- 225. Essenberg, J. M. 1926. Complete sex-reversal in the viviparous teleost Xiphophorus helleri. Biol. Bull., 51: 98-111. Fontaine, M., S. Baraduc & J. Hatey 1953. Influence de la thyroxination sur la teneur en glycogene du foie des poissons teleos- teens. Compt. rend. Soc. Biol., 147: 214- 216. Fontaine, M., F. Querriere & A. Raffy 1957. Action de l’hypophysectomie sur le me- tabolisme respiratoire de l’anguille (An- guilla anguilla L.). Compt. rend. Soc. Biol., 151: 232-233. Fontaine, Y.-A. 1957. Diminution du pouvoir thyreotrope de l’hypophyse apres thyroi'dectomie chez un Mammifere (le rat) et un Teleosteen (L’anguille). Compt. rend. Acad. Sci., 245: 2538-2541. Fortune, P. Y. 1955. Comparative studies of the thyroid func- tion in teleosts of tropical and temperate habits. J. Exper. Biol., 32: 504-513. Frieders, F. 1954. The effects of thyroid-inhibiting drugs on some tropical fish. Catholic Univ. Amer., Biol. Stud., 31: 1-37. Fromm, P. O. & E. P. Reineke 1956. Some aspects of thyroid physiology in rainbow trout. J. Cell. & Comp. Physiol., 48: 393-404. Gaiser, M. L. 1952. Effets produite par l’administration pro- longee de thiouree et de thyroxine chez Lebistes reticulatus. Compt. rend. Soc. Biol., 146: 496-498. Gerbilsky, N. L. & M. G. Saks 1947. Postembryonic development of sturgeon (Acipenser stellatus) as affected by thy- roxine. Doklady Akad. Nauk. U.R.S.S., N.S., 55: 663-666. Gibson, W. C. 1954. A study of the effects of thyroxine on the early embryonic development of Brachy- danio rerio (Hamilton and Buchanan). Thesis, Oklahoma Agricultural & Mechan- ical College (State Univ. of Oklahoma), Stillwater, Oklahoma, pp. 1-60. Glasser, O., E. H. Quimby, L. S. Taylor & J. L. Weatherwax 1952. “Physical Foundations of Radiology,” Paul B. Hoeber, Inc., New York. Goldberg, R. C, I. L. Chaikoff, S. Lindsay & D. D. Feller 1950. Histopathological changes induced in the normal thyroid and other tissues of the rat by internal radiation with various doses of radioactive iodine. Endocrinol., 46: 72-90. Goldsmith, E. D . 1949. Phylogeny of the thyroid: descriptive and experimental. Ann. N. Y. Acad. Sci., 50: 283-313. Goldsmith, E. B., R. F. Nigrelli, A. S. Gordon, H. A. Charipper & M. Gordon 1944. Effect of thiourea upon fish development. Endocrinol., 35: 132-134. Gorbman, A. & O. Berg 1955. Thyroidal function in the fishes Fundulus heteroclitus, F. majalis, and F. diaphanus. Endocrinol., 56: 86-92. Gordon, M. 1950. Fishes as laboratory animals. In: E. J. Farris, ed., “Care and Breeding of Labo- ratory Animals,” John Wiley & Sons, New York. 216 Zoologica: New York Zoological Society [46: 16 Grobstein, C. 1953. Endocrine and developmental studies of gonopod differentiation in certain Poeciliid fishes. I. The structure and development of the gonopod in Platypoecilus macula- tus. Univ. Calif. Publ. Zool., 47: 1-21. Grobstein, C. & A. W. Bellamy 1939. Some effects of feeding thyroid to imma- ture fishes (Platypoecilus). Proc. Soc. Exp. Biol. & Med., 41: 363-365. Grosso, L. L. & H. A. Charipper 1958. The effect of thiourea on embryos of Lebistes reticulatus by immersion of the maternal organism. Anat. Rec., 132: 449. Ham, A. W. 1953. “Histology,” 2nd Ed., J. B. Lippincott Co., Philadelphia. Harris, P. J. 1959. A study of thyroid function in Fundulus heteroclitus. Biol. Bull., 117: 89-99. Hasler, A. D. & R. K. Meyer 1942. Respiratory response of normal and cas- trated goldfish to teleost and mammalian hormones. J. Exper. Zool., 91: 391-403. Hatey, J. 1950. Action de la thiouree sur la metabolisme glucidique de la carpe ( Cyprinus carpio L.). Compt. rend. Soc. Biol., 144: 955- 957. Herzfeld, E., P. Mayer-Umhofer & F. Scholz 1931. Fische als Test-objekte fur pharmakolo- gische Versuche. II. Wirkung von Schild- driisen derivaten und Blutschutz. Klin. Wschr., 10: 1908-1910. Hoar, W. S. 1951. Hormones in fish. Univ. Toronto Biol. Ser. 59, Publ. Ontario Fisheries Research Publ., 71: 1-51. 1957. Endocrine organs. In: M. E. Brown, ed., “The Physiology of Fishes,” Vol. I., Aca- demic Press, New York. pp. 245-285. Hoar, W. S. & G. M. Bell 1950. The thyroid gland in relation to the sea- ward migration of Pacific salmon. Canad. J. Res., D„ 28: 126-136. Hoar, W. S., M. H. A. Keenleyside & R. G. Goodall 1955. The effects of thyroxine and gonadal ster- oids on the activity of salmon and gold- fish. Canad. J. Zool., 33: 428-439. Hoar, W. S., D. Mackinnon & A. Redlich 1952. Effects of some hormones on the behavior of salmon fry. Canad. I. Zool., 30: 273- 286. Honma, Y. 1960. Studies on the morphology and the role of the important endocrine glands in some Japanese cyclostomes and fishes. (In Jap- anese, with English summary.) Publ. Dept. Biol., Univ. Niigata [Niigata, Ja- pan]: 1-139. Honma, Y. & S. Murakawa 1955. Effects of thyroxine and thiourea on the development of chum salmon larvae. Jap. J. Ichthyol., 4: 83-93. 1957. Effect of thiourea on the development of goldfish larvae ( Carassius auratus Linne juv.). Jap. J. Ichthyol., 6: 121-127. Hopper, A. F. 1950. The effect of mammalian thyroid powder and thiouracil on growth rates and on the differentiation of the gonopod in Lebistes reticulatus. Anat. Rec., 108: 554. 1952. Growth and maturation response of Le- bistes reticulatus to treatment with mam- malian thyroid powder and thiouracil. J. Exper. Zool., 119: 205-217. 1961. The effect of feeding mammalian thyroid powder on growth rates of immature guppies. Growth, 25: 1-5. Iakovlova, I. V. 1949. “The independence of the activity of the thyroid gland from the thyrotropic func- tion of the hypophysis in post-embryonic development of acipenserines.” Doklady Akad. Nauk, U.R.S.S., 60: 281-284. (In Russian) Immers, J. 1953. The influence of the sexual cycle on the metabolism of glycogen in the liver, gon- ads and skin of the stickleback ( Gaster - osteus aculeatus L.) and the minnow (Phoxinus laevis Ag. ) . Arkiv Zool, Stock- holm, Ser. 2, 4: 327-339. Ito, M. 1952. Studies on melanin. Tohoku J. Exp. Med., Suppl. 1, 55: 1-104. James, E. S. 1939. The morphology of the thymus and its changes with age in the neotenous am- phibian ( Necturus maculosus) . J. Morph., 64: 455-481. Jones, R. W., W. C. Gibson & C. Nickolls 1951. Factors influencing mitotic activity and morphogenesis in embryonic development. I. The effects of thyroxine and thiouracil on the development of Brachydanio rerio (zebra fish). Anat. Rec., Ill: 509. 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 217 Jones, R. W. & M. N. Huffman 1957. Fish embryos as bio-assay material in testing chemicals for effects on cell divi- sion and differentiation. Trans. Amer. Micr. Soc., 76: 177-183. Kajishima, T. 1960. Analysis of gene action in the transparent scaled goldfish, Carassius auratus. II. The effects of pituitary and thyroid on gene action. Embryologia, 5: 127-138. Krockert, G. 1936. Die Wirkung der Verfiitterung von Schild- driissen und Zirbeldrusensubstanz an Le- bistes reticulatus (Zahnkarpfen). Zeitschr. f. gesamte exper. Med., 98: 214-221. Landgrebe, F. W. 1941. The role of the pituitary and the thyroid in the development of teleosts. J. Exper. Biol., 18: 162-169. Langford, H. G. 1957. Production of exophthalmos in Fundulus heteroclitus (var. bermudae ) by triiodo- thyronine and desiccated thyroid. Endo- crinol., 60: 390-392. La Roche, G. 1949. Influence of the thyroid gland in the de- velopment of the salmon. Anat. Rec., 103: 480. 1950. Analysis of the effect of thyroid prepara- tions on the young river salmon (parr). Anat. Rec., 106: 213. 1951. Effect of thyroid treatment on the connec- tive tissue of Salmonidae. Anat. Rec., 109: 316. La Roche, G. & C. P. Leblond 1952. Effect of thyroid preparations and iodide on Salmonidae. Endocrinol., 51: 524-545. 1954. Destruction of thyroid gland of Atlantic salmon ( Salmo salar L.) by means of radio-iodine. Proc. Soc. Exp. Biol. & Med., 87: 273-276. La Roche, G., C. P. Leblond & G. Prefontaine 1950. Effets de l’hormone thyroidienne sur le saumon de 1’Atlantique: stade parr. Rev. Canad. de Biol., 9: 101-103. Leathem, J. H. & C. H. Howell 1950. Thyroid activity and liver histology. Anat. Rec., 106: 216. Leblond, C. P. & H. E. Hoff 1944. Effect of sulfonamides and thiourea de- rivatives on heart rate and organ mor- phology. Endocrinol., 35: 229-233. Leloup, J. & M. Fontaine 1960. Iodine metabolism in lower vertebrates. Ann. N. Y. Acad. Sci., 86: 313-353. Leloup, J. & M. Olivereau 1950. Production d’exophthalmie par la thiouree chez un teleosteen marin : Dentex vulgaris Cuv. Compt. rend. Soc. Biol., 144: 772- 774. Lillie, R. D„ F. S. Daft & W. H. Sebrell 1941. Cirrhosis of the liver in rats on a deficient diet and the effect of alcohol. Public Health Repts., 56: 1255-1258. Lushbaugh, C. C. 1957. Vertebrate radiobiology (The pathology of radiation exposure). Ann. Rev. Nuclear Sci., 7: 163-184. Lynn, W. G. & H. E. Wachowski 1951. The thyroid gland and its functions in cold-blooded vertebrates. Quart. Rev. Biol., 26: 123-168. Maqsood, M. 1952. Thyroid functions in relation to repro- duction of mammals and birds. Biol. Revs,. 27: 281-319. Marine, D. & C. H. Lenhart 1910. Observations and experiments on the so- called thyroid carcinoma of brook trout ( Salvelinus fontinalis ) and its relation to ordinary goitre. J. Exper. Med., 12: Sll- SS?. Matty, A. J., D. Menzel & J. E. Bardach 1958. The production of exophthalmos by an- drogens in two species of teleost fish. J. Endocrinol., 17: 314-318. May, L. G., R. W. Moseley & J. C. Forbes 1946. Effect of thiourea on body fat and liver glycogen of rats. Endocrinol., 38: 147- 151. McClosky, W. T., R. D. Lillie & M. I. Smith 1947. The chronic toxicity and pathology of thiouracil in cats. J. Pharmacol., 89: 125- 130. Metcalf, R. G., R. J. Blandau & T. B. Barnett 1954. Pathological changes exhibited by animals exposed to single doses of x radiation. In: H. A. Blair, ed., “Biological Effects of External Radiation,” McGraw-Hill Book Co., New York. pp. 11-57. Muller, J. 1953. Uber die Wirkung von Thyroxin und Thyreotropem Hormon auf den Stoff- wechsel und die Farbung des Goldfisches. Zeitschr. f. vergleich. Physiol., 35: 1-12. Nigrelli, R. F., E. D. Goldsmith & H. A. Charipper 1946. Effects of mammalian thyroid powder on growth and maturation of thiourea-treated fish. Anat. Rec., 94: 523. 218 Zoologica: New York Zoological Society [46: 16 Oguro, C. 1956. Some observations on the effect of estro- gen upon the liver of the three-spined stick- leback, Gasterosteus aculeatus L. Annot. Zool. Japon., 29: 19-22. Olivereau, M. 1957. Radiothyroidectomie chez l’Anguille (An- guilla anguilla L.). Arch. Anat. Micr. Morph. Exper., 46: 39-59. Olivereau, M. & J. Leloup 1950. Variations du rapport hepatosomiques chez la roussette (Scy Ilium canicula L.) au cours du development et de la repro- duction. Vie et Milieu, 1: 377-420. Pflugfelder, O. 1959. Beeinflussung der Thyreoidea und anderer Organe des Haushuhnes durch Kalium- perchlorat, mit vergleichenden Unter- suchungen an niederen Wirbeltieren. Roux’ Archiv f. Entwickl. d. org., 151: 78-112. PlCKFORD, G. 1952. Hormonal regulation of liver size in fish. Anat. Rec., 112: 429-430. 1957. The thyroid and thyrotropin. In: G. E. Pickford and J. W. Atz, “The Physiology of the Pituitary Gland of Fishes,” New York Zoological Society, New York. Rasquin, P. & L. Rosenbloom 1954. Endocrine imbalance and tissue hyper- plasia in teleosts maintained in darkness. Bull. Amer. Mus. Nat. Hist., 104: 359- 426. Rizzo, L. 1950a. Modificaziene sperimentali dello sviluppo in Gambusia holbrooki Grd. Boll. Mus. 1st. Biol. (Geneva), 23: 1-12. 1950b. Dissimmetrica sensibilita ai fattori or- monali dello sviluppo. Richerche in Gam- busia holbrooki Grd. Boll. Mus. 1st. Biol. (Geneva), 23: 23-29. Robertson, O. H. 1949. Production of the silvery smolt stage in rainbow trout by intramuscular injection of mammalian thyroid extract and thyro- trophic hormone. J. Exper. Zool., 110: 337-355. Rosen, D. E. 1960. Middle- American poeciliid fishes of the genus Xiphophorus. Bull. Florida State Mus., 5: 1-242. Samokhvalova, G. V. 1935. The influence of the x-rays on the sex gland and the secondary sexual characters in Lebistes reticulatus. Trans. Dynamics of Development (U.S.S.R.), 10: 213-229. (In Russian, with English summary.) Scott, J. L. 1953. The effects of thiourea treatment upon the thyroid, pituitary and gonads of the zebra fish, Brachydanio rerio. Zoologica, 38: 53-62. Sellers, E. A. & R. Wen You 1951. Propylthiouracil, thyroid and dietary liver injury. J. Nutrition, 44: 513-535. Sembrat, K. 1954. Effect of the thyroid gland on the skin of Teleostei. Acta physiol. Polonica, 5 : 647- 648. 1956. Influence of the thyroid gland on the skin of teleosts. Zool. Polonica, 7: 3-24. Sklower, A. 1927. Ueber den Einfluss von Schilddriisen und thymusfiitterung auf die Korperlange und das Gewicht von Forellenbrut. Zeitschr. f. Fischerei, 25: 549-552. Smith, D. C. & G. M. Everett 1943. The effect of thyroid hormone on growth rate, time of sexual differentiation and oxygen consumption in the fish, Lebistes reticulatus. J. Exper. Zool., 94: 229-240. Smith, D. C., S. A. Sladek & A. W. Kellner 1953. The effect of mammalian thyroid extract on the growth rate and sexual differentia- tion in the fish, Lebistes reticulatus, treat- ed with thiourea. Physiol. Zool., 26: 117- 124. Smith, S. B. 1949. The effects of thyroxine and related com- pounds on young salmon and trout. The- sis, University of British Columbia, Van- couver, B. C. pp. 1-29. Sokol, H. W. 1957. Cytophysiological studies on the teleost hypophysis ( Fundulus heteroclitus and Lebistes reticulatus). Thesis, Radcliffe College, Boston. Solberg, A. 1938. The susceptibility of the germ cells of Oryzias latipes to x-irradiation and recov- ery after treatment. J. Exper. Zool., 78: 417-434. Spellberg, M. A. 1954. “Diseases of the Liver,” Grune & Strat- ton, New York. Stolk, A. 1955. Changes in the pituitary gland of the vivi- parous cyprinodonts Xiphophorus (Xiph- ophorus) helleri Heckel and Lebistes reticulatus (Peters) after thiouracil treat- ment. Koninkl. Nederl. Akad. Weten- schap., Ser. C, 58: 179-189. 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 219 1956a. Changes in the pituitary gland of the cy- prinid Tanichthys albonuhes Lin. with thy- roidal tumor. Proc. Kon. Ned. Akad. We- tenschap., 59: 38-49. 1956b. Changes in the pituitary gland of the cich- lid Cichlasoma biocellatum (Regan) with thyroidal tumor. Proc. Kon. Ned. Akad. Wetenschap., 59: 494-505. 1959. Effect of thiouracil and thyroxine on de- velopment and growth of cutaneous mel- anoma in killifish hybrids. Nature, 184: 562-563. SVARDSON, G. 1943. Studien ii’oer den Zusammenhang zwischen Gesclilechtsreife und Wachstum bei Le- bistes. Medd. Undersokaanst. Sotvattens- fisk., Stockholm (21). pp. 1-48. Tavolga, M. C . 1949. Differential effects of estradiol, estradiole benzoate and pregneninolone on Platy- poecilus maculatus. Zoologica, 34: 215- 237. Tavolga, W. N. 1949. Embryonic development of the platyfish {Platypoecilus) , the swordtail ( Xipho - phorus), and their hybrids. Bull. Amer. Mus. Hist., 94: 161-230. Terao, A. 1922. “Effect of feeding thyroid on goldfish.” .1. Fish., Tokyo, 17: 257-259. (In Japa- nese). Tullis, J. L. 1959. The pathological anatomy of total body irradiation. In: C. F. Behrens, ed., “Atomic Medicine,” 3rd Ed., Williams and Wilkins Co., Baltimore, Md. pp. 131-159. Vilter, V. 1946. Action de la thyroxine sur la metamor- phose larvaire de l’anguille. Compt. rend. Soc. Biol., 140: 783-785. Vivien, J . 1952. Influence de la castration precose et de 1’irradiation par les rayons x sur la differ- entiation des caracteres sexuels second- aires chez le Xiphophore. Compt. rend. Acad. Sci., 234: 2394-2395. 1953a. Effets du radio-phosphore, 32 P, sur le gonades des Cyprinodontes vivipares. Compt. rend. Acad. Sci., 236: 535-538. 1953b. Sterilisation totale des gonades apres traite- ment par le phosphore radio-actif, chez les Cyprinodontes: Lebistes et Xiphophores. Compt. rend. Acad. Sci., 236: 2172-2174. 1953c. Essai d’interpretation du mode d’action du phosphore radioactif sur le complexe hypophyso - genital des Cyprinodontes. Compt. rend. Soc. Biol., 147: 1462-1464. Vivien, J. & M. L. Gaiser 1952. Action d’un traitement longue duree par la thiouree sur la structure et le fonction- ment de la thyroi'de et de l’hypophyse chez le Lebistes reticulatus, modification du rythme de la croissance staturale et de processus de la differenciation genitale puberale. Compt. rend. Acad. Sci., 234: 1643-1645. Warner, E. D. 1952. Some effects of thiouracil in the German brown trout. Trans. Wisconsin Acad. Sci., 41: 169-175. Welander, A. D., L. R. Donaldson, R. F. Foster, K. Bonham & A. H. Seymour 1948. The effects of roentgen rays on the em- bryos and larvae of the Chinook salmon. Growth, 12: 203-242. Wolf, L. E. 1931. The history of the germ cells in the vivi- parous teleost Platypoecilus maculatus. J. Morph. & Physiol., 52: 115-153. WOODHEAD, A. D. 1959. Variations in the activity of the thyroid gland of the cod, Gadus callarias L. in re- lation to its migrations in the Barents Sea. Parts I. and II. J. Marine Biol. Assoc., U. K., 38: 407-422. Zahnd, J. P. 1959. Modifications hepatiques liees au cycle ovarien chez deux poissons ovovivipares : Xiphophorus helleri et Lebistes reticula- tus. Arch. Anat. Micr. Morph. Exper., 48: 231-259. Zaks, M. G. & M. A. Zamkova 1947. “On the role of the thyroid gland in the embryogenesis of vertebrata.” Fiziol. Zh., 33: 449-462. (in Russian). Zarski, E. 1927. La glande thyroi'de des mammiferes pro- duit-elle des changements dans la struc- ture histologique de la peau des teleosteens ( Misgurnus fossilis et Tinea vulgaris )? Compt. rend. Soc. Biol., 97: 1683-1684. 220 Zoologica: New York Zoological Society [46: 16 EXPLANATION OF THE PLATES Plate I Fig. 1. Radiothyroidectomized platyfish, 6 months old, at bottom left, together with normal broodmates— male above and fe- male below, right. Note the hunched shape and drooping fins of the thyroidectomized fish. X %. Photo by Sam Dunton, New York Zoological Society. Fig. 2. Fish harvested from an experiment in which radioiodine-treated fish were treated subsequently with potassium iodide, or were fed mammalian thyroid. All fish were broodmates and 8 were started in each group. A, Radioiodine-treated; B, Untreat- ed controls; C, Radioiodine-treated, fed thyroid; D, Radioiodine-treated, Kl-treated. Notice the normal size, longer tails and paler coloring of the thyroid-fed fish, in comparison with the untreated controls. Also note the greater survivorship among the Kl-treated in comparison with the fish treated solely with radioiodine. X % Fig. 3. Fish harvested from an experiment in which radioiodine-treated fish subsequent- ly were treated with KI. All fish were broodmates. The initial groups were: A, 5 radioiodine-treated; B, 4 radioiodine- treated; Kl-treated; C, 5 untreated con- trols. The upper fish in group A is shown alive in Fig. 1. Notice the dumpy body shape in all of the radioiodine-treated fish. X I Fig. 4. Fish from three experiments in which young platyfish were treated with P32. Un- treated controls appear on the left and the corresponding broodmates treated with P32 on the right. The P32-treated fish are nor- mal in appearance, and on the average are as large as the controls. X % Plate II Fig. 1. Stripping film I131 radioautograph, show- ing exposure of emulsion grains over food particles in the pharynx of a platyfish. X 100 Fig. 2. Thyroid follicle in the kidney of a radio- thyroidectomized platyfish. This follicle is the only example of renal thyroid tissue found in a radioiodine-treated specimen. X 400 Fig. 3. Regenerated thyroid tissue above the bul- bus arteriosus in a radioiodine-treated fish. This female specimen was 7 months old and had been treated with I131 when 37 days old. X 100 Fig. 4. Regenerated thyroid tissue above and around the ventral aorta in a radioiodine- treated fish 20 months old. This female had been exposed to I131 when 50 days old. In this fish, the regenerated thyroid tissue was almost goitrous in proportion, al- though the follicles retained their distinct identity from one another. X 100 Fig. 5. Fibrous adhesions developed between the aorta and pericardium in an I131-treated fish. The thyroid area above the pericar- dium and between the gill chambers is filled with loose stromal tissue. X 100 Fig. 6. Aorta surrounded by thick gelatinous fibrous tissue in a young platyfish treated with radioiodine 24 days prior to fixation. X 200 Plate III Fig. 1. Kidney of a normal platyfish 10 months old. The dark masses of cells between the tubules are lymphoid tissue; lighter masses consist of nucleated erythrocytes. X 100 Fig. 2. Kidney of a radioiodine-treated brood- mate of the normal fish in Fig. 1, at the same age. The kidney tubules are smaller primarily because the fish was much smaller. Chiefly to be noted is the ex- treme lack of lymphoid cells. X 100 Figs. 3, 4. Examples of “concretions” (arrows) formed within kidney tubules of radioio- dine-treated platyfish. Some concretion- like masses not within tubules may repre- sent degenerate tubules. X 100 Fig. 5. Spleen of a normal 6-month-old platyfish. Pigment-laden macrophages are the source of the black spots scattered within the organ. X 100 Fig. 6. Small, deeply basophilic regenerating kid- ney tubules in an I131-treated platyfish. X 400 Fig. 7. Extremely shrunken spleen (arrow), which lacks all lymphoid elements, of a 6.5- month-old radioiodine-treated fish. The spleen seems more excessively shrunken than it is because the whole animal was subnormal in size. X 100 1961] Baker-Cohen: Role of Thyroid in Development of Platyfish 221 Plate IV Fig. 1. Liver of a normal platyfish. Capillaries are well delineated by the nucleated ery- throcytes within them. X 100 Fig. 2. Extremely vacuolated liver and abdominal tissue in a radiothyroidectomized platyfish. The liver is at the right, pancreatic and ab- dominal connective tissue to the left, and part of the spleen shows at the upper left. X 100 Fig. 3. Ceroid-filled liver in a radiothyroidectom- ized platyfish. A rounded mass of ceroid- filled macrophages appears next to the blood vessel at the right of center, and the liver cells themselves are packed with ceroid globules. X 100 Fig. 4. Vacuolated liver in a radiothyroidectom- ized fish, showing part of the gall bladder at the upper left, with gall ducts entering the liver. X 100. Fig. 5. Pancreas of a normal platyfish. Two loops of intestine are visible at the upper left and lower right. Vacuoles (fat cells?) and large blood vessels appear among the pan- creatic masses. X 100 Fig. 6. Pancreas of a radiothyroidectomized platy- fish. The pancreatic tissue is compressed by the great amount of abdominal vacuo- lation. X 100 Plate V Fig. 1 . Ovarian development in young radioiodine- treated fish. Left: normal ovary of a 4- month-old fish. Right: ovary of an En- treated broodmate. These sections through the largest diameter of the ovary show the many fewer ova present in the En- treated fish. X 67 Fig. 2. Testicular development in young radioio- dine-treated fish. Left: normal testes of a 4-month-old fish. Many germinal cysts are present. Right: testes of a radioiodine- treated broodmate. No germinal cysts yet have developed. X 286 Fig. 3. Testes of a normal adult male platyfish, 5.5 months old. Many spermatophores are fully formed and grouped in the ducts (bottom right). X 100. Fig. 4. Testes of a radiothyroidectomized fish, 6 months old. No gonial cysts have devel- oped in this specimen. X 286 Fig. 5. Ovary of a normal female platyfish, 8 months old. Ova at all stages in develop- ment may be seen. X 67 Fig. 6. Single ovum found in a radiothyroidectom- ized fish, 6.5 months old. No yolk deposi- tion has occurred. A portion of intestine lies to the left of this poorly developed ovary. X 100 Plate VI Fig. 1. Ovary of a radiothyroidectomized fish, 8 months old. It contained only the three oocytes shown. These ova are in the reticu- lated stage that normally precedes the ap- pearance of yolk granules. The oviduct appears at the left. X 67 Fig. 2. Testes of a radiothyroidectomized fish, 6.3 months old. In the left lobe of this testis a few germinal cysts had developed and reached early meiotic stages, but the right lobe showed none. X 100 Fig. 3. Nearly mature testis in a radioiodine-treat- ed male fish, 6 months old. This exception- al specimen had no positively identifiable thyroid tissue, and the anal fin was elon- gated, but totally undifferentiated (see PI. VII, Fig. 4). X 100 Fig. 4. Mature testis and heavily vacuolated liver in a second radioiodine-treated male fish, 7 months old. This fish had been addition- ally treated with KI. Its anal fin was totally undifferentiated (See PI. VII, Fig. 4), and thyroid tissue could not be positively iden- tified. X 100 Fig. 5. Cross-section through the anterior pituitary of a normal platyfish. Compare with Figs. 6 and 7, below, and with PI. IX, Fig. 7. X 100 Fig. 6. Cross-section through the anterior pitui- tary of a radioiodine-injected adult platy- fish, 15 months old and 6.5 months post- injection. This fish was almost completely thyroidectomized. The large growth of paler basophiles create a striking difference from the normal pituitary, as seen in Fig. 5. X 100 Fig. 7. Cross-section through the anterior pitui- tary of an untreated adult platyfish, 24-29 months old. The prominent overgrowth of paler basophiles in this case was accom- panied by a quiescent and possibly regres- sive thyroid. The latter may have been the result of the relatively old age of the fish. X 100 Plate VII Fig. 1. Gonopodium of a normal adult male pla- tyfish. Compare with Text-fig. 4, and with Figs. 2-4, below. X 25 Fig. 2. Gonopodium of a radiothyroidectomized fish which also was thyroid-fed. Even though the preservation is poorer than that of the fin in Fig. 1, due to fixation of the fish in Bouin’s fluid, it may be seen that all of the terminal hooks and spines are pres- ent and normally formed. X 25 Fig. 3. Gonopodium of a P32-treated fish. Normal differentiation of all elements is present. X 25 222 Zoologica: New York Zoological Society [46: 16: 1961] Fig. 4. Elongated but undifferentiated anal fin of a radioiodine-treated male fish, 10 months old. This fish had no demonstrable thyroid and had infantile testes with no germinal cysts (see Fig. 7, below). X 25 Fig. 5. Testicular ducts of a P32-treated male pla- tyfish, 5.5 months old. No germ cells are present, but the duct system is normally complex. The gonopodium of this fish is shown in Fig. 3, above. X 100 Fig. 6. Testes of a 22-day-old platyfish, bom to a female injected with radioiodine 19 days before the birth (arrows). This young fish appeared to have a normal thyroid and no other abnormalities. The testes of this fish are almost as large as those of some radio- thyroidectomized fish 6 or more months old. Pancreatic tissue appears to the right of the testes, which are centrally located and suspended from the peritoneum. At this age, the two testes are entirely sepa- rated; later, in the course of normal de- velopment, they become fused. X 100 Fig. 7. Testes of a radiothyroidectomized fish, 10 months old. No gonial cysts were present. These testes are but little larger than those of the 22-day-old male in Fig. 6. The elon- gated, but undifferentiated, anal fin of this fish appears in Fig. 4, above. X 100 Plate VIII Fig. 1. Ovary of a normal pregnant female pla- tyfish. Young oocytes appear at the right, the oviduct at the top, and the neural tube of an early embryo at the arrow. X 100 Fig. 2. Ovary of a radiothyroidectomized, thyroid- fed female. All stages in egg development may be seen in this normally matured ovary. X 100 Fig. 3. Embryo contained in the ovary of a radio- thyroidectomized, thyroid-fed platyfish. In this sagittal section, organs such as the eye, brain, kidney, gills and liver may be recognized. X 100 Fig. 4. Section through the pharyngeal area of an- other embryo contained in the same fish. The arrow points to thyroid follicles. The heart appears at the lower right, and part of the brain at the extreme upper left. X 400 Fig. 5. Ovary of a platyfish treated when young with P32. Oocytes at various stages in de- velopment are visible — from the small, darkly basophilic one at the upper left, to the fully yolked egg at the top center. X 100 Fig. 6. Ovary, containing only atretic follicles, of a 17-20 month old platyfish, injected with I131 8.5 months earlier. The oviduct is the darkly stained structure in the cen- tral area. X100. Fig. 7. Ovary of another 17-20 month old platy- fish, injected with I131 8.5 months earlier. In this fish, primary gonial cells appear to be proliferating from the lining of the ov- arian cavity, in a manner suggestive of early spermotogenesis (see text) . The fem- inine germinal elements of this ovary ap- peared to be represented only by atretic follicles. X100 Fig. 8. Higher power view of the cells prolifer- ating from the ovarian lining in the above specimen. The cells are grouped in balls, as found in the primary gonial cysts of the young testis. X400 Plate IX Fig. 1. Pathological nodule found in the abdom- inal cavity of a male platyfish, 13 months old, which had been injected with I131 3.5 months earlier. This mass appeared to be attached to the testis, of which a portion may be seen at the upper right. X 100 Fig. 2. Tremendous proliferation of epithelial- like cells in the dorsal pericardium and among the transverse pharyngeal muscula- ture of a platyfish injected with I131 41 days earlier. This female was 10 months old when killed. X100 Figs. 3-7. Pathological effects seen in a female platy- fish which appeared to represent the most extreme case of hypothyroidism encoun- tered in this study. The fish was 28-29 months old and had been injected with I131 19.5 months before the time of sacri- fice. Fig. 3. Kidney, showing shrunken condition, edematous subcapsular spaces and lack of lymphoid tissue. X25 Fig. 4. Kidney, under higher power, showing al- most complete absence of lymphoid tissue, shrunken tubules and cysticity. Pigment deposits and whorled “nests” are also in evidence. X 100. Fig. 5. Spleen, very shrunken, owing to reduction of lymphoid elements, and with large pig- ment deposit in the center. X 100 Fig. 6. Liver, showing vacuolation, fibrous cysts and “nests.” The large, rounded mass at the upper left is a large fibrous cyst within the liver. These changes are similar to those of cirrhosis. X100 Fig. 7. Pituitary, with an overgrowth of paler basophilic elements of such magnitude that it might well be classed as a tumor. Com- pare with the normal pituitary shown in PI. VI, Fig. 5, which is a section taken at the same level. X100 BAKER-COHEN PLATE I THE ROLE OF THE THYROID IN THE DEVELOPMENT OF PLATYFISH BAKER-COHEN PLATE II THE ROLE OF THE THYROID IN THE DEVELOPMENT OF PLATYFISH BAKER-COHEN PLATE III wMmmm THE ROLE OF THE THYROID IN THE DEVELOPMENT OF PLATYFISH V BAKER-COHEN PLATE IV THE ROLE OF THE THYROID IN THE DEVELOPMENT OF PLATYFISH BAKER-COHEN PLATE V THE ROLE OF THE THYROID IN THE DEVELOPMENT OF PLATYFISH BAKER-COHEN PLATE VI THE ROLE OF THE THYROID IN THE DEVELOPMENT OF PLATYFISH BAKER-COHEN PLATE VII THE ROLE OF THE THYROID IN THE DEVELOPMENT OF PLATYFISH BAKER-COHEN PLATE VIII THE ROLE OF THE THYROID IN THE DEVELOPMENT OF PLATYFISH BAKER-COHEN PLATE IX $s SKsiwS&i s THE ROLE OF THE THYROID IN THE DEVELOPMENT OF PLATYFISH [1961] Zoologica: Index to Volume 46 223 INDEX Names in bold lace indicale new genera, species or subspecies/ num- bers in bold face indicale illustra- tions; numbers in parentheses are the series numbers of papers con- taining the plates listed immediately following. A Acanthorhodeus atremius, 25 Acheilognathus lanceolata, (9) PI. I X Rhodeus amarus, 101, (9) PI. I labira, (9) PI. I Agraulis vanillae vanillae, 2, 105, 113, 116 Alligator mississippiensis, 83 Arenaeus mexicanus, 142 Ateles belzebufh, 168 geoffroyi, 168 B Beilschmiedia tovarensis, 32 Bothridium parvum, 162 C Cacajao rubicundus, 168 Callinectes arcuatus, 141 toxotes, 142 Callilhrix jacchus, 168 Cebus albifrons, 168 apella, 168 capucinus, 168 nigrivittatus, 168 Cercocebus torquatus, 168 Cercopithecus aethiops, 168 diana, 168 l'hoesti, 168 mitis, 168 neglecius, 168, (15) PI. II talapoin, 168 Chasmocarcinus Iatipes, 155 Chasmophora macrophlhalma, 154 Colobus polylcomos, 168 Comopiihecus hamadryas, 168 Cronius ruber, 143 Cymopolia leucasii, 156 D Dione juno juno, 2, 6, 10, 105, 116 Dryadula phaetusa phaelusa, 2, 23, 106, 116 Dryas iulia iulia, 2,3, 11, 23, 105, 113, 116, (1) PI. I Durio carinatus, 79 E Erylhrocebus patas, 168 Eunectes murinus, 86 Euphylax dovii, 144 robusius, 145 Euryplax polita, 154 Eurytium tristani, 149 tristani minor, 149 G Galago crassicaudatus, 168 Ganua motleyana, 79 Geonoma vaga, 32 Gorilla gorilla, 168 H Heliconius aliphera aliphera, 2, 10, 105, 113, (11) PI. I doris doris, 2, 3, 105 erato hydara, 2, 3, 23, 105, 113 isabella isabella, 2, 3, 105, 116 melpomene euryades, 2, 3, 10, 105, 116 numala ethilla, 2 ricini insulana, 2, 105, 116 sara thamar, 2, 105, 116 wallacei wallacei, 2 Heteractaea peterseni, 152 Hexapanopeus beebei, 148, (13) PI. I costaricensis, 146 nicaraguensis, 147 orcutti, 147 sinaloensis, 147 Hexapus williamsi, 156 Hylobates Iar, 168 L Lagothrix cana, 168 infumata, 168 poppigii, 168 Lemur fulvus, 168 M Macaca irus, 168, (15) PI. II maura, 168 nemestrina, 168 X Macaca silenus, 168 radiata, 168 silenus, 168 Mandrillus leucophaeus, 168 sphinx, 168 Medaeus lobipes, 145 spinulifer, 146 Menippe obtusa, 151 Micropanope (?) maculatus, 151 polita, 150 xantusii, 150 N Nycticebus coucang, 168 O Ocotea wackenheimeii, 31 P Panopeus bermudensis, 149 purpureus, 149 Pan troglodytes, 168 Papio cynocephalus, 168 Paraxanthias taylori, 151 Passiflora auriculata, 4, 6 foetida, 4, 7 laurifolia, 4, 7 lonchophora, 4 quadriglandulosa, 4 rubra, 4, 6 serrato-digitata, 4, 6 tuberosa, 4, 11 vespertilio, 4 Perodicticus potto, 168, (15) PI. I Persea caerulea, 32 Philaethria dido dido, 2 Pilumnus limosus, 151 pygmaeus, 151 stimpsonii, 152 Pilhecia monacha, 168 Pongo pygmaeus, 73, 74, 168 Porlunus (Achelous) affinis, 139 brevimanus, 139 iridescens, 141 pichilinquei, 139 tuberculaius, 140 Porlunus (Portunus) acuminatus, 137 asper, 138 panamensis, 138 xantusii, 137 Presbytis enlellus, 168 obscurus, 168 Pseudorhombila xanthiformis, 154 Python molurus, 84 reticulatus, 84 sebae, 84 Q Quadrella nilida, 152, 153 R Rana curlipes, 103, 104, (10) PI. I Rhodeus amarus, (9) PI. I X Acheilognathus tabira, 101 ocellalus, 25 X Acanthorhodeus atremius, (2) PI. I sericeus amarus X Acheilognathus tabira, (9) PI. I S Saguinus leucopus, 168 nigricollis, 168 oedipus, 168 Saimiri sciureus, 168 Salmo gairdneri, 49, (4) Pis. I- VI Speocarcinus californiensis, 155 granulimanus, 154 ostrearicola, 155 Stealornis caripensis, 27, 29, 31, 37, 38, (3) Pis. I & II T Tanqua tiara, 161 Testudo vicina, 83 Thamnophis elegans, 57, 60, 61, 64, 66 Trattinickia rhoifolia, 31 U Uca, 91 minax, 89, (8) PI. I pugilator, 89, (8) PI. I pugnax, 89, (8) PI. I Urogale everetti, 168 V Varanus gouldii, 161 X Xanthodius stimpsoni, 146 Xiphophorus maculatus, 181, 187, 190, 194, 196, (16) Pis. MX variatus xiphidium, 125, 126, (12) Pis. I & II Z Zalacca conferta, 79 NEW YORK ZOOLOGICAL SOCIETY GENERAL OFFICE 30 East Fortieth Street, New York 16, N. Y. PUBLICATION OFFICE The Zoological Park, New York 60, N. Y. OFFICERS PRESIDENT VICE-PRESIDENT Fairfield Osborn Laurance S. Rockefeller SCIENTIFIC STAFF: John Tee-Van , General Director William G. Conway. . Director , Zoological Park Christopher W. Coates . . Director, Aquarium ZOOLOGICAL PARK Joseph A. Davis, Jr. . . Associate Curator, Mammals Grace Davall ...... Assistant Curator, Mammals and Birds William G. Conway . . Curator, Birds Herndon G. Dowling . Curator, Reptiles Charles P. Gandal. . . Veterinarian Lee S. Crandall. .... General Curator, Emeritus William Beebe Honorary Curator, Birds AQUARIUM James W. Atz Curator Carleton Ray Associate Curator Ross F. Nigrelli Pathologist & Chair- man of Department of Marine Biochem- istry & Ecology C. M. Breder, Jr Research Associate in Ichthyology Harry A. Charipper. . . Research Associate in Histology Sophie Jakowska ..... Research Associate in Experimental Biology SECRETARY TREASURER George W. Merck David H. McAlpin Klaus D. Kallman. . . . Research Associate in Genetics Louis Mowbray Research Associate in Field Biology GENERAL William Bridges . , Editor & Curator, Publications Dorothy Reville . . Editorial Assistant Sam Dunton Photographer Henry M. Lester. . Photographic Consultant DEPARTMENT OF TROPICAL RESEARCH William Beebe Director Emeritus Jocelyn Crane ........ Assistant Director David W. Snow ....... Resident Naturalist John Tee-Van Associate William K. Gregory .... Associate AFFILIATE L. Floyd Clarke. Director, Jackson Hole Biological Research Station EDITORIAL COMMITTEE Fairfield Osborn, Chairman James W. Atz William G. Conway William Beebe Lee S. Crandall William Bridges Herndon G. Dowling Christopher W. Coates John Tee-Van - ... . ... . V ..„ _ ., -