ZOOLOGIC A SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME 43 • 1958 • NUMBERS 1-14 PUBLISHED BY THE SOCIETY The ZOOLOGICAL PARK, New York NEW YORK ZOOLOGICAL SOCIETY GENERAL OFFICE 30 East Fortieth Street, New York 1 6, N. Y. PUBLICATION OFFICE The Zoological Park, New York 60, N. Y. TREASURER David H. McAlpin OFFICERS PRESIDENT VICE-PRESIDENTS SECRETARY Fairfield Osborn Alfred Ely Harold J. O’Connell Laurance S. Rockefeller SCIENTIFIC STAFF: John Tee-Van General Director James A. Oliver. . .Director, Zoological Park Christopher W. Coates. .Director, Aquarium ZOOLOGICAL PARK Grace Davall Assistant Curator, Mammals and Birds Joseph A. Davis, Jr.. .Assistant Curator, Mammals William G. Conway . . Curator, Birds James A. Oliver ... .Curator, Reptiles Charles P. Gandal . . Veterinarian Lee S. Crandall General Curator Emeritus William Beebe Honorary Curator, Birds AQUARIUM Janies W. Atz Associate Curator Carleton Ray Assistant to the Director Ross F. Nigrelli Pathologist & Chair- man of Department of Marine Biochem- istry & Ecology Myron Gordon Geneticist C. M. Breder, Jr Research Associate in Ichthyology Harry A. Charipper. . .Research Associate in Histology 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 Jocelyn Crane Assistant Director David W. Snow Resident Naturalist Henry Fleming Entomologist John Tee-Van Associate William K. Gregory. . . .Associate AFFILIATES L. Floyd Clarke Director, Jackson Hole Biological Research Station SCIENTIFIC ADVISORY COUNCIL A. Raymond Dochez Caryl P. Haskins Alfred E. Emerson John S. Nicholas W. A. Hagan EDITORIAL COMMITTEE Fairfield Osborn, Chairman James W. Atz William G. Conway William Beebe Lee S. Crandall William Bridges James A. Oliver Christopher W. Coates John Tee-Van Contents Part 1. April 4, 1958 PAGE 1. Social Behavior of the American Buffalo ( Bison bison bison). By Tom McHugh. Plates I-III, Text-figure 1 1 Part 2. August 27, 1958 2. The Iguanid Lizard Genera Urosaurus and Uta, with Remarks on Related Groups. By Jay M. Savage. Text-figures 1-6 41 3. Feeding Habits of the Northern Water Snake, Natrix sipedon sipedon Lin- naeus. By E. E. Brown 55 4. Uptake and Turnover of a Single Injected Dose of I131 in Tadpoles of Rana clamitans. By Nancy Weber Kaye & Elizabeth E. Le Bourhis. Text-figures 1-3 73 5. Oral Incubation in the Cichlid Fish Geophagus jurupari Heckel. By Melvin J. Reid & James W. Atz. Plate 1 77 6. The Specific Distinctness of the Fiddler Crabs Uca pugnax (Smith) and Uca rapax (Smith) at Their Zone of Overlap in Northeastern Florida. By Richard E. Tashian & F. John Vernberg. Plate I. 89 Part 3. November 20, 1958 7. A Practical Method of Obtaining Blood from Anesthetized Turtles by Means of Cardiac Puncture. By Charles P. Gandal. Text-figure 1 93 8. Observations on the Breeding in Captivity of a Pair of Lowland Gorillas. By Warren D. Thomas. Plates I-III; Text-figures 1 & 2 95 9. The Morphology of Renicola philip pinensis, n. sp., a Digenetic Trematode from the Pheasant-tailed Jacana, Hydrophasianus chirurgus (Scopoli). By Horace W. Stunkard, Ross F. Nigrelli & Charles P. Gandal. Plate I; Text-figures 1-3 105 Part 4. December 31, 1958 PAGE 10. Aspects of Social Behavior in Fiddler Crabs, with Special Reference to Uca maracoani (Latreille). By Jocelyn Crane. Plate I; Text-figures 1-5. 113 11. A Catalog of the Type Specimens of Fishes Formerly in the Collections of the Department of Tropical Research, New York Zoological Society. By Giles W. Mead 131 12. The Influence of Environment on the Pigmentation of Histrio histrio (Linnaeus). By C. M. Breder, Jr., & M. L. Campbell. Plates I-III.. . . 135 1 3. Studies on the Histology and Histopathology of the Rainbow Trout, Salmo gairdneri irideus. I. Hematology: Under Normal and Experimental Con- ditions of Inflammation. By Eva Lurie Weinreb. Plate 1 145 14. Radiobiology of the Newt, Diemictylus viridescens. Hematological and Histological Effects of Whole-body X Irradiation. By Sophie Jakowska, Ross F. Nigrelli & Arnold H. Sparrow. Plates I-III; Text-figure 1.. . . 155 t ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME 43 • PART 1 • APRIL 4, 1938 • NUMBER 1 PUBLISHED BY THE SOCIETY The ZOOLOGICAL PARK, New York Contents PAGE 1. Social Behavior of the American Buffalo ( Bison bison bison). By Tom McHugh. Plates I-III, Text-figure 1 1 1 Social Behavior of the American Buffalo ( Bison bison bison) Tom McHugh1 Jackson Hole Biological Research Station, Moran, Wyoming (Plates I-III; Text-figure 1 ) Contents I. Introduction 1 Sources of Information 1 Methods 3 Acknowledgments 4 II. Fundamentals of Behavior 4 Stances 4 Grooming Behavior 5 Feeding Behavior 5 Vocalizations 6 Reception 6 Play 6 Investigative Behavior Patterns 7 Charging 8 III. Herd Coordination, Integration and Movement. . . 8 Locomotion 8 Coordination during Movements 10 The Daily Round 12 Home Range of Free-ranging Herds 13 Integration of the Herd 14 IV. Herd Composition 14 Bull Groups and Cow Groups 14 Subgroups 16 Clans 16 V. The Dominance Hierarchy 17 Display of Dominance 17 Situations in which Dominance Was Displayed. .17 Structure and Dynamics of the Dominance Hier- archy in the Wildlife Park 18 Dominance in Wild Herds 21 Interspecific Dominance 22 VI. The Rut and Sexual Behavior 23 Development of the Rut 23 Sniffing and Extending Neck with Upcurled Lip. 24 The Tending Bond 24 Bellowing and Snorting 26 Wallowing and Horning 27 Battles 27 Precopulatory and Mounting Behavior 29 Bull Groups during the Rut 30 VII. Reproduction and Family Relations 30 Parturition 30 Relationships between Cows and Calves 32 Relationships between Calf and Group 34 VIII. Ecological Relations of Buffalo 35 Relationships with Other Animal 35 Interrelation with Vegetation and Soil 36 Summary 37 Literature Cited 38 I. INTRODUCTION HE American buffalo ( Bison bison bison ) not only shaped the life of the Plains Indians but also figured more prominently in American history than any other animal. A vast literature has grown up around the 1Present address: 13001 North Parkway, Cleveland 5, Ohio. buffalo, but there is still no adequate scientific study of its social behavior. This paper aims to add to the present limited knowledge in that field and to compare its contents with historical liter- ature. In the gathering of data on which this paper is based, I observed both free-ranging and con- fined herds of buffalo through the seasons. Man- nerisms and basic behavioral patterns of the animals themselves were noted as well as inter- actions with other buffalo in the coordination, integration and movement of the herds. Interac- tions between various herd members were also recorded to determine the type of social organi- zation. The herds were further studied to de- termine their composition. Reproductive be- havior was observed during the rut and the sub- sequent calving season. Due to the limitations of space, large sections of data had to be condensed into a few sentences. These conclusions, abrupt as they may seem at times, nevertheless rest on a substantial founda- tion of repeated observations recorded in field notes. I have used “buffalo” throughout in prefer- ence to “bison” because of common usage. The term “herd” refers to all the buffalo in any one geographic area. Each herd is in turn composed of smaller units caller “groups.” Sources of Information Observations. Most of the observations re- ported here were made on three buffalo herds, with additional data gathered from six other herds. The various herds are listed in approx- imate order of their importance as contributions to this paper in the following discussions of their size, origin, and range: 1 SMITHSONIAN INSTHVt APR z t 1551* 2 Zoologica: New York Zoological Society [43: 1 Yellowstone National Park: The Yellowstone Park Herd contained approximately 1,360 ani- mals when a census was made by air in 1951. These buffalo inhabit grassland areas which are divided into four main sections by natural bar- riers of forests and mountains. The four herds remain fairly distinct with only occasional inter- mingling: Hayden Valley Herd 440 Lamar Valley Herd 380 Pelican Valley Herd 340 Lower Firehole River Herd 200 The present Yellowstone Park Herd developed from the intermingling of introduced animals with the “native herd.” The introduced group was started in 1902 with 18 cows from the Pablo- Allard Herd of Montana and three bulls from the Goodnight Herd of Texas. These increased to 764 by 1925. In the meantime, a decrease in poaching permitted the native herd to grow from a remnant of 22 in 1902 to an estimated 100 to 125 animals by about 1925. Natural intermin- gling between the native and introduced (Lamar Valley) herds began in the early 1920s and was well developed by 1929. Remnants of the native herd plus an introduc- tion of 36 animals from the Lamar Herd led to the start of the Hayden Valley Herd in 1936. The Hayden Herd was particularly valuable in this study because it furnished the greatest num- ber of observations and because it was the only “undisturbed” herd studied. It had never been managed by reduction or artificial feeding since its start in 1936. Hayden Valley (Text-fig. 1) is a triangular-shaped, sagebrush-grassland valley of about 36 square miles located with its broad base along the Yellowstone River between Lake and Canyon. Yellowstone Park herds were observed during periods from March through November, 1951; in January and August, 1952; and in March, July, August and November, 1953. Systematic observations in Hayden Valley were possible only from May through October or November. Inacessibility at other times was due to snow cover and the closing of all park roads. One trip was nevertheless made into Hayden Valley in March, 1953, in two snow- planes. The Lamar Herd was followed during the winter months from November to May. Ac- cess to this herd is possible by a road which is kept open throughout the year. Jackson Hole Wildlife Park: The herd in the Jackson Hole Wildlife Park at Moran, Wyoming, was composed of 22 animals in 1951. About the same number was maintained in later years by removal of surplus buffalo. The herd was kept in the fenced “Display Area” of 133 acres from June to September. Tourists drove through this area to observe and at times disturbed the ani- mals. The buffalo were moved into a fenced “Winter Area” of 380 acres from approximately September to June. Hay and prepared concen- trated stock feed were provided from September or October through part of May— a period when available natural forage was scarce or buried under as much as four feet of snow. This Wild- life Park Herd was established in 1948 with 20 animals brought in from the Lamar Herd of Yellowstone Park. Observations were made dur- ing periods from February through November, 1951; in February and May, 1952; and in March and November, 1953. Wind Cave National Park: The Wind Cave Herd of approximately 500 animals was en- closed in a fenced area of 28,056 acres near Hot Springs, South Dakota. U. S. Route 85A passes through part of the park. The original 14 animals were shipped from the New York Zoological Park. Observations on this herd were made dur- ing periods from May through August, 1952, and from June through October, 1953. Approximately 350 buffalo were observed in the Wichita Mountains Wildlife Refuge near Cache, Oklahoma, during July and August, 1953. The herd was kept in a fenced area of 54 square miles. It developed from 15 animals originally shipped from the New York Zoological Park. The 700 buffalo in the Crow Indian Reserva- tion live on 34,000 acreas of grassland at an elevation of 6,000-7,000 feet in the foothills of the Big Horn Mountains of Montana. They are enclosed by one mile of heavy log fence and the precipitous cliffs adjoining the canyon of the Big Horn River. The Forest and Range Division of the Indian Service manages the herd for the benefit of the Crow tribe. The herd originated from buffalo shipped from Yellowstone National Park (Lamar) and the National Bison Range. It was observed during periods in August and September, 1953. Additional incidental observations were made on herds in the following four areas: National Bison Range at Moiese, Montana; Custer State Park at Hermosa, South Dakota; R. B. Marquiss’ Little Buffalo Ranch near Gillette, Wyoming; and Fort Niobrara National Wildlife Refuge near Valentine, Nebraska. Literature. The amount of historical literature on buffalo exceeds that of scientific treatises. This is because the pioneers in the West were virtually the only observers before the millions of buffalo on the Great Plains were slaughtered. 1958] McHugh: Social Behavior of the American Buffalo 3 Such historical material contains many casual observations on behavior and can at times be very useful but can also be most misleading if not carefully evaluated. I have cited pertinent historical evidence in this paper, regarding it as supplementary to actual observations and not necessarily of great scientific value. Methods Observations. Stalking under cover was used most often in approaching the herds in Hayden Valley and occasionally in other areas. Several devices aided observations. An automobile was used either as a blind or as a convenient and maneuverable base in Wind Cave, the Wichita Refuge, the Wildlife Park, the National Bison Range and the Crow Reservation. In winter in the Wildlife Park a canvas blind was used for warmth and concealment. Two platforms in trees about 15 feet from the ground facilitated viewing aggressions of buffalo feeding on hay. Most observations were made with the aid of 7 X 50 binoculars or a 27 to 60 power spotting scope. Telephoto lenses up to a focal length of 16 inches were used to record various activities on 16 mm. motion picture film. Later analysis of the film in a viewer supplemented the original observations. Movements of herds were noted on mimeo- graphed maps of Hayden Valley, the Wildlife Park, Wind Cave and the Wichita Refuge. A hand counter was used at times to make a census of herds. Identification of individual buffalo. Natural differences in the physical features of the various animals in a herd served to distinguish them as individuals. All of the 16 animals of the Wild- life Park Herd were identified by sight in this fashion. Identification of individuals in other herds was more difficult because more animals were involved. Individuals were often followed during a continuous daily observation but were seldom traced from day to day. Damaged or misshapen horns were the surest features for identifying individuals in large herds over long periods of time. Even the slight natural variations in horns served to identify individuals in the small, confined Wildlife Park; the main characters of horn structure in six mature cows were still obvious after a two-year lapse in obser- vations. The different positions of ear tags on these animals served as a further absolute check on identity. Natural distinctions between young animals of the Wildlife Park Herd were occasionally in- adequate for individual recognition, making sup- plementary marking with white paint necessary. Two methods were used to mark six calves, two yearling heifers and a pair of two-year-old heifers. In the first, arrows were fitted with rubber balls covered with layers of absorbent cloth. The ball was dipped in paint and fired at the buffalo from distances of 50 to 100 feet. The herd be- came irritable after one or two animals were marked and soon stampeded away. In the second technique, hay was used to lure the buffalo close to the slatted doors of a shed. A stick was tipped with absorbent cloth which was dipped in paint and thrust between the slats to daub different patterns on their heads. This method was simpler and quicker than the bow and arrow but could be used only during periods when no natural food was available. Aging of buffalo. Ages of buffalo were esti- mated by techniques considered adequate for the purposes of this paper. These were largely an expansion of the methods described by Hornaday (1889). Further refinement of tech- nique would be desirable for detailed population studies. Much additional research could well be done on herds where ages are accurately known and easily determined from brands on all ani- mals. This condition applies in the National Bison Range and the Wichita Mountains Wildlife Refuge. Such brands are difficult or impossible to read in winter because of concealment by thicker fur. Variations in the location and scar- ring of these brands are also helpful in identify- ing individuals. As with cattle (Pope, 1919) , the age of buffalo may be estimated by the development and sub- sequent wear of the permanent incisor teeth on the lower jaw. Age up to and including five years is estimated by the appearance of each pair of incisors. Beyond five years the age is approxi- mated by the levelling or wear on the incisors. The correlation with cattle is similar but was not determined exactly. Age may also be estimated by counting the number of growth rings on the horns, but any determination based on the number of rings should be increased by approximately three years to allow for the juvenile period when no obvious rings are formed. An alternate method is to es- timate the amount of horn growth in this period by comparison with the horn of a known three- year-old. Aging by such rings can be only an estimate. Most determinations of age for the purposes of this paper were reached by a visual examina- tion of the physical features of the buffalo, chiefly the size and shape of the horns. Bulls were aged accurately up to five or six years and 4 Zoologica: New York Zoological Society [43: 1 approximately up to about fourteen years. Cows were aged accurately up to three years. Plate I, Figures 1 & 2, demonstrate the tech- nique of aging by size and shape of horns more completely than any written description. Aging was most easily accomplished when the horns were viewed head-on and was more difficult in profile views. Complementing the horns as in- dicators of age were several other physical fea- tures of the buffalo. These included relative size, profile of the body and head, size of the hump and the amount of hair on the crown, forehead and forelegs. The size of the chin “bell” was a helpful indicator that was sometimes indefinite since some animals wore it off in grazing. Acknowledgments This project was assisted by several grants-in- aid from the Jackson Hole Biological Research Station of the New York Zoological Society. Great appreciation is due my adviser, Professor John T. Emlen, for direction, encouragement and advice in the drafting of this paper, which was originally submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy at the University of Wisconsin. I am also indebted to James R. Simon for help in the Jackson Hole area. Research in Yellowstone National Park was carried on with the aid and cooperation of Super- intendent Edmund B. Rogers and his staff. In- dividuals giving special information and help were Chief Park Naturalist David de L. Condon, Head Animal Keeper David W. Pierson, Biolo- gist Walter H. Kittams, W. Verde Watson, W. Leon Evans, Stanley McComas, John S. Bauman, Frank T. Hirst, Joseph Way and Lee L. Coleman. Ready cooperation was extended by Earl R. Semingsen and H. R. Jones of Wind Cave Na- tional Park; by Gordon Powers and Clark Stan- ton of the Bureau of Indian Affairs; by the Crow Indian tribe; by Ernest J. Greenwalt, Claud A. Shrader and Richard J. Hitch of the Wichita Mountains Wildlife Refuge; by John Schwartz, Cy Young and George E. Mushbach of the Na- tional Bison Range; and by Eliot Davis and K. C. (Sunny) Allan of Grand Teton National Park. The following persons helped with various items of information: Raymond J. Parker with his valuable observations in the Wichita Refuge, Tom Smith, Robert K. Enders, Margaret Alt- mann, Warren Garst, Richard Held, Martin W. Schein, Catherine Crocker, Cleveland Grant, John King, Charles McClaren, Watson Beed, Paul Hoppe, William Anderson, Hugh Wilmar and R. B. Marquiss and sons. The following persons read the manuscript and offered helpful suggestions: John T. Emlen, James R. Simon, David de L. Condon, Robert A. McCabe and John Schwartz. II. FUNDAMENTALS OF BEHAVIOR Stances Loafing positions of the buffalo included standing, lying flat on the side with legs and head outstretched on the ground and lying on the belly with legs tucked under or alongside. In the latter pose the head was sometimes rested on the ground. The eyes were partially or completely closed at times in all of these positions. Attitudes of “fear” were reactions to disturb- ance by strange objects, usually human beings. The buffalo stopped and stared for several sec- onds with ears brought forward and head di- rected toward the disturbance. This happened before running away or midway in the retreat. In a more accentuated alarm position, the head was raised above the level of the hump. Another stance involved the positions of the tail. Switching of the tail back and forth flushed insects from the rear of the buffalo. Yet frequent tail-switching also occurred in a variety of situa- tions when there was no apparent function of insect removal. It was prominent during playful battles and other types of play such as chasing and bounding. Elevation and switching of the \/ tail also occurred during the violent battles of the rut. Tail-switching was general during nurs- ing, more frequently by the calf than by the cow. It was seen among calves during herd movements when they were hesitating between staying with the calf subgroup or moving on with the cows. It was observed in a cow as she resisted each approach of a tending bull. Tail-switching may have been a “displacement activity” (Tinbergen, 1952:24-6), particularly in the latter two cases. The tail was raised and stiffly held 0° to 90° above the horizontal— most frequently at least 45 "—during the following circumstances: ( 1 ) In some trotting, running or bounding, such as during playful chases, stampedes and short charges; (2) While moving toward or investigat- ing unfamiliar objects, such as the vulva of a cow, a new bull or bull group, a new calf, human beings and a dead buffalo; (3) During moments of tenseness or excitement, such as moving through the herd in rut, before each attempted mount, “mock battles” (Section VI), disturb- ance from human beings, pauses in or before battles, play, anger and irritation. In the stance of bucking, the buffalo kicked its legs up or out to the side. Bucking usually involved the hind legs, singly or together. It occurred frequently during play. 1958] McHugh: Social Behavior of the American Buffalo 5 Arched back involved the humping of the back up into an arc. Bulls arched their backs during vicious battles, during some “mock battles” (Sec- tion VI) and while walking among the rutting herd or, more typically, in a bull subgroup dur- ing the rut. Arched backs during the last two oc- casions were usually accompanied by tail lifting and bellowing. Grooming Behavior Buffalo commonly rubbed their heads, necks and sometimes their sides on stumps, large low branches and trunks of trees. They also oc- casionally used smaller bushy pine trees, pine branches, boulders, sage-brush bushes, earth and earthen banks and snow banks. On trees, rubbing smoothed but did not remove the bark, although persistent rubbing could conceivably wear through the bark. Rubbing trees removed tufts of shed winter fur, yet this was not the sole purpose of rubbing since it occurred during all seasons. In addition to rubbing against trees, buffalo often horned lodgepole pine by stripping bark with the ends of their horns. The horning was sometimes accompanied by eating of bark and rubbing. Buffalo preferred horning the bark or branches of previously horned trees rather than starting on fresh material. Most horning occurred during the rut (Section VI). Buffalo wallowing usually consisted of one to three actions : a sniffing of the ground, a prelimi- nary pawing, followed by rolling on the ground. Although the first two actions were sometimes omitted, rolling never was. In rolling, the buffalo lay down on one side and then kicked the legs so as to roll onto the back. Buffalo were never seen to roll completely over. The action of the legs, particularly the forelegs, and the rolling of the body stirred up much dust. Wallowing was preceded or followed by horning or rubbing the head in the earth, and a type of “neck-crooking” where the neck was stretched and flexed and the horns occasionally scratched against the back. Some buffalo wallowed once on one side, stood, then wallowed on the other side. One or two rolls as an incomplete wallow sometimes pre- ceded the arising of a lying buffalo. Most wallow- ing was done in places where previous wallowing had broken the sod. Such wallows were scattered thickly over the entire range, as noted by aerial survey. Buffalo also wallowed in eroded or other natural bare areas, on prairie dog mounds, in wet mud holes, occasionally on snow and during the rut on any area. Wallowing among calves, first seen at the age of 1 3 days, lacked the skill of adult wallowing for at least the first month. The calves lay down and stretched or raised their legs slightly but did not roll thoroughly. Although both sexes wallowed during all seasons, wallow- ing was most noticeable among bulls during the rut (Section VI). I was not able to correlate rubbing or wallow- ing with the peak of concentration of mosquitos or black flies in Hayden Valley. The correlation of other authors (Garretson, 1938: 34; Good- win, 1939: 367; Hornaday, 1889: 413; Roe, 1951: 100; Soper, 1941: 385-6, 394) may have been genuine or may have been confused with the increase in wallowing among bulls during the rut. Feeding Behavior Buffalo grazed most commonly in meadows but also in open lodgepole pine and aspen to some extent. Browsing was infrequent. The Wildlife Park Herd fed on aspen bark only rarely. Buffalo in the Crow Reservation barked aspen extensively only during one severe winter when the range was depleted. Both cows and bulls in the Wildlife Park and Hayden Valley infrequently stripped and ate bark from lodge- pole pine at all seasons. When there was a thin snow cover of only a few inches, buffalo clipped off the exposed forage within one inch of the snow. In deeper snow, the buffalo pushed their noses into the snow and then cleared a trench down to the forage by swinging the head. The side of the head and even the neck were used to clear very deep snow. Pawing with either front hoof aided this nosing only infrequently. Considerable feed- ing by nosing snow resulted in shortened hair around the mouth and crusted snow conditions produced raw noses. Buffalo fed by these tech- niques in snow up to four feet deep, leaving large pits as evidence of their work. Buffalo went to water at least once each day in all herds observed. Water in these areas was always rather accessible in springs, streams, lakes or rain-flooded puddles. The Wind Cave Herd sometimes stayed near water for one or more days when watering holes were scarce. The Superintendent of the National Bison Range found that the most heavily grazed areas were near water holes and the most lightly used areas away from water. He established new water holes to effect more even use of the range. When water holes froze in the winter, buffalo ate snow. If both snow and open water were available, however, they preferred water. When ice covered small puddles, they broke through with their noses or front hooves. They had no aversion to drinking the very brackish water found in stagnant pools. 6 Zoologica: New York Zoological Society [43: 1 Both the Wind Cave and Wildlife Park Herds regularly visited the salt blocks that were put out for them. Other herds occasionally ate the mud from natural salt licks. Vocalizations Observations on several herds showed that vocal expression consisted mostly of grunting and some snorting and sneezing sounds. The variation in grunts was extensive, difficult to de- scribe, and was correlated with a variety of responses from other buffalo. The frequency of grunting increased during group movements, such as stampedes or direct movements, and was noticeable when the Wildlife Park Herd was tem- porarily split by a fence. During feeding on hay, the dominant cow in the Wildlife Park frequently uttered threat grunts as she advanced on nearby buffalo. These subordinate animals yielded to her. She also gave the threat grunt when other buffalo or human beings approached within a few feet of her new calf. This threat grunt was quite similar to the usual buffalo grunt, such as the one she used to call her calf. Threat grunts were also heard from one bull in the Wildlife Park and two others in Yellowstone Park. They would be noticed in wild herds only under ideal condi- tions of close observation. During periods of intense play, buffalo in- frequently uttered a loud short sneeze, a loud belching snort, or a bellow-like grunt. Calves or juveniles who were hit by the full force of a charge from another buffalo uttered a loud screaming or bellowing grunt. Buffalo ground their teeth together to produce a squeaking noise. Waving of hands stimulated this from cows enclosed in a small corral in the Lamar Valley. Bellowing and snorting will be discussed later (Section VI). Reception Buffalo apparently relied most on an acute sense of smell for detecting danger, raising their heads to test the air when new odors drifted in. Whenever scenting human beings, the Hayden Herd stampeded without hesitation even though the source of scent was not visible. They reacted to scent carried at least one mile, but Inman (1899: 246) stretched this distance to four miles. Two authors stated that buffalo scented water when it was “five miles or more” away (Garret- son, 1938: 46) or “miles distant” (Inman, 1899: 247). Vision was evidently not as acute nor relied upon as much as scent. It should not be under- estimated, however. Buffalo in the Crow Reserva- tion spotted our moving jeeps when they were more than a mile distant. Buffalo in the National Bison Range paid no attention to horses yet be- came restless as soon as they spotted a horse and rider. One herd immediately became restless when a horse and rider appeared on the skyline 0.8 mile distant (oral comment from Supt. John Schwartz). The Hayden Herd paid little atten- tion to people walking more than 0.5 to 0.8 mile from them if neither noise nor scent were de- tected. They would become alarmed by a closer approach. Closer approach often went unde- tected, however, if there was no movement or slight camouflage from sagebrush or tree shadows. Sometimes the buffalo would moment- arily stare, only to start grazing again. Small groups of buffalo were stalked in open view by moving only when all were grazing or facing away and by “freezing” whenever any animal looked up. Play Play among buffalo consisted of battling, mounting, frolicking, or, more generally, a com- bination of these. Bucking was quite typical of all play. From an anthropomorphic viewpoint, these activities were assumed to be pleasurable and non-utilitarian, two features discussed by Beach (1945: 524) as charateristics of playful behavior. This play also fitted the description given by other authors. Schein & Fohrman (1955) stated that playing in cattle “is an activity engaged in solely for the sake of the activity it- self and not for the normal end result of the activity.” Thus, battles among buffalo would occur for the sake of the battle itself and with- out the emotion of competition over a specific object. Mountings occurred without the specific emotion of a sexual drive. Racing occurred with- out the specific emotion of fleeing from an enemy or reaching a goal. Such reasoning makes a sat- isfactory hypothesis for play, yet the actual recognition of these activities in the wild is often virtually impossible without relying on a sub- jective interpretation. Numerous observations revealed that play was most frequent among calves, frequent among yearlings, common among two-year-olds, and occasional in older buffalo. There was more play —chiefly battling— among bulls than among cows in the group two to three or more years old. Most play occurred during the period of evening dusk. Play was also common during direct and semi- direct group movements (including stampedes) and just after feeding, particularly in the Wild- life Park. (Play occurring during dusk may have followed feeding, but the weaker light was ad- judged to be a more important stimulus than 1958] McHugh: Social Behavior of the American Buffalo 7 removal of hunger since the period of evening play was more intense than that after any diurnal feeding period.) Buffalo battled in play by butting and twisting their crowns together or by hooking horns with some pushing back and forth or circling. Calves butted their small horns against the opponent’s shoulder, neck or head (Plate II, Figure 3). Many battles were started when one opponent approached another, shaking its head or bucking up and down on its front legs. Pairs of yearlings and calves “shadow-butted” by facing each other with heads low and then quickly shifting side- ways so that each “butted” the space his op- ponent had just vacated. Since battles occurred between a variety of sexes and ages, many op- ponents were unevenly matched, yet viciousness or severe routs were rare. The latter two qualities characterized battles that were not considered play (Section VI). Occasional battles involved just one buffalo who butted a trunk of a tree or a springy sapling that “fought back.” Mounting as play occurred without regard to sex, age or dominance. It consisted either of at- tempted mounting where the chin was placed on the partner’s rump or of complete mounting where the buffalo clasped its forelegs around the flanks of its partner and rode for several seconds (Plate II, Figure 4). Attempted mounting was sometimes a complete act with no intention shown of proceeding further. Unsheathing of the penis or thrusting actions were very rare. Mount- ing among calves was first noted when a 15-day- old calf made three feeble and unsuccessful at- tempts. A 20-day-old calf mounted another for four seconds. Such mounting during play was probably not based specifically on sexual moti- vation (Carpenter, 1942: 198; Nissen, 1951: 439). Mounting and battling were often coupled in play as one frequently initiated or followed the other. Frolicking included the aimless racing of one or several animals, playful stampedes and “race- and-hide.” Most aimless racing was done in the bounding or galloping gaits as one animal or small groups dashed back and forth, in circles, among the group, or in short charges. The bound- ing of one animal away from the group usually resulted in the quick following of another and then the aimless racing of the pair. Some juve- niles raced in these games for so many minutes that they finished with tongues hanging out and panting heavily. Calves as young as one day trotted, ran or bounded in 15-50 foot circles from their mothers. Some racing took the form of playful stampedes as the entire group or subgroup dashed aimlessly about. In still another form of racing, race-and-hide, single animals galloped for some distance and then abruptly lay down. The pattern somtimes repeated, and there were indications that some animals lay down near certain other individuals. Grooming behavior— mainly wallowing with occasional rubbing— occurred so commonly with play that it should be considered an accessory to play or sometimes a form of play itself. Investigative Behavior Patterns Curiosity was well developed in buffalo. They thoroughly investigated new or strange objects such as an observation platform, recently shed elk antlers, a skeleton, horses, a porcupine, prairie dogs, and a human being lying or sitting after the buffalo were used to seeing him stand- ing. Juveniles were more curious than adults. Herd members moved in to investigate, sniff and lick new calves. Older calves were particularly curious about new calves. Exploring further manifested the development of curiosity. It was uncommon and restricted mainly to the Wildlife Park Herd. Subgroups moved several hundred feet away from the main herd to investigate various objects and then re- turned. Entire herds also explored. This explor- ing may have been a type of play since it was often accompanied by activities discussed pre- viously as play. In earlier times, investigative behavior may have played a part in the creation of what hunters called a stand at which entire herds were slaughtered by methodical shooting from under cover. There are numerous historical references to this (Allen, 1876: 212; Dodge, 1877: 135-7; Garretson, 1938: 44, 116-7; Hornaday, 1889: 466-9, 1922: 143; Inman, 1899: 261-2; Mayer, 1934, II: 36-7; et al.). Three instances of shooting in present day herds are of interest in this connection. Rangers in Yellowstone Park successively shot all seven members of a cow group and another bull group by approaching so the buffalo could neither see nor smell them. The groups did not stampede away when their members started to drop. Some cows in the cow group started to walk gradually away, but the motion was slow and without panic. When a hunter in the Crow Reservation shot from cover, the buffalo did not run away but milled around and sniffed the first slaughtered animal. Mr. Q. Marquiss killed 20 animals each year in his herd of 150 by driving up to it and shooting the desired animals one by one. The herd was not afraid of cars or human beings and did not run away during the shooting, remaining instead to mill about the dead buffalo. Zoologica: New York Zoological Society [43: 1 In discussing stands such as these, several authors mentioned the attraction of buffalo to dead animals or blood. Mayer (1934, II: 36) stated, . . they would begin what we call ‘mill- ing.’ They would nervously smell the wounded animal, then hook her with their horns, then smell her again, bewildered . . .” Hornaday (1889: 476) noted that “They cluster around the fallen ones, sniff at the warm blood . . .” Allen (1876: 212) told how “The hunter must drive away the stupid creatures to prevent the living from playfully goring the dead!” Dodge (1877: 135-6) mentioned how “Attracted by the blood they collect about the wounded buffalo.” Similar incidents in present-day herds are interesting. Buffalo milled around and sniffed the spot where an elk calf was born. Another herd milled about, looked at and sniffed the carcass of an old bull killed in battle. Butchering of buffalo in the Wildlife Park was twice impeded by the herd, which pressed close to sniff and lick the viscera and carcass. During one of these butcher- ings, the two older bulls pawed the snow and engaged in a vicious battle. Buffalo in the Lamar Herd mauled and gored a dead mature elk. If Mr. Q. Marquiss’ slaughterers did not immedi- ately move in to each killed buffalo, the herd milled about and horned it. When buffalo were killed from cover in the Crow Reservation, the group remained about and sniffed the carcasses. The grass in the National Bison Range surround- ing a buffalo carcass was very well trampled. All buffalo sniffed another area where blood was spilled on the ground. Bulls pawed, rolled in and horned this spot. Charging Buffalo are potentially dangerous animals and should always be approached with caution. On six occasions I was faced with situations in which danger was latent, and was charged at five other times. Each charge was a bluff, ending with the buffalo stopping a few feet distant. There is no reason to believe that this would be a constant pattern, however. III. HERD COORDINATION, INTEGRATION AND MOVEMENT Locomotion Gaits. Buffalo progress by means of four dif- ferent gaits, walking , trotting , galloping and bounding. The detailed description of gaits in various animals was pioneered by Muybridge (1907) , who analyzed them with a series of pho- tographs. I have relied heavily on his work, using his definitions combined with my own observa- tion and analysis of motion picture film of mov- ing buffalo. The descriptions of the four gaits are de- pendent upon a few definitions: The step of a moving buffalo is an act of raising, thrusting and returning one hoof to the ground so that it can reassume the complete or partial support of the body. The normal stride of a buffalo consists of a sequence of four steps as each hoof is lifted and again placed on the ground. For brevity, the sequence of steps is listed by the number of each hoof, according to the following system: 1. Left anterior 2. Right anterior 3. Left posterior 4. Right posterior In walking, the sequence of hoof movements is 3, 1, 4, 2. This was a standard sequence for all quadrupeds investigated by Muybridge. The support during each stride of an average walk falls on two or three hooves. The slowest walk, such as the “hesitant gait” of bulls during the rut, involves the movement of only one hoof at a time so that the support during each stride falls on three or four hooves. In the fastest walk, an amble, the buffalo supports its weight alternately on one hoof and then two hooves. Trotting is the next faster gait in buffalo. The sequence for limb movements is 1-4, 2-3 with each diagonal pair of limbs moved more or less simultaneously. During each stride, the buffalo advances twice with all four feet off the ground. Galloping or running is the most rapid gait. The sequence here is 1, 2, 3, 4, making the type of gallop of buffalo a “transverse gallop.” Other animals showing this same type are horse, goat, cat, bear, raccoon and hog (Muybridge, 1907: 156), in contrast to a lesser number of animals showing the “rotary gallop” (1, 2, 4, 3). Some slower speeds of galloping in buffalo show a forward-backward rocking motion. Bounding is a less commonly seen gait in which the buffalo springs ahead by the more or less simultaneous flexing of all four legs. It is similar to bounding in mule deer. This bounding type of gait was observed mainly under two conditions. It was frequent during periods of play, accompanying such actions as battling, butting, mounting, frolicking and bucking (Sec- tion II) . Buffalo often bounded in their playful chases or races and it was first seen among calves at an age of two months. Buffalo also bounded when frightened or suddenly surprised. The ap- proach of a human being, especially if abrupt, resulted in the fleeing buffalo using this gait. The animal characteristically stopped at a distance of 20 to 100 feet, looked back, and then con- tinued in a walk, trot or gallop. 1958] McHugh: Social Behavior of the American Buffalo 9 Speed. Herds of buffalo chased by a jeep in the Crow Reservation maintained a steady speed of 30 miles an hour. The jeep could not go faster over the uneven ground of these meadows, and on one occasion a group spurted from behind and passed jeep, travelling at about 35 miles an hour. Cottam & Williams (1943) measured the speed of a three-year-old bull on level ground as 30 miles an hour and stated that older and younger animals were slower. Howell (1944) timed the top speed of a two- or three-year-old buffalo at 32 miles an hour. Buffalo generally increased their speed on downhill sections of a path when proceeding steadily and directly toward some goal, such as water or food, or when stampeding. If the group was walking at the time, they trotted or galloped downhill and for a short distance out onto the level. If galloping, they increased their pace still more. Buffalo moving toward a goal sometimes in- creased their speed as they neared their destina- tion. This happened with animals returning to a herd after being temporarily separated and with those joining new herds. It also occurred in groups moving toward water or food. Trails. The trails of buffalo have been de- scribed in a glowing and often exaggerated fash- ion. Garretson (1938: 57) states that “The buf- falo was the best natural engineer the world has ever known . . .” Hornaday (1889: 417) be- lieves that “The trail of a herd ... is usually as good a piece of engineering as could be executed by the best railway surveyor . . .” The more reserved opinions of Allen and Dodge better describe the extensive system of trails which I saw in Hayden Valley. Allen (1876: 63) says: “. . . a buffalo trail can be de- pended upon as affording the most feasible road possible through the region it traverses.” Dodge points out that “. . . though a well-defined buf- falo trail may not be a good wagon road, one may rest well assured that it is the best route to be had.” The extensive network of trails in Hayden Valley included good routes from any one sec- tion of the valley to any other. The trails were almost invariably the most practical routes available, but they were not necessarily the best engineered routes. Main trails traversed rolling country with a minimum of wasted climbing. They cut through corners of timber for a more direct path. They veered around the edge of marshy areas containing quicksand. Fords of streams and creeks had gradual approaches and shallow water. Direct approaches were made to the only feasible crossings of an almost impass- able canyon. Trails cut about one-half mile through large areas of timber in a fairly direct line from one meadow to another. The buffalo took a gradual route up hills if such a route was essentially direct, but they had no aversion to steep climbs. If there was a choice between a longer, more gradual incline and a shorter route through a steep-walled ravine, the buffalo took the latter. The route over the Mary Lake Divide was so steep that a horse with rider could negotiate it only by following a series of switchbacks superposed on the buffalo trail. The only impractical choice of route by the buffalo that I observed occurred on the trail over the southern divide in Hayden Valley. An 1,800-foot section of this was very difficult to traverse because of fallen timber. Another side trail, seldom used by the buffalo, was 700 feet longer but completely avoided the tangle of fallen timber. Trails through timber had many shunts and were occasionally hard to detect. Trails through meadows were usually more clear-cut and easily followed, although seldom-used trails sometimes disappeared in large meadows. Buffalo sometimes started new trails from eight inches to several feet to the side of the old ones. These were started under the following conditions: (1) Trails on hills became trampled and eroded to considerable depth. Erosion in particular produced a trail that was deep yet very narrow, and thus difficult to walk in. Similar narrow trails also occurred in soft, marshy areas. When these trails became deeper than about one foot, the buffalo started new parallel paths. (2) Trails pushed through areas with thin top- soil over a rocky subsoil soon eroded down to the rocks. Such rocky trails were abandoned for new parallel routes. Buffalo can at times wear down trails with surprising speed. On one occasion when the soil was muddy, 265 buffalo passed over one trail and thereby cut it one-half to two inches deeper and from 12 to 15 inches wide on this single trip. Buffalo soon trampled a path through rotten logs lying across trails, yet large, recently fallen logs were either crossed or bypassed. Maneuvering in snow. Snow accumulated in Hayden Valley to a depth up to four feet in the ravines, level meadows and forests but was largely blown or melted from the ridges and some south-facing hills. The buffalo spent most of their time on such bare or near-bare areas and also on thermal areas, where the snow was melted by the heat of the ground. 10 Zoologica: New York Zoological Society [43: 1 They showed a distinct aversion to travelling through deep snow. A group of five bulls chased by a snow plane refused to go into the deeper snow of a ravine; a bull dashed six feet in front of the plane to take the shortest route to a bare spot. In summer the buffalo in Hayden Valley characteristically took flight when the observer was 1,000 feet distant, yet they were not this wary in winter. In one case a group of 26 cows rushed through snow six inches deep within 30 to 50 feet of the author rather than move through snow farther away that was one to two feet deep. One bull galloped through the cold, shallow water of a stream rather than plunge through snow two or more feet deep on the surrounding ground. A complex system of well-packed trails through deep snow connected the ridge tops so that the buffalo could move almost anywhere in the valley without plunging into deep snow. The warm water of Alum Creek melted snow for one to two feet along the bank and this bare area was used as a trail. In making new trails, the buffalo tended to follow in the hoofprints of the leader, but older trails lost most of this pitted pattern. Buffalo headed directly toward the proper trail when leaving a bare area, suggesting that they were well aware of the system of trails. Although the buffalo avoided plunging into unbroken snow whenever possible, they were still able to negotiate deep snow. When an air- plane frightened a group of 25 in Hayden Valley, they blindly ran off a bare spot over a cornice twenty feet deep. The group floundered badly, and one buffalo even clambered over the back of another. They still plodded through. In an- other case, two bulls struggled through snow four feet deep. The lead bull broke trail by climb- ing up on the snow with his front feet and then plunging down. Travel was very slow and was interrupted with frequent periods of rest of a minute or so. On another occasion, a cow group of 30 moved single file or occasionally two abreast for 4,000 feet across unbroken snow 40 inches deep in a large meadow. The extremely slow speed of the cows under such conditions would allow Indians to approach closely and slaughter them, as recorded by Catlin (1876, I: 253) and Seton (1929, I: 271). In March, 1945, rangers trailed a group of 54 buffalo that moved six miles through snow 42 inches deep from the Nez Perce meadows to Violet Springs in Hayden Valley (recorded by Chief Park Naturalist David Condon). Swimming. Buffalo seldom hesitated to enter water yet were usually hesitant about crossing swift, deep streams. When the Wind Cave or Wichita Herds waded out to drink in a lake, occasional animals swam out into deeper water. Buffalo in the Wichita Refuge also swam across ponds in their movements even though wading or walking around them would have added little extra travel. Buffalo in Yellowstone Park regu- larly waded and swam the flood-swollen Lamar River in spring. Coordination during Movements Leadership. Leadership during travels of a herd was apparent when one or a small group of animals initiated or directed the movement, usually from in front of the herd. Leadership during movements of the confined Wildlife Park Herd of 16 buffalo is tabulated in Table 1. Three cows accounted for 66% of the leadership, with no cow acting as the exclusive leader. All indi- viduals led the herd at least once, and leadership sometimes varied during each movement. Fre- quent changes in leaders also occurred in free- ranging herds. Table 1 . Leadership During 43 Movements of the Herd of 16 Buffalo in the Wildlife Park Cow % of trips in lead in herd Age Position of dominance in hierarchy among cows (from Table 4) F3 27% 7 Third FI 23% 5 First F6 16% 4 Sixth Highest percentage for any remaining herd member was 7%. Leadership depended upon cooperation. The lead buffalo characteristically looked back to see if the rest of the group was following and usually advanced alone no more than about 100 feet. In the occasional cases where the group balked and did not follow, the leader either turned back to become a follower or waited several minutes until some individuals began to follow. During situations of disturbance by human beings, one animal was able to initiate move- ment in the entire group by moving in one di- rection in a steady walk. Leadership under such conditions was not balked. Leadership was not readily detectable in many group movements, particularly during grazing. This mass-action type of travel, where the group moved as one unit without any obvious leader, seemed to be based on several factors. Con- tagious behavior was important and especially noticeable when all animals were facing in the same direction and moving at the same rate. 1958] McHugh: Social Behavior of the American Buffalo 11 Dominant animals in the center or rear of the group seemed to direct the procession by mov- ing all subordinates ahead of them. The leader or leaders were not conspicuous because of the close coordination of their actions with the mass of the group. Movement was also based on a high degree of cohesiveness within the group. The advance of a group of buffalo from the main herd was more apt to result in the remainder following and joining than in separation into two or more groups or the return of the first group. Large groups of 100 or more sometimes moved by the successive and repeated closing of gaps between subgroups. Some movements were so random that leadership probably never existed. Stampedes. Stampedes involved the sudden and usually simultaneous trotting or, more typically, galloping of a group of buffalo. Clouds of dust rose up around and sometimes com- pletely enveloped the stampeding group. The noise of a stampede resembled a continuous rumble, the rush of wind, or the rain-like patter of hooves. Stampedes through timber were char- acterized by the very clamorous breaking of branches and hitting of hooves against fallen trees. The following notes on stampedes are based on numerous observations in all herds studied. In addition, I spent thirteen days in the Crow Reservation in Montana with the express pur- pose of starting and filming as many buffalo stampedes as possible. During the latter part of this period stampedes were started by parties of hunters who chased the buffalo in jeeps to carry out the annual program of herd reduction. Stampedes were started by the sudden trotting or galloping of one or a group of buffalo, with the herd usually taking the direction of the first group. Some stampedes started so abruptly that the initiating group was not evident. Stampedes were often caused by the disturb- ance of a wary group. They also occurred in all herds for reasons of seemingly insignificant im- portance. For example, stampedes were started by the sudden running of one animal which voluntarily decided to join the group from the outskirts or which was scared by a sudden snap of a twig. Subgroups stampeded to rejoin the main group when there was no apparent reason for such a burst of speed. The sudden flight of elk, deer or a pair of mallards near the herd caused the buffalo to stampede in the same direction, even though they may never have seen the reason for the flight. Such stampedes as these, covering only a few hundred feet, were shorter than those result- ing from the approach of human beings toward a wary herd. At the start of a stampede, all the buffalo in the area usually massed together quickly and then galloped away. After this initial burst, the herd either stopped and looked or continued moving. The continuation of the stampede was made at a gallop, a trot or a fast walk. The same massed pattern was held, or, more commonly, the herd strung out as a column of groups, a long continuous column or single file. Stampedes initiated by the presence of human beings, were not a blind fleeing from the dis- turbance. Rather, the path of these stampedes was regulated mainly by the end-point, which was usually a small timber-secluded meadow or timber itself. One stampede observed in the National Bison Range was, however, quite directionless. A se- vere hailstorm in August, 1954, completely cov- ered the ground with stones up to half an inch in diameter. A group of 150 buffalo was loosely spread out into many smaller subgroups when the storm arrived. They quickly massed and then galloped at intervals in an aimless fashion. De- spite rather frequent stampeding, they circled in no more than a quarter square mile. Playful stampedes observed in the Wildlife Park were also rather directionless. Three incidents were observed where animals fell under the herd during a stampede. The first occurrence was observed by Gordon Powers when a motion picture company was attempting to film a stampede in the Crow Reservation. Stationary lines of vehicles and of men were supposed to funnel the herd past the camera. As a group of jeeps started to drive the herd of 400, the buffalo headed directly for the line of men. When they were only 50 feet away, the four lead animals suddenly turned back. They were trampled under the rest of the herd, which could not turn so abruptly. None was killed, but one animal walked away stiffly as though badly bruised. A similar incident occurred when a stampeding herd suddenly caught sight of a group of hunters in the Crow Reservation. Three lead animals were trampled as they turned abruptly. In another case, a buffalo fell in front of a stampeding herd. Other animals jumped over or passed by this buffalo as it made one complete roll and got up again. Wariness. Wariness varied greatly in different herds. The Wildlife Park herd was in constant contact with people and would allow approach within three to five feet. The Wind Cave animals allowed approach of cars or people within 25 12 Zoologica: New York Zoological Society [43: 1 feet. The Wichita herd tolerated the approach of cars within five feet or less but persons could approach only to within 30 feet. Disturbance by cars or people caused the herds to move no more than a few feet. Buffalo in the National Bison Range permitted the gradual approach of cars to within five feet or less, yet would usually stampede away if the occupants got out. Further discussion of wariness hinges on a definition of flight distance. It is here defined as the distance between an approaching dis- turbance and the herd, measured at the time when the herd takes flight. The Crow and Hayden Herds had little con- tact with human beings. Either herd stampeded several miles when disturbed by cars or people. The flight distance for the Crow Herd was one mile or more on some occasions. It lessened, however, after the fifth day of chasing and hunt- ing. This herd used to be relatively tame and allowed approach of people to within 15 to 50 feet. Gordon Powers of the Bureau of Indian Affairs noted that the herd grew progressively wilder in the past seven years as jeeps were used to chase and hunt it. Flight distance varied sea- sonally in the Hayden Herd. In the summer it was 1,000 to 1,500 feet, and the herd would also move away upon catching human scent. In November, when there was no snow, the flight distance was 300 to 600 feet and the retreat of the herd was slower and covered less distance than in summer. It was still less in winter, some- times as low as 200 feet. The herd was slower in starting to move away, particularly if on a bare spot surrounded by deeper snow. Other herds in Yellowstone Park were less wary than the Hayden Herd. Both the Lamar and Firehole River Herds had more contact with cars or people and also had shorter flight dis- tances than the Hayden Herd. Historical literature shows instances of wild or wary herds (Audubon & Bachman, 1849: 47; Catlin, 1876, I: 254-5; Parkman, 1903: 91-2, 404-5; Long, 1905, XV: 255-6, XVI: 140, 288; Hornaday, 1889: 536; Inman, 1899: 57; and Roe, 1951: 131-8). There are also records of relatively tame or unwary herds (Dodge, 1877: 120-1 and 1885: 283; Hornaday, 1889: 389, 404-5, 465, 471; Long, 1905, XVI: 153; Park- man, 1903: 423-5; and Roe, 1951: 131-8). The herds became notably wilder as their numbers decreased near the end of the period of slaughter (Hornaday, 1889: 430; Roe, 1951: 139-140). This survey of historical literature on wariness shows that probably none of the original free- ranging herds were as tame as those in the Wildlife Park, Wind Cave or the Wichita Refuge. There are records of wariness comparable to the wilder Crow or Hayden Herds. The Daily Round The daily round, consisting of behavior dur- ing an entire day, was divided into two main types of activity: (1) Feeding behavior which included short periods of walking to select and eat natural forage or hay, and (2) non-feeding behavior which included much loafing behavior such as lying and standing, and also ruminating, traveling, watering, licking salt, playing and grooming. Feeding usually commenced sometime after dawn, but the exact time varied. Throughout the day, cycles of feeding alternated with those of non-feeding. This is quite similar to grazing be- havior in cattle (Hancock, 1953: 3). Feeding activity sometimes declined during periods of high temperature around noon. Groups con- tinued feeding into complete darkness at times. A period of feeding and traveling occasionally occurred at any time during the night in all herds. These periods of nocturnal activity were relatively short when compared with the greater amount of time spent loafing. Nocturnal periods of feeding were usually singular and there was no evidence of any cyclic rhythm. It should be realized, however, that nocturnal observations were too meagre to support any general con- clusion. The daily radius (Leopold, 1947) was that dis- tance traveled by a group of buffalo in one 24-hour period. The average daily radius during the rut for five group-days of travel in Hayden Valley was two miles, with a range from 1.5 to 2.4 miles. The maximum daily radius was re- corded during a stampede over the southern divide, when the group traveled at least nine miles. The average daily radius for 20 group-days of travel during the rut in the Wichita Refuge was 2.1 miles, with a range from 0.5 to 5.8 miles per day. The average daily radius during the rut in the Wildlife Park was 1.7 miles, but the 133 acres of this park probably reduced the radius to less than that for wild herds. The daily radius for herds in areas where water holes were scarce would be expected to be considerably greater than any of the above aver- ages. In late summer, the Wind Cave Herd some- times made daily round trips to water of at least six miles. Hancock (1953: 8) surveyed the literature for daily radius in cattle. A free-ranging herd 1958] McHugh: Social Behavior of the American Buffalo 13 Text-fig. 1. Range of Buffalo in Hayden Valley. traveled 3.0 miles per day, while herds enclosed in 1.0 to 5.5 acres moved from 1.4 to 1.7 miles per day. Home Range of Free-ranging Herds Text-figure 1 pictures the home range of the Hayden Herd. It was approximately 12 square miles in summer and extended to 36 square miles in winter (October-November through April- May). This extension in winter coincided with a cover of snow and an absence of tourist traffic on the main highway. The Hayden range was largely confined by lodgepole pine forests on all sides. The herd nevertheless often penetrated part way into this border. It moved into all small meadows that jutted out into the timber from the main meadow, and also went through the timber to reach isolated meadows. The timbered border on the north and south of the valley still proved to be an effective barrier against egress to new range. The highway on the east stopped most egress or ingress in summer, yet there may have been some interchange between the Hayden Herd and the Pelican and Lamar Herds toward the east in winter. The only known interchange of any extent occurred to the west across a 3.6 mile trail through the timber over the Mary Lake Divide. Interchange over this trail took place between the Hayden Herd and the Nez Perce Herd, which was part of the Lower Firehole River Herd. The Lamar Herd moved each year about 10-25 miles from meadows at 8,000-9,000 feet to the lower Lamar Valley at 6,500 feet with the advent of fall snows. It returned to the higher meadows again with the disappearance of snow in May or June. 14 Zoologica: New York Zoological Society [43: 1 Integration of the Herd The high degree of cohesiveness of the Wild- life Park Herd was demonstrated when the herd was artificially split. A fence divided the herd into a group of five cows, one yearling and one bull, and another group of five cows, three year- lings and one bull. The two groups regularly lay down as one herd in spite of the fence. There was much exchange grunting between them. They also paced back and forth along the fence, a type of behavior which was not seen at other times. Their daily schedule of activity was almost identical. Other groups of buffalo also lay to- gether as one on each side of the fence separating Wind Cave from Custer State Park. This high degree of cohesiveness in the Wild- life Park was remarkable in view of the great amount of aggression between members of the herd, particularly during feeding on hay (Sec- tion V). Mutual toleration must have overbal- anced aggressive behavior in this group, as shown objectively by its tendency to aggregate (Collias, 1944:89). A high degree of cohesiveness was shown by individuals who became temporarily separated from the main group through differential move- ments. When they discovered their isolation, they abruptly ran back to join the group. Herds also displayed a high degree of co- hesiveness during drives and stampedes. If one animal broke through a line of men driving them, the rest of the herd would quickly follow. During direct movements or stampedes, the path chosen by the first member of each subgroup or by the first member of the herd was followed closely by every succeeding buffalo, even though men shouted from a short distance away. Interesting in this respect are Dodge’s (1877: 121-2) refer- ences to the derailing or the stopping of trains by the insistent passage of herds. Different degrees of cohesiveness were ob- served during movements of herds. Herds were most compactly bunched during direct move- ments and stampedes. Either of these movements was also characterized at times by a herd strung out for more than a mile in small groups or a column of several animals abreast. Direct move- ments were frequently single file. Single file was the usual pattern on trails through snow and through timber, and was also common on trails through broad meadows, yet it did not depend on the presence of a trail. The cohesiveness of a herd depended on a balance of centrifugal (dispersive) and centri- petal (gregarious) forces. Grazing was a cen- trifugal factor, since the distance between mem- bers of a herd characteristically increased when the herd arose to graze. Centripetal factors in- cluded the occurrence of localized shade or bare areas that brought herd members closer together when lying in them. Other factors were moder- ately dense or loosely spaced timber, and rain or falling snow. Herds were usually more closely grouped when loafing while standing or lying. Disturbance from foreign objects such as man caused a massing of the herd. Stampedes and some direct movements were also centripetal factors. IV. HERD COMPOSITION Bull Groups and Cow Groups I classified groups of buffalo into two types according to their composition: (1) Bull groups contained males with an infrequent female. (2) Cow groups contained a majority of females and a smaller number of males. During the rut, the increase in group size and the fusion of some bull groups into the cow groups did not create a third classification. Such groups were still classed as cow groups since their nature was essentially matriarchal. The bulls here were largely followers and had little effect on group coordination. Composition. Table 2 analyzes the size of bull groups. They were small groups of one to 12 members. The average size in Yellowstone Park increased slightly from 3.3 to 4.7 during the calving season (May) and decreased greatly to 1.2 during the rut when most bulls entered the cow groups. Most members of bull groups were four years old or older and the groups included some three-year-olds and occasional two-year- olds. Two- to four-year-old bulls were more com- monly distributed in the groups of three or more members than in smaller groups. Solitary bulls were sometimes several miles from any other buffalo. The only cows attached to bull groups were rare cases of one to four old, barren cows. Table 3 analyzes the size and composition of cow groups. Group size (average about 23) re- mained rather constant during the non-breeding season but varied more during the calving sea- son. There was a great increase in size during the rut as many groups in Hayden Valley co- alesced and were joined by bull groups. Cow groups during the non-breeding season were composed of cows, yearlings, calves, two-year- old bulls, some three-year-old bulls, and rare bulls four or more years old. Cow groups virtually never had an equal socionomic sex ratio (that ratio within groups— Carpenter, 1952: 236). The mean of males in these groups was 24% in the Lamar in January- 1958] McHugh: Social Behavior of the American Buffalo 15 Table 2. Size of Bull Groups S Lamar, Jan. -Mar. 33 18.2 30.3 18.2 6.1 18.2 3.0 0 6.1 3.3 Lamar, May 24 16.7 16.7 16.7 4.2 8.3 8.3 4.2 25.0 4.7 Wind Cave, May 21 28.6 14.3 9.5 19.0 14.3 4.8 9.5 0 2.4 Hayden, rut 42 81.0 19.0 0 0 0 0 0 0 1.2 March and 31% in May, 17% in Wind Cave in May, and 44% in Hayden Valley during the rut. All of these percentages were calculated using data for animals two or more years old. Much of the inequality in sex ratios is accounted for by bull groups living outside these cow groups. Numerous observations on cow and bull groups in several areas were examined to for- mulate the following trends; behavior in the two types of groups is contrasted in each category: Leadership. Older bulls tended to lead bull groups more often than younger ones, but there was great variation. Cow groups were led prin- cipally by older cows and seldom by younger cows or bulls. Leadership changed among these cows. Irritability. Cow groups were more wary, more watchful and showed a greater flight distance than bull groups. Flight distances in Yellowstone Park were usually under 300 feet for bull groups and 200 to 1,000 feet or more for cow groups. Some bull groups were so obstinate that ap- proach within ten feet and tossing of stones did not move them. Winter range. The winter range of bull groups from all four Yellowstone Park Herds extended into areas in which cows were not seen. This included the Mary Bay thermal area for the Pelican Herd, the area around Mud Volcano and the Yellowstone River for the Hayden Herd, meadows along the Madison River for the Fire- hole River Herd, and the Hellroaring Drainage, lower Lamar Valley and upper Lamar Valley above Soda Butte for the Lamar Herd. Daily round. Cow groups showed greater uni- formity of activity during any one time interval than bull groups; that is, almost all of the mem- bers were more or less simultaneously engaged in grazing, loafing or the activity of the moment. Intermingling. Cow and bull groups occa- Table 3. Size and Composition of Cow Groups Group size Group composition 00 Number of groups censused Mean grouping tendency Group and period observed Number of grouj observed Mean Standard deviation Range Mean size Cows two or more years old Yearlings Calves Bulls two-three years old Bulls four or more years old % bulls <4 years old Lamar, Jan. to Mar. 18 23.0 11.4 10-50 12 20.3 12.1* 4.4 0 3.8 0 0 Lamar May (calving) 15 23.6 18.4 4-63 10 17.3 8.6{ 4.8 3.9 3.7 0.2 1.0 Wind Cave (calving) 17 21.9 21.2 3-76 14 16.8 7.4 4.0 3.9 1.1 0.4 2.1 Hayden during rut 36 175.3 108 19-480 3 115.2 39.3 20.0 25.3 16.3 14.3 9.2{ *19.3% two-year-olds, f 19.5% two-year-olds. {Computed from census of 8 groups. 16 Zoologica: New York Zoological Society [43: 1 sionally intermingled during the non-breeding season for a few hours or less. Intermingling during the rut was, of course, frequent. Integration. At times during the entire sea- son, various groups joined together into larger units, and separated into smaller ones either by differential drifting or by direct withdrawal. This splitting and joining occurred in both cow and bull groups and even involved solitary bulls. There were indications that the stability of bull groups with one or two members was greater than in groups of larger bulls and cows. One pair of bulls remained together from December 10 to March 15 in the Lamar Valley. Cohesion. Bull groups (with the exception of some pairs) showed a low degree of cohesive- ness, and cow groups a high degree. The average distance separating members of bull groups was greater and more variable than that for cow groups. Cow groups moved as more compact units while bull groups sometimes strung apart for 100 to as much as 2,000 feet, the lead animal disappearing from sight before the last buffalo started to move. Closed groups. There were a few indications that some bull groups and possibly some cow groups were closed groups. Such observations were unusual compared with the frequent obser- vations of random coalescing and splitting of cow or bull groups. Among bull groups there were three cases where they approached closely yet did not join and two cases where newcomers were challenged. Included in these cases was a pair of two-member bull groups in the Lamar Valley which remained within a few hundred feet of each other for three months, occasionally grazed within sight of each other, yet did not join. Cow groups showed fewer indications of being closed, yet the following three incidents are of interest: A group of 33 cows moved single file through a scattered group of 48, and a group of 27 trotted through a group of 16, both without any loss or gain in members (Hayden Valley, March, 1953). One group of six cows in the Lamar Herd picked up three new members while wandering through a group of 14. Subgroups Wild groups of all sizes and types readily broke into smaller subgroups that were spatially distinct from other clusters in the group. Some subgroups contained a random cross-section of ages and sexes. Others, particularly in larger cow groups, showed tendencies toward bulls, barren cows, cows-with-calves, calves and year- lings, with some two-year-olds. The bull sub- groups tended to be on the edges of the herds, and the last animals in herd movements were regularly bulls. The presence of bulls on the flanks and in the rear of herds was noted by several authors (Allen, 1876: 55; Aubrey, 1908: 216; Audubon, 1849: 41; Bradbury, 1904: 147; Inman, 1899: 231;Parkman, 1903: 423). There was no evidence that these peripheral bulls were sentinels, as proposed by Inman (1899: 231). The more distinct subgroups in the Wildlife Park Herd were delineated by repeated observa- tions of such close groupings over a period of time. Maps of the patterns of lying in the herd aided this study. There were three main sub- groups: juveniles, mature cows and the older bulls. The subgroup of cows-with-calves was much more distinct than that of pre-partum ma- ture cows. The subgroup of older bulls (two animals) commonly stayed aloof from the cow group, yet joined them during the rut. When these two bulls were divided temporarily by a fence, the five-year-old repeatedly returned to the fence in search of the six-year-old. The at- tachment between two yearling heifers was the closest in the herd. Subgroups of new calves were first noted at an age of two to four weeks. Calves alternated between calf subgroups, their cows, or a cow-with-calf subgroup. There was no evi- dence of closed groups in the Wildlife Park Herd. Clans The preceding description of subgroups and closed groups contains no more conclusions than my observations permitted. Four authors went farther, however, and stated that subgroups were blood relations, family groups, or clans (Grin- nell, 1904: 129; Inman, 1899: 234;Seton, 1929: 693; Soper, 1941: 391-2), while Mayer (1934, II: 36) thought that small groups remained “separate and apart” in spite of coalescing and splitting. All of these authors failed to give con- clusive evidence. I have noted several cases where yearlings and two-year-olds, particularly heifers, followed cows, but such groupings could not be construed as clans. Garretson (1938: 59) did not believe the clan theory and Dodge (1877: 123) called the for- mation of subgroups “entirely accidental.” From the present study of group composition these last two authors would appear to be correct. I believe, as merely a personal conclusion, that subgroup or group formation was flexible and depended little on blood relations beyond the age of one year. 1958] McHugh: Social Behavior of the American Buffalo 17 V. THE DOMINANCE HIERARCHY Display of Dominance The herd in the Jackson Hole Wildlife Park was studied intensively for information on social organization. The outcome of interactions be- tween individuals delineated the social structure of this herd. It was a linear type of dominance hierarchy with dominant individuals exercising a virtual constancy of success in interactions. Sub- ordinates recognized dominants quickly and avoided them. I distinguished the herd members as individ- uals by physical differences or painted markings. Their interactions were divided into passive dominances and aggressions. Out of 1 ,027 inter- actions recorded, 72.8% were passive domi- nances and 27.2% were aggressions. Passive dominances were the more gentle of the two types of interactions. They were char- acterized by a lack of any obvious show of force or threat. Typically, the dominant individual walked toward and displaced the subordinate with no aggressive action. The most subtle pas- sive dominance took place when one animal avoided another while moving through the herd. Aggressions occurred when the dominant in- dividual displaced a subordinate by force or by threat. Threat usually involved an intention movement (Tinbergen, 1951: 79) of an aggres- sive act. The most gentle form of threat involved a mere look toward the subordinate, causing it to move away. The most common form was horn swinging, where the horns were swung up and down or sideways toward the subordinate with occasional contact. Partial or complete charges were also used to displace subordinates. The dominant animal appeared to be trying to impale or gore the subordinate, but actual contact was uncommon. Occasional violent charges threw the victim for several feet or tossed it as much as twenty feet. One mature bull hooked a cow on his horns and tossed her over his back, where she fell to the ground ( in a corral at the Lamar Buffalo Ranch). She was able to walk away, but subsequent autopsy revealed two broken ribs and a punctured, collapsed lung. Battles were tabulated as aggressions. They consisted of hooking of horns and pushing back and forth that lasted a few seconds or several minutes. The loser usually either moved away or was pushed back a few feet, yet many battles were indecisive. They were started by dominant or subordinate individuals and occurred between younger bulls and between calves, yearlings and two-year-olds. Most battles in the Wildlife Park occurred between one-year-old bulls and two- year-old cows. This may have been correlated with a later change in rank between these two groups. Battles between cows were rare. Another indication of dominance was the in- tention movement of mounting, seeming to occur without any sexual motivation. This was tallied as an aggression if the subordinate moved out when the dominant placed its chin on the rump of the subordinate. Both the intention movement for mounting and genuine mounting also oc- curred as play, apparently with no regard for dominance. Attempts by a subordinate to mount a dominant resulted in a prompt reversal of the same behavior, playful battles, or other forms of play. This mounting during play produced no withdrawal by the subordinate and was not tallied as an aggression. Situations in Which Dominance Was Displayed More than 90 per cent, of the interactions in the Wildlife Park were recorded during displace- ments on the feeding grounds. These buffalo were fed on hay from November to May, ap- proximately. Practically no natural food was available during most of this period. Since the hay was spread in piles over at least 100 feet of ground, there was a surplus of piles at all times except in the first few minutes of distribution. Most buffalo fed from separate piles of hay. They continually shifted every few minutes, even though there was little difference between vari- ous piles. A shift of one dominant animal might eventually cause a shift in more than half the herd as the move was passed down the line. Dominance was also recorded during inter- actions over special objects as follows: (1) an item of curiosity, (2) a spot for lying, (3) a wallow, (4) a water hole or puddle, (5) a tree for rubbing or horning, (6) choice grass, (7) a salt lick, (8) shade under a small group of trees (dominants clustered in the shade and subordi- nates were in the sun), (9) sniffing the vulva of a cow, (10) a cow in heat. Dominance was evident during group move- ments as well. Most commonly, a dominant ani- mal in the rear or middle of the group pushed subordinates in front or stopped those directly behind. When two dominants stopped on a packed snow trail, subordinate animals wishing to press ahead had to flounder around through deep snow (March, 1951, Wildlife Park). Domi- nance was also occasionally recorded during grazing. The preceding paragraphs discuss dominance that was usually traceable to interaction over a certain object, yet much aggressive dominance / 18 Zoologica: New York Zoological Society [43: 1 also occurred apparently independent of any in- citing situation. In these cases, one member of a herd suddenly and inexplicably interrupted its feeding to charge another member. This hap- pened most frequently between a cow and ju- venile, occasionally between two cows. Structure and Dynamics of the Dominance Hierarchy in the Wildlife Park Factors influencing dominance. Table 4 illus- trates some factors that determine dominance. Distinct differences in size and weight insured dominance. Seniority in age usually insured dominance, yet one seven-year-old cow was sub- ordinate to two five-year-olds. Within groups of similar size, bulls were dominant over cows; ag- gressiveness or certain undetermined factors were more important than slight differences in size. Reversals. Out of 726 interactions in the spring of 1951, there was only one temporary reversal when F4* displaced F3 by a swing of the horns. Occasional reversals were attempted by cows which horned either Ml or M2, yet these bulls did not yield. The two-year-old heifers did this most frequently. It was classed as play. Schein & Fohrman (1955) recorded 248 re- versals in 4,935 interactions among 163 dairy cows. Permanent changes in the dominance hier- archy resulted from the differential growth of bulls and cows. Previous to July and August, 1951, M3 and M4 were slightly smaller than F8, F9 and F10. During those two months, M3 *Nomenclature of individuals refers to sex and posi- tion in hierarchy. Thus, M2 was the second in position in the hierarchy of males and F4 was fourth among the females. Table 4. Dominance Hierarchy in the Wildlife Park, March 14 Through May 4, 1951 (Buffalo listed in order of decreasing dominance; M = male and F = female) Individual Total number of interactions % of total that were victorious % of total that were aggressions % of victorious interactions that were aggressions Age in spring’ 51 Individual differences within each group Ml 29 100% 28% 28% 6 Ml larger in size and weight than M2. M2 36 94% 25% 26% 5 Ml stronger than M2 in battles. FI 112 94% 22% 23% 5 All in this cow group smaller in size and weight F2 100 75% 10% 11% 5 than Ml and M2. F3 142 70% 8% 11% 7 No distinct difference in size and weight among F4 133 76% 23% 29% 5 these cows with the exception of F7, which was F5 112 38% 18% 7% 5 slightly thinner and shorter in height than all F6 126 54% 12% 9% 4 others. F7 84 40% 26% 44% 4 F8 84 37% 30% 45% 2 Smaller in size and weight than Fl-7. F9 77 40% 51% 68% 2 No noticeable differences in size or weight within F10 56 18% 32% 60% 2 this group. F8 stronger than F9-10 and F9 stronger than F10 in battles. All stronger than following subordi- nates in battle. (The progeny of Fl-7 born in Wildlife Park.) M3 98 32% 43% 68% 1 Smaller in size and weight than F8-10. M4 78 12% 44% 78% 1 No noticeable differences in size within this group. M3 stronger than M4 in battles. (The progeny of Fl-7 born in Wildlife Park.) Fll 65 5% 45% 67% 1 Smaller in size and weight than M3-4. FI 2 54 0 52% — 1 No noticeable difference in size and weight between these two, although FI 2 was at least two weeks older than Fll. Fll stronger than F12 in battles. (The progeny of Fl-7 born in Wildlife Park.) 1958] McHugh: Social Behavior of the American Buffalo 19 Table 5. Dominance Hierarchy in the Wildlife Park, February 4 to 9, 1952 FI dominates 13: F12 Fll F10 F9 F8 F7 F6 F5 F3 M4 M3 F4 F2 F2 dominates 10: F12 Fll F10 F9 F8 F7 F6 F5 F3 F4 F4 dominates 10: F12 Fll F10 F9 F8 F7 F6 F5 M4 M3 M3 dominates 10: F12 Fll F10 F9 F8 F7 F6 F3 M4 F2 M4 dominates 9: F12 Fll F10 F9 F8 F7 F6 F3 F2 F3 dominates 9: F12 Fll F10 F9 F8 F7 F6 F5 F4 F5 dominates 9: F12 Fll F10 F9 F8 F7 F6 M4 M3 F6 dominates 6: F12 Fll F10 F9 F8 F7 F7 dominates 5: F12 Fll F10 F9 F8 F8 dominates 4: F12 Fll F10 F9 F9 dominates 3: F12 Fll F10 F10 dominates 2: F12 Fll Fll dominates 1: F12 F12 dominates 0. and M4 grew to be approximately the same size as these two-year-old heifers. The exact time schedule for the advance in dominance of M3 and M4 is not available, since it occurred during the summer and fall when interactions were un- common. By September 12, both M3 and M4 had advanced above F8, F9 and F10. By October 9, both had moved above F6 and F7. By February 4, 1952, M3 and M4 had ad- vanced still farther up the hierarchy, as shown in Table 5. Both M3 and M4 were dominant over all cows except FI, F4 and F5. The irregu- lar gain in dominance of M3 and M4 produced eight triangles (Table 6). By the following sum- mer the advance of M3 and M4 over all cows eliminated these triangles. The dominance hierarchy was checked during November, 1953, to finish a survey of almost three years. During that period there was no permanent reversal among the two bulls or the top seven cows. There was a dominance hier- archy in the remaining five mature cows, all progeny of the first seven. None had moved ahead of their parents. Dominance hierarchy among calves. Domi- nance among calves developed slowly. Although they were alert to advances from adults within three weeks after birth, no dominance hierarchy among calves was detected during the first two months. The first definite signs of dominance were noted at an age of four months. At that time, the only male calf in a group of six was dominant over others and there were interactions among female calves. I was not able to determine the complete dominance hierarchy among calves until Febru- ary, 1952. Even during this period of feeding on hay, when there were numerous interactions among other adult herd members, interactions between calves were infrequent. Table 7 shows the hierarchy among calves. The male calf was dominant over all females and accounted for 41.4% of the interactions in the entire calf group. There was no correlation between the position of dominance of the calf and its seniority in the calf group, small yet noticeable size differences, or the position of dominance -of its mother. There was one reversal out of 70 calf interactions when D displaced B with an aggression. Table 6. Triangle Situations in the Dominance Hierarchy of Table 5. FZ /\ FF F3 /\ /y F3 / \ A? 3 FS FZ /W-e F4 FZ /\ W* FF F3 /\ FW-* F4 F3 /\ AV-6 FF 20 Zoologica: New York Zoological Society [43: 1 Table 7. Dominance Among Six Calves in the Wildlife Park Herd, February 4 to 9, 1952 (Computed from 70 interactions) A BCDEF Sixth (5/16) M Second B CDF Second (4/29) F Fifth C EF Fourth (5/5) F Third D CE First (4/21) F First E BF Fifth (5/15) F Fourth F D Third (5/1) F Sixth There were no triangles among the four calves in the herd one year previous on March 15, 1951. On this basis, one would expect the five triangles in the 1952 calf group to straighten out in a few weeks. Factors affecting frequency of interactions. The number of interactions among certain ani- mals or during certain time intervals varied with the factors discussed in the following paragraphs. Table 8 shows the tendency for dominant cows to have more interactions with the cows immediately below them than with the less domi- nant cows. “Probability” indicates the probabil- ity of getting such a skewed distribution if a column of values was picked at random. It was computed by chi-square. The distribution of interactions for FI and F2 are statistically sig- nificant. While the remaining columns are not, they still show the same distributional trend. Schein & Fohrman (1955) statistically analyzed similar interactions in a herd of dairy cows. They found that “fight contests involved cows closer together on the social rank scale than do threat or butt contests.” The total number of interactions and the per- centage of aggressive interactions for each indi- vidual are listed in Table 4. Bulls Ml and M2 had fewer interactions than any other herd mem- ber. With this exception, the more dominant ani- mals tended to have more interactions. This is in line with findings on pigeons (Masure & Allee, 1934: 327), chickens (Allee, 1951: 135) and canaries (Shoemaker, 1939: 404). An analysis of the percentage of aggressions shows just the reverse tendency. The percentages increased dis- tinctly among the less dominant juveniles. Hunger and palatability of food caused varia- tions in the frequency of interactions. Table 9 shows the pattern for a typical day of feeding on hay. There was a high peak of interactions and aggresssions during the first 20 minutes of feed- ing. Fewer interactions and no high peak occurred during the afternoon feeding period, when the herd was not as hungry. This positive correlation between hunger and number of in- teractions was further verified on days when a surplus of hay remained from the previous day’s feeding. When fresh hay was put out under such conditions, there were fewer interactions than usual. In addition to hay, the herd was also fed concentrate on some days. The buffalo preferred the concentrate, as evidenced by their quick withdrawal from the hay to feed on it. Even though the concentrate was fed about one hour after the hay, the initial peak of inter- actions sometimes doubled the initial peak for hay on the same day. The interactions over con- centrate were also more aggressive. Various disturbances also increased the num- ber of interactions in a herd. They increased greatly in any group enclosed in a corral or pen. Interactions were increased by disturbance from the presence of a strange object, such as a car, a recently shed elk antler, a human being or a person concealed under a buffalo hide. One cow in the Wildlife Park Herd increased the fre- Table 8. Interactions Between Mature Cows in Wildlife Park, March 14 to May 4, 1951 Cow FI F2 F3 F4 F5 F2 20 + F3 25 + 22 + F4 9- 5_ 17 + F5 9- 14 + 21 + 17 + F6 13 - 3- 15 + 12 + 9 + F7 5- 6- 6- 8- 6- Mean 13.5 10.0 14.8 12.3 7.5 Probability .005 * .005 * .04 .18 .62 + = Above the mean for column. — = Below the mean for column. 1958] McHugh: Social Behavior of the American Buffalo 21 Table 9. Interactions per Ten-minute Interval while Wildlife Park Herd Was Feeding on Hay (Feeding Time: 10 A.M.) quency of interactions by her numerous ag- gressions which resulted in a chain reaction of interactions. The herd was noticeably more calm after the cow was removed. Quality of dominance. Rough or gentle dis- plays of dominance varied with age and sex. Comparatively gentle displacements were made by bulls over cows, by a mother over her calf, among juveniles, among calves, and occasionally among bulls outside the rut. The dominant buf- falo used moderate force, and the subordinate moved away slowly for no more than a few feet. The subordinate sometimes returned to feed with the dominant. More forceful displacements were made by cows over yearlings, by a cow to a strange calf, among cows, and among bulls dur- ing the rut. The dominant was more aggressive and the subordinate moved faster, farther, and did not return. There were no obvious behavior characteris- tics correlated with rank in the hierarchy. Subor- dinates showed no unusual amount of fear. All buffalo showed a keen awareness of the identity of the animals near them. The approach of a dominant from almost directly behind the sub- ordinate resulted in a passive submission with hardly a glance at the dominant. Derived dominance. Some buffalo derived a higher status in the dominance hierarchy by as- sociating with a more dominant buffalo. Calves thus benefited from the positions of their mothers. The calf was elevated to her status when close to her and if tolerated by her. The mother sometimes pushed close subordinates away from her calf. Most of these instances of derived dominance occurred during hay feeding. Derived dominance was also recorded once dur- ing the rut in Hayden Valley: A seven-year-old bull stopped chasing a cow when she started to follow an older bull closely. Dominance in Wild Herds The extent of development of dominance hierarchies in wild herds could not be determined 22 Zoologica: New York Zoological Society [43: 1 because enough individuals could not be recog- nized. Animals in wild groups frequently ex- hibited passive dominances and aggressions. The frequency of these interactions never exceeded half those recorded for the Wildlife Park Herd during hay feeding. It was usually considerably less. The highest frequency occurred in a cow group of 15 that was feeding in deep snow (Lamar Valley, January 24, 1952). Displace- ments were frequent as each buffalo moved into an area recently cleared of snow by a sub- ordinate. The more dominant buffalo would be expected to have a particular advantage in a severe winter. A dominance hierarchy was noted in the above cow group and in a bull group of eight (Madison River, March, 1953) and of five (Hayden Valley, March, 1953). A wild group of 55 that was fed on hay in a Yellowstone Park corral showed numerous passive interactions. But the fact that groups of two to four cows were regularly observed feeding from the same hay pile would seem to indicate a lack of a completely developed hierarchy. Interspecific Dominance Free-living buffalo sometimes interacted with other species on the same range. These species included elk, antelope, mule deer, bighorn sheep and coyotes. In the Wildlife Park, buffalo fed during the winter in the area with a herd of 29 elk and some pronghorn, mule deer, whitetail deer, moose and horses. In free-living or en- closed herds the number of interspecific interac- tions greatly exceeded the number of intraspe- cific ones. The buffalo were dominant over all these species with only occasional reversals among elk and horses. Elk. In the Wildlife Park, month-old buffalo calves were always successful in observed at- tempts to displace six-point elk bulls. Interac- tions involving younger calves were not ob- served. The buffalo herd completely dominated the hay feeding circuit and could push the elk from any section. They occasionally stampeded back and forth on the hay feeding circuit in order to force the elk into deeper snow. They also chased the elk in a steady, rapid walk during grazing in summer. Buffalo and elk sometimes fed within three feet of each other, but the elk soon became wary of the buffalo and moved farther away. Although elk nervously watched buffalo and usually dodged all charges, the violent charge of a cow buffalo threw one yearling elk three feet into the side of the hay shed. Five- or six-point elk bulls were dominant over a buffalo cow or yearling in rare cases. The displacement always occurred on the edge of the buffalo herd. Such reversals were considered temporary since the aggression of any buffalo never failed to move the largest elk. The buffalo herd chased the elk herd on three occasions when a new elk calf was present. The speed of the chase was so fast that the elk calf dropped behind within a few hundred feet. The buffalo then milled about the calf. In the first incident the calf was bruised, cut behind the ear, and bleeding at the mouth, yet still able to walk back to its herd. The buffalo herd bruised the calf again on the next day. It had no broken bones yet was so badly mauled it could not stand. It died in spite of careful nursing by the attendants of the Park. A second calf was res- cued just before the buffalo herd reached it. They milled about and smelled the spot where the calf was born. When this calf moved out with the elk herd a few days later, the buffalo followed and slowly chased the elk. Groups of elk and buffalo in Yellowstone Park moved within 30 feet of each other and even intermingled. The buffalo occasionally charged the elk, but the wary elk withdrew quickly. Buf- falo disrupted the elk trapping program since elk would not enter a trapping corral if buffalo were inside or nearby. One incident was observed in the Lamar Valley where the buffalo killed an elk calf they found in hiding. Rush (1942: 225) and Chapman (1937: 148) mention deaths of elk in Yellowstone Park due to buffalo. Superintendent John Schwartz noted a few cases of dominance of elk over buffalo in the National Bison Range. One elk raised on a bottle showed an attachment for two buffalo bulls at an age of eighteen months despite the presence of other elk. When two and a half years old, he rounded up a group of 15 buffalo cows and bugled. Another bull elk assembled a harem of cow elk during the rut and belligerently chased any buffalo that wandered near. This same elk rammed a tine of his antler through the palate of a yearling buffalo. The buffalo bled profusely and died from this injury within an hour. Pronghorn. Buffalo in the Wildlife Park occa- sionally charged pronghorns (antelope) which wandered near the herd. They killed one eight- month-old pronghorn buck. Pronghorns in Wind Cave passed near or through buffalo groups, and the latter occasionally charged them, but the alert and agile pronghorns easily out-maneuvered these attacks. A two-month-old buffalo calf chased one pronghorn buck for almost a hundred feet. The walking of a buffalo group within 150 feet of resting pronghorns caused the pronghorns 1958] McHugh: Social Behavior of the American Buffalo 23 to rise and watch the group. Bryant (1885: 132) observed in Wyoming that “The deer and ante- lope are compelled to frequently shelter them- selves from the attack of wolves under the strong protection of buffalos and you sometimes see herds of buffalos and antelopes mingle in grazing together.” Mule deer. One charge of a buffalo cow in the Wildlife Park tossed a mule deer ten feet. The attack left the deer lying on the ground, but it arose and rapidly moved away when the cow charged a second time. Another charge by a cow knocked a mule deer to the ground. Moose. When a moose calf was placed in the Wildlife Park in August, the buffalo sometimes chased it. The buffalo killed the calf when it was seven months old. One rib was broken and the chest was pierced with a hole two inches in diameter (oral comment from James R. Simon). Horses. Two draft horses were dominant over all cow and yearling buffalo during interactions at a salt lick in the Wildlife Park in October, 1951. The horse made no passes toward the buffalo, which moved slowly away as the horses came in to the salt. In March, 1953, the draft horses had lost some rank with the buffalo and were dominant over only a minority of cows and all yearlings. All buffalo were dominant over one saddle horse. Head Animal Keeper David Pierson recorded one death of a horse in the Lamar Valley in 23 years of observation. This horse was ranging free in the valley at the time. He also noted three other horses and one mule that were gored by buffalo. Bighorn sheep. Cy Young observed three cases in the National Bison Range where an older bighorn ram associated with a buffalo bull dur- ing the summer. Dominance hierarchies. The interspecific dominance hierarchy among the big game spe- cies in the Wildlife Park was determined for the period February-May, 1951. With the exception of reversals noted below, the order was as fol- lows: adult human beings, buffalo, elk, mule deer, pronghorn, moose and whitetail deer. The exact order of the last two species was not known. The man who fed these animals maintained his dominant position by the use of a pitchfork. Tame elk, mule deer and pronghorns enjoyed derived dominance over the buffalo when they followed this man closely. The elk exercised dominance by charges, by swinging antlers or by rapid downward strikes of the forelegs. The mule deer involved in this hierarchy in- cluded two tame bucks and a tame and a wild doe. They used short charges or rapid downward strikes with the forelegs to display dominance. Bucks were dominant over does. The mule deer bucks dropped below the pronghorn buck in dominance when they shed their antlers. The horns of the pronghorn were becoming larger and harder at the same time. The reversal was not absolute; both deer occasionally resisted the horning of the pronghorn buck and drove him away. The pronghorn buck later gained domi- nance over the tame mule deer doe. All pronghorns were tame. Pronghorn bucks dominated pronghorn does and other species by horning them. One buck was dominant over a castrated buck of the same age. VI. THE RUT AND SEXUAL BEHAVIOR Development of the Rut A marked increase in or the onset of certain behavior patterns in bulls marked the beginning of the rut or rutting season. These activities in- cluded sniffing of vulvas, tending of cows, bellowing, wallowing, horning, vicious and non- vicious battles and incomplete and fertile mount- ings. Section IV discussed the changes in group composition during the rut— the great increase in average group size, the union of groups in Hayden Valley and the almost complete dis- solution of the bull groups as they mingled with the cow groups. This pattern of aggregation dur- ing the rut was not as well marked in Wind Cave or the Wichita Refuge. Groups here were larger than at other seasons but combined into one gigantic group only rarely. The fusion of smaller groups on the Great Plains and elsewhere into larger masses during the rut was mentioned by several authors (Bradbury, 1904; Branch, 1929: 7 ; Catlin, 1876-1: 249,11: 13;Hornaday, 1889: 415-6; Long, 1905-XV: 246; Soper, 1941: 390). The rutting season in Hayden Valley ran from about June 15 to September 30, with less activity during the first and last two weeks. The rut in Wind Cave ran from about June 23 to September 14. The sporadic birth of calves outside the main calving season (see Section VII) would depend on occasional rutting activity at any time of the year. There was more general activity in the herd during the rut than during other seasons. This was most obvious among bulls tending cows but was also typical to a lesser degree of other bulls and the rest of the herd. Peaks in rutting activity occurred just after dawn and at dusk. Any dis- turbance that altered the organization of the herd 24 Zoologica: New York Zoological Society [43: 1 also increased the rutting activity. This included traveling and the approach of a car into the center of a herd. During periods when the herd lay down, bulls tending cows often remained standing. Bulls bellowed at times during the entire night. Sniffing and Extending Neck with Upcurled Lip Sniffing of cows’ vulvas by bulls was observed during all months but increased greatly during the rut. Extending of neck with upcurled lip commonly but not always followed the sniffing (Plate III, Figure 5). There were indications that this was omitted after sniffing cows which showed no sign of heat. Some bulls remained with and tended one cow after sniffing several. In other cases, a bull displaced a subordinate which was tending a cow, sniffed the cow, and then moved on through the group to inspect others. Bulls made methodical checks by sniffing all cows of a group in their paths. New bulls sniffed cows after first entering the group. Lying cows were prodded into standing so that the bull could sniff them. Urination by the cow was a stimulus for sniffing by nearby bulls, and sometimes the extended neck and upcurled lip of one bull after sniffing was a stimulus for another close one to move in and sniff. Interactions resulted when as many as three bulls sniffed the same cow con- secutively. The subordinate bull or bulls moved back only a few feet and still kept the neck extended and lip upcurled. In sniffing, the bull often moistened his nose from the liquids exuding from the vulva or licked this organ. If the cow was urinating, he wet his nose or licked his tongue in the stream of urine and then sometimes continued to lick the vulva after the stream stopped. When sepa- rated from cows, some bulls even wet their noses with or sniffed cow’s urine on grass and then extended their necks with upcurled lips. This sniffing activity was not restricted to bulls or the stimuli of vulvas. Cows extended their necks with upcurled lips at times after sniffing or licking the vulva or stream of urine from another cow, a rotted skeleton, the bloody urine on snow from a cow about to calve, human urine on a tree, a new calf— either her own or a strange one— the vulva and stream of urine from her own calf, her own urine on grass or the torn scrotum of a bull. Both male and female calves (45 days or older) and yearlings also extended their necks with upcurled lips after sniffing urine or the vulva of another buffalo. A two-year-old bull extended his neck with upcurled lips after sniff- ing a new calf. The Tending Bond Technique. The bond between the bull and cow during the rut is here called the tending bond. The bull kept alongside or behind a par- ticular cow at a close distance— usually from one to five feet but sometimes touching the cow. When more than one tending pair was operating in a group, each pair was generally on the edge of the group with the bull maneuvering so as to keep the cow peripheral. The tending pair also occurred in the center. Only two instances were observed in which the tending pair completely separated from the group. Bulls took either a passive or an aggressive part in the tending bond. Some followed the cows closely, turning around or moving forward with each move of the cow. Others guided the movements of the cows at times by swinging their heads or by moving back and forth in front of the cows. The only two observed cases of extreme control by the tending bull involved young bulls three or four years old. They isolated or completely restrained selected cows by run- ning alongside or ahead and cutting back in front of them whenever they started to run back toward the herd. In spite of the efforts of the bull, the tending bond was still essentially matriarchal. The cows usually moved anywhere by momentarily dodg- ing from under the pressure of the bulls. For example, during stampedes the cows moved off and the tending bulls followed close alongside. Of regular occurrence were cases where one cow dashed several hundred feet from the group, followed by at least one bull. As many as eight bulls chased after a cow in this fashion. One chase in the National Bison Range continued intermittently for three hours, at the end of which time the cow appeared completely ex- hausted, with tongue hanging out. Other cows regularly resisted the advances of the bulls by threatening with motions of their horns, by horning the bulls or by kicking out with one hind leg. The bulls either ignored these attacks or moved one or two feet farther away. Exceptions to the customary tending bond of one bull and one cow were rare. Cows were sometimes shared consecutively yet exclusively by more than one bull as a more dominant bull displaced the current bull by threat or battle. This tending bond lasted for no more than a few days— usually much less. Among buffalo it can be said that there is, therefore, a temporary monogamous mateship. Some bulls were promis- cuous because they successively served several cows in any one season. While tending cows, many bulls showed in- 1958] McHugh: Social Behavior of the American Buffalo 25 tolerance toward surrounding group members of either sex which ventured close. Most intoler- ance was displayed toward nearby bulls, which were repelled by threat, aggressive charges or battle. The tending bull also kept away even the calf or yearling of the cow he was tending. Other behavior accessory to the tending bond is dis- cussed later (attempted mountings, bellowing, wallowing, horning and sniffing the vulva of the tended cow or other nearby cows). Age of tending bulls. Almost all of the tending in Hayden Valley was done by bulls ranging in age from six to at least fourteen years, with bulls eight or more years old being most active. Bulls one to five years old usually were not successful at tending due to pressure from other bulls. Two- year-old bulls often failed because many were subordinate to cows, and this was even more true of one-year-olds. Yet tending bonds among younger buffalo still occurred. There were various combinations among yearling or two-year-old bulls and heifers and those between three-year-old bulls and fe- males two or more years old. Such yearling heifers probably were not in heat. It was also doubtful whether the other females were in heat, since older bulls paid no attention to them. These tending bonds among younger animals were comparatively short-lived and were often char- acterized by more resistance than usual on the part of the female. In the Wichita Refuge, most tending was done by bulls six to eleven years of age with some done by five-year-olds. (There was only one bull older than eleven years in this refuge). Tending in the National Bison Range was done mostly by bulls four or more years old, slightly by three-year-olds, and rarely by two- year-olds. (Observations of John Schwartz and Hugh Wilmar). Yearlings tended momentarily when they got a chance, as older bulls left tem- porarily or were absent. The composition of this herd was unbalanced since there were very few bulls five or more years old. Interesting in this area was the pattern of tending in the exhibition herd of about seven cows and two bulls. In 1952, a bull 13 or 14 years old dominated the nineteen- year-old bull and did all the tending. In 1953, the twenty-year-old white bull dominated a five- year-old bull and did all the tending. The despot six-year-old bull in the Wildlife Park tended any cow that was in heat. The five- year-old bull was able to tend or mate only when more than one cow was in heat. The preceding notes on ages of tending bulls are interesting when compared with data on sexual maturity and physical growth. At least one bull was sexually mature at an age of 14 months, since a domestic cow bred by this bull aborted midway during pregnancy. The testis of a twenty-one-month-old bull killed in January in Yellowstone Park showed no spermatogenesis, but this does not necessarily mean that the bull could not breed with the advent of the rut. The testis from an eighteen-month-old bull killed in the Crow Reservation in September also showed no spermatogenesis. A three-year-old and pos- sibly a two-year-old sired eight calves conceived in the Wildlife Park in 1948. Both two-year-olds and three-year-olds have successfully bred cows in or near the National Bison Range (oral com- ment from George Mushbach, John Schwartz and Cy Young). Although bulls attained near maximum growth in size by the fifth or sixth year, small yearly increments in growth continued for a few years afterward. Atypical tending bonds. Loose tending bonds were discussed previously in connection with tending among younger buffalo. Other loose bonds occurred at all times during the rut and included all ages of bulls. Loose bonds were marked by close tending that lasted for only a few minutes, by intermittent tending where the bull tended for a few minutes and then moved off to graze nearby for a while, and by lax tending where the bull did not follow closely nor try to control the movements of the cow. Many of these loose bonds involved cows which were not in heat, as determined by the non- swollen condition of the vulva. Three unnatural tending bonds were observed : On September 20 in Hayden Valley a three-year- old bull closely tended a five-month-old calf for at least 25 minutes. On October 12 near Old Faithful a two-year-old bull closely tended a five-month-old calf for at least 15 minutes. In both of these cases the bulls attempted many mountings. On July 16 a two-year-old bull closely tended a yearling bull for at least four hours in the Wichita Refuge and attempted mounting with penis unsheathed. This was not classed as play, due to the duration and intensity of the tending. Heat in cows. Heat in cows was determined by the condition of the vulva and the tail. The swelling of the vulva in cattle is slight during proestrum, becomes more pronounced during heat, and subsides rapidly after heat has passed (Asdell, 1946: 346; Dukes, 1937: 647-651). The vulvas of buffalo cows swelled into a globular mass that protruded an inch to an inch and a half. The switching of the tail sometimes brushed the lips to expose the red mucosa, 26 Zoologica: New York Zoological Society [43: 1 although during heat it was held slightly out from the vulva and seldom switched. Cows with swollen vulvas were vigorously tended by bulls and those with little or no swelling were usually tended loosely or not at all. A further symptom of heat in cattle is a ten- dency to mount other cows (Dukes, 1937; Schein & Fohrman, 1955). One two-year-old buffalo heifer in heat in the National Bison Range disrupted the mounting attempts of her tending bull by almost continuously attempting to mount him. She kept this up for at least 30 minutes. In two other cases, a buffalo cow in heat mounted a nearby cow, and another in heat mounted her tending bull. Bellowing and Snorting Nature of bellowing. Bellowing, produced only by bulls, was an extreme variation of the grunt. It sounded like a guttural, growling roar and was unlike bellowing of a domestic bull. The simultaneous bellowing of more than three or four buffalo resolved into a continuous roar. The bellowing bull opened his mouth, stuck out his tongue for a few inches, and contracted his abdominal muscles so that his belly rose slightly. The sound of bellowing was audible for at least three miles in still air and for lesser dis- tances under windy conditions. Audubon & Bachman (1849: 38) reported this distance as at least ten miles and Hornaday (1889: 416), five miles. I located some groups by the sound of bellowing rather than by sight and assumed that straggling bulls or herds did likewise. Bellows varied in length from a third of a second to more than five seconds (single breath) , in intensity from a very soft purring noise to a loud roar, and in quality. Some of this variation was correlated with the emotional state of the moment. For example, a rapid interchange of loud bellows frequently preceded vicious battles. Exchange bellows were longer and louder be- tween bulls or from a tending bull than those from lone bulls, grazing bulls or those which did not elicit competitive bellowing from others. Bulls often continued to bellow even while engaged in activities such as running or trotting in a stampede, ruminating, grazing and wallow- ing. i Circumstances of bellowing. Bellowing was frequent during the rut and sporadic or rare at all other times, with a gradation of activity be- tween those two periods. Bulls bellowed under the following circumstances: (1) while tending a cow— loose tending bonds were sometimes characterized by weak bellowing or a lack of it; (2) while approaching another bull previous to meeting or battle; (3) while moving through a bull subgroup or the herd, accompanied at times by sniffing of vulvas; (4) while following the trail of the herd or moving toward it; (5) in answer to the bellow of another bull, resulting commonly in a constant interchange of bellow- ing (one tending bull answered distant rumbles of thunder, and another bull answered my crude imitation of a bellow); (7) while another bull was approaching. Some bellowing occurred dur- ing other situations, such as standing or lying, that did not fit the above categories. Bellowing subsided partially or completely during periods of lying in the group. Bellowing was induced among bulls in the Wichita Refuge and once in the Wildlife Park by the observer’s approach in an automobile to 25 to 35 feet of a bull tending a cow. The bull looked directly at the car and increased the length and loudness of his bellows for the next 30 to 45 seconds. A few bulls which were not tending cows responded similarly. Analysis of this incident seems to indicate that the bellow was threat behavior. This was cor- roborated by instances in which one or more bulls stood and listened to the increase of bellow- ing of another bull and then retreated from him. Bellows were never heard from two-year-old bulls and were uncommon from three-year-olds; those from four-year-old bulls were shorter than those from older bulls. Older bulls had lower pitched bellows than younger bulls. Bellowing was comparatively rare among the five- and six- year-old bulls in the Wildlife Park. This may have been correlated with the fact that domin- ance was well established between the two, and the younger seldom challenged the older during the rut. Snorting. Snorting was produced by the slight, quick contraction of the abdominal mus- cles to force blasts of air through the nostrils. This sound carried at least 4,000 feet in still air but was difficult to hear under windy conditions unless at close range. Snorting sometimes alter- nated with, followed or preceded bellowing. It was heard infrequently, and was given most commonly by one bull approaching another pre- vious to meeting but also occurred under the same circumstances listed previously for bel- lowing. It was quite typical from a lone bull heading toward and entering a cow group. Tend- ing bulls in such a group usually looked toward the snorting bull. This bull often took over an- other cow by threat alone. One old bull in Flay- den Valley shortened the interval between snorts from three seconds to one second as he moved 1958] McHugh: Social Behavior of the American Buffalo 27 closer to another bull during the rut. Neither was tending a cow at the time. Wallowing and Horning Nature of wallowing. The great increase in wallowing by bulls during the rut was attributed more to conflict occurring as a result of the rut than to irritation from insects. Wallowing most typically consisted of one to three acts: (1) sniffing of the ground; (2) vigorous pawing with the forehooves; (3) rolling. Rolling and particularly pawing also occurred independently. The bull customarily moved a short distance backward with each pawing. He often un- sheathed his penis and urinated a thin stream during the pawing and during and after the rolling. Such urination was quite typical of any hostile situation during the rut. The bull also dug his horns into the ground at times and rubbed his head in the earth after pawing. Cows showed no noticeable increase in wal- lowing during the rut. They were observed to urinate pulsatingly and weakly on only three occasions previous to wallowing. In one instance, the act appeared to be a direct reaction to my crawling into the herd under a buffalo hide. Circumstances of wallowing. Bulls wallowed under the following circumstances: (1) as two or more bulls approached each other previous to meeting or battle— lone bulls which wandered within sight of each other often wallowed imme- diately and simultaneously after catching sight of each other; (2) while tending a cow; (3) while following the trail of a herd, moving toward and entering a herd, or walking through the herd; (4) during “mock battles” (discussed later); (5) previous to a possible meeting with human beings (three occasions) ; (6) in a wallow in which a cow had been lying, had urinated, or had wallowed a few minutes previously (several bulls sometimes wallowed successively in such a spot); (7) just previous to getting up after lying. These were the usual occasions; a few others were rare. Bulls used established wallows most of the time but tore up the sod to form new or tem- porary wallows when a regular wallow was not immediately available. For example, a tending bull wallowed close to his cow, and one bull ap- proaching another wallowed wherever he was. Bulls pawed vigorously without rolling during the first six circumstances listed above. Pawing was also done by bulls or cows during the fol- lowing circumstances: (1) by either opponent during a pause in battle; (2) by a mature cow enclosed in a corral when I motioned toward her from a distance of ten feet; (3) by a cow with a one-hour-old calf when I started to ap- proach her; (4) by a lone cow three hours and five minutes previous to the birth of her calf; (5) by a cow with a two-hour-old calf as a larger cow moved in and licked the calf; (6) by a four- year-old bull as I imitated a bellow from under a buffalo hide; (7) by bulls previous to charging as a result of the approach of human beings (Allen, 1876: 64; Chapman, 1937: 145; Dodge, 1877; Garretson, 1938: 51; Inman, 1899: 249). A survey of these occasions for wallowing or pawing shows that most of this behavior oc- curred previous to, during or in anticipation of hostile situations. Such situations presumably involved the simultaneous activation of the drives of attack and escape. The resulting be- havior may thus be considered “displacement activities” (Tinbergen, 1952: 24-6), which here will be called “displacement wallowing” and “dis- placement pawing.” Both were a part of groom- ing behavior and thus were “irrelevant” during hostile situations. Displacement wallowing differed at times from its genuine “example” of grooming behavior (Section II). It was incomplete since it consisted of pawing alone or was done in areas other than regular buffalo wallows, such as on unbroken, grassy meadows. It was also more vigorous, par- ticularly the pawing. As such it may have been recognizable as threat by other buffalo. Horning. Horning of lodgepole pine trees by bulls increased during the rut. The bulls de- barked the trees (Plate III, Figure 6), gouged their horns into them to make gashes as deep as three-fourths of an inch, uprooted smaller trees and thrashed their horns in the branches. Some also butted against or rubbed on the trees. Battles Preliminaries. Preliminary to battle, a bull usually indulged in bellowing, snorting, wallow- ing (pawing and/or rolling) and the hesitant gait. Bulls approaching each other at close range used this gait, a modification of walking where but one foot was slowly advanced at a time. All of these preliminaries functioned at times as threat, as evidenced by the fact that one bull yielded to another without battle. If these dis- plays of threat were given by a bull tending a cow, they functioned to warn invading bulls, which in turn were announced by giving such displays. Bulls also threatened others away by using aggressions such as short charges or shak- ing of the head. Since the response to aggression was often active retreat or passive avoidance, vicious battles were uncommon. Vicious battles. Vicious battles differed from 28 Zoologica : New York Zoological Society [43: 1 most fighting in their brisk and violent move- ments, so much so that chunks of sod flew into the air or clouds of dust cloaked the action. These battles started as one bull slowly ap- proached another, as one shook his head toward a close opponent, as one charged toward another and attempted to horn his flanks or crash head- on, or as both bulls were slowly butting and pushing back and forth. Vicious battles consisted of quick horn jabs with thrusts to the side or upward after contact, forceful head-on butting, violent charging back and forth with heads to- gether, and swift routs where one bull was driven backward for several feet (Plate III, Figure 7). There were two instances in which a bull mo- mentarily fell under the powerful charge of the victor. Battles ended as both bulls stopped and remained close together, as both bulls separated, or as the loser backed away from the constant jabs of the victor. Contestants were usually at least four years old. Battles commonly lasted from one to 35 seconds with most ranging from five to ten seconds. A few lasted several minutes with repetition of fighting at intervals., Most battles occurred during daylight hours, yet some took place at night, as shown by the fact that George Mushbach noted new injuries on certain bulls early in the morning. Catlin (1876—11: 13) observed “desperate battle” while a herd was swimming a river. The shock of these encounters was partially absorbed by the cushion of long hair on the crown. These hairs measured eight to ten inches long on bulls three to eight years old, while Hornaday (1889: 404, 414) measured some up to 22.5 inches. A .30/06 bullet fired into the head of a bull in the Crow Reservation from a distance of 30 feet bounced back from the matted hair and was recovered as a twisted piece of lead. Casualties. Several casualties of battle were recorded. A fourteen-year-old bull was dis- covered within two hours after death that prob- ably resulted from battle, since he was found in the middle of a small group that was battling viciously and frequently. The only apparent in- juries were a bloody wound just below the eye that did not penetrate the skull and a hernia just to one side of the penis. This hernia was seven inches in diameter, protruded 2.5 inches, and had a bloody cut from a horn jab. Three other bulls found dead in Hayden Val- ley may have been killed in battle. Rangers Kittams and Coleman discovered two dead ma- ture bulls lying about 20 feet apart in early August, 1939. I found a dead bull at least 16 years old that was killed or died about August 23, 1951. George Mushbach found three successive casualties of rutting battles in the National Bison Range. All were lying when discovered, yet could rise and walk short distances with great diffi- culty. The viscera of one were hanging out. All died within a few hours or were so seriously in- jured that they were shot. In the same range, John Schwartz found a dead bull during the rut. The bull probably died after battle, since he was badly wounded, with the hide torn open in sev- eral places and a puncture in one rear leg. David Pierson saw one bull with a crippled rear leg being killed during a rutting battle in August in the Lamar. The victorious bull lunged for 40 feet with the crippled bull on his horns. He also reported that during the hay-feeding period in the Lamar, two mature bulls first suc- cessively and then collectively battled and killed another old bull. Two other deaths in the Lamar were recorded in winter by Baggley (1933). Both of the bulls, “rather old and not in good shape,” came to the feed grounds after being absent for a long time. Both were killed by a mass attack of younger bulls (“seven or eight” in one case). Baggley further states that “such occurrences are not uncommon in the buffalo herd even on the summer range.” A mature cow battled a two-and-one-half- year-old bull in January in the Wildlife Park. The bull was not cornered, yet chose to remain and fight back in several battles. He died later from numerous bruises and a punctured lung resulting from a horn jab. Seton (1929: 689) reports one bull that killed another in a battle in an enclosure in the Phila- delphia Zoo in July. Fremont (1845: 26) saved one old bull from certain death in July by dis- persing 18 or 20 bulls which were attacking him and which had already knocked him down sev- eral times. Hornaday (1922: 296-7) observed one bull kill another in an enclosure in a zoo by puncturing the lungs with horn jabs. He also reports that fatal fights sometimes occurred in the large 27,000-acre reserve of the Corbin Blue Mountain Forest Association. Plate 105 from Catlin (1876, I) depicts a rutting scene with three battles in progress and one dead bull. Sev- eral authors believed, however, that death from battle in the wild was non-existent or rare (Allen, 1876: 46; Audubon & Bachman, 1849: 38; Bradbury, 1904; Garretson, 1938: 36; Goodwin, 1939: 366; Hornaday, 1922: 282-3; Inman, 1899: 12; Soper, 1941: 390). Injuries yield further evidence concerning the violence of vicious battles. Such injuries were noted in bulls butchered during the annual herd reduction in the Lamar. I counted from one to 1958] McHugh: Social Behavior of the American Buffalo 29 three healed fractures of the ribs in 23 per cent, of the bulls over four years of age. Other ob- servers noted a similar proportion of healed frac- tures in bulls and occasionally cows (David Pierson, John Schwartz, Cy Young). Examina- tion during this butchering also revealed that most older bulls had scars on their sides from horn jabs. A few had abscessed sores resulting from horn jabs. Injuries were also observed in the wild. David Condon discovered one injury probably from battle in the Nez Perce area (Yellowstone Park) on August 22, 1953. A lone, large, mature bull had one horn broken off next to the skull with much blood in this region. Cy Young observed a similar injury during the rut in the National Bison Range. In the same area during the rut, George Mushbach saw occasional lame bulls. They walked with difficulty and had fresh bloody scars. Occurrence. Vicious battles took place under a variety of circumstances. Some involved a bull tending a cow in heat. The victor moved to tend the cow but was occasionally displaced by still a third bull as a result of battle or threat. Other battles occurred without the presence of a cow in heat. “Battle subgroups” formed infrequently. Such a subgroup of several bulls four or more years of age was characterized by numerous battles. One battle subgroup collected immedi- ately after two bulls battled viciously 100 feet from the edge of the herd. Vicious battles were limited mostly to the rutting period, while less violent, non-vicious battles occurred during all seasons among bulls. These latter battles increased in frequency during the rut and were prevalent among bulls younger than four years. There was no obvious distinc- tion between these battles and those discussed previously as play (Section II). Mock battles. These were so called because they contained many of the preliminaries to genuine battles. Mock battles were seen on five occasions during all seasons between the two older bulls in the Wildlife Park. They were also observed four times in the National Bison Range, once in Hayden Valley and once in Wind Cave. They consisted of bellowing and tail-lifting but usually no fighting. Such battles also exhibited one or more of the following activities: bound- ing, aimless charges, frolicking, bucking, hesi- tant gait, arched back, pawing or wallowing, horning and butting a tree. Precopulatory and Mounting Behavior Amatory behavior. Licking of the cow’s fur by the bull was classed as amatory behavior. This was rare, however. In the most extreme case of such behavior (August 4, 1953, Wichita Refuge), both bull and cow licked each other for at least 90 minutes previous to copulation. This same pair also butted each other head-on gently, and on five occasions the cow mounted the bull for a few seconds with no thrusting. Another case involved the same cow which was discussed previously as mounting her tending bull almost continuously (National Bison Range). She also licked the sexual organs of this bull. lncompleted mountings. Attempted mount- ings by bulls tending cows were of regular oc- currence. These showed three degrees of com- pleteness: 1 . The bull merely swung his head toward the cow, usually without being in a position for proper mounting. The cow turned away. 2. The bull put his chin on the cow’s rump and the cow moved out from underneath. His penis was sometimes unsheathed. 3. The bull mounted the cow with his forelegs around her rump, but she dodged out. Un- sheathing was typical here. Some bulls made short panting sounds just before attempting to mount, lasting at times until the bull put his chin on the cow’s back. The sounds were so soft that they would not be heard more than 20 to 100 feet from the bull. They seemed to forewarn the cow, which some- times prematurely dashed away from the bull as a result. Completed mountings. The following data on completed mountings were compiled from six observations (mountings lasted for so few sec- onds that many were probably missed because another section of the herd was being watched) . The bull swung up onto the cow by using his chin as a lever on her rump. He then pressed his forelegs together in front of her hips and his head against her side. He remained mounted for four to ten seconds and thrust steadily. The mounting ended either when the bull voluntarily dropped off or when the cow walked out as the bull released his grip. After copulation, the cow characteristically moved a few feet ahead and voided a pulsating stream of milky urine. She also kept her tail lifted up in the air, lowering it gradually during the next few hours. Cows suffered occasional injuries from the rough treatment of the bulls during mounting. Close observation in the National Bison Range (Hugh Wilmar) revealed fresh, bloody wounds on several cows. The injuries were always on the 30 Zoologica: New York Zoological Society [43: 1 same area on the flanks, where the front hooves of the bull would strike in the mounting position. As soon as one bull mounted his cow, other bulls in the herd converged on the pair and fol- lowed them for some time afterward. All of these investigating bulls were not tending cows at the time and were generally younger than six years. As many as nine bulls followed one tending pair as it left one group and wandered toward an- other. The tending bull remained with the cow in all observed cases and sometimes repelled nearby bulls with threat, short charges or vicious battles. Masturbation was observed on four occasions when a bull unsheathed his penis, moved his hind quarters as though thrusting in copulation, and ejaculated. Bull Groups during the Rut Even though most groups of older bulls joined the cow groups during the rut, a few parties of one or two older members still remained apart. Some of these bulls stayed in approximately the same area for several days and probably never joined the cow groups. Others circulated more freely between cow groups and bull groups. Stray bulls lay down or grazed from 100 feet to a mile distant from the main herd and also straggled after the traveling herd. They followed the trail of the herd several hours later by sight, sound or scent, and often joined the herd. Some bulls passed near the herd, looked or listened mo- mentarily, and then passed on by. Bulls also voluntarily left the herd, and some returned again within a few hours. There were four observations where a bull voluntarily left the herd after a battle and two observations where a bull which was not tending a cow kept another bull outside the herd by threat. There was no indication that such “out- cast” bulls could not enter the same group a few hours later or enter any other group. The unsubstantiated convictions of several authors (Goodwin, 1939: 366; Inman, 1899: 231, 235; Seton, 1929: 690-1; Soper, 1941: 389, 392-3) that lone bulls are outcasts driven from the herd is further refuted by the evidence that bulls cir- culated freely during the rut. The view of Dodge (1877) is more logical: “The old bulls do un- doubtedly leave the herd . . . but I am disposed to believe this to be due to a misanthropic abne- gation of society on the part of these old fellows, to whom female companionship no longer pos- sesses its charm, rather than to their being driven out by the younger bulls, as is generally be- lieved.” Allen (1876: 55) and Garretson (1938: 37) also corroborated this interpretation. It is my opinion that such isolated bulls remained temporarily or permanently separate from the cow groups as a result of choice or physiological reasons rather than aggression of other bulls. (Permanent isolation of a few such bulls was suspected yet never adequately proved). Many of these isolated bulls during the rut were quite tolerant of human intrusion. Some ran 50 to 100 feet after first catching sight of human beings but then paid little attention. A few were extremely stubborn and could not be moved from their location with considerable noise or disturbance, such as shouting or rock throwing. VII. REPRODUCTION AND FAMILY RELATIONS Parturition Calving season. The main calving season in the Lamar Valley ran from April 15 to May 31 with a greater concentration of births in the cen- tral two weeks of this period. The season was similar in the Wildlife Park and Wind Cave. A scattering of calves nevertheless arrived outside of this main season in most of the herds studied. A few calves were born from June through October in all Yellowstone Herds, the Crow Re- servation, Wind Cave and the National Bison Range. Audubon & Bachman (1849: 47) re- ported that one cow found ready for calving in August “was an extraordinary circumstance at that season of the year.” Rush (1932) autop- sied 185 cows in the Lamar in December, 1931, and noted a “great variance in the size of the fetuses,” indicating an extensive calving season. Birth of calves during the winter season of November to March occurred at least five times in the Yellowstone Herds and once in the Niobrara National Refuge. Both Roe (1951: 94, 98) and Aubrey (1908:133) mentioned rare instances of winter calving. Seton (1929: 672) described one calf that was born and survived during January, 1884, when the temperature was — 38°F. Reproductive data on cows. Cows were sexu- ally mature at two years and calved at three years, according to data in Table 10 and obser- vations in all herds. Cy Young and George Mushbach, however, noted rare cases in the Na- tional Bison Range and on neighboring ranches where two-year-old cows gave birth to calves. Negus (1950) states that cows do not bear young until they are four years old, but his findings are inaccurate, since they were based on inconclusive evidence from the Wildlife Park Herd. Table 10 also shows that incidence of preg- 1958] McHugh: Social Behavior of the American Buffalo 31 Table 10. Records of Pregnancy in 125 Cows Autopsied in October 1950-1953 (National Bison Range 1950-1953) Age of cow Size of sample Per cent, pregnant 2 12 92% 3 5 80% 4 2 100% 5 6 100% 6 7 86% 7 2 100% 8 9 78% 9 5 80% 10 14 85% 11 1 100% 12 7 85% 2-12 70 87% 15-18 22 77% 19-24 19 58% 25-35 14 21% nancy declined gradually after an age of 12 years and markedly after 24 years. Two cows, however, were lactating and one was pregnant in a group of five cows more than 30 years of age killed in the National Bison Range in 1951. Burns (1953: 128) reports that slaughter of the herd in Wainwright Park, Alberta, showed that several of the cows earmarked 40 years earlier were accompanied by calves. Most cows bore one calf every year, although David Pierson observed one set of twins within about five days after birth in the Lamar Valley in 1953. Barrenness in cows was not necessarily an individual trait that appeared yearly. Evidence for this lies in an analysis of both pregnancy and lactation during slaughtering. Eighteen per cent, of 61 cows (National Bison Range, 1941-1953) and 24 per cent, of 206 cows in the Lamar Herd (Rush, 1932) were lactating and yet not preg- nant when autopsied on October-December. The few out-of-season pregnancies discussed previ- ously might account for a very minor part of these percentages. The gestation period is reported as nine months (Brown, 1936), 270 to 300 days (Burns, 1953: 199), and 914 months (Seton, 1929: 695). Ratios for calves. Calf ratios are defined as that percentage of a group of buffalo bearing calves. They are recorded in Table 11. Sexing of 1,465 calves during the fall and early winter of twelve years (National Bison Range, 1941-1953) yielded a male: female ratio of 50.4: 49.6. The same ratio for cattle is listed as 52:48 (Crew, 1925:255). In spite of this almost equal sex ratio for buffalo, three of the more divergent years yielded distinctly unequal ratios: 132 calves in 1950 had a ratio of 57:43, 100 in 1949 had 55:45, and 109 in 1942 had 42:58. Preliminaries to parturition. The pregnant cow was restless and wandered in short trips away from the herd for one to sometimes several days previous to calving. Physical changes pre- vious to parturition included a viscous, mucous discharge from the vagina, swelling of the vulva into a heart-shaped, flaccid mass, and filling of the udder. One pregnant cow was observed at close quarters in the Wildlife Park (field notes from James R. Simon). Three hours and 24 minutes previous to parturition she humped her back in a strained fashion while standing. There was a leakage of about three ounces of a thin, almost clear fluid from the vagina following this labor and again after 20 minutes. Five minutes later she pawed and rolled and then horned a small pine tree quite viciously. One pregnant cow was harassed for five hours previous to calving by a three-year-old bull which licked and nosed her vulva. Other herd members occasionally sniffed the vulva of cows soon to calve. Parturition. Cows gave birth to calves either when separated completely from cow groups or when in them. In the latter case, the groups were usually smaller and composed of several cows which either were pregnant or possessed young calves. Some authors similarly concluded that cows gave birth in or out of the herd (Aududon & Bachman, 1849: 37, Hornaday, 1889: 425; Roe, 1951: 98) while others main- tained that cows regularly separated from the herd to give birth (Aubrey, 1908: 133; Grinnell, 1904: 132; Seton, 1929: 695; Soper, 1941: 391; Stone & Cram, 1902: 69). The following data on the birth of calves were compiled from notes on two calves observed in parturition (field notes of James R. Simon). The two calves were born when the cow was flattened on her side with legs and neck out- stretched. The neck was strained dorsally and the upper hind leg was kicked violently upward and forward a few times. One calf was born in 20 minutes and the other in 27 minutes. Each cow broke the umbilical cord as she rose from the ground after parturition. The amnion ruptured about the head of one calf but tightly enclosed that of the other. Both cows im- mediately devoured these membranes and a part of the umbilical cord to within a few inches of the calf. The umbilical cord shriveled and dried several hours after birth. 32 Zoologica: New York Zoological Society [43: 1 Table 11. Calf Ratios in Various Herds Locality Date Calf ratio No. adults sampled Source of data Percentage of Entire Herd with Calves Hayden Valley Sept. 4, 1951 25% 384 Ground observ. (McHugh) Hayden Valley Mar. 1, 1950 22% 301 Aerial survey (YNP) Hayden Valley Mar. 7, 1949 18% 312 Aerial survey (YNP) Lamar Valley May 29, 1945 20% 495 Aerial survey (YNP) Wood Buffalo Park Feb. 1948 8% Fuller 1950 Wood Buffalo Park Nov. 1932-33 15% 605 Soper 1941 Percentage of Cows Two or More Years Old with Calves Hayden Valley Jun. 15, 1951 63% 70 Ground observ. (McHugh) Hayden Valley July 11, 1951 65% 92 Ground observ. (McHugh) National Bison Range Fall 1952 78% 119 Census (NBR 1952) National Bison Range Fall 1951 64% 128 Census (NBR 1951) National Bison Range Fall 1950 75% 177 Census (NBR 1950) National Bison Range Fall 1949 64% 229 Census (NBR 1949) National Bison Range Fall 1948 63% 250 Census (NBR 1948) National Bison Range Fall 1947 72% 233 Census (NBR 1947) National Bison Range Fall 1943 68% 204 Census (NBR 1943) National Bison Range Fall 1942 55% 199 Census (NBR 1942) National Bison Range Fall 1941 61% 183 Census (NBR 1941) Lamar Valley Winter 1932 69% 250 Tunnicliff Marsh 1935 Lamar Valley Winter 1931 55% 469 Tunnicliff Marsh 1935 Percentage of Cows Three or More Years Old with Calves National Bison Range Oct. 1953 95% 87* Census (NBR 1953) *Only seven cows older than nine years in this group. Expelling of the afterbirth was observed in only one cow, which expelled these tissues about one hour after birth and devoured them. Early life of the calf. The cows licked their calves almost constantly for at least the first 10 to 25 minutes after birth. The wet hair on the calf dried within an hour. One calf first stood up after 18 minutes and another after 28 minutes following birth. The first attempts at standing by new calves usually soon resulted in collapse again, mainly due to the weak, bowed legs. One calf tottered over on its chin and another somersaulted back on its rump. Not only did the calves have difficulty rising, but some also showed a lack of coordination in lying for the first time. The first attempts at walk- ing involved small, spider-like steps and a waver- ing gait. The earliest attempts at bucking occurred 70 minutes after birth. Within 180 minutes one calf playfully ran in circles and bucked around its mother. Another calf kicked its hind legs in the air at an age of about 100 minutes. All calves tried to nurse soon after standing but most failed at first, since they pushed their mouth up between the front legs. A close ob- servation showed that one calf continually moved its tongue in and out while trying to nurse in front and in back. Another failed in an early attempt between the hind legs because it probed too low. No cow made an effort to direct the nursing at- tempts of her calf. As each calf grew older, various transforma- tions were evident. At an age of about four to six days the calves started to graze. At about nine to ten weeks the reddish-orange pelage started to darken. The moult to a dark brown color was largely complete five weeks later. At two to three months the grunting started to lose its nasal quality and become more mature- sounding. Relationships between Cows and Calves Cohesion. For the first few days the calves re- mained particularly close to their mothers, often running to keep just ahead of the moving cows. The calves were closer to the cows during direct movements than at other times. Calves younger 1958] McHugh: Social Behavior of the American Buffalo 33 than two or three weeks generally lay down within a few feet of their cows, while older calves often lay down farther away, sometimes in calf subgroups. Up to an age of eight to twelve months cohesion between cow and calf was sufficiently evident to identify each pair during most periods of the day. After this age the attachment between the two weakened considerably, particularly with bull calves. Aggressive displays by a cow toward her calf-yearling were rare before the birth of the new calf but became frequent and firm within a few days after the new calf arrived. Mild ag- gression toward other nearby yearlings and two- year-olds also became more frequent. A few calves— more heifers than bulls— followed their cows for more than a year and infrequently for more than two years. Recognition between calf and mother de- pended upon scent, sight or sound. Cows on seven occasions identified their own calves mainly by scent when they sniffed closely one or two strange calves and then sniffed and stayed with the last calf, their own. Cows also com- monly sniffed their calves when returning to them after grazing. Instances of recognition by scent were rare for calves older than one month. Cows also identified their calves by sight or sound when they proceeded directly to their own calf or chased away a strange calf. Recognition by grunts without aid of sight showed that some grunts were distinctive. This was observed on four occasions between a cow and her calf and on one occasion between two calves. Calves identified their mothers largely by sight. They searched for and then moved directly toward their mother, avoiding strange cows after one glance. Scent and sound also appeared to play a part in some recognitions, but my observa- tions here are incomplete. The cohesion between mother and calf some- times resulted in the calf’s controlling the mother’s movements. Cows moved faster or farther at times due to the sudden advances of their calves. At other times, the calf detained its mother by failing to get up. Cohesion between mother and calf was further demonstrated by the start or intensification of grunting between the two under the following circumstances: (1) by a mother when her calf strayed away; (2) by a mother during moments of danger, such as the approach of a foreign animal; (3) by a calf when its mother strayed away; (4) by either calf or mother when one was separated artificially from the other; (5) by mothers or calves during herd movements, with a distinct increase in grunting synchronized with the movement; (6) by mothers when a wide river was to be crossed (Aubrey, 1908: 133); (7) by either in answer to the other, with the possibility of an interchange of grunts; (8) by either prior to nursing (discussed below). The customary result of grunting under these con- ditions was a closer grouping between mother and calf. Defense of the calf. Mothers never abandoned their calves or hesitated to defend them against approaching animals or human beings by quick charges or slow advances. Mothers were seen to make twelve attacks against human beings, five against horses, two against ravens, one against a pronghorn and one against a porcu- pine. Many authors have noted this characteristic of the buffalo cow to defend her calf against marauding animals and human beings (Aubrey, 1908: 133; Audubon & Bachman, 1849: 37; Allen, 1876: 58; Garretson, 1938: 39; Grinnell, 1904: 133; Hornaday, 1889: 432; Inman, 1899: 58, 135, 154, 205; Seton, 1929; and Stone & Cram, 1902). The picture is not complete, how- ever, without note of the numerous records where the cow abandoned her calf when the calf either fell behind on a long chase or was roped from the herd (Bradbury, 1904: 84; Dodge, 1877: 124; Garretson, 1938: 61; Grinnell, 1904: 133; Hornaday, 1889: 396, 400; Inman, 1899: 62, 67, 76-80). Coues (1897: 177) and Inman (1899: 141) recorded one instance each where a cow later returned to look for her calf, which had dropped behind the herd on a long chase. I observed no cases in wild herds where a bull defended a cow or calf. In fact, such defense would have been quite unlikely since older bulls did not join the cow herds until most calves were at least five weeks old (discussed further in Section IV). This lack of defensive behavior by bulls is surprising in view of the numerous rec- ords of it in the literature (Branch, 1929: 7; Gat- lin, 1876; I: 255; Dodge, 1877: 124-126; Gar- retson, 1938: 39; Hornaday, 1889: 433; Inman, 1899: 248; Seton, 1929: 695; Soper, 1941: 391). Interesting in this connection is one case in the Wildlife Park where a lone calf was trapped inside a corral. The entire herd, includ- ing two older bulls, clustered about the calf and could not be chased away. This incident and the lack of personal observations on defense of calves by bulls accords well with Grinnell’s (1904: 134) comment: . . it is true that a group of buffalo, if one of their number is at- tacked or threatened by wolves while they are close together, will all rally to the general de- 34 Zoologica: New York Zoological Society [43: 1 fense, and will stand by each other. But that bulls make it their business to defend calves . . . I do not believe.” Calves aided their own defense in three ways : (1) they often quickly moved close to their cows when approached by a strange animal or human being; (2) they occasionally counter-at- tacked; (3) they uncommonly hid themselves. A two- to four-day-old calf that was captured and placed in a corral charged and butted when- ever pursued by a human being. The butting was so gentle that no injuries resulted. Hornaday (1889: 396) described one captured calf that firmly butted horses and any person who ap- proached it. Four other calves that were tied by the neck charged any approaching persons and butted hard enough to knock a man off his feet (Inman, 1899: 134, 138). One calf hid in foliage in much the same manner as an elk calf (James R. Simon). This calf was missing when the herd in the Wildlife Park was driven by several men from the winter area to the display area on June 7. The cow was allowed to return into the area 28 hours later. She went directly to her calf, which was lying close to a log. A thorough search had failed to reveal the calf, and it did not stir in spite of the fact that one person had stepped ten feet from it. Similar reports of calves hiding are uncommon (Allen, 1876: 67; Catlin, 1876, II: 50, 255; Coues, 1897: 177; Grinnell, 1904: 132, 146). Nursing. In the customary position for nurs- ing, calves faced to the rear while alongside the cows, but some calves also infrequently suckled from between the hind legs. Calves after six days of age occasionally roughly jabbed the udder. Up to an age of three months, calves suckled for eight to ten minutes at a time. This period dropped gradually to about four minutes at five months and slightly more than three minutes at six months. An examination of cows slaughtered in the Lamar (David W. Pierson; Rush, 1932: 372), coupled with field observations, indicated that most cows suckled their young for seven to eight months. Hornaday (1889: 426) men- tioned nine or more months, but Roe’s (1951: 98) figure of three or four months is incredible. Eight instances in which a yearling nursed a mature cow were observed in Yellowstone Park and Wind Cave between May 7 and July 14. Duration of the suckling was two to eight min- utes with a mean of four minutes. These year- lings were allowed to suckle unmolested until the end of these periods. None of the cows had calves, although at least two of them were preg- nant and would calve soon. The cow-with-calf subgroup. Cows with their calves often came together to form one or sev- eral subgroups within the groups. They moved either within or at a short distance from the edge of the group. Sometimes calves initiated these smaller units, acting as a nucleus to which all mothers returned after grazing. Cow-with-calf subgroups were most obvious through July but continued at least into September. With the arrival of the first calf in each cow group, the flight distance of the group became greater than at any other time. The group moved away more quickly and for a greater distance when disturbed. Withdrawals were usually initi- ated by cows with calves, but sometimes by an- imals in advanced pregnancy. The increase in wariness was most noticeable in the Wildlife Park, where the herd that normally tolerated ap- proach within a few feet stampeded whenever it caught sight of human beings at distances up to 500 feet. This wariness subsided greatly by the end of May. During the entire summer, however, stam- pedes away from human beings in Hayden Val- ley were often led by a cow with a calf. Stam- pedes away from people in the Crow Reservation were led in about 85 per cent, of 34 cases by mature cows, most of which possessed calves. Relationships between Calf and Group During the first few hours after the birth of a calf, several group members characteristically came over to investigate. They looked at the calf, sniffed it and licked its fur. These curious buf- falo were of all ages and both sexes, and included cows which already had calves. They seldom in- vestigated the calf in this manner after its first day. The mother generally tolerated these buf- falo, yet sometimes chased them away if they were subordinate to her in the group hierarchy. At two or three weeks the calf’s cohesion for other calves began to compete with that for its mother. This was shown by the frequent oc- currence of calf subgroups, an increase in com- munication between calves, and instances in which a calf voluntarily returned to the calf subgroup when both mother and calf were tem- porarily separated from the herd. Grunting by calves and occasionally by older animals in juve- nile subgroups increased when one or a small group was detached or was separating from an- other. The grunting often resulted in one sec- tion rejoining the other. Of interest in this re- gard is Allen’s (1876: 207) description of how an Indian, disguised under a buffalo hide, “bleat- ed like a calf” to decoy a herd of buffalo into a 1958] McHugh: Social Behavior of the American Buffalo 35 pen. Other Indians on foot and on horse helped drive the herd. Calves also had interactions with strange buf- falo. Four instances were observed where calves of ages of 30 minutes, one day, two days, and 20 days, started to nurse the wrong cow. On three occasions a very young calf was seen to follow an animal other than its mother for a few min- utes. One thirty-minute-old calf followed a strange cow that was sniffing it. Its mother tried to retrieve the calf by driving off other cows but did not succeed until the strange cow lost inter- est in the calf. Again, when a yearling bull sniffed and then started to walk away from a calf about 5 Vi hours old, the calf followed. The bull trotted away and then butted the calf, which persisted in following. The cow also followed, but the calf did not return to her until the group stopped after trotting about 100 feet. Still another calf about one day old started to follow David Pier- son when he was riding a horse through a Lamar group. The calf followed so persistently that it had to be roped and taken back for release near its cow. These three incidents are interesting when compared with historical accounts of capturing young calves by separating them from the herd and then letting them voluntarily follow the horse and rider into camp (Audubon & Bach- man, 1849: 47; Catlin, 1876: 255-6; Coues, 1897: 176; Garretson, 1938: 40, 61; Hornaday, 1889: 398). Hornaday and Audubon thought that following was induced by letting the calf suck on one’s finger, while Catlin recommended breathing into the nostrils of the calf. VIII. ECOLOGICAL RELATIONS OF BUFFALO Relationships with Other Animals An animal as large and, under some condi- tions, as numerous as the buffalo inevitably af- fects the lives of birds and other mammals shar- ing the habitat. Such an influence works both ways, and we here examine the relationships between the buffalo and other animals. Buffalo birds. “Buffalo birds” is the term loosely but generally applied to the various birds that gathered about buffalo. Listed in order of decreasing abundance, they included cowbirds ( Molothrus ater). Brewer’s blackbirds ( Eupha - gus cyanocephalus), starlings ( Sturnus v. vul- garis), redwings ( Agelaius phoeniceus) and magpies ( Pica pica hudsonia). Friedmann (1929: 289) mentions that mo- tions of the buffalo flush grass-inhabiting insects and thus render them far more available to the birds. Birds were frequently seen feeding on flushed insects within a few inches of the heads or feet of grazing or walking buffalo. The birds frequently lit on and rode the backs of moving buffalo, sometimes crouching low and firmly grasping the fur to keep from being jostled off. They readily clung to the backs of walking buffalo but flew off when the gait became faster. Buffalo sometimes swung their heads around toward the birds to dislodge them. Older bulls were most tolerant— five cowbirds were seen roosting on the back of one old bull. Three Brewer’s blackbirds roosted on the backs of buf- falo during a severe rain storm. In winter birds sank low into the fur on the buffalo’s back and fluffed out their feathers. Seton (1929: 685) re- corded a cowbird that slept at night in a hollow in the fur behind the horn of an old bull. The buffalo birds also fed on invertebrates that were attracted to the buffalo. One cowbird walked all over the head, horns and body of a sleeping bull to feed for 35 minutes on swarms of flies. Other birds occasionally picked in the fur, presumably taking insects, and fed in the grass around resting buffalo. Magpies sometimes tried to pick open sores, but the buffalo prevent- ed this by rolling the body or swinging the head. Audubon & Bachman (1849: 46) described magpies picking the scabs or sores of buffalo emaciated by a hard winter. The relationship between buffalo and buffalo birds is thus a mutualism. The bird benefits from an increased food supply and a roost, while the buffalo benefits to a much lesser degree by the removal of flies or external parasites. In no case did the birds warn the buffalo of danger, so far as could be seen. Ravens ( Corvus corax ) followed buffalo at times, particularly during the calving season. They closely approached newborn calves, pos- sibly in search of the eyes, a fondness for which they have often demonstrated (Aldous, 1942). They also scavenged on the remnants of the afterbirth. Prairiedogs. Buffalo in Wind Cave frequented the towns of prairiedogs ( Cynomys ludovician- us) more than the surrounding grassland. They may have preferred the numerous bare areas around the mounds for lying and wallowing or the greater variety of forbs in the town. The prairiedogs disappeared into their burrows dur- ing moments of great activity in the herd, such as during battles, wallowing or rapid herd move- ments, but came out again and fed without show- ing any fear when the buffalo were grazing or lying quietly. Resting buffalo sometimes covered 36 Zoologica: New York Zoological Society [43: 1 and thus sealed off a prairiedog burrow for as long as two hours. Wallowing caused the most disturbance in prairiedog towns, since the buffalo often selected the bare areas around the burrows. Their paw- ing, horning and wallowing tore up the craters and sometimes filled in the burrows. Prairiedogs rebuilt the craters but could not keep repairs ahead of the constant destruction in some of the more heavily-used wallowing areas. As the width of the wallow increased it was used even more frequently. The prairiedogs usually aban- boned burrows in such areas (King, 1955). Predators. Grizzly bears commonly hunted over meadows in Hayden Valley. They usually came out of the timber at dusk and returned within a few hours after dawn, yet were seen during all hours of the day, particularly in spring or fall. When a female grizzly and three cubs hunted and captured a hidden elk calf within 700 feet of a buffalo herd of 78, the herd paid no attention to the bears. Catlin (1876-1: 254), Goodwin (1939: 369) and Soper (1941: 403) noted a similar indifference to wolves close to the herd. No definite grizzly kills were ever dis- covered in Hayden Valley, although the bears obviously had devoured one old bull, the cause of whose death was unknown, and another bull killed in battle. Approximately one per cent, of the buffalo in Hayden Valley had one shriveled rear leg, an injury that could have been caused by bears. Four buffalo were seen with claw- mark scars on their flanks or legs, probably from grizzlies. Way (1951) presented circumstantial evi- dence for the killing of a grizzly by a buffalo in the Lamar Valley of Yellowstone Park on about June 7. The badly mutilated carcass of a mature female grizzly bear was discovered three or four days after death. Coyotes occasionally lay within a few feet of buffalo herds, closer in winter than in summer. It is conjectured that they caught mice that were trapped in pits in the snow left by the feeding herd, but there is no observational evidence to support this theory. Joffe (1931) discovered the remains of a calf in the Lamar Valley that had been devoured by 10 or 12 coyotes. Whether it had been killed by the coyotes or died from natural causes was not determined. Buffalo killed a coyote on two occasions in the National Bison Range, according to Cy Young. As the coyote tried to move through the middle of the herd, the buffalo closed in and trampled and horned it. Miscellaneous associates. Piles of buffalo dung increased the invertebrate population, for they were the only droppings in the area that remained moist inside and underneath for several months. They were frequented most by numbers of Cole- optera and Diptera. Unusually large droppings, some of them ten inches tall, were used as sing- ing perches by Western meadowlarks and pos- sibly other prairie birds. Buffalo chips, as the dry dung is called, were an important source of fuel for pioneers on the treeless prairie. Tufts of shed buffalo hair commonly clung to low-hanging pine branches. Buffalo fur was found in a red squirrel nest and was probably used in the nests of many other birds and mam- mals. Interrelation with Vegetation and Soil Buffalo wrought their most conspicuous change on the vegetation of Hayden Valley by horning the lodgepole pine (Plate III, Figure 6). In a 12 by 150-foot sample strip of 68 lodge- pole pine, 51.4 per cent, of the trees was horned. The proportion of the circumference horned ranged from two per cent, to 100 per cent, with a mean of 49 per cent. The vertical range of the horning was from 0 to 62 inches with a mean of 1 1.4 to 42 inches. Of those trees horned, 14.3 per cent, died and another 8.6 per cent, were completely debarked within the past few months and probably would die shortly. Trees not horned in this sample area were usually in closely-spaced clusters of two or more, a pattern of grouping that made access by the buffalo diffi- cult. The converse of this is also true— solitary trees were most apt to be horned. Numerous low branches did not hinder the horning. This sample strip of horned trees was located at the edge of a much-frequented meadow near the junction with a main trail. There was an above-average amount of horning. In localized strips such as this, the reproduction of lodgepole pine was completely stopped and the tree line in rare cases was moved back. But a survey of the extensive border of pine in the entire valley showed that the over-all effect on the reproduc- tion of pine was minor. Moss (1932: 405-6) re- ported that the buffalo was doubtless an impor- tant factor in checking the invasion of prairie by aspen. Buffalo also helped to disseminate seeds. The fur, particularly the long hair on the head and legs, of several buffalo in Wind Cave was thick- ly clotted with cockleburs. The seeds of St. Johnswort ( Hypericum perforatum) were also found clinging to buffalo fur. Superintendent John E. Schwartz of the National Bison Range believes that this weed was spread through that area by buffalo. 1958] McHugh: Social Behavior of the American Buffalo 37 Buffalo affected the soil with their droppings, trails, wallows, and feeding patterns. Droppings served as fertilizer and were of greater importance in areas of heavy concentra- tion, such as on tops of ridges. Trails were cut deeply into the sod by the con- stant trampling, as deep as 28 inches below the sod for one trail in Hayden Valley. Such trails then started other disturbances in the soil; they eroded still more deeply, particularly on steep slopes. They occasionally acted as drainage can- als to lower the water table in uphill areas and thus eventually produce a change in vegeta- tion. Trails cut near the top of steep, sandy hills initiated washing and slippage of sand to create barren areas. Buffalo also caused erosion in other ways. Wallows on hillsides sometimes started eroding gullies. Concentration of buffalo on south-fac- ing, wind-blown ridges in winter produced al- most complete denudation of vegetative cover. Such areas then eroded badly. Summary A. Behavior of buffalo was studied in free-rang- ing and fenced herds in Yellowstone Na- tional Park, the Jackson Hole Wildlife Park, Wind Cave National Park, and other areas. B. Some general characteristics of buffalo were noted: They most typically grazed all then- forage, browsing only infrequently. Forage in winter was reached by clearing the snow with a swinging motion of the head, only uncommonly by pawing with the forehoof. Vocal expression consisted of grunting, bel- lowing and occasionally some snorting and sneezing sounds. Buffalo relied most on acute sense of smell for detecting danger. Play among buffalo involved battling, mounting, frolicking, bucking, or a combination of these. C. In locomotion, buffalo use four principal gaits— walking, trotting, galloping and bound- ing. In bounding, they spring ahead by the more or less simultaneous flexing of all four legs. Their trails were considered to be the most practical routes for cross-country travel by human beings, though not necessarily the most precisely engineered routes. For ex- ample, such trails customarily took a gradual climb up most hills yet cut steeply up and down others in order to achieve a more direct route. D. The leader of a moving herd was usually a cow, but several individual cows changed in this position so that there was no exclusive leader. No leader was apparent in some mass- action types of travel. E. Groups of buffalo were classified into two types according to composition. 1. Bull groups contained mature males with an infrequent female. They were small clusters of one to 12 members, most of whom were four or more years old. 2. Cow groups contained a majority of fe- males and a smaller number of males, mostly younger bulls. They averaged 23 members during the non-breeding season but increased in size during the rut, when many cow groups coalesced and were joined by bull groups. Cow groups during the non-breeding season were composed of cows, yearlings, calves, two-year-old bulls, some three-year-old bulls, and rarely bulls four or more years old. 3. Cow and bull groups broke into smaller subgroups of similar or random age and sex classes. These were spatially distinct from other subgroups in the area. F. Observation of interactions between indi- viduals of the Jackson Hole Wildlife Park Herd revealed a linear type of dominance hierarchy with dominant individuals exercis- ing a virtual constancy of success. 1. Such interactions were classified into pas- sive dominances (72.8%) and aggressions (27.2%). Aggressions included at times, as an indication of dominance, the inten- tion movement for mounting. 2. Permanent changes in the hierarchy re- sulted from the differential growth of bulls and cows. Two yearling bulls grew to about the same size as three two-year-old heifers during the summer of 1951. Pre- viously subordinate to all older cows, these two bulls then advanced above all cows in the next few months. Their irregular gain in dominance produced eight tri- angles, which finally straightened out. 3. Dominance among calves developed slow- ly, with the first definite signs noted at an age of four months. At about nine months, five calves demonstrated a pentagonal hier- archy. 4. The interspecific dominance hierarchy in the Wildlife Park showed the following order: adult human beings > buffalo > elk > mule deer > pronghorn > moose or whitetail deer. 38 Zoologica: New York Zoological Society [43: 1 G. The rut was delineated by a marked increase or the onset of the following activities among bulls: sniffing of vulvas, tending of cows, bel- lowing, wallowing, horning, vicious and non- vicious battles and incomplete or fertile mountings. 1. Bulls customarily extended the neck with upcurled lip after sniffing a vulva. 2. The rutting bond between bull and cow was called a tending bond and resulted in a temporary monogamous mateship. Heat in cows was detected by swelling of the vulva and a slight elevation of the tail. 3. Bulls bellowed frequently during the rut, typically as a display of threat. 4. Wallowing by bulls was attributed to ten- sion between bulls rather than to irritation from insects. It was thus considered a “dis- placement activity” occurring during situa- tions involving the simultaneous activation of the drives of attack and escape. 5. The vicious battles of the rut differed from other fighting in their brisk and violent movements. Injuries were of regular oc- currence and occasional deaths were re- corded. 6. Bulls circulated freely among herds. Lone bulls were apparently not outcasts but vol- untary isolates. H. Most cows were sexually mature at two years and calved at three, although a very few calved at two. Incidence of pregnancy de- clined gradually after an age of 12 years. Cows gave birth to calves when isolated from cow groups or when with them. Mothers de- fended their calves vigorously. I. Relations with the environment. 1. Buffalo influenced various birds and mam- mals living near them, and these animals sometimes affected buffalo in return. “Buf- falo birds,” the species gathering about buffalo, included cowbirds, Brewer’s blackbirds, starlings, redwings and mag- pies. Buffalo damaged prairiedog mounds by repeated wallowing on them. 2. Buffalo wrought their most conspicuous change in the vegetation of Hayden Val- ley by horning the bark of lodgepole pine. They also influenced the soil with their droppings, trails, wallows and concentra- tion of grazing in certain areas. Literature Cited Aldous, Shaler E. 1942. The white-necked raven In relation to agriculture. Fish and Wildlife Service, Research Report 5. Washington: U. S. Govt. Printing Office. Allee, W.C. 1951. Cooperation among animals with human implications. New York: Henry Schuman. 233 pp. Allen, J. A. 1876. The American bisons, living and extinct. Mem. Mus. Comparative Zoology, 4( 10) : 1-246. Asdell, S. A. 1946. Patterns of mammalian reproduction. Ithaca: Comstock Publ. Co. 437 pp. Aubrey, Charles 1908. Memories of an old buffalo hunter. Forest and Stream, 71(4) : 133-4; 71(5): 173-4; 71(6) : 2 16-7. Audubon, John James, & Rev. John Bachman 1849. The quadrupeds of North America. Vol- ume 11:32-55. New York: V. G. Audubon. 334 pp, Baggley, George F. 1933. The survival of the fittest. Yellowstone Nature Notes, 10(9-10) :39. Beach, F. A. 1945: Current concepts of play in animals. American Naturalist, 79:523-541. Bradbury, John 1904. Travels in the interior of America, 1809- 1811. Volume V of Early western travels 1748-1846, edited by Reuben Gold Thwaites. Cleveland: Arthur H. Clark Co. 320 pp. Branch, Douglas E. 1929. The hunting of the buffalo. New York: D. Appleton and Co. 240 pp. Brown, C. Emerson 1936. Rearing wild animals in captivity, and gestation periods. J. Mammalogy, 17(1): 10-13. Bryant, Edwin 1885. Rocky Mountain adventures. New York: Worthington Co. 452 pp. Burns, Eugene 1953. The sex life of wild animals. New York: Rinehart & Co. 290 pp. Carpenter, C. R. 1942. Societies of monkeys and apes. Volume VIII of Biological symposia: Levels of integration in biological and social sys- tems, edited by Robert Redfield. Lan- caster: Jacques Cattell Press. 240 pp. 1958] McHugh: Social Behavior of the American Buffalo 39 1952. Social behavior of non-human primates. Structure et Physiologie des Societes Ani- mates par Centre National de la Recher- che Scientifique, Paris, 13(7) : 227-246. Catlin, George 1876. Illustrations of the manners, customs, and condition of the North American Indians. Two volumes. London: Chatto & Windus. 530 pp. Chapman, Wendell and Lucie 1937. Wilderness wanderers. New York: Charles Scribner’s Sons. 318 pp. Collias, N. E. 1 944. Aggressive behavior among vertebrate ani- mals. Physiol. Zoology, 17(1) :83-123. Cottam, Clarence, & C. S. Williams 1943. Speed of some wild animals. J. Mammal- ogy, 24(2): 262-263. Coues, Elliott 1897. The manuscript journals of Alexander Henry and of David Thompson (1799- 1814). Volumes I and II. New York: F. P. Harper. 916 pp. Crew, Francis Albert Ely 1925. Animal genetics. Edinburgh: Oliver & Boyd. 420 pp. Dodge, Colonel Richard Irving 1877. The hunting grounds of the great west. (Identical to Plains of the great west). London: Chatto & Windus. 440 pp. 1885. Our wild Indians. Hartford: A. D. Worth- ington & Co. 653 pp. Dukes, H. H. 1937. The physiology of domestic animals. Ithaca: Comstock Publ. Co. 695 pp. Fremont, Brevet Capt. J. C. 1845. Report of the exploring expedition to the Rocky Mountains in the year 1842, and to Oregon and North California in the years 1843-1844. Washington: Blair and Rives. 583 pp. Friedmann, Herbert 1929. The cowbirds. Baltimore: Charles C. Thomas. 421 pp. Garretson, Martin S. 1938. The American bison. New York: N. Y. Zoological Society. 254 pp. Goodwin, George C. 1939. The bison. Chapter XX in North Ameri- can big game, edited by Alfred Ely, et al. New York: Charles Scribner’s Sons. 533 pp. Grinnell, George Bird 1904. The bison. A chapter in Musk-ox, bison, sheep and goat, by Caspar Whitney, George Bird Grinnell and Owen Wister. New York: The Macmillan Co. 284 pp. Hancock, John 1953. Grazing behaviour of cattle. Common- wealth Bureau of Animal Breeding and Genetics, 21(1) : 1-13. Hornaday, William T. 1889. The extermination of the American bison, with a sketch of its discovery and life his- tory. Report of the National Museum, 1886-1887, pages 367-548. 1922. The minds and manners of wild animals. New York: Charles Scribner’s Sons. 328 pp. Howell, A. B. 1944. Speed in animals. Chicago: Univ. of Chi- cago Press. 270 pp. Inman, Colonel Henry 1899. Buffalo Jones’ forty years of adventure. Topeka: Crane and Co. 469 pp. Joffe, Joseph 1931. 1 93 l’s first buffalo baby. Yellowstone Nature Notes, 8(4) :25. King, John A. 1955. Social behavior, social organization, and population dynamics in a black-tailed prairiedog town in the Black Hills of South Dakota. No. 67 of Contributions from the Laboratory of Vertebrate Biol- ogy, Univ. of Michigan. Leopold, Aldo 1947. Game management. New York: Charles Scribner’s Sons. 481 pp. Long, Major 1905. Account of an expedition from Pittsburgh to the Rocky Mountains, performed in the years 1819, 1920. Volumes XV-XVI of Early western travels 1748-1846, edited by Reuben Gold Thwaites. Cleveland: Ar- thur H. Clark Co. 647 pp. Masure, Ralph H. & W. C. Allee 1934. The social order in flocks of the common chicken and the pigeon. Auk, 51(3): 306- 327. Mayer, Frank H. 1934. Running buff. Field and Stream, 38(11): 28-9, 48, 60; 38(12) :36-8, 67-9. Moss, E. H. 1932. The vegetation of Alberta. J. Ecology, 20:380-415. Muybridge, Eadweard 1907. Animals in motion. London: Chapman and Hall. 264 pp. National Bison Range 1941-1953. Refuge narrative reports. On file at Moiese, Montana. 40 Zoologica: New York Zoological Society [43: 1: 1958] Negus, Norman C. 1950. Breeding of three-year-old females in the Jackson Hole Wildlife Park buffalo herd. J. Mammalogy, 31(4): 463. Nissen, Henry W. 1951. Social behavior in primates. Chapter 13 in Comparative psychology, third edition, edited by Calvin P. Stone. New York: Prentice-Hall, Inc. 525 pp. Parkman, Francis 1903. The Oregon trail. Boston: Little, Brown & Co. 479 pp. Pope, George W. 1919. Determining the age of cattle by the teeth. Farmer’s Bulletin 1066, U. S. Dept, of Agriculture. Roe, Frank Gilbert 1951. The North American buffalo. Toronto: Univ. of Toronto Press. 957 pp. Rush, W. M. 1932. Bang’s disease in the Yellowstone Na- tional Park buffalo and elk herds. J. Mam- malogy, 13(4): 371-2. 1942. Wild animals of the Rockies. New York: Harper and Brothers. 296 pp. Schein, Martin Warren, & Milton H. Fohrman 1955. Social dominance relationships in a herd of dairy cattle. Brit. J. Animal Behaviour, 3(2): 45-55. Seton, E. T. 1929. Lives of game animals. Volume III, Part II, and Volume II, Part I. New York: Garden City. Shoemaker, Hurst H. 1939. Social hierarchy in flocks of the canary. Auk, 56(4): 381-406. Soper, J. Dewey 1941. History, range and home life of the northern bison. Ecological Monographs, 11(4): 347-412. Stone, Witmer, & William Everett Cram 1902. American animals. New York: Double- day. 318 pp. Tinbergen, N. 1951. The study of instinct. Oxford: Clarendon Press. 228 pp. 1952. “Derived” activities; their causation, bio- logical significance, origin, and emancipa- tion during evolution. Quart. Rev. Biol., 27(1) : 1-32. Way, Joe J. 1951. Survival of the fittest. Yellowstone Nature Notes, 25(4): 46. EXPLANATION OF THE PLATES Plate I Fig. 1. A series of heads from cows slaughtered in the Lamar in January, 1952. Each age indicated on the photograph is represented as that at the approaching calving season (April-May). Fig. 2. A series of heads from bulls slaughtered in the Lamar in January, 1952. Each age indicated on the photograph is represented as that at the approaching calving season (April-May). Plate II Fig. 3. Two calves butting head-on in play in Yellowstone Park. (Photograph by cour- tesy of Yellowstone National Park). Fig. 4. One young bull mounting another in play. (Photograph by courtesy of Yellowstone National Park). Plate III Fig. 5. A bull extending neck with upcurled lip after sniffing a stream of urine from the adjacent cow. Fig. 6. A bull debarked this lodgepole pine by homing it during the rut in Hayden Valley. Fig. 7. Two bulls engaged in a vicious battle in the National Bison Range. (Photograph copyright by Walt Disney Productions). MCHUGH PLATE I FIG. 2 SOCIAL BEHAVIOR OF THE AMERICAN BUFFALO (BISON BISON BISON) PLATE II MCHUGH FIG. 4 SOCIAL BEHAVIOR OF THE AMERICAN BUFFALO (BISON BISON BISON) FIG. 3 I'i JiM MCHUGH PLATE 111 FIG. 6 FIG. 5 FIG. 7 SOCIAL BEHAVIOR OF THE AMERICAN BUFFALO (BISON BISON BISON) 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-PRESIDENTS SECRETARY TREASURER Fairfield Osborn Alfred Ely Harold J. O’Connell David H. McAlpin Laurance S. Rockefeller SCIENTIFIC STAFF: Zoological Park and Aquarium John Tee-Van General Director GENERAL ZOOLOGICAL PARK Leonard J. Goss Assistant Director and Veterinarian John L. George. . . . .Associate Curator of Mammals William G. Conway . . Curator of Birds Grace Davall Assistant Curator, Mammals and Birds James A. Oliver Curator of Reptiles Charles P. Gandal. . . Associate Veterinarian Lee S. Crandall General Curator Emeritus William Beebe Honorary Curator, Birds AQUARIUM Christopher W. Coates . Director James W. Atz Associate Curator Carleton Ray Assistant to the Director Ross F. Nigrelli Pathologist & Chair- man of Department of Marine Biochem- istry & Ecology Myron Gordon Geneticist C. M. Breder, Jr Research Associate in Ichthyology Harry A. Charipper. . .Research Associate in Histology Homer W. Smith Research Associate in Physiology William Bridges Editor & Curator, Publications 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 Henry Fleming Entomologist John Tee-Van Associate William K. Gregory Associate AFFILIATES C. R. Carpenter Co-ordinator, Animal Behavior Research Programs L. Floyd Clarke Director, Jackson Hole Research Station SCIENTIFIC ADVISORY COUNCIL A. Raymond Dochez Alfred E. Emerson W. A. Hagan Caryl P. Haskins K. S. Lashley John S. Nicholas EDITORIAL COMMITTEE Fairfield Osborn, Chairman James W. Atz William Beebe William Bridges Christopher W. Coates William G. Conway Lee S. Crandall John L. George Leonard J. Goss James A. Oliver John Tee-Van ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME 43 • PART 2 • AUGUST 27, 1958 • NUMBERS 2 & 3 PUBLISHED BY THE SOCIETY The ZOOLOGICAL PARK, New York Contents PAGE 2. The Iguanid Lizard Genera Urosaurus and Uta, with Remarks on Related Groups. By Jay M. Savage. Text-figures 1-6 41 3. Feeding Habits of the Northern Water Snake, Natrix sipedon sipedon Lin- naeus. By E. E. Brown 55 2 The Iguanid Lizard Genera Urosaurus and Uta , with Remarks on Related Groups Jay M. Savage Department of Biology, University of Southern California, Los Angeles 7, California (Text-figures 1-6) IN the course of studies leading toward a re- evaluation of the supraspecific units within the lizard family Iguanidae it seems expedi- ent to pause from time to time and make avail- able portions of the completed work. It is hoped that these progress reports can eventually be correlated with additional material to form a comprehensive revision of the family. This paper is the second of a series pertaining to the systematics of the Iguanidae. Statement of the Problem Until recently all North American iguanids with a well-developed gular fold, a large rostral scale, non-imbricate supralabials, imbricate supercilliaries, large and regularly arranged up- per head shields, a large interparietal scale, a well-defined ear-opening, sternal fontanels and with the parietal organ consistently piercing the parietal bone posterior to the suture between the frontal and parietal bones, have been placed in the genus Uta Baird & Girard, 1852. Mittleman (1942) suggested that this assemblage was arti- ficial and purported to demonstrate that the group was actually comprised of four distinct genera, Petrosaurus Boulenger, 1885; Strepto- saurus Mittleman, 1942; Urosaurus Hallowell, 1854; and Uta Baird & Girard, 1852. Although the differences used by Mittleman to separate these groups were of questionable significance, support for the division of Uta was provided by his ideas of the phylogeny of the North Ameri- can iguanids. This supposed natural subdivision of the Iguanidae traditionally has included the following genera: Callisaurus Blainville, 1835; Crotaphytus Holbrook, 1842; Ctenosaura Wieg- mann, 1828; Dipsosaurus Hallowed, 1854; Eny- aliosaurus Gray, 1845; Holbrookia Girard, 1851; Iguana Laurenti, 1768; Phrynosoma Wiegmann, 1 828; Sator Dickerson, 1919 \Sauro- malus Dumeril, 1856; Sceloporus Wiegmann, 1828; Uma Baird, 1858; Uta Baird & Girard, 1852. According to Mittleman’s system there were two main evolutionary lines represented in this group, both stocks being derived from the genus Ctenosaura. One line was composed of the relatively primitive genera Dipsosaurus and Sauromalus (and presumably Enyaliosaurus and Iguana) and the more highly specialized Calli- saurus, Holbrookia, Uma and Crotaphytus. Also placed in this section were two of the compo- nent genera, Petrosaurus and Streptosaurus, re- moved from Uta by Mittleman. These two gen- era were supposed to be derived from Crota- phytus. The second major stock included Phry- nosoma, Sceloporus, Sator, Urosaurus and Uta. The last three genera were considered by Mit- tleman to be independent derivatives of Scelo- porus. Stejneger & Barbour (1943) and Smith & Taylor (1950), in their checklists of the lizards of the United States and Mexico, adopted Mit- tleman’s arrangement of the “utas,” while Smith (1946, p. 92) presented a somewhat modified phylogeny of North American iguanids that is nevertheless in basic agreement with Mittle- man’s work. Many herpetologists, some perhaps influenced by the above acceptance of Mittle- man’s nomenclature, have followed his conclu- sions. On the other hand, other workers have been inclined to follow Oliver (1943, p. 106), who was loathe to recognize Mittleman’s genera because so few characters separate them, and have retained all the species within a single genus. Schmidt (1953) and Stebbins (1954), among others, adhered to the latter view. It is obvious from the above discussion that Mittleman’s classification hinges more upon his SMITHSONIAN CCD SNSTITI mnw -btp 3 196* 42 Zoologica: New York Zoological Society [43: 2 interpretation of the phylogeny of the North American iguanids than upon marked struc- tural differences between the several species groups. If his conception of the evolution of these lizards is correct, it would appear that rec- ognition of four genera of “utas” is necessary. If, however, his interpretation of the group’s phylogeny is erroneous and no additional mor- phologic features can be discovered to support his divisions, Oliver’s conclusion that but one generic unit is involved must be accepted. The problem, therefore, is: (1) to determine if any characteristics will separate the groups included in the genus Uta prior to Mittleman’s study and (2) to evaluate the relationships between these groups and other iguanid genera. Plan of Analysis My interest in this problem was originally aroused during preliminary examination of skel- etal material being assembled for studies on the Iguanidae. At that time it was noted that there were remarkable differences between several species of Uta ( sensu latu ) in the nature of the sternum and associated structures. If these dif- ferences proved to be constant for each species group, it was thought that they might validate generic segregation. Consequently, since the external features used to distinguish between the several groups of “utas” were of doubtful sig- nificance, the present analysis has centered around a review of their comparative osteolo- gies. A survey of external differences has also been undertaken in order to determine if these substantiate differences in internal character- istics. It became apparent early in the study that the principal difficulties of the problem lay in the allocation of the genera Urosaurus and Uta. Once these genera had been properly placed, the position of Petrosaurus and Streptosaurus can be readily understood. For this reason, a comparison of Urosaurus and Uta forms the first part of this report. The second section deals with the status of Petrosaurus and Streptosaurus. A third section considers a recent attempt to classify these lizards on the basis of ecologic characteristics. The final portion of the paper is concerned with the general relationships be- tween the “utas” and other iguanid lizards. Information for this report has been derived from preserved material of all genera and species mentioned. In addition skeletons of the following species, prepared by the Bolin Method (Bolin, 1936), have been examined: Callisaurus draco- noides (4), Crotaphytus collaris (1), Crotaphy- tus wislizeni (2), Ctenosaura hemilopha (1), Dipsosaurus dorsalis (3), Phry nosoma corona- turn (4), Phrynosoma platy rhinos (1), Sator angustus (2), Sator grandaevus (2), Sauromalus ater (1), Sauromalus obesus (1), Sceloporus magister (5), Sceloporus occidentalis (10), Sceloporus orcutti (1), Uma notata (1), Uro- saurus graciosus (2), Uta mearnsi (1), Uta slevini (1), Uta stansburiana (5). Also available was additional skeletal material or cleared and stained specimens of: Amblyrhynchus cristatus (5), Anolis garmani (2), Anolis leucophaeus (1) , Brachylophus fasciatus (1), Conolophus pallidus (1), Conolophus subcristatus (10), Ctenosaura cicanthura ( 1 ) , Ctenosaura pectinata (2) , Ctenosaura similis (3), Cyclura carinata (1), Cyclura cornuta (3), Cyclura stejnegeri ( 1 ) , Holbrookia maculata ( 1 ) , Holbrookia tex- ana (1), Iguana iguana (15), Leiocephalus psammodromus (2), Sauromalus hispidus (1), Sauromalus varius ( 1 ) , Uma inornata ( 1 ) , Uma scoparia (1), Urosaurus auriculatus (1), Uro- saurus bicarinatus (1), Urosaurus nigricaudus (1) , Urosaurus ornatus (1) and Uta thalassina (2) . The pertinent morphological points have been uncovered by dissection on specimens of Enyaliosaurus quinquecarinatus, Urosaurus mi- croscutatus and representatives of most of the major subpopulations of Uta stansburiana and its insular allies. Comparison of Urosaurus and Uta Mittleman (1942, pp. 109-112) presented what purport to be extensive differential diag- noses of Petrosaurus, Streptosaurus, Urosaurus and Uta. Oliver (1943, p. 106) pointed out that few of the listed features satisfactorily dis- tinguished between these groups and that none of them clearly indicated the existence of more than one genus. Mittleman (1942, p. 106) stated that there were no differences in the bony structures of these lizards that would separate them from one another or from Sator and Scelo- porus. Preliminary examination of skeletal mate- rial tended to dispute this latter assertion and my study has been aimed at discerning whether osteological features distinguished the nominal groups of “utas.” The lizards placed in the genera Urosaurus and Uta resemble one another rather closely in the structure of the skull, vertebral column, girdles and limbs. The two groups are profound- ly divergent, however, in the condition of the sternal plate and associated structures. These differences, supported to some extent by ex- ternal structures, convince me that two genera can be recognized. The differences between Urosaurus and Uta in sternal anatomy are sum- marized as follows: 1958] Savage: Iguanid Lizard Genera Urosaurus and Uta 43 Text-fig. 1. Sternal plates and asso- ciated structures in iguanid lizards from ventral view. A. Diagram of uro- saurine sternum from specimen of Urosaurus graciosus. B. Diagram of utiform sternum from example of Uta stansburiana. Abbreviations denote the following structures: C. clavicle; I. in- terclavicle; S. sternal plate; R. xiphi- sternal rod; L. lateral xiphisternal rib; T. terminal xiphisternal rib. Urosaurus.— Sternal plate rather long and rel- atively narrow, with the posterior margin taper- ing almost to a point; xiphisternal rods originat- ing near center line of sternal plate, long, being much longer, when measured from the sternal plate to origin of terminal xiphisternal ribs, than sternal plate is wide; lateral xiphisternal ribs present. Uta— Sternal plate relatively short and broad, with the posterior margin truncate and forming a wide base; xiphisternal rods originating at lateral edges of posterior margin of sternal plate, short, being much shorter, when measured from sternal plate to origin of terminal xiphisternal ribs, than sternal plate is wide; no lateral xiphi- sternal ribs, only terminal ones. These sternal characteristics have been ob- served in osteological material of the generic types, Urosaurus graciosus Hallowell, 1854, and Uta stansburiana Baird & Girard, 1852, and verified in examples of all major groups within these nominal genera. Text-fig. 1 illustrates the sternum in these groups. The apparent defer- ences in the nature of the interclavicle-clavicle relationships shown in these figures are not con- stant throughout the two groups. Elsewhere in the family Iguanidae the shape and relative position of these elements are frequently of value in generic determination. The condition of the sternum has not previ- ously been employed to characterize genera of the Iguanidae and some question may arise as to the validity of a distinction made upon this feature. Conceivably, the observed differences could be due to modifications of a single sternal type within a single genus. To confirm the gen- eric significance of the sternal condition, an examination of this structure was made on ex- amples of the majority of North American iguanids, exclusive of the myriad forms within the genus Sceloporus. This examination re- vealed that not only is the type of sternum con- sistent within every genus but that the condition of the structure appears to have considerable phylogenetic significance. Although the sternal plates and associated structures of some other American iguanids superficially resemble the condition found in Uta, within the section of the family closely allied to Urosaurus and Uta the sternum seems indicative of two evolution- ary lines. In this regard, the distribution of the two sternal types in this section of the Iguanidae is informative : Urosaurine Sator Sceloporus Urosaurus Utiform Callisaurus Holbrookia Phrynosoma Uma Uta Tn the material examined, the typical uro- 44 Zoologica: New York Zoological Society [43: 2 saurine sternum in found only in the three gen- era listed above, although some Sceloporus tend to have rather short xiphisternal rods. These groups have usually been held by previous work- ers to be closely allied on the basis of external features, and the sternal arrangement fully sup- ports this view. It seems likely that Urosaurus is best understood as a specialized off-shoot of Sceloporus from which it differs primarily in the presence of a fully developed gular fold and the absence of a scapular foramen, no gular fold but a scapular foramen being present in Sceloporus. Uta appears to be a specialized genus related to the highly adapted, but appar- ently more primitive, Callisaurus-Holbrookia- Uma series. Uta is probably best considered as a recent derivative of this stock and as such can only be distantly allied to Urosaurus, which it resembles in several external features. In view of the evidence of complete morphological sepa- ration and the probability of different origins it would seem that Urosaurus and Uta ought to be retained as distinct genera. Because the principal characteristics diagnos- tic of the genera Urosaurus and Uta are internal ones, it seemed worthwhile to attempt to de- termine if there were any external features that distinguish them. Such characters might be useful for rapid identification or in artificial keys. Careful examination of examples of all the species groups within the two genera reveals that a single scutellational character can be used for generic recognition. This feature in- volves the arrangement of the scales in the nasal region and makes it necessary to digress at this point from the major problem of the paper to treat a matter of terminology. Unfortunately herpetologists have seldom attempted to standardize the nomenclature of the head scales in lizards. General agreement has been reached in dealing with the compara- tively simple arrangement of snakes, but since the size, shape and position of the scales vary from lizard group to lizard group, the subject has become surprisingly complex. In the case in point, for example, Smith (1946, p. 276) and Mittleman (1942, p. 123), in naming the head scales of Urosaurus and Uta, use entirely dif- ferent terms for what appear to be positionally homologous units. It is understandable under these circumstances why previous workers have overlooked scale characters that readily sepa- rate these two genera. Because of the great diversity in the number and disposition of head scales in different lizard families, it does not seem possible or desirable to instigate a univer- sally applicable nomenclature. Nevertheless, it does seem worthwhile to employ a standard set of terms within familial or subfamilial limits. While this cannot be accomplished as yet with the iguanids because of our lack of understand- . ing of the suprageneric groupings within the | family, I have attempted to standardize the scale nomenclature for the nasal region in the genera j allied to Urosaurus and Uta. Subsequent work i will probably find that this system can be applied without too much difficulty to more distantly related and less specialized genera. Terminology for Scales in Nasal Region Rostral— Scale at tip of snout, bordering upper lip. Internasals.— Usually a double pair (anterior and posterior) of scales lying between the nasals on top of the snout. Nasals— Scales in which the nostrils are pierced. Postrostrals — These are the subnasals of Uro- saurus as defined by Mittleman (1942) and the postrostrals of Uta and Sceloporus accord- ing to Smith (1946); the two series are homologous in position, bordering the nasal along its anterior and lower margins and on occasion separating the rostral from the an- terior internasals. Supranasals.— Scales bordering the upper mar- gins of the nasals and separating the nasals from the internasals. Subnasals.— Scales bordering the lower edge of the nasals and lying posterior to the postros- trals; they separate the nasals from the supra- labials. Postnasals.— Scales bordering the posterior mar- gin of the nasals and separating them from the loreals or canthals. The set of terms here defined and figured (Text-fig. 2) can be applied to the scales in the nasal region of the following genera (names as listed by Mittleman) : Petrosaurus, Sator, Scelo- porus, Streptosaurus, Urosaurus and Uta. The system is not satisfactory when working with gen- era with a more or less homogenous complement of head scales as in Crotaphytus, Ctenosaura, Dipsosaurus, Enyaliosaurus, Iguana and Sauro- malus. Callisaurus, Holbrookia, Phrynosoma and Uma have the head scales somewhat inter- mediate between the Urosaurus-Uta group and the IguanaAike lizards. There is probably little need to distinguish between the scales in the nasal region at such a fine level as is profitable in Urosaurus and Uta in this Phrynosoma-Hol- brookia line. The genera Urosaurus and Uta are distin- guished by: Urosaurus having the nasals and internasals in contact and lacking supranasals 1958] Savage: lguanid Lizard Genera Urosaurus and Uta 45 Text-fig. 2. Scutellation of nasal region in iguanid lizards. A. Dorsal view of snout of Urosaurus. B. Dorsal view of snout of Uta. The following symbols indicate the pertinent scales: l. internasal; N. nasal; P. postrostral; Q. subnasal; R. rostral; S. supranasal; W. postnasal. (Text-fig. 2, A); Uta has the anterior internasals separated from the nasals by definite supranasal scales (Text-fig. 2,B). Although Urosaurus is distinct from members of the Uta stansburiana group in having two distinct abdominal patches in the adult males and sometimes in the female and never any axillary or shoulder dark spots (in Uta stansburiana and its allies no well-de- fined abdominal patches of color are present in either sex, although the belly may be suffused with blue, gray or black in adult males and there is usually a dark blue or black axillary spot and frequently a dark spot anterior to the shoulder insertion), these differences break down when other members of the genus Uta are considered. In Urosaurus and Uta the rostral scale may or may not be in contact with the internasals. If the rostral and internasals are separated the postrostrals lie between them. This arrangement appears to be quite variable in populations of Uta but it is more consistent within specific limits in Urosaurus. In my material usually 90 per cent, or more of the specimens of a single form have the same internasal-rostral relation- ship. In Urosaurus bicarinatus (A. Dumeril, i856), not all subspecies seen, Urosaurus nigri- caudus (Cope, 1864) and Urosaurus microscu- tatus (Van Denburgh, 1894), the rostral usually meets the anterior internasals. Urosaurus auri- culatus (Cope, 1871), Urosaurus graciosus Hallowell, 1854, and Urosaurus ornatus (Baird & Girard, 1852), not all subspecies seen, usually have the postrostrals preventing a contact be- tween the rostral and internasals. In related genera these external features are variable. Sator has supranasals and. definite ab- dominal color patches in adult males. Supra- nasals may be present or absent in Sceloporus although apparently consistently present or absent within species limits, and the abdominal color patches are regularly present in many species in adult males and sometimes in females. In several species these color patches are totally absent. As previously indicated, the scales in the nasal region of Callisaurus, Holbrookia, Phrynosoma and Uma are relatively small and cannot be recognized as definite postrostrals, supranasals or internasals. Definite abdominal color patches are present in males (and some- times in females) of all these genera except Phrynosoma. The Status of Petrosaurus and Streptosaurus Since Urosaurus and Uta have been shown to be distinct, the position of Petrosaurus and Streptosaurus, two nominal genera formerly in- cluded in Uta, can now be analyzed. According to Mittleman, Petrosaurus was derived from Crotaphytus and Streptosaurus from Petro- saurus. There can be no question regarding the close relationship of Petrosaurus and Strepto- saurus but their supposed affinity to Crotaphytus is, on the basis of data accumulated in the pre- paration of this report, subject to considerable doubt. Evidence showing why Petrosaurus and Streptosaurus cannot be closely related to Cro- taphytus will be presented later in this paper. The first problem at hand is to determine the generic status and differences between Petro- saurus thalassinus (Cope, 1863) and the doubt- fully valid form Petrosaurus reprens (Van Den- burgh, 1895) on the one hand and the two nominal species of Streptosaurus, mearnsi (Stej- neger, 1894) and slevini (Van Denburgh, 1922) on the other. Mittleman (1942, pp. 110-111) attempted to segregate these two species-groups 46 Zoologica: New York Zoological Society [43: 2 on the basis of numerous characters. Many of the features listed by him were given under both genera, however, and an analysis of the others indicates that few of the differences hold. The features employed by Mittleman are given below: Petrosaurus Streptosaurus 1. Larger ventrals 1. 2. Enlarged, strongly 2. keeled caudal scales 3. Three rows of 3. enlarged supraoculars 4. Anterior gular fold 4. (pregular) well- developed 5. Lateral fold poorly 5. developed 6. Dorsal pattern of 6. bands, no neck ring 7. No abdominal color 7. patches 8. No palatine bones 8. Smaller ventrals Caudals weakly keeled Two rows of enlarged supraoculars Anterior gular fold poorly developed Lateral fold well- developed Dorsal pattern with- out bands, a neck ring Abdominal color patches in males Palatine bones present These features are considered in the order given above. ( 1 ) There appears to be a definite difference in the size of the ventral scales in the two groups but this character is of questionable generic significance. (2) The differences in the degree of keeling of the caudal scales are evi- dent but hardly generic in character. (3) The number of rows of enlarged, supraoculars is a consistent feature. (4) The condition of the pregular fold in life is variable and the degree to which it appears to be developed in preserved material is not consistent within any available sample. This character is therefore useless for distinguishing between the two groups. (5) The same remarks given for the pregular fold apply to the condition of the lateral fold. (6) The dorsal bands and nuchal collar are present in both Petrosaurus and Streptosaurus but differ in degree of intensity. The dark nuchal collar of Streptosaurus appears to be the same pattern element as the anterior dorsal band in Petro- saurus. The posterior bands are prominent in the latter genus but, although present in the former group, they are obscured to some extent by the darker body coloration. (7) Both groups have the same type of abdominal coloration in adult males although the predominant color in both living and preserved material is blue in Streptosaurus and blackish in Petrosaurus. These colored areas appear to constitute definite abdominal patches although the color tends to suffuse over the entire ventral surface. Laterally the suffusion of darker color is similar to the condition found in adult male Uta stansburiana; however, this latter group does not have ex- tensive coloration superimposed on the ventral abdominal surfaces. (8) A palatine bone is present on both sides of the skull in all speci- mens of either group seen by me, although the bone is somewhat thinner than in less specialized iguanids. It is obvious that the two presumed genera are distinct from one another in a few minor details of scutellation and coloration and that none of the observed differences seem indicative of two distinct generic groups. Examination of skeletal material of the several species shows that they are essentially similar. The only points of difference between them are in the relative proportions of a few elements. Because of the absence of trenchant distinguishing features and the fact that all workers, including Mittleman, have considered these taxa to be closely related, inclusion of these lizards in a single genus seems appropriate. Fully supporting this conclusion are the facts of morphology and the distribu- tional pattern of the several species. The form mearnsi has a range from Riverside County, California, south in Baja California, Mexico, to the region of Santa Rosalia. Thalassina occupies the southern portion of the peninsula from Comundu (about 75 miles south of Santa Rosalia) southward. If the form reprens is recognized, it would occupy the northern por- tion of the range given for thalassina. Although additional evidence is needed to verify this point, it is likely that thalassina is more primitive than mearnsi. The insular form, slevini, is obviously of close affinity with mearnsi. It is restricted to Isla Angel de la Guarda and adjacent islets in the Gulf of California. The question now arises as to the generic placement of these three species. Since the pec- toral apparatus as well as the scutellation and general morphology of these forms are of the utiform type, the genera allied to Sceloporus need not be considered. The condition of the vertebrae, the nasal structure, the scapulocora- coid foramina (Text-fig. 5) and the sternum of these three forms are totally different from these features in Crotaphytus and its relatives. Conse- quently these genera also need not be discussed. (See the section on classification at the end of this paper for additional information on the affin- ity between these species and Crotaphytus postu- lated by Mittleman). These eliminations leave only the genera associated with Uta as possible congenitors of mearnsi, slevini and thalassina. Within this series, only Uta approaches the three in osteological and other morphological features and it is here that the relationship apparently lies. The following summary of characteristics 1958] Savage : lguanid Lizard Genera Urosaurus and Uta 47 will separate Uta stansburiana and its allies from the giant Baja California forms: Uta: enlarged supraoculars in a single series; well-defined median and lateral frontonasal scales; scales along gular fold much larger than gulars; no definite dorsal bands, no nuchal collar; neural spines well-developed, higher than long; usually three sternal ribs (rarely four) . Petrosaurus: enlarged supraoculars in two or three series; no definite median and lateral frontonasals; scales along gular fold same size as gulars; back with definite dark bands, a nuchal dark collar; neural spines low, longer than high; four sternal ribs. The differences between Uta and Petrosaurus are slight and no single feature in itself is par- ticularly significant. However, the total char- acter-complements of the two groups are rather divergent and a decision as to the most pro- pitious allocation of the species involved is dif- ficult. Although there is considerable merit in recognizing Petrosaurus as a phylogenetic line distinct from Uta, the obvious close relationship between the two and the kind of differences separating them lead me to conclude that the evolutionary picture can best be explained by placing them in a single genus. Recognition of two subgeneric categories within the genus Uta, one for stansburiana and its immediate allies and a second ( Petrosaurus ) for the mearnsi group, may be a useful way to emphasize the differences between the two evolutionary lines in the genus. In this regard it should be noted that the shoulder spot of Uta stansburiana and related forms appears to represent the remnant of the nuchal collar of Uta mearnsi, Uta sievini and Uta thalassina. In these latter forms, the dark blue or black lateral and abdominal suffusions of the male are most intensive in the axillary region, and the axillary spot in the Uta stans- buriana section is probably a retention of the anterior portion of this densely pigmented area. The Ecologic Genus and the Present Problem The concept of the genus adopted in the present report agrees in principle with that given by Mayr (1942, pp. 282-286). Because of the nature of the material studied, the de- grees of relationship and difference between the several groups are based upon morphologic characteristics, although it is clearly understood that other kinds of characters may be used, and ought to be used when available, in generic definition. The genera accepted in this account are therefore convenient but natural groupings of species separated from other such units by discontinuities in morphologic variation. A radical conception of the genus in terms of ecology has recently been adopted by one herpetologist (Lowe, 1955a; 1955b) and ap- plied to the problem of the generic status of Uta and its allies. Lowe holds the extreme posi- tion that genera can be recognized on the basis of ecologic divergence alone, without support from any other kind of characteristics. The dif- ficulties arising from the rigid application of this idea are too numerous to consider at this time, but may be summed up as follows: (1) any two species, if different from each other in ecology, regardless of similarities in morphology, phy- siology or other features, may be recognized as distinct genera; (2) all species having the same or very similar ecologies, regardless of genetic relationships or differences in other features, may be placed in the same genus. Lowe and Norris (in Lowe, 1955a) utilized this concept as the basis for a classification of the lizards formerly placed in the genus Uta. They maintained Mittleman’s arrangement of these species because of supposed differences between and similarities within the groups in- volved. According to these authors the species can be arranged as follows: Genus Petrosaurus Subgenus Petrosaurus Subgenus Streptosaurus Genus Urosaurus Genus Uta Petrosaurus and Streptosaurus were placed together because of their cliff-dwelling propen- sities. Urosaurus was retained as a distinct genus because the species within the group are, ac- cording to Lowe and Norris, plant-dwellers and plant-climbers. Uta is supposedly distinguished from the other two genera by living on the ground. The genera are thus recognized because they occupy different ecologic niches. The primary reason why Mittleman’s classi- fication of these lizards has not been generally accepted lies in his failure to present convincing evidence that the several groups were morpho- logically different from one another. The most striking morphologic feature listed by him as separating Uta and Urosaurus, for example, was the homogeneous dorsal scutellation of the former and the differentiation of the paraver- tebral scales in the latter. This character fails to hold for Urosaurus microscutatus and some examples of Urosaurus nigricaudus which have a homogeneous complement of dorsal scales. 48 Zoologica: New York Zoological Society [43: 2 The system adopted by Lowe and Norris can only have merit if it is based upon consistent ecologic features that do not vary within the several groups established by Mittleman. Unfortunately this is not the case. Firstly, several species placed within the nominal genera do not have the ecologic mode of life attributed to them by Lowe and Norris. Urosciurus micro- scutatus and Urosaurus nigricaudus are typically found on rocks and boulders, often in association with Uta ( Petrosaurus ) mearnsi or Uta ( Petro - saurus ) thalassina and only rarely in bushes or other plants. Secondly, there is considerable variation in the habitats occupied by different individuals or populations within many species. The supposedly ground-living Uta stansburiana is frequently found in low bushes or on rocks or boulders. Many of the insular forms of Uta are more or less restricted to this latter habitat. It may be assumed that the species in the other generic groups also exhibit some variation in habitat selection. Finally, it ought to be pointed out that members of the related genera Sator and Sceloporus are found in all three habitats attributed to Petrosaurus, Urosaurus and Uta. Within the limits of Sceloporus, various species tend to be inhabitants of trees and bushes, or are typically found on the ground or in rocky and boulder regions. Other members of this genus may occur in two or three of these habi- tats. The two species of Sator are unselective in habit, individuals of the same form being commonly found in all three situations. If ecologic characteristics alone were used in set- ting up the genera in this section of the Iguani- dae, all of the taxa mentioned above would have to be placed in a single genus since no clear-cut distinction can be made between them. If all other characters were disregarded, it would be possible to re-align the species into several genera on the basis of habitat preference, but genera erected on this criterion would be ex- tremely artificial. Either of these alternatives, particularly in the light of the morphologic data presented in this report, illustrates the tangles that ensue from application of a strictly ecologi- cal concept of the genus. The statement by Lowe and Norris (appar- ently based upon their evaluation of ecologic features) that Petrosaurus is not closely related to either Crotaphytus (as postulated by Mittle- man) or to Uta, needs no further comment here. Remarks on the Classification of North American Iguanids The principal argument advanced by Mittle- man (1942) for the division of the genus Uta into four genera was his idea of the phylogeny of the several species groups. His system of clas- sification was based upon the assumption that j the North American iguanids form a natural group of genera and that this stock includes two related but divergent evolutionary lines. Mittle- man suggested that the genus Ctenosaura repre- sents the primitive ancestor from which both lines evolved. One of these stocks contained the genera (in approximate order from primitive to advanced) Dipsosaurus, Sauromalus, Calli- saurus, Holbrookia, Uma, Crotaphytus and the nominal genera Petrosaurus and Streptosaurus. The other group included Phrynosoma, Scelo- porus, Sator, Urosaurus and Uta. Smith (1946, p. 92) retained Mittleman’s basic arrangement but added Leiocephalus Gray, 1827, to the Phrynosoma-Uta line. In the preceding sections of this report, in- formation is presented to substantiate Mittle- man’s concept of Urosaurus and Sator as allies of Sceloporus. However, all other data accumu- lated during an investigation of this problem are in strong contradiction to Mittleman’s and Smith’s basic classification of northern iguanids. Evidence at hand clearly indicates that the con- sideration of the North American iguanids as a natural inter-related group is without factual foundation. Because my views are in sharp con- trast to those of Mittleman it has been neces- sary to present a summary of tentative conclu- sions regarding the relationships of these lizards. Conclusions are based upon available informa- tion in the literature (especially Boulenger, 1885; Cope, 1900; Camp, 1923) and on a preliminary evaluation of skeletal and other morphological features. The classification outlined is therefore a tentative one to be modified in its details by later work. The main lines of evolution, how- ever, appear to be clearly recognizible, and it is hoped that my arrangement will stand scrutiny better than that proposed by Mittleman. Insofar as can be determined at this time, the so-called Nearctic iguanids form two diverse groups that can be only distantly related. These two sections are distinguished by marked dif- ferences in vertebral and nasal structures and include several genera not usually recognized as being allied to Nearctic forms. No species inter- mediate in significant characters has been found to bridge the gap between the two lines. Since a thorough revision of the entire family would be necessary to establish the exact status of the suprageneric groups, no attempt has been made to place them in a definite classificatory category. One of the primary divisions in the Iguanidae, represented by a number of genera in North America, is a stock characterized as follows: 1958] Savage: lguanid Lizard Genera Urosaurus and Uta 49 Vertebrae: each dorsal vertebra provided with zygosphenes and zygantra in addition to the zygapophyses. Nasal structure: nasal organ of the relatively simple S-shaped type, concha present (Dipso- saurus- type of Stebbins, 1948, p. 209). This section, hereafter referred to as the iguanine group, includes the following genera: Amblyrhynchus Bell, 1825 Brachylophus Cuvier, 1829 Conolophus Fitzinger, 1843 Crotaphytus Holbrook, 1842 Ctenosaura Wiegmann, 1828 Cyclura Harlan, 1824 Dipsosaurus Hallowell, 1854 Enyaliosaurus Gray, 1845 Iguana Laurenti, 1768 Sauromalus Dumeril, 1 856 Although I have not been able to examine the nasal structure of Amblyrhynchus, Brachy- lophus, Conolophus and Cyclura, these genera have the typical iguanine vertebrae with zygosphenes and zygantra. Their agreement with other members of the group in this regard and their close similarity in basic features make it probable that they possess S-shaped nasal organs. Additional genera may be added to this section when their skeletons and nasal struc- tures have been studied. The second group, essentially North American in distribution, is characterized by: Vertebrae: dorsal vertebrae without zygo- sphenes and zygantra. Nasal structure: nasal organs of the sink-trap type, no concha (Uma- type of Stebbins, 1948, p. 205). This section, hereafter called the sceloporine line, contains: Callisaurus Blainville, 1835 Holbrookia G irard, 1851 Phry nosoma Wiegmann, 1828 Sator Dickerson, 1919 Sceloporus Wiegmann, 1828 Uma Baird, 1858 Urosaurus Hallowell, 1854 Uta Baird & Girard, 1852 Illustrations of the differences in vertebral and nasal structure are given in Text-figs. 3 & 4. It should be noted that in some species of Phryno- soma and Sceloporus, a vertical facet is present on each side of the neural lamina at the anterior end of the vertebrae in the same position where Text-fig. 3. Anterior portion of dorsal region of vertebrae in iguanid lizards. A. Diagram of scelo- porine vertebra of Uta mearnsi. B. Diagram of iguanine vertebra of Crotophytus wislizeni. The let- ter Z lies adjacent to one of the zygosphenes. zygosphenes are developed in iguanine lizards. There are no zygantra in species with these facets, and the latter structures do not appear to be morphologically similar to true zygo- sphenes, which are horizontally flattened and markedly projected anteriorly from the base of the neural spine. In addition to the primary differences listed above, the two groups differ from one another in several general tendencies that hold for a majority of genera. Iguanine line— Teeth usually on pterygoid (usually on palatine as well in Crotaphytus wis- lizeni) ; the small parietal foramen usually pierced in frontal or in suture between frontal and parietal bones; parietal bone thick; pectoral girdle usually with primary and secondary cora- coid foramina, scapular and scapulocoracoid foramina also present (Text-fig. 5); head scutel- lation essentially a homogenous group of small scales not arranged into definite series; inter- parietal scale small, not markedly larger than adjacent head scales; usually a mid-dorsal crest of enlarged scales. Sceloporine line.— Never any palatal teeth; usually a large parietal foramen pierced in a thin membraneous parietal bone; never any sec- ondary coracoid foramen in pectoral girdle, scap- ular foramen often absent (Text-fig. 5); head scutellation usually a heterogeneous mixture of enlarged and smaller scales arranged in definite series; interparietal scale usually enlarged, much larger than adjacent scales; never a mid-dorsal crest of enlarged scales, although paravertebral scales may form an enlarged series. Table 1 indicates the distribution of these features in the individual genera. The evolutionary significance of the develop- ment of the specialized vertebrae with zygo- sphenes and zygantra and the divergent types of nasal structure are not certainly known. The 50 Zoologica: New York Zoological Society [43: 2 Text-fig. 4. Structure of nasal organs in iguanid lizards. A. Lateral view. B. Dorsal view. C. Cross-sectional view. Figures on the left of the sink-trap nasal organization of Callisauriis draconoides typical of the sceloporine line. Figures on right of the S-shaped nasal organization of Dipsosaurus dorsalis typical of the iguanine line. Abbreviations indicate the most important parts as follows: C. concha; E. external naris; I. internal naris; N. nasal cavity; O. vestibule; P. palatine fold; S. nasal septum; W. nasal passage. All figures after Stebbins (1948) . vertebral modification which provides for two additional points of contact and support between vertebrae probably has something to do with the large size attained by most iguanine lizards. The zygantra are significantly reduced in size in Crotaphytus, the genus including the species having the smallest adult size within the section. Stebbins (1948, p. 213), the original discoverer 1958] Savage: Iguanid Lizard Genera Urosaurus and Uta 51 of the differences in nasal structure, has con- sidered at length the possible functional signif- icance of the sink-trap nasal arrangement. He concludes that this feature is an evolutionary adaptation to intensification of the problem of cleansing inspired air in arid enviroments and under circumstances where the lizard buries it- self in the soil. The S-shaped nasal structure of the iguanine line is interpreted by Stebbins to represent a somewhat specialized stage inter- mediate in its adaptation for an arid enviroment between the relatively unmodified structures of other lizards and the complex condition in the sceloporines. The genus Leiocephalus suggested by Smith (1946, p. 92) as a possible ally of Sceloporus has an unmodified nasal organization totally un- like that found in either the iguanines or scelop- orines. Leiocephalus does not appear to be particularly closely related to any of the genera considered in this report. It does not seem advisable at present to specu- late on the relationships of the iguanine lizards, due to lack of adequate material. Final decisions on the phylogeny of the sceloporine section must also await additional research. However, a tentative scheme of relationships within the latter group has been drawn up and is presented in Text-fig. 6. Table 1. Characteristics of Iguanine and Sceloporine Lizards1 Group Pterygoid Teeth Parietal Parietal Foramen Pectoral Foramina s S-C c C' Head Scales Inter- Parietal Scale Mid- Dorsal Crest Iguanines: Amblyrhynchus + (B) T sm. F or F-P X X X X H sm. + Brachylophus + T sm. F or F-P — — — — H sm. + Conolophus + (B) T sm. F or F-P X X X X H sm. + Crotaphytus + T sm. F or F-P X X X X H sm. — Ctenosaura + T sm. F or F-P X X X X H sm. + Cyclura + T sm. F or F-P X X X X H sm. + Dipsosaurus + (B) T sm. F X X X o H sm. + Enyaliosaurus + T — — — — H sm. + Iguana + T sm. F or F-P X X X X H sm. + Sauromalus + T sm. F or F-P X X X X H sm. — Sceloporines: Callisaurus t l.P X X X o h i. Holbrookia — t 1. P X X X o h i. — Phrynosoma — T sm. F-P X X X o H l.-sm. — Sator — t l.P o X X o h 1. — Sceloporus — t l.P X X X o h 1. — LJma — t l.P X X X o h 1. — Urosaursus — t l.P o X X o h 1. — Uta — t l.P o X X o h 1. — 1 The following list indicates the meaning of the symbols utilized in the table: + = present — = absent (B) = according to Boulenger (1883) T — thickened t = thinned sm. = small 1. = large F = frontal bone P = parietal bone F-P = suture between frontal and parietal bones S = scapular S-C = scapulocoracoid C = coracoid (primary) C' = coracoid (secondary) X = present O = absent H = homogeneous fa = heterogeneous 52 Zoologica: New York Zoological Society [43: 2 Text-fig. 5. Pectoral girdles of iguanid lizards in lateral view. A. Uta mearnsi. B. Crotophytus wislizeni. The letters S and C indicate the suprascapula and scapulocoracoid respectively. The numbered structures are: 1. scapular foramen, 2. scapulocoracoid foramen, 3. primary coracoid foramen, 4. secondary coracoid foramen. Two major subdivisions are recognized within this group, based upon the type of sternal ar- rangement. Within the line having a utiform sternum, two distinct stocks are indicated. One of these is represented by the highly specialized genus Phrynosoma, which lacks xiphisternal ribs, has bony spines projecting from the skull and exhibits a very peculiar hyoid apparatus. The other group contains the highly specialized genera Callisaurus, Holbrookia and Uma and the less specialized but probably more recently evolved genus Uta. Within Uta, the subgenus Petrosaurus appears to be most primitive al- though highly adapted for a rock habitat. The genera Urosaurus, Sator and Sceloporus are closely allied and differ from other sceloporines in having a urosaurine type of sternum. Scelo- porus presumably is the most primitve genus, with the other two groups apparently derived from it. Generic Descriptions The genera Urosaurus and Uta have never been adequately characterized. To rectify this situation these groups are briefly described below: Both genera share the following features in common: skull not produced posteriorly into a projection or spines; premaxillary teeth conical; anterior maxillary teeth simple, posterior maxil- lary teeth weakly triconodont; mandibular teeth simple anteriorly, weakly triconodont pos- teriorly; no teeth on palatine or pterygoid; parietal organ piercing the parietal bone pos- terior to frontoparietal suture; parietal very thin in region about parietal foramen; vertebrae without zygosphenes and zygantra; no scapular foramen, a scapulocoracoid foramen, a primary coracoid foramen, no secondary coracoid foramen; one or two sternal fontanels; three or four sternal ribs; xiphisternal ribs present; no parasternal ribs. Nasal organ of the sink-trap type; no concha. Rostral a well-developed scale; supralabials not imbricate, supercilliaries imbricate; inter- parietal scale large; tympanum present; auricu- lar scales enlarged; a distinct gular fold but no gular pouch or pocket; digital lamellae not ex- panded to form pads, strongly keeled; toes with- out lateral fringes of small scales; no mid-dorsal crest of enlarged scales; some scales in para- vertebral region are usually enlarged in Urosaurus. Genus Urosaurus Hallowell, 1854 Type of genus.— Urosaurus graciosus Hallo- well, 1854, by monotypy. Distinctly different from all other iguanids in the characters mentioned above and in: ( 1 ) pec- toral girdle of urosaurine type; lateral xiphis- ternal ribs present; (2) no supranasal scales. Included species. — Urosaurus auriculatus (Cope, 1871); Urosaurus bicarinatus (Dumeril, 1856); Urosaurus clarionensis (Townsend, 1890); Urosaurus gadovi (Schmidt, 1921); Uro- saurus graciosus Hallowell, 1854; Urosaurus irregularis (Fischer, 1882); Urosaurus micro- scutatus (Van Denburgh, 1894); Urosaurus nigricaudus (Cope, 1864); Urosaurus ornatus (Baird & Girard, 1852); Urosaurus unicus (Mit- tleman, 1941). 1958] Savage: lguanid Lizard Genera Urosaurus and Uta 53 Text-fig. 6. Phylogenetic diagram indicating the suggested relationships between the genera of sceloporine line. Genus Uta Baird & Girard, 1852 Type of genus.— Uta stansburiana Baird & Girard, 1852, logotype by subsequent designa- tion of A. E. Brown, 1908, p. 117. Distinct from groups with which it might be confused as indicated above and in having: (1) pectoral girdle of utiform type; no lateral xiphisternal ribs; (2) supranasal scales sepa- rating nasals from internasals. Included species.— Subgenus Petrosaurus Bou- lenger, 1885: Uta mearnsi Stejneger, 1894; Uta slevini Van Denburgh, 1922; Uta thalassina Cope, 1863 (type of subgenus by monotypy). Subgenus Uta Baird & Girard, 1852: Uta concinna Dickerson, 1919; Uta martinensis Van Denburgh, 1905; Uta nolascensis Van Den- burgh & Slevin, 1921; Uta palmeri Stejneger, 1890; Uta squamata Dickerson, 1919; Uta stans- buriana Baird & Girard, 1852 (type of sub- genus); Uta stellata Van Denburgh, 1905. Acknowledgements It is with a great deal of pleasure that I record my indebtedness to the following persons, who have been most helpful in providing material and discussing various aspects of the problem. Mr. Bayard H. Brattstrom, Adelphi College; Mr. D. Dwight Davis, Chicago Natural History Mu- seum; Dr. Theodore Downs, Los Angeles County Museum; Dr. Max K. Hecht, Queens College and the American Museum of Natural History; Dr. Robert C. Stebbins, Museum of Vertebrate Zoology, University of California; Dr. Charles F. Walker, Museum of Zoology, University of Michigan. The Revolving Research Fund of the Ameri- can Society of Ichthyologists and Herpetologists, then under the chairmanship of Fred H. Stoye, gave financial assistance that made possible ex- amination of collections in Ann Arbor, Chicago and New York. Finally I wish to thank Mr. Charles E. Shaw of the Zoological Society of San Diego for much valuable skeletal material, for the opportunity to examine his extensive collection of skulls from lizards of the genera Crotaphytus and Sauronialus and for most generously reading over the completed manuscript. His comments and criticisms of the study based upon his wide knowledge of these lizards is greatly appreciated. Literature Cited Bolin, Rolf Ling 1936. A method of preparing skeletons of small vertebrates. Sci., 82 (2132): 446. Boulenger, George Albert 1885. Catalogue of the lizards in the British Museum (Natural History). Second edi- tion. Taylor and Francis Ltd., London, 2: xiii -f- 497, 24 pis. Brown, Arthur Edward 1908. Generic types of Nearctic Reptilia and Amphibia. Proc. Acad. Nat. Sci. Phila- delphia, 1908: 112-127. Camp, Charles Lewis 1923. Classification of the lizards. Bull. Ameri- can Mus. Nat. Hist., 48, art. 11; 289-481, text-figs. 1-112. Cope, Edward Drinker 1900. The crocodilians, lizards and snakes of North America. Ann. Rep. United States Nat. Mus. for 1898, pt. 2: 153-1270, pis. 1-36, 347 text-figs. 54 Zuologica: New York Zoological Society [43: 2 Lowe, Charles Herbert 1955a. A new subspecies of Urosaurus graciosus Hallowell with a discussion of relation- ships within and of the genus Urosaurus. Herp., 11, pt. 2: 96-101, 1 text-fig. 1955 b. Generic status of the aquatic snake Tham- nophis angustirostris. Copeia, 1955, no. 4: 307-309. Mayr, Ernst 1942. Systematics and the origin of species. Columbia Univ, Press, xiv + 334, 30 text- figs. Mittleman, Myron Budd 1942. A summary of the iguanid genus Uro- saurus. Bull. Mus. Comp. Zool., 91, no. 2: 105-181, pis. 1-16, text-figs. 1-11. Oliver, James Arthur 1943. The status of Uta ornata lateralis Bou- lenger. Copeia, 1943, no. 2: 97-107, 1 text-fig. Schmidt, Karl Patterson 1953. A check list of North American amphibi- ans and reptiles. 6th ed. American Soc. Ich. Herp., viii + 280. Smith, Hobart Muir 1946. Handbook of lizards. Comstock Publ. Co., Ithaca, New York, xxi + 557, 135 pis., 136 text-figs., 41 maps. Smith, Hobart Muir, & Edward Harrison Taylor 1950. An annotated check list and key to the reptiles of Mexico exclusive of the snakes. Bull. U.S. Nat. Mus. no. 199: v + 253. Stebbins, Robert Cyril 1948. Nasal structure in lizards with reference to olfaction and conditioning of the in- spired air. American J. Anat., 82, no. 2: 183-222, 9 text-figs. 1954. Amphibians and reptiles of western North America. McGraw-Hill Book Co., New York, xxii + 528, 104 pis. 52 text-figs. Stejneger, Leonhard Hess, & Thomas Barbour 1943. A check list of North American amphibi- ans and reptiles. 5th ed. Bull. Mus. Comp. Zool., 93, no. 1: xix + 260. 3 Feeding Habits of the Northern Water Snake, Natrix sipedon sipedon Linnaeus1, 2 E. E. Brown Biology Department, Davidson College, Davidson, North Carolina Introduction THE present paper summarizes the food content of 207 stomachs of the northern water snake, Natrix sipedon sipedon Lin- naeus, from central New York and northern Michigan, and presents additional material on the feeding habits of this species. These stomachs became available to me be- tween 1933 and 1938. Most of the New York stomachs are from the general vicinity of Ithaca, in the central part of the state. In Michigan, work centered at the University of Michigan Biological Station at Douglas Lake and was confined mostly to the two northernmost coun- ties of the Lower Peninsula and adjacent islands. We now have a fair amount of information on the food of the northern water snake. Surface ( 1906) reported upon an unstated number (ap- parently about 30) of Pennsylvania stomachs. Uhler, Cottam & Clarke (1939), reporting upon 30 stomachs from the George Washington Na- tional Forest in western Virginia and adjacent West Virginia, found that fishes made up 61 per cent, and amphibians 35 per cent, of the total food volume. King (1939), with 48 stomachs from the Great Smoky Mountains National Park, found fish remains in 29 stomachs and amphibian remains in 17. Raney & Roecker (1947), examining 59 stomachs from western New York streams, found fishes to comprise 96 per cent, of total volume and amphibians 4 per cent. Lagler & Salyer (1947) provided from Michigan a valuable report on 106 stom- achs from trout streams, with 7 per cent, con- xExtracted in part from a doctoral thesis at Cornell University. Contribution from the University of Michigan Bio- logical Station and the Zoological Laboratories, Cornell University. taining trout (19 per cent, of volume) and 73 per cent, containing forage fishes (59 per cent, of volume). They also reported 18 stomachs from inland lakes and 64 from fish hatchery situations. Hamilton (1951b) examined 23 stom- achs of Natrix sipedon insularum and reported the food to be equally fish and amphibians. Additional observations from more limited num- bers of stomachs have been recorded by Ever- mann & Clark (1920) , Boyer & Heinze (1934) , Conant (1938), McCauley (1945), Barbour (1950) and Neill (1951). Minor notes on food items are scattered widely through the liter- ature. Methods Food was secured from specimens by dissec- tion, by voluntary regurgitation and by manu- ally-induced regurgitation. Used carelessly, the latter method might lead to gross inaccuracies, but a high degree of proficiency in its use may be attained, especially with less “muscular” forms such as water snakes and garter snakes. By this method, in several instances, food items smaller than 0.2 cc. in volume were removed from snakes that were needed alive. Sampling checks by x-ray and by dissection suggested that in the majority of cases material that could not be detected and removed by manually-induced regurgitation was too far liquefied to be of value in any event. Findings are presented in the form of (1) percentage frequency of a given item in the total number of food items taken, (2) percent- age of stomachs in which a given item was found, and (3) the percentage of total food vol- ume (both actual and estimated) attributed to a given type of food item. The first two fre- quencies were usually found to give somewhat comparable, but not similar, results. Actual volume means just what it says. Esti- 55 56 Zoologica: New York Zoological Society [43: 3 mated volume, however, means that if an item were 80 per cent, digested, the lost 80 per cent, of volume was restored in the record. In one sense, this procedure would seem to give a much greater degree of accuracy than if the item were credited only with the actual 20 per cent, of its original volume. In practice, the two methods of recording volume were found to give com- parable, but again not similar, results. Workers in food studies do not agree as to the relative importance of the various frequen- cies versus volume. Each factor is obviously important, and each obviously does not tell the whole story (for snakes of all sizes). Percentage frequency of a given item appears to tell more regarding the food preference, or food availa- bility, for the average snake of the study sample. Volume appears to emphasize the predation impact of the study sample on the total avail- able food mass. It may at times over-emphasize the unusual. The two points of view may sometimes give vastly different pictures of the food habits of an animal. In the present study, 23 per cent, of the food items taken consisted of minnows, but minnows comprise only 7 per cent, of the vol- ume of food taken. Lake lampreys were taken less than 1 per cent, of the time, yet they make up 13 per cent, of total volume This study in- volves contents of 207 stomachs with a total volume of 1,372 cc. However, if only 10 selected stomachs were missing from the study material, total volume would be reduced to 653 cc. In some food habit studies there appears to be a real need for a formula that will adequately evaluate both frequency of occurrence and vol- ume, and perhaps other factors as well— if, in- deed, such evaluation is practicable. Habitats Represented Typical small, rocky streams of south-central New York are represented by 120 stomachs (Table 1). Minnows, darters, suckers and scul- pin comprise 72 per cent, of the food items taken, although they make up slightly less than half of the total volume of food. Lake lampreys are important in volume (23 per cent.) but are represented by only two food captures. Game fishes were taken only twice (one fingerling each of brown trout and small-mouth bass). Amphib- ians are represented by 15 per cent, of the food captures. These were mostly frogs, toads and small salamanders ( Eurycea ) . Small New York lakes are represented by only 15 stomachs containing 23 food items (Table 2). Slightly more than half the items taken were amphibians (mostly frogs, toads and tadpoles)., The fish most frequently taken was a catfish (Ameiurus). However, game fishes (Perea and Lepomis), representing 9 per cent, of the food captures, made up 30 per cent, of the food by volume. Michigan lakes of the Cheboygan region are reasonably well represented by 48 stomachs containing 60 food items (Table 3). Minnows and darters together made up 52 per cent, of the items taken, although their volume was only 10.5 per cent, of the total. Game fish (Perea) represent 15 per cent, of the food captures. Al- though the burbot (Lota) was taken only twice and Necturus four times, their volumes (17 and 45 per cent, respectively) were impressive. Great Lakes beaches of the Cheboygan region, and of Bois Blanc Island, Hog Island and Garden Island, are represented by 19 stom- achs (Table 4). These stomachs contain only sculpin (89 per cent, of food captures) and frogs (11 per cent.) . Michigan bog ponds are represented by only 4 stomachs with 11 food items (Table 5). These few suggest a considerable dependence upon amphibian food in this type of habitat. Table 1. Food of 120 Specimens from Streams of Central New York Volume of food taken ( % No. of food No. of stomachs in items taken which the food occurred Actual Estimated (% of 129) (% of 120) (777 cc.) (886 cc.) Minnows [27.9 27.5 [ 7.7 8.3 Darters 7?J21.0 19.1 ,«J 31 3.0 Suckers ( Catostomus ) 72]15.5 16.7 48 ] 35.4 36.0 Sculpin ( Cottus ) [ 7.8 8.3 l 1-4 1.8 Catfish 2.3 2.5 9.3 9.0 Lamprey 1.5 1.7 23.0 20.2 Game Fishes 1.5 1.7 1.2 1.1 Unidentified Fish 7.0 7.5 1.6 3.3 Amphibians 15.5 16.7 17.3 17.3 1958] Brown: Feeding Habits of Natrix sipedon sipedon 57 Table 2. Food of 15 Specimens from New York Lakes No. of food items taken (% of 23) No. of stomachs in which the food occurred (% of 15) Volume of food taken (%) Actual Estimated (122 cc.) | (141 cc.) Catfish ( Ameiurus ) 21.7 26.7 18.6 16.9 Game Fishes 8.8 13.3 30.4 28.5 {Perea, Lepomis) Sculpin ( Cottus ) 4.4 6.7 0.9 1.0 Unidentified Fish 13.0 20.0 0.2 0.4 Amphibians 52.0 60.0 50.0 53.6 Table 3. Food of 48 Specimens from Michigan Lakes Volume of food taken (% No. of food No. of stomachs in items taken which the food occurred Actual Estimated (% of 60) (% of 48) (407 cc.) (491 cc.) Minnows 36.7 35.4 9.1 9.4 Darters 15.0 18.7 1.4 1.6 Sculpin ( Cottus ) 5.0 6.25 2.2 2.0 Catfish 6.7 6.25 0.3 0.4 Troutperch ( Percopsis ) 5.0 6.25 2.8 2.7 Burbot (Lota) 3.3 4.2 17.4 20.3 Game Fishes (Perea) 15.0 8.3 14.1 12.7 Amphibians 13.3 16.7 52.8 50.8 Table 4. Food of 19 Specimens from Michigan Great Lakes Beaches No. of food No. of stomachs in Volume of food taken (%) items taken (% of 28) which the food occurred (% of 19) Actual (50 cc.) Estimated (60 cc.) Sculpin (Cottus) Frogs 89.3 10.7 89.0 15.8 78.0 22.0 81.0 19.0 Table 5. Food of 5 Specimens from Michigan Bog Ponds No. of food No. of stomachs in Volume of food taken (% ) items taken (% of 11) which the food occurred (% of 5) Actual (17 cc.) Estimated (20 cc.) Frogs & Tadpoles Mudminnow (Umbra) 91.0 9.0 80.0 20.0 90.3 9.7 89.0 11.0 Summary of Food Items A broad picture of water snake food is ob- tained by combining the 207 stomachs from all the above habitats (Table 6). Seventy-nine per cent, of the total 251 food captures (68 per cent, of volume) are seen to involve fish forms; 21 per cent. (32 per cent, by volume) involve am- phibians. Minnows, darters, suckers or sculpin were taken in 61 per cent, of the captures. Catfish and game fishes were each taken in 5 per cent, of the captures. Discussion of Food Findings Fish— All evidence presently available testi- fies to the prominence of fish in the food of the northern water snake— 50 to 96 per cent. It is likely that virtually every species occurring in favorable habitats with the snake may at times fall prey. Minnows, darters, suckers and sculpin 58 Zoologica: New York Zoological Society [43: 3 Table 6. Food of All 207 Specimens Combined No. of food items taken No of stomachs in Volume of food taken (% ) which the food occurred Actual Estimated (% of 251) (% of 207) (1,372 cc.) (1,598 cc.) Minnows 23.1 24.1 7.1 7.5 Darters 14.3 15.4 2.2 2.2 Suckers ( Catostomus ) 8.0 9.7 20.0 19.9 Sculpin ( Cottas ) 15.5 15.0 4.3 4.7 Catfish 4.8 4.8 7.1 6.7 Troutperch (Percopsis) 1.2 1.5 .8 .8 Burbot (Lota) .8 .96 5.1 6.2 Lamprey ( Petromyzon ) .8 .96 13.0 11.2 Mudminnow (Umbra) .4 .48 .12 .2 Game Fishes 5.2 3.9 7.6 7.0 Unidentified Fish 4.8 5.8 .9 1.9 Amphibians 21.1 21.2 31.7 31.7 (doubtless reflecting the availability of these forms) appear to be the fishes most frequently captured, especially in stream habitats, although they do not loom so large in total bulk. It is of considerable interest that, even in the stomachs from Michigan trout streams reported by Lagler & Salyer (1947), 72 per cent, of the food cap- tures involved forage fishes (56 per cent, of vol- ume), while only 6 per cent, involved trout (19 per cent, of volume) . In studies involving habitats of the large lake lamprey, this animal may be expected to rank well from the standpoint of total volume. How- ever, it is likely to be an important food item only for large snakes and then only during the limited period of spawning. How large a snake must be to capture or swallow a lake lamprey has not been determined. Only 7 per cent, of the snakes of this study were as large (900 mm.) as the two that had taken lampreys. It has sometimes been asserted that dead fishes make up a large part of the water snake’s diet, but there is little evidence to support this belief. Several of the fish which I recovered from stomachs apparently had been picked up dead. Dead fish are readily eaten, although some of my captive specimens seemed to prefer fresh or only slightly decayed fish. Alexander’s (1943) and Lagler’s (1943) observations concerning the preference of snapping turtles for fresh meat may be of interest here. However, the water snake can, and does, readily capture live fish, and the taking of large numbers of dead fish is probably exceptional and fortuitous in the average habitat. (See page 60). Trembley (1948), ably seconded by Conant, has done the cause of reptilian conservation and common sense a real service in raising a voice from the heart of the Pennsylvania “bounty country,” pointing out the possible utility of water snakes in the ecology of ponds and lakes. Amphibians— In a general way, the food of the water snake may be said to consist of fishes and amphibians, with the latter occupying a substantial second place. In the present series of 207 stomachs, amphibian material was repre- sented in 21 per cent, of the food captures, in 21 per cent, of the stomachs and in 32 per cent, of total food volume. Frogs and toads together seemed to play a more important part than did salamanders (especially in some lakes), with the latter becoming more important in the food of very young snakes in stream habitats. Tadpoles were taken sparingly in all habitats. The taking of very large salamanders has now been reported a number of times for Necturus (Gentry, 1941; Creaser, 1944; Lagler & Salyer, 1947) and Cryptobranchus (Welter & Carr, 1939; Anon., Penn. Angler, May, 1935), in ad- dition to the instances cited in this paper. Be- cause of the large average size of these animals, they are more likely to be taken by sizeable snakes and to be more conspicuous in the volu- metric results of food studies than their infre- quent capture might warrant. Conant (1938) mentions specimens of Natrix sipedon insularum that would not eat frogs. I have had specimens of N. s. sipedon that seemed unaccustomed to frogs, although it was found that an amphibian meal, even though forced, would often convert these specimens to a diet of either type available. Dunn (1935) was not successful in inducing garter snakes to eat the pickerel frog (Rana palustris). This frog was not found in the snake stomachs of the present study, although some snakes were collected from pickerel frog habi- tats. However, Gentry (1944) found 8 young 1958] Brown: Feeding Habits of Natrix sipedon sipedon 59 pickerel frogs in a specimen of N. s. sipedon and Hamilton (1951a) recorded it from Thamnophis. Among captive snakes that were good feeders, I have had some specimens of both the northern water snake and the eastern garter snake that would, and some that would not, take the pick- erel frog experimentally. Pickerel frogs were accepted by several different snakes on about a dozen occasions. One water snake captured and swallowed a small pickerel frog, then took another away from a garter snake that was at- tempting to swallow it. The frogs were retained and digested in all cases. However, these frogs were usually accepted less eagerly than were other species, and they seemed to be mouthed with a noticeable degree of gentleness and cau- tion during the process of swallowing. After such a frog had been swallowed, the snake usu- ally went through the motions of rubbing the sides of its head against nearby objects. Newts are not a popular water snake food, although snakes and newts are at times abundant in the same habitat. Fitch (1941) and Fox (1952) reported larvae of Taricha from Pacific coast garter snakes. Hubbard (1903) was unable to induce Thamnophis elegans to take adult Taricha, and Fox (1951, 1952) reported it only from the race atratus and from two specimens of Thamnophis sirtalis of the San Francisco area. Hamilton (1951b) reported a specimen of Diem- ictylus from Natrix s. insularum, but did not find it (1951a) in eastern Thamnophis. My own best water snake feeders in captivity could not be induced to take this salamander. One old specimen, accustomed to accepting instantly anything shaken in her direction, grabbed a newt, quickly gulped it down several inches, then suddenly changed her mind and ejected the salamander with such violence that it was flung to a distance of about four feet. The back of the newt was covered with the milky-looking secretion from skin glands. Crayfish— It is likely that the importance of crayfish as water snake food has been uninten- tionally exaggerated. Some of the generaliza- tions about crayfish in the literature seem to be indirectly traceable to Ortmann (1906) and to Atkinson (1901). Unfortunately, the statements in these sources refer to more than one snake species, and seem to be incapable of definite interpretation with regard to the northern water snake. Ditmars (1912) recorded a definite instance of crayfish as northern water snake food. Conant (1938) reported crayfish from Ohio snakes but did not say how often he had found them. The major studies of the food habits of this snake have not included crayfish in the findings. I myself found none, although many stomachs were examined from habitats in which crayfish were abundant. The best feeders among my captive snakes could not be induced to take this type of food. The fact that Barbour (1950) and Neill (1951) report crayfish from sipedon in the mountain region farther south suggests that there may be some regional differences in the importance of this animal as water snake food. Other Vertebrates— That small mammals may very occasionally be taken by the water snake is indicated by Surface’s (1906) record of mead- ow mouse and shrew. Uhler, Cottam & Clarke (1939) reported mammal hairs from two stom- achs, and Lagler & Salyer (1947) reported a rodent trace. Gloyd (1928) did not succeed in interesting captive specimens in warm-blooded prey. I also was unsuccessful here. Conant (1938) found a small northern water snake in a stomach of a snake of the same spe- cies. Uhler, Cottam & Clarke (1939) found a few snake scales in one specimen, and Lagler & Salyer (1947) recognized a fragment of shed skin. Conant & Bailey (1936) reported a fence lizard taken by a captive snake. I have noted only indirect tendencies toward “cannibalism” —when two snakes were attempting to swallow the same food, or when one snake had crawled over fish and therefore carried the odor of the food. I know of no records of the taking of birds by the northern water snake. However, it would not be surprising if this does occur in rare cir- cumstances. Other Invertebrates— Minor amounts of ma- terial representing various other invertebrates have been reported from this snake, to the extent of 2.4 per cent, of the volume of the 30 stomachs of Uhler, Cottam & Clarke (1939) and 1.6 per cent, in the 106 trout stream stomachs of Lagler & Salyer (1947). These items have included young or adults of various insects (Coleoptera, Odonata, Plecoptera, Orthoptera, Diptera, Lepi- doptera), earthworms, leech and millipede. Breckenridge (1944) reported a spider. Arthro- pod material should always be examined criti- cally, since some of it may be traceable to the stomachs of vertebrates that have themselves been preyed upon. King (1939) reported a slug from one stom- ach, and Lagler & Salyer (1947) an aquatic snail. Mr. William C. Wise, of Quentin, Leba- non County, Pennsylvania, informed me (letter) of finding a water snake “captured” by a large aquatic snail. The snail, in retracting its oper- culum, caught the fore part of the snake’s head between shell and operculum. The snake appar- ently could not free itself and smothered. A 60 Zoologica: New York Zoological Society [43: 3 photograph of this specimen appeared in the June, 1939, issue of Pennsylvania Game News. Mr. Frederick Tresselt of Hunting Creek Fish- eries, Thurmont, Maryland, told me (conversa- tion) in 1938 that at least a dozen times he had observed water snakes “caught” by large “Japa- nese snails” in his goldfish ponds. It appears likely that in such cases the water snake has actually attempted to feed upon the snail. This snake-snail relationship has not been checked experimentally. Food of Young Snakes It is desirable to know how the food of very young snakes compares with the broad findings. Of the 207 stomachs, 73 were from snakes known, or estimated, to be in their first year of life (207-380 mm. in length). The findings from these are listed in Table 7. According to these data, fishes are still the most important type of food. Minnows, darters and amphibians together comprise 4 of every 5 food items taken. The amphibians included very small frogs and the slender northern two-lined salamander. Most of the animals captured were quite small, and 90 per cent, of these young snakes contained only a single food item. It was interesting that even these young snakes did not hesitate to take catfish. This practice occasionally results disastrously for a water snake, but it is frequently managed safely. The food of young snakes is important but it should not be over-emphasized. It appears im- portant because, if the snakes of the present study are a representative sample, about 40 per cent, of all water snakes would be members of this first-year class. However, on the basis of the same sample, I estimate at present that snakes of this size would consume only about 8 per cent, of all the food taken by water snakes. Food Variation under Special Conditions Fish hatcheries often offer situations where access may be had to dense populations of few species and where fish may be captured with greater ease than elsewhere. Under such cir- cumstances one might expect to find snakes gorged with the species at hand, and stomach contents would not yield a completely natural picture of food habits. Lagler & Salyer (1947) examined many stomachs of snakes taken at trout rearing stations in Michigan. Slightly more than half of these (56 per cent.) contained the fish being propagated. The occasional practice of planting large num- bers of hatchery-reared fish without sufficient scattering may also offer excellent opportunity for predation by water snakes. According to A. S. Hawkins, a party of local sportsmen planted trout along the Stein Kill near Chatham, New York, in early August, 1934. Little more than an hour later Harry Carr, a member of the party, returned to one of the points of planting. He found a water snake containing 5 two-inch fin- gerlings. Conant (1938) mentions a number of in- stances of the eating of dead fish by Natrix sipedon insularum. Around Lake Ontario and in the Finger Lakes region of New York water snakes would be expected to take advantage of the extensive dying off of the alewife (Pomolo- bus pseudo-harengus) , during such seasons as this occurs. G. F. MacLeod told me of seeing a number of snakes on a cove at the north end of Seneca Lake gorging themselves on these fish as the latter drifted in to shore. Isolated pools during drought conditions, pools from which metamorphosing frogs are emerging, the presence of spawning lampreys in the spring, assemblages of breeding frogs and toads, are examples of other local conditions that might temporarily influence water snake diet. Capture of Food Wilde (1938) found that chemical sense, op- erative through the tongue, lips and organs of Jacobson, is extremely important in feeding reactions of Thamnophis s. sirtalis. Fox (1952) considered odor particularly and also sight to be used in food recognition by his garter snakes of the Thamnophis elegans group. Methods used by water snakes in capturing prey will not be fully understood until a thorough study has been made of the relative importance of the various senses employed. Present information seems to indicate: 1. That the sense of touch may be extremely important in a large proportion of under-water hunting. 2. That sight may be of some importance in daylight under-water hunting with reference to the detection of near moving objects. 3. That a submerged snake, especially if it is moving, apparently does not see objects above the surface. 4. That near moving objects are readily de- tected in terrestrial operations. 5. That the extent of the role of chemical sense in under-water hunting is problematical. This sense is efficient on land, at least with re- spect to some odors, but its degree of utility under natural conditions is uncertain. Many ob- servers have noted the confusion into which a group of hungry water snakes may be thrown when dead fish is placed in their cage. I have 1958] Brown: Feeding Habits of Natrix sipedon sipedon 61 Table 7. Food of 73 Young Specimens (All Habitats) Volume of food taken (%) No. of food No. of stomachs in items taken which the food occurred Actual Estimated (% of 86) (% of 73) (72 cc.) (91 cc.) Minnows (25.6 27.4 [26.0 28.0 Darters 80] 30.2 31.5 71] 27.0 25.0 Amphibians [24.4 24.6 [18.0 19.7 Sculpin ( Cottas ) 7.0 6.8 10.0 9.8 Suckers ( Catostomus ) 4.6 5.5 7.0 6.0 Catfish 3.5 4.1 1.7 1.6 Troutperch (Percopsis) 1.2 1.4 5.6 5.0 Game Fish ( Micropterus ) 1.2 1.4 5.0 3.9 Unidentified Fish 2.3 2.7 0.3 1.0 had captive specimens which would tear to bits paper that had been wrapped around fish; ones which, after I had handled fish, would greedily swallow a finger as far as anatomy would permit. It is difficult to describe anything that might be called typical procedure in the taking of fish, and I am aware that it may be neither advisable nor possible to classify hunting methods. How- ever, for organizational convenience, this sub- ject is dealt with under several headings. Groping or Exploratory Method of Hunting — Abbot (1884) early called attention to this method. He remarked about the haphazard way in which the snake hunted, not seeming to single out any particular fish in a group. It “opened its mouth and left the rest to luck.” Evans (1942) carefully described a number of instances of hunting by water snakes of sev- eral species. In most cases the snake swam or drifted in the water near the surface. “The head was submerged and the mouth kept open wide as it swept through the water from one side to the other in a continuous series of figure eights, the entire body following the path of the head.” Although this exploratory method would seem particularly suitable to night fishing, Evans’ account points out that it may also be used by day. Stoner (1941) made an observation that may apply to this method, but it appears incapable of exact interpretation. Kellogg & Pomeroy (1936), in their maze experiment, gained the impression that a snake “felt” its way through the maze by pushing against the sides with its nose. In 1937 and 1938 I made many observations on water snakes ac- tively fishing in tanks and aquaria. Motion picture films of some of these operations were also studied, and the actions of the snake gave a strong impression that it was “groping” for the prey. Often it did not move toward nearby fish easily within reach (even though the water was clear and visibility presumably good). The mouth might or might not be open but it fre- quently was. With rather deliberate movements, the snake “felt” around in an almost aimless manner, first in one direction then in another. However, the instant any part of the head or neck touched a fish, a wild grab was made in that direction. The efficiency of the method is surprising. If a fish was grasped, the snake often bent its head around at a sharp angle and pressed the prey against its body till it secured a firmer grip. An apparent modification of the groping method of hunting is that of exploring under rocks and other objects on the bottom. Many observers have encountered evidence of this activity. The large proportion of sculpin, dart- ers and two-lined salamanders found in snake stomachs seems to support the idea that this type of hunting may be very important in suit- able habitats. Uhler, Cottam & Clarke (1939) reported a case in which a brook trout was caught beneath a rock. Direct Attack Method —Were the sense of sight seems to play a dominant part. Fishing tactics may at times be a mixture of this and the exploratory method, and probably some of the following examples might be interpreted as illustrating either type of procedure. DeKay (1842) mentioned a water snake that was seen to fall from a bush into a stream and seize one of a number of chubs that were swim- ming by. In describing the capture of a Notropis procne by a snake, Cope (1869) said, “approaching cautiously, he struck right and left below the surface, as the minnows passed him, but often fell short.” On the Cayuga Lake inlet about 1916, A. H. Wright saw a water snake swim out from shore, seize a 14-15-inch lamprey on its nest and drag it back to shore. 62 Zoologica: New York Zoological Society [43: 3 In the summer of 1926 on a small tributary of Wolf Creek, Wyoming County, New York, P. W. Claassen and others were watching a brook trout about six inches long. Suddenly, from the bordering vegetation a few inches away, a water snake lunged, grasped the trout amid- ships and swam off with it. S. C. Vanderbilt, of Clyde, New York, for several minutes watched a water snake in the edge of aquatic vegetation as it seized tadpoles that swam by. The snake was killed and found to contain 16 tadpoles. While observing spawning minnows near Ithaca, New York, in May, 1936, W. J. Koster saw three small snakes fishing among the min- nows “without success, although they made quite a few lunges at the fish.” C. W. Creaser observed a snake 15 feet out from shore on Burt Lake, Michigan, striking among members of a school of minnows. The snake was neither touching bottom nor anchored to any object in the water. W. J. Koster, sitting motionless on the bank of Danby Creek, near Ithaca, in May, 1936, watched a water snake crawl ashore nearby. “It was just about settled, apparently to bask in the sun, when two Notropis cornutus began splash- ing in very shallow water. The snake immedi- ately lifted its head, which had been about an inch from the ground, and turned in the direc- tion of the disturbance. After watching for several minutes it crawled into the water and attempted to catch a fish.” Raney & Roecker (1947) observed the band- ed water snake “actively chasing and capturing fishes” in Erie County, New York. In Delaware County, New York, on the inlet of Silver Lake, I watched a two-foot snake at- tempting to capture fish at midday in August, 1 935. The snake was in a pool eight or ten inches deep. The caudal end of the body extended under a log, perhaps for anchorage. The rest of the body was moving around in the water in a manner that appeared to be partly exploratory, partly directed. Of the dozen or so small suckers, horned dace and mad toms in the pool, one or more were almost continually approaching the snake, apparently in a state of curiosity. When one came within range the snake would make a lunge for it. Although this occurred a num- ber of times during the two or three minutes before the snake became alarmed, the fish were a little too quick each time. The procedure clearly demonstrated, however, how an unwary fish might easily fall victim. The fish never ex- hibited wild excitement or dashed about in the pool. They gave the impression of mild curiosity toward the snake. When it lunged, they would simply dart a short distance out of reach, often “gathering around” again within a few moments. The interesting feature of this incident is not the behavior of the snake but rather that of the fish. It suggests that, if game fish behave in the same manner in the presence of a snake, their speed might be of little advantage to them. There is no present indication that water snakes actually pursue fish. Swimming speeds of the snake are far too slow, as compared with that of almost any fish. Under the circumstances here indicated, the competing characteristics of the two animals probably are the speed of lunge of the snake and the alertness and speed of take-off of the fish. Deep Water Hunting— While most hunting by water snakes is probably done in relatively shal- low water, they are capable of working at greater depths. The finding of Lota maculosa in stom- achs suggests, but by no means proves, this. At Hook Point on Douglas Lake, Michigan, in 1932, Sol R. Baker and a group of students saw a water snake emerge some distance from shore with a struggling Necturus and swim to land. At the point where the snake appeared, the depth was somewhere between 10 and 25 feet. From the same lake I have another account of a snake that was seen to swim straight out from a rockpile at the water’s edge (near the Biologi- cal Station boathouse). In about ten minutes it came ashore again with a Necturus. The water in which the prey was secured could scarcely have been less than 8 to 10 feet in depth. Again at Douglas Lake (1933), F. C. Gates saw a snake swim to shore from beyond the rim of the beach shelf carrying a live fish. The increase in depth beyond the rim of the shelf is extremely rapid. But whether the fish was secured at the surface or near the bottom is not known. Is Prey Brought Ashore For Swallowing?— The prey may, or may not, be brought ashore to be swallowed. If its size, compared with that of the snake, is large enough to require consid- erable time and effort for overpowering and swallowing, it is almost certain to be dragged ashore. A small snake would probably find this necessary with almost any prey. So also would a large snake with a good-sized fish, lamprey or Necturus. However, the swallowing of a min- now by a large snake may be a matter of only a few seconds, and I have seen a snake swallow a good-sized green frog under water in a large outdoor tank. On dozens of occasions I have observed this with minnows or other small fish in aquaria or tanks. Capture of Prey on Land.— This probably in- volves a relatively small proportion of the food 1958] Brown: Feeding Habits of Natrix sipedon sipedon 63 of the average water snake. Sight seems to be the sense of prime importance here, with chemi- cal senses playing a secondary role at times. The feeding of frogs to water snakes or garter snakes in large enclosures is likely to illustrate well the comparative safety of the prey, just so long as it remains motionless. But the moment it leaps, the snake may take notice and be in hot, though at times clumsy, pursuit. If the frog is a power- ful and persistent jumper, it may outdistance, or “lose,” the snake, but if it hesitates between jumps, it is likely to be overtaken. G. J. Lever- see told me that he once observed a water snake on the bank of a stream in Greenbrier County, West Virginia. A frog, which apparently had been sitting very near the snake, leaped for the water. So instantaneous was the snake’s reaction, that the frog was grasped in mid-air. He thought that the snake had been aware of the frog’s presence. Method of Swallowing Food.— Observation of water snakes in captivity is likely to give the impression that the method of swallowing is altogether haphazard and that the prey is as likely to go down tailfirst as headfirst, and at times even sidewise. However, of 200 swallow- ings of fish, frogs, toads and salamanders in nature— checked by examination of stomach contents-the prey went down headfirst in 80 per cent. (160) of the cases, tailfirst in 18.5 per cent, and otherwise in 1.5 per cent. Fish seem most likely to go down headfirst, sala- manders least likely to do so. For details see Table 8. Table 8. Head-tail Orientation of Food in 200 Swallowings by Water Snakes in Nature Food Type Headfirst Tailfirst Otherwise Fish 140 27 2 Salamanders 9 24 0 Frogs and toads 11 8 1 Totals 160 37 3 Percentages (80% ) (18.5%) (1.5%) Since a fish is a form of food that would pass down more easily headfirst than tailfirst, I once supposed that it might be possible to check on the learning ability of the water snake by com- paring the proportion of food swallowed head- first with the age of the snakes involved. How- ever, the actual condition seems to be about the opposite of what might at first be expected. A snake swallowing food is usually dealing with a more or less elongated object. It has of neces- sity to work the object around in its mouth and begin swallowing from an end. If one end does not seem to go down readily, the other is often tried. A small snake in its first year of life may find any minnow that it attempts to swallow such a tight fit that it will go down only headfirst. On the other hand, a larger snake may have a relatively easier time with the swal- lowing process, and is more likely to be able to pass the food down oriented in whatever way it is first tried. This explanation is supported by actual findings. Of 73 young snakes less than 400 mm. in length, 90 per cent. (66) had swal- lowed the prey headfirst. Of the 127 older speci- mens, 74 per cent. (94) had taken the prey headfirst. Diurnality Versus Nocturnality According to Walls (1931 and 1942) Natrix possesses a reasonably typical diurnal colubrid type of eye. However, the highly efficient “grop- ing” method of fishing fits this snake admirably for nocturnal activity as well, apparently with- out the aid of sight. That the activity of Natrix s. sipedon may be to a high degree both diurnal and nocturnal is now well established by rec- ords in the literature and by the experience of many observers. Variations in the degree of diurnality or noc- turnality in a given region are probably largely a matter of weather conditions, season and temperature factors in the available habitat. Swanson (1952) noted in Venango County, Pennsylvania, that the cooler streams were sel- dom frequented by this snake. He noted further that along moderately cool Big Sandy Creek, water snakes were common by day but almost never seen at night. Along Carp Creek, a cool woodland trout stream (water temperatures 52°- 56° F. at times of observation) near the Uni- versity of Michigan Biological Station, I found no water snakes except at the stream’s mouth where it enters Burt Lake. The late Dr. George E. Nichols, who fished the stream for many years, said that he had seen Natrix only at a small stagnant pool adjacent to the creek at one point. These observations merely agree with those of Lagler & Salyer (1947) for cold, shaded streams. After hot (90° F.) mid-summer days at Ithaca, New York, I found water snakes ac- tive at night in water temperatures of 12° -IV F. In my experience this snake is most active at temperatures between 70° and 80° F., tending to seek shelter at air temperatures much above 80° F., and with activity ceasing altogether at temperatures in the low fifties. It will no doubt be found that this snake is more diurnal in the northern portions of its range and during spring and fall, more nocturnal farther south and during mid-summer months. 64 Zoologica: New York Zoological Society [43: 3 Maximum Size of a Meal Here we are not dealing with the average size of meals in nature, but with the maximum size under the most favorable circumstances pos- sible. The size of the meal is expressed in terms of percentage of the weight of the snake, previ- ous to the meal. The question is, what per- centage of its own weight in food may a snake take (voluntarily accept) at one time? Uhler, Cottam & Clarke (1939) reported a water snake swallowing a bullhead whose weight was 40 per cent, of that of the snake. In 7 snakes collected in nature that appeared to contain unusually large amounts of food, I found that the food varied from 11 to 37 per cent, of the weight of the snake, with a mean of 20 per cent. Nineteen empty snakes that were good feed- ers in captivity3 were gorged until they refused to take more food. The food taken ranged from 11 to 43 per cent, of the weight of the snake, with a mean of 26 per cent. ± 2.35 and a stand- ard deviation for the distribution of 10.2. It is interesting to note that a snake may at one meal take food amounting to 40 per cent, of its own weight. It seems reasonably certain that this is a near maximum figure for this water snake— a figure which may occasionally be reached but which is probably seldom exceeded. The individual in this series that took a meal amounting to 43 per cent, of its weight later disgorged part of the food, apparently because it had taken too much for comfort. This same individual on three other occasions took maxi- mum meals amounting to 32, 34 and 35 per cent, of its weight. Another snake on two occa- sions took meals of 36 and 39 per cent. Most of the other specimens took somewhat smaller meals. It was noted that following these maximum feedings the forward region of the alimentary canal was always left unobstructed. When a snake had fed until it refused further food, its stomach and esophagus were usually gorged to a point 1 to Wi inches (25 to 44 mm.) caudad of the heart. This meant that the heart and at least part of the highly vascular and functional anterior end of the lung were permitted free- dom of movement. In eight cases the length of the unencumbered anterior end of the body equalled 15 to 23 per cent, of the total length of the snake. 3A11 “captive” snakes referred to in this paper in connection with various feeding observations were kept out of doors on the ground. Some were in a roofless wire enclosure, others were in a roofed enclosure that had all sides open. All snakes were exposed to natural conditions of temperature, sunshine and moisture. The enclosures contained logs and boards under which specimens could seek shelter. Average Size of a Meal There is no doubt that under unusually favor- able conditions water snakes will gorge them- selves. Most hatchery men have seen examples of this. Lamson (1935), without stating the source of the record, says, “as many as sixty fingerling trout have been taken from the di- gestive tract of a single snake.” Blatchley ( 1891 ) found seven Rana pipiens in one. I have fed as many as 42 black-nosed dace at one time to a large specimen before the snake finally refused to take more. However, there is much doubt that the aver- age water snake meal in nature consists of such proportions. This doubt is based upon the con- tents of the 207 northern water snake stomachs from nature. Of this number, 183 (88 per cent.) contained only one food organism. The con- tents of the remaining 24 stomachs (12 per cent.) varied from 2 to 7 food organisms, aver- aging 2.8 per stomach. The mean for the entire group of 207 was 1.2 organisms per stomach. It was relatively unusual for a single food or- ganism to be of maximum meal size. In my experience, after a captive snake had been fed a “maximum” meal, it refused to take additional food for three to five days. In other words, it would take no more until gastric di- gestion, at least, was apparently complete. On the other hand, snakes that were fed small or moderately sized meals (i.e., a medium sized frog for a large snake) would continue to feed every day, or at least every other day, almost indefinitely. It is not desired to give the impression that snakes will not feed if they already contain some food. Snakes containing food are not always found lying quietly away under cover. I have collected individuals that were active and ap- parently hunting, although they contained food. Individuals have also been collected containing two or more food organisms that had been cap- tured many hours apart. However, these findings merely further suggest that the average snake probably takes moderate meals at fairly frequent intervals, rather than gorging itself to capacity when it feeds. Maximum Amount Eaten During a Given Period The time demanded by this type of work made it impossible for me to make observations on large numbers of individuals. Nevertheless, records are available for 11 of the best feeders on hand during the summer of 1938. These specimens were so accustomed to human beings and to being fed that they practically never re- fused to eat unless they were already gorged. 1958] Brown: Feeding Habits of Natrix sipedon sipedon 65 Four of the 1 1 were young approaching the end of their first year of life, 2 were second-year individuals, 1 was in its third year, and the re- maining 4 were adults in their fifth or sixth year. The feeding was carried on during July and Aug- ust, which months had the highest mean tem- peratures (73° F.) for the summer. All food taken by the snakes was weighed. All of the specimens were fed fish, except for number 11 which was fed entirely on frogs. In order to make food consumption the maximum possible, each snake was offered all that it would take at each feeding. During 2 of the 9 feeding periods for specimen number 10, it refused to take food, apparently because of shedding complications. The period of experimental feeding lasted from 25 to 55 days in various cases. Final weighing of the specimens was done 4 to 6 days after the last feeding. Feeding data on these specimens are summarized in Tables 9 and 10. The gross amount of food consumed is of interest. For the first-year specimens this aver- aged 61 per cent, of the original weight of the snake per week, or 247 per cent, per month (of four weeks) . For the four adult specimens food consumption averaged 43 per cent, of the origi- nal weight of the snake per week or 174 per cent, per four-week month, with the frog-eater (num- ber 11) consuming a slightly higher percentage than the three fish-eaters. The few specimens of intermediate age exhibited lower, but rather con- sistent, consumption of 30 per cent, per week and 121 per cent, per month. Specimen number 9 may be considered a con- servative example among the adult fish-eaters. This snake consumed 360 grams (12.7 ounces or 153 per cent, of the original weight of the snake) of fish in 28 days. This weight of fish is the approximate equivalent of any one of the following: 164 medium black-nosed dace, or 89 three-inch common shiners, or 34 four-inch horned dace, or 24 three and one-half inch carp, or 1 8 five-inch brook trout, or 6 seven-inch com- mon suckers. Snake number 11 was actually fed 15 frogs of varying sizes during the 25 days. These weighed 507 grams and would be the approxi- mate equivalent of either 14.5 fair-sized Rana pipiens or 6.8 large Rana clamitans. There is certainly much variation among cap- tive snakes in their inclination to feed. Evidence suggests that snakes in the wild state also vary in proportionate amounts of food consumed. This is reflected in varying growth rates among snakes of the same age in the same wild habitat. Nevertheless, the feeding habits of the best cap- tive feeders may provide a rough optimum index to conditions in the wild state. The statement is occasionally made that cap- tive snakes are probably much better fed than are those in the wild state. If winter feeding is disregarded there is reason to doubt this. Snakes in poor physical condition are often seen in cap- tivity. But in my experience the more usual occurrence is for wild specimens to be in ex- cellent condition (aside from occasional heavily parasitized individuals) and to contain extensive visceral fat deposits, even immediately follow- ing hibernation. Table 9. Basic Data on Experimentally-fed Snakes Sex Original Length (mm.) Original Weight (gm.) Times Fed Duration of Feeding Period (wks.) (Young Ending First Year) 1 F 256 5.3 11 7 2 M 278 5.7 10 7 3 F 288 6.7 11 8 4 F 320 10.4 13 8 (Ending Second Year) 5 F 516 32.7 10 7 6 M 475 25.2 6 3% (Ending Third Year) 7 M 607 53.2 8 3% (Adults Ending Fifth or Sixth Year) 8 F 870 237. 9 4 9 F 887 235. 9 4 10 M 770 114. 7 4 11 F 945 255. 10 3y7 66 Zoologica: New York Zoological Society 143: 3 Table 10. Feeding Record of Experimentally-fed Snakes Food Eaten Gross (gm.) In % of Orig. Wt. of Snake Total Per Wk. Per Mo. Per Wk. Per Mo. (Young Ending First Year) 1: 26.6 3.8 15.2 71 286 2: 23.1 3.3 13.2 58 232 3: 30.4 3.8 15.2 57 227 4: 50.1 6.3 25.2 60 242 (Avgs.: ) 32.5 4.3 17.2 61 247 (Ending Second Year) 5: 62.8 9. 36. 27 110 6: 31.2 8.4 33.6 33 133 (Avgs.:) 8.7 34.8 30 121 (Ending Third Year) 7: 58.5 16.1 64.4 30 121 (Adults Ending Fifth or Sixth Year) 8: 403. 101. 403. 42 170 9: 360. 90. 360. 38 153 10: 197. 49. 197. 43 173 11: 507. 127. 507. 50 199 (Avgs.:) 366. 92. 366. 43 174 Growth during Period of Experimental Feeding Growth data for the same 11 snakes during the period of experimental feeding are summar- ized in Table 11. Although the number of specimens is still small, the data exhibit sufficient consistency to suggest that they have some value. Increases in gross length averaged about the same for the first-year specimens as for the adults: 7 mm. per week or 28 mm. per month (of four weeks). (This, of course, means that young would double in length much more rap- idly than would adults). Although the rate of gross length increase is about the same for young and adults in this series, other data indicate that in still older snakes (from 7th or 8th year on) the rate would be somewhat slower. On the other hand, increase in gross weight is seen to be much more rapid in adults than in young. However, increase in weight, in terms of percentage increase over the original weight of the snake, was approximately twice as rapid in the first-year specimens as in the adults. Young individuals of the first three groups consumed about 3 grams of food for each gram of gain in weight. Among the adults, the three fish-eaters averaged, about 50 per cent, greater consumption (4.5 grams) of food for each gram of gain in weight, while the highest consumption (6.9 grams) per gram of gain was registered for the single frog-eater in this particular series. It is impossible to say whether this last item has any significance. Lacking more extensive data on food con- sumption, on ages of specimens and on the effect of hibernation upon weight, I do not con- sider it advisable to attempt to calculate, from the weight of specific snakes, the bulk of fish that may have gone into the make-up of those snakes. Also, it should be noted that the speci- mens recorded here were given maximum feed- ing. Parallel observations might well have been made on specimens subjected to more moderate feeding. Rate of Gastric Digestion When we think of “digestion” in a snake we usually have in mind gastric digestion. When food material leaves the stomach, it usually has become liquefied and its presence is no longer evident from the exterior. The time required for gastric digestion will depend upon at least three important factors: environmental temperature, size of meal and 1958] Brown: Feeding Habits of Natrix sipedon sipedon 67 Table 11. Growth Record of Experimentally-fed Snakes Gain in Wt. (gm.) % Increase over Original Wt. Food Eaten Gain in Length (mm.) Total Per Wk. Total Per Wk. Per Mo. Gain (gm.) Total Per Wk. Per Mo. 1 6.7 .96 126 (Young Ending First Year) 18 72 3.97 60 8.75 35.0 2 6.3 .9 111 16 64 3.66 24 3.5 14.0 3 9.3 1.16 139 17 68 3.27 46 5.85 23.4 4 15.6 1.9 150 19 76 3.21 82 10.4 41.6 (Avgs.:) 1.23 17.5 70 3.53 7.1 28.5 5 19.8 2.8 61 (Ending Second Year) 9 36 3.17 38 5.43 21.7 6: 11.7 3.15 46 12 48 2.67 20 5.38 21.5 (Avgs.:) 2.97 10.5 42 2.92 5.4 21.6 7 20.2 5.6 41 (Ending Third Year) 1 1 44 2.9 11 2.96 11.8 8 75. 18.7 (Adults Ending Fifth or Sixth Year) 32 8 32 5.3 24 6. 24. 9 85. 21.2 36 9 36 4.2 39 10. 40. 10 49. 12.2 43 11 44 4.0 20 5. 20. 11 73. 20.4 29 7 28 6.9 23 7. 28. (Avgs.:) 18.1 ~9 T5 5.1 ~ 28. size of snake. Both of the last two are important because the actual stomach of a snake is a rela- tively short portion of the alimentary tract. When a large meal is taken, some food may occupy the esophageal region anterior to this. But the food remains apparently unchanged until space is available for it in the stomach proper. Although investigations of the rate of diges- tion at various controlled temperatures were not carried out, some observations were made under “normal” mid-summer temperatures at Ithaca, New York. One important difficulty was the determination of the end-point of gastric diges- tion, i.e., the point at which all solid material in the stomach had become liquefied. It was found that this could be determined approxi- mately for most foods by careful manual exami- nation of the gastric region4 at intervals through- out the period of digestion. In a number of cases the observer’s findings were checked by X-ray examination to determine his degree of accuracy. It was noted that about two days might be suggested as an average length of time re- quired for gastric digestion of a moderately large 4Very careful palpation will usually disclose the posi- tion of the gall bladder and pancreas. This is, of course, a useful landmark as it marks the level of the caudal end of the stomach. summer meal. W. A. Kenyon informed me (letter) that garter snakes on which he made observations (1925) seemed to require 2 to 3 days for gastric digestion of a frog. In ten cases water snakes were fed one frog each, the weight of the frog varying from 16 to 38 per cent, of the weight of the snake (mean: 28 per cent.). The time required for gastric digestion ranged from 42 to 66 hours (mean: 50.4 hours) at mean temperatures5 of 75° to 76° F. In three cases snakes were fed large meals of fish. The weights of the meals were equal to 30 to 43 per cent, (mean: 36 per cent.) of the weight of the snake. The time required for gastric digestion ranged from 42 to 60 hours (mean: 49 hours) at mean temperatures of 74° to 75° F. In other miscellaneous observations the rate of gastric digestion tended to be slower at lower temperatures, more rapid at higher tempera- tures (e.g., 73 to 80 hours in four instances at 63° to 64° F.; 29 to 50 hours in four at 79° to 82° F.). r,Mean temperature, as here used, is the average of readings taken at six-hour intervals throughout the 24 hours of a day. 68 Zoologica: New York Zoological Society [43: 3 It was pointed out by Benedict (1932) that probably little, if any, digestion in snakes pro- ceeds at temperatures below 50° to 60° F. This was nicely illustrated for the observer by a large water snake that swallowed a lamprey just at the beginning of an unusually cool period during late May, 1936. After five days the lamprey was removed from the snake’s stomach. It was still only about half digested. The mean temperature for this period had been 55° F. At these lower temperatures snakes also show less inclination to feed with, however, consider- able variation in different specimens. Among good feeders at Douglas Lake in September, 1937, one refused food at 60° F. Several others fed fairly readily at temperatures from 57° to 60° F. The lowest temperature at which any fed was 52° F. Two accepted a fish with some hesitancy and swallowed it laboriously. (One of these same snakes refused food at 54° F. on another occasion). When exposed to a tempera- ture of 45° F. for one hour, no specimens would feed. Collections of snakes taken during, or im- mediately following, periods of cool weather may be expected to show an extremely low per- centage of freshly-taken food— if any is present. Post-Gastric Digestion Few observations were made on the duration of post-gastric digestion, and these were made at unfortunately low temperatures (62° to 65° F.) at Douglas Lake, Michigan, in the early fall of 1937. However, they suggest certain pro- bable features. The crude “marker” method was used. A wad of indigestible mouse fur was in- serted within the body cavity of the small fish (minnow) that made up each meal. The time elapsing from the completion of gastric digestion until the first appearance of fur in the excreta was then recorded. In one instance all the marker material was passed in one fecal sample. In several others, periods varying from 22 to 38 hours elapsed between first and last passages of the marker. This method presents the obvious danger (especially in poorly fed snakes) that the marker may be retained within the bodv after it has reached the colon and is actually ready to be passed. This difficulty may be avoided by feeding a second meal soon after the “marker meal” has left the stomach. In five instances the time required for post- gastric digestion (in the above sense) varied from 40 to 113 per cent, of that required for gastric digestion, averaging 71 per cent. In terms of temperatures in the middle 70s ( F. ) , with a gastric digestion time of about two days, roughly a day and a half would be added to this for post-gastric digestion, thus amounting to three and one-half days for total digestion or total passage of the alimentary canal. Extensive variation either way is to be expected. Season of Feeding Activity During several seasons at Ithaca, New York, I was able to make fairly extensive collections of water snakes throughout the entire season of activity of the snakes. The food findings from these collections suggested that during the aver- age summer in this region, June, July and Aug- ust were all months of heavy feeding. Food taken during all the remaining months combined (April, May and September) tended to total considerably less than that for any one of the three mid-summer months. Control Measures Snakes should, of course, be guarded against in fish hatchery situations. Lagler (1939) has pointed out that intensive efforts during one or two seasons of collecting may effectively reduce local water snake populations for several suc- ceeding years. This idea was borne out by experi- ence at the Cornell University hatchery some years ago. It was also supported by my own extensive collecting in certain limited snake habi- tats along streams of the Ithaca region. Frederick Tresselt, of Thurmont, Maryland, told me (conversation) some years ago that he had decided success in trapping water snakes on his 17 acres of goldfish ponds. His traps were of a cylindrical minnow-trap type, constructed of wire mesh, about 2 feet long by 10 inches diam- eter, and with funnel entrances (one removable) at the ends. (This trap was rather similar to that described by Fitch in 1951 if the projecting bibs be left off of the latter) . Traps were set in shal- low water with the surface of the water cutting across the entrances, so that a snake could swim in with its head above water. Snakes would ex- plore and enter these traps even though they were unbaited. Mr. Tresselt believed that he trapped more than 1,000 snakes in a dozen traps in 1936 and about 500 snakes in more traps in 1937. In 1938 snakes were relatively scarce. Summary Contents of the stomachs of 207 New York and Michigan water snakes (Natrix sipedon sipe- don) were tabulated according to frequencies, volume and habitats of collection. Fishes com- prised 79 per cent, of the food items taken (with minnows, darters, sculpin and suckers predomi- nant), amphibians 21 per cent. Food is listed for 73 young snakes of the first year. Minnows, darters and amphibians together comprised 80 1958] Brown: Feeding Habits of Natrix sipedon sipedon 69 per cent, of the food items captured. Capture of prey involves so-called groping and direct attack methods. Prey may be taken in relatively deep water. Prey may, or may not, be brought ashore for swallowing. Food organisms were swallowed headfirst in 80 per cent, of 200 cases. This snake appears to be well fitted for either diurnal or nocturnal feeding activity, apparently being more diurnal in cooler habitats, more noc- turnal in warmer ones. It seems to be most active at temperatures between 70° and 80° F. Food amounting to 40 per cent, of the weight of the snake may be taken at one time, but a meal is usually considerably smaller. Two hun- dred and seven stomachs in nature contained 1 to 7 food items, but averaged only 1.2 items per stomach. Examples are presented of amounts of food taken experimentally by good feeders during a period of a number of weeks. Four first-year specimens consumed food averaging 61 per cent, of the original weight of the snake per week, or 247 per cent, per month. Corre- sponding figures for 4 adult snakes were 43 per cent, of the original weight of the snake per week or 174 per cent, per month. Increases in gross length during this experimental feeding period averaged about the same for first-year specimens as for the fifth and sixth year adults: 7 mm. per week or 28 mm. per month. Increase in gross weight was much more rapid in the adults than in the young. However, weight in- crease in terms of percentage increase over the original weight of the snake was approximately twice as rapid in the young. Young individuals consumed about 3 grams of food for each gram of gain in weight. The adults consumed half again as much (4.5 grams) per gram of gain in weight. Moderate meals required about 2 days for gastric digestion to be completed at mid- summer temperatures. Post-gastric digestion re- quired somewhat less time. In the central New York region June, July and August were the months of heavy feeding by snakes in nature, with much more moderate food consumption during late April, May and September. Acknowledgements The frequent references herein to the observa- tions of others convey some impression of the large number of persons to whom I am indebted. I profited particularly through association with, and counsel from, the late Dr. Frank N. Blan- chard, Dr. A. H. Wright and Dr. W. J. Ham- ilton, Jr. The latter also kindly read a first draft of much of the present material. I am also grateful to Dr. Geo. R. LaRue for many courte- sies extended at the University of Michigan Biological Station and at Ann Arbor; to Dr. C. E. Palm for the use of outdoor enclosures of the Department of Entomology of Cornell Uni- versity; to Dr. Emmeline Moore for kind con- sideration extended during two summers’ work with the New York State Conservation Depart- ment Biological Survey; to Dr. C. L. Hubbs, Dr. C. W. Creaser, Dr. W. J. Koster and Dr. W. H. Stickel for help on occasion with identification of fishes; to Miss Ruth Gilreath, Dr. F. J. Hinds, Dr. Theodora Nelson, Miss Ethel B. Finster and Dr. L. C. Pettit for the contribution of many specimens; to Miss Evelyn S. Hoke for motion pictures of feeding snakes; to Dr. E. C. Show- acre for x-ray help; to Dr. M. Graham Netting for a valuable bibliographic suggestion; to Dr. J. A. Oliver for urging the publication of this material without further delay. Bibliography Abbott, Charles C. 1884. A naturalist’s rambles about home. New York: D. Appleton & Co. Adams, Charles C. & T. L. Hankinson 1928. The ecology and economics of Oneida Lake fish. Roosevelt Wild Life Annals, 1 (3,4): 235-548, figs. 175-244, 20 tables, 4 pi., map. Alexander, Maurice M. 1943. Food habits of the snapping turtle in Con- necticut. Jour. Wildlife Mgt., 7 (3): 278- 282, 1 fig., 1 table. Atkinson, D. A. 1901. The reptiles of Allegheny County, Penn- sylvania. Ann. Carnegie Mus., 1: 145-157. Barbour, Roger W. 1950. The reptiles of Big Black Mountain, Har- lan County, Kentucky. Copeia, 1950 (2): 100-107, 5 tables. Benedict, Francis G. 1932. The physiology of large reptiles. Carnegie Inst, of Washington, Publ. 425: x + 539 pp„ 106 figs., 121 tables. Blatchley, W. S. 1891. Notes on the batrachians and reptiles of Vigo County, Indiana. Jour. Cincinnati Soc. Nat. Hist., 14: 22-35. Boyer, Dorothy A. & A. A. Heinze 1934. An annotated list of the amphibians and reptiles of Jefferson County, Missouri. Trans. Acad. Sci. St. Louis, 28 (4): 185- 200. Breckenridge, W. J. 1944. Reptiles and amphibians of Minnesota. Minneapolis: Univ. Minnesota Press, xiii + 202 pp., 52 figs., 45 maps. 70 Zoologica: New York Zoological Society [43: 3 Burt, Charles E. & May Danheim Burt 1929. Field notes and locality records on a col- lection of amphibians and reptiles chiefly from the western half of the United States. II. Reptiles. Jour. Washington Acad. Sci., 19 (20) : 448-460. Cagle, Fred R. 1942. Herpetological fauna of Jackson and Union Counties, Illinois. Amer. Midland Nat., 28 (1): 164-200, 15 figs. Clark, Hubert Lyman 1903. The water snakes of southern Michigan. Amer. Nat., 37: 1-23. Conant, Roger 1938. The reptiles of Ohio. Amer. Midland Nat., 20 (1): 1-200, 1 table, 38 maps, 26 pis. Conant, Roger & Reeve Bailey 1936. Some herpetological records from Mon- mouth and Ocean Counties, New Jersey. Occas. Papers Mus. Zool., Univ. Michigan, 328: 1-10. Cope, E. D. 1869. Synopsis of the Cyprinidae of Pennsyl- vania. Trans. Amer. Phil. Soc„ 13: 35 1 - 410. Creaser, Charles W. 1944. The amphibians and reptiles of the Uni- versity of Michigan Biological Station area in northern Michigan. Papers Michigan Acad. Sci., Arts and Letters, 29: 229-249. DeKay, James E. 1842. Natural history of New York: Zoology of New York, pt. 3. Albany. Ditmars, Raymond L. 1912. The feeding habits of serpents. Zoologica 1 (11): 195-238. Dunn, Emmett Reid 1935. The survival value of specific characters. Copeia, 1935 (2): 85-98. Eaton, E. H. 1928. The Finger Lakes fish problem. In: A biological survey of the Oswego River System. Suppl. 17th Ann. Rept., N. Y. St. Conservation Dept., 1927, pp. 40-66. Evans, Philip D. 1942. A method of fishing used by water snakes. Chicago Nat., 5 (3): 53-55. Evermann, Barton Warren & Howard Walton Clark 1920. Lake Maxinkuckee, a physical and bio- logical survey. Indiana Dept. Cons., Publ. No. 7, vol. 1: 1-660 pp.; vol. 2: 1-512 pp. Fitch, Henry S. 1941. The feeding habits of California garter snakes. California Fish and Game, 27 (2) : 2-32, 2 figs., 1 table. 1951. A simplified type of funnel trap for rep- tiles. Herpetologica, 7 (2) : 77-80. Fox, Wade 1951. Relationships among the garter snakes of the Thamnophis elegans rassenkreis. Univ. California Pub. Zool., 50 (5): 485-529, illus. 1952. Notes on feeding habits of Pacific Coast garter snakes. Herpetologica, 8(1): 4-8. Gentry, Glenn 1941. Herpetological collections from counties in the vicinity of the Obey River drainage of Tennessee. Jour. Tenn. Acad. Sci., 16 (3): 329-332. 1944. Some predators of the Flintville State Fish Hathcery. Jour. Tenn. Acad. Sci., 19 (3): 265-267. Gloyd, Howard K. 1928. The amphibians and reptiles of Franklin County, Kansas. Trans. Kansas Acad. Sci., 31: 115-141. Grant, Chapman 1935. Natrix sipedon sipedon in central Indiana, its individual and sexual variations. Amer. Midland Nat., 16 (6): 921-931. Greeley, J. R. 1928. Fishes of the Oswego Watershed. In: A biological survey of the Oswego River System. Suppl. 17th Ann. Rept., N. Y. St. Conservation Dept., 1927. Hamilton, W. J., Jr. 1951a. The food and feeding behavior of the garter snake in New York State. Amer. Midland Nat., 46 (2): 385-390. 1951b. Notes on the food and reproduction of the Pelee Island water snake ( Natrix sipe- don insularum Conant and Clay). Can- adian Field-Nat. 65 (2): 64-65. Hankinson, T. L. 1917. Amphibians and reptiles of the Charles- ton region. Trans. Illinois Acad. Sci., 10: 322-330. Hay, Oliver Perry 1892. The batrachians and reptiles of the State of Indiana. 17th Ann. Rept., Indiana Dept. Geol. and Nat. Resources. Hubbard, Marian E. 1903. Correlated protective devices in some California salamanders. Univ. California Publ. Zool., 1 (4): 157-170. Hubbs, Carl L. & A. I. Ortenburger 1929. Fishes collected in Oklahoma and Arkan- sas in 1927. Publ. Univ. Okla. Biol. Surv., 1 (3): 47-112. 1958] Brown: Feeding Habits of Natrix sipedon sipedon 71 Kellogg, W. N. & Wardell, B. Pomeroy 1936. Maze learning in water snakes. Jour. Comp. Psych., 21 (3) : 279-295. Kenyon, Walter A. 1925. Digestive enzymes in poikilothermal ver- tebrates: An investigation of enzymes in fishes, with comparative studies on those of amphibians, reptiles, and mammals. Bull. U. S. Bur. Fish., 41: 181-200. King, Willis 1939. A survey of the herpetology of the Great Smoky Mountains National Park. Amer. Midland Nat., 21 (3): 531-582, 9 figs., 6 tables, map. Lagler, Karl F. 1939. The control of fish predators at hatcheries and rearing stations. Jour. Wildlife Mgt., 3 (3): 169-179, 1 table. 1943. Food habits and economic relations of the turtles of Michigan with special reference to fish management. Amer. Midland Nat., 29 (2): 257-312, 9 figs., 9 tables. Lagler, Karl F. & J. Clark Salyer, II 1947. Food and habits of the common water- snake, Natrix s. sipedon, in Michigan. Papers Mich. Acad. Sci., Arts and Letters, 31: 169-180, (1945). Lamson, George Herbert 1935. The reptiles of Connecticut. Connecticut Geol. and Nat. Hist. Surv., Bull. No. 54, 35 pp., 12 pis. McCauley, Robert H., Jr. 1945. The reptiles of Maryland and the District of Columbia. Hagerstown, Md.: Pub. by the author, ii + 194 pp., 12 tables, 46 maps, 48 photos. Minton, Sherman, Jr. 1944. Introduction to the study of the reptiles of Indiana. Amer. Midland Nat., 32 (2): 438-477. Neill, Wilfred T. 1951. Notes on the role of crawfishes in the ecology of reptiles, amphibians and fishes. Ecology, 32 (4): 764-766. Nichols, J. T. 1916. Water snakes swallowing fish. Copeia, 31:44. Ortmann, Arnold E. 1906. The crawfishes of the State of Pennsyl- vania. Mem. Carnegie Mus., 2: 343-524. Raney, Edward C. & Robert M. Roecker 1947. Food and growth of two species of water snakes from western New York. Copeia, 1947 (3): 171-174. Stoner, Dayton 1941. Feeding behavior of a water snake. Sci- ence, 94 (2442): 367. Surface, H. A. 1899. Removal of lampreys from the interior waters of New York. Fourth Ann. Rept., Comm. Fisheries, Game and Forests: 191-245. 1906. The serpents of Pennsylvania. Pennsyl- vania Dept. Agr., Monthly Bull. Div. Zool., 4 (4 & 5): 114-202, 20 figs., 27 pis. Swanson, Paul L. 1952. The reptiles of Venango County, Pennsyl- vania. Amer. Midland Nat., 47 (1): 161- 182, map. Trapido, Harold 1943. Observations on the feeding habits of some water snakes. Chicago Nat., 6 (2): 42. Trembley, Francis J. 1948. The effect of predation on the fish popula- tion of Pocono Mountain lakes. Proc. Pennsylvania Acad. Sci., 22: 44-49. Uhler, F. M., C. Cottam & T. E. Clarke 1939. Food of snakes of the George Washington National Forest, Virginia. Trans. North Amer. Wildlife Conf., 4: 605-622. Walls, Gordon L. 1931. The occurrence of colored lenses in the eyes of snakes and squirrels, and their probable significance. Copeia, 1931 (3): 125-127. 1942. The vertebrate eye and its adaptive radia- tion. Cranbrook Inst. Sci., Bull. No. 19: xiv + 785 pp., 197 figs., 1 pi. Welter, Wilfred A. & Katherine Carr 1939. Amphibians and reptiles of northeastern Kentucky. Copeia, 1939 (3): 128-130. Wilde, Walter S. 1938. The role of Jacobson’s organ in the feed- ing reaction of the common garter snake, Thamnophis sirtalis sirtalis (Linn.). Jour. Exp. Zool., 77: 445-464, 5 tables, 1 pi. j NEW YORK ZOOLOGICAL SOCIETY GENERAL OFFICE 30 East Fortieth Street, New York 1 6, N. Y, PUBLICATION OFFICE The Zoological Park, New York 60, N. Y. OFFICERS PRESIDENT VICE-PRESIDENTS Fairfield Osborn Alfred Ely Laurance S. Rockefeller SCIENTIFIC STAFF: John Tee-Van General Director James A. Oliver. . . Director , Zoological Park Christopher W. Coates. . Director, Aquarium ZOOLOGICAL PARK Grace Davall Assistant Curator, Mammals and Birds William G. Conway. .Curator, Birds James A. Oliver. . . .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 Myron Gordon .Geneticist C. M. Breder, Jr. . . . . .Research Associate in Ichthyology Harry A. Charipper . . . Research Associate in Histology Homer W. Smith. . . . .Research Associate in Physiology SECRETARY TREASURER Harold J. O’Connell David H. McAlpin 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 A ssistant Director David W. Snow Resident Naturalist Henry Fleming ....... Entomologist John Tee- Van ....... .Associate William K. Gregory Associate AFFILIATES L. Floyd Clarke Director , Jackson Hole Biological Research Station SCIENTIFIC ADVISORY COUNCIL A. Raymond Dochez Caryl P. Haskins Alfred E. Emerson K. S. Lashley W. A. Hagan John S. Nicholas EDITORIAL COMMITTEE Fairfield Osborn, Chairman James W. Atz William G. Conway William Beebe Lee S. Crandall William Bridges James A. Oliver Christopher W. Coates John Tee-Van ZOOLOGICA SCIENTIFIC CONTRIBUTIONS OF THE NEW YORK ZOOLOGICAL SOCIETY VOLUME 43 • PART 3 • NOVEMBER 20, 1958 • NUMBERS 4 TO 9 PUBLISHED BY THE SOCIETY The ZOOLOGICAL PARK, New York Contents PAGE 4. Uptake and Turnover of a Single Injected Dose of I131 in Tadpoles of Rana clamitans. By Nancy Weber Kaye & Elizabeth E. Le Bourhis. Text-figures 1-3 73 5. Oral Incubation in the Cichlid Fish Geophagus jurupari Heckel. By Melvin J. Reid & James W. Atz. Plate 1 77 6. The Specific Distinctness of the Fiddler Crabs Uca pugnax (Smith) and Uca rapax (Smith) at Their Zone of Overlap in Northeastern Florida. By Richard E. Tashian & F. John Vernberg. Plate 1 89 7. A Practical Method of Obtaining Blood from Anesthetized Turtles by Means of Cardiac Puncture. By Charles P. Gandal. Text-figure 1 93 8. Observations on the Breeding in Captivity of a Pair of Lowland Gorillas. By Warren D. Thomas. Plates I-III; Text-figures 1 & 2 95 9. The Morphology of Renicola philippinensis, n. sp., a Digenetic Trematode from the Pheasant-tailed Jacana, Hydrophasianus chirurgus (Scopoli). By Horace W. Stunkard, Ross F. Nigrelli & Charles P. Gandal. Plate I; Text-figures 1-3 105 4 Uptake and Turnover of a Single Injected Dose of I131 in Tadpoles of Rana clamitans 1,2 Nancy Weber Kaye3 & Elizabeth E. Le Bourhis Department of Zoology, Barnard College, Columbia University, New York (Text-figures 1-3) Introduction IN 1941 Gorbman & Evans first demon- strated by means of radioautographs that the thyroid of an anuran tadpole, Hyla regilla, can accumulate, in organic combination, radioactive iodine placed in the water as iodide. Dent & Hunt (1952), also employing radio- autographs, have mapped the radioiodine distri- bution in Hyla versicolor. Later, Hunt & Dent (1957) studied, by quantitative techniques, the radioiodine uptake and turnover in frog tadpoles in whose environmental water radioiodide was placed. Money, Lucas & Rawson (1955) and Saxen et al. (1957) have conducted similar studies with tadpoles of Rana pipiens and Xen- opus laevis respectively. None of the observations on thyroidal radio- iodine accumulation in radioiodide-immersed animals gives a clear idea of the metabolism of a single tracer dose of the isotope. Since ab- sorption from the environment must occur in immersed animals, various additional physio- logically variable factors are interposed between the thyroid function one wishes to measure and the tracer iodine. It is not possible, with any degree of assurance or precision, to express the thyroidal I131 in terms of per cent, uptake of a given dose since the amount of I131 which actually enters the animal is difficult to evaluate and varies from one animal to the next. Fur- thermore, the effect of number and/or size of financial support for this work came from grant number G-1177 from the National Science Foundation. 2Physical facilities were provided by the Department of Zoology, Barnard College, Columbia University, New York. sPredoctoral Fellow, National Science Foundation, 1956-57. animals per unit volume of environmental water and of the amount of “carrier” iodide in this water is not easily determined. For these reasons the present study was under- taken. The purpose of these observations was the determination of the fate of a single dose of carrier-free radioiodide injected into the tadpole. Materials and Methods Premetamorphic tadpoles of Rana clamitans of 6-8 cm. in length with hind legs of less than 6 mm. were used (Taylor & Kollros, 1946). Prior to experimentation they were kept in tap water and fed on Elodea and corn meal. During the experimental period the tadpoles were kept in tap water, 7 animals per three liters, at 22° C. They were not fed. Into each animal a single dose of 5 microcuries of carrier-free radioiodine (I131) in a 0.05 cc. volume was injected intra- peritoneally, the needle being inserted through the tail musculature. At time intervals of 2, 10, 20, 40, 72 and 120 hours after injection, 6 of the 7 animals in each group were anesthetized by immersion in 0.1% M.S.-222 (Tricaine-methane-sulfonate, Sandoz). Each tadpole was then dissected into a planned number of parts, each of which was placed in a closed vial containing 1 cc. of 0.7 % NaCl. The weight and the I131 content of each part was recorded so that the localized concentration of iodine could be followed over a period of time. Measurements of radioactivity were made in a well-type gamma ray scintillation counter (and were corrected for physical decay of the isotope) . Results Per cent, uptake of the injected I131 was cal- culated for the total animal, thyroid and eyes. 74 Zoologica: New York Zoological Society [43: 4 Concentration was calculated as per cent, uptake per mg. of tissue X 100 (for convenience) for the thyroid, eyes, ventral white skin, tail (both skin and musculature), carcass of head, dark skin of carcass, coiled intestine and remaining viscera. Thyroids were dissected out under the binocular microscope. The small size of the glands obliged us to include small pieces of car- tilage, which were necessarily weighed along with the glands. In Text-fig. 1 the per cent, uptake curve of the eyes, as a typical body tissue, reaches a peak at about the same time as the per cent, uptake curve of the whole animal, whereas the thyroid curve reaches its peak at 72 hours. Although total thyroid uptake is low, the con- centration per mg. of tissue is exceedingly high as compared with the other body tissues in Text-fig. 2. It is interesting to note in Text-fig. 3 that the curves of I131 concentration for the various body tissues fall into three groups: tail, eyes and car- cass of head in one group; ventral white skin and dark skin of carcass in a second group; and coiled intestine and remaining viscera in a third group. Discussion In this study iodine accumulation by the thy- roid slowly reached a level of 2 % of the injected dose at 72 hours and remained at this value until the end of the experiment. This level in the thyroid is reached only after the values for the other body tissues have decreased, as indi- cated by the curve for the pair of eyes in Text- fig. 1. Very possibly the thyroid continues to take up radioiodine from the blood stream after the iodine has been distributed throughout the body. Unpublished data of Gorbman & Dundee show that Eurycea, a neotenic form of salaman- der, also has a low iodine turnover, so that at the end of 7 days the thyroid still contains about 50 % of its peak value of 2.5 % achieved at about 24 hours. Text-fig. 2 clearly shows that the thyroid is the most potent of all tissues in concentrating radioiodine. Since the dissected thyroid piece I includes the weight of the thyroid, as well as the cartilage upon which it rests, the values obtained for concentration are considerably lower than the real values. The peak for thyroid, which is at least 7 times greater than any other value, comes at a time when the majority of other body tissues have reached an equilibrium (Text-fig. 2). Other experiments with anuran tadpoles have not provided information concern- ing the relative concentrations of iodine in the thyroid and the remainder of the body. The chemical form of the thyroidal iodine in tad- poles is as yet unknown since chromatographic studies have not been done. However, radio- j autographs indicate that it is bound in an organic form (Dent & Hunt, 1952). Money, Lucas & Rawson (1955), using an alternative technique of immersing Rarta pipiens tadpoles of different ages for a definite period of time in a solution of I131, find a value of close to 10-20 % of environmental I131 localized in the block of thyroid tissue at the end of five days (MLR Fig. 2). When they injected tadpoles lower thyroidal I131 accumulation was found (5-10 %), which still exceeded the values we have observed in Rana clamitans. Such data illustrate the differences to be found between single administrations of tracer I131 and con- tinuous absorption from the environment. The fact that continuous absorption varies with stage and therefore cannot be used as a standard value is shown also in the work of Saxen et al. (1957). Text-fig. 1. Uptake for whole body, thyroid and eyes. Each point on the curve rep- resents the average of deter- minations done on six ani- mals. 1958] Kaye & Le Bourhis: 7131 in Tadpoles of Rana clamitans 75 Text-fig. 2. Concentration of I131 in the thyroid and other body tissues. Concentration calculated as per cent, uptake per mg. tissue X 100. Average values determined the same as in Text- fig. 1. The apparently continued increase in total body I131 to 65 % at 20 hours after injection (Text- fig. 1) is at first difficult to interpret. The only apparent explanation is that up to that time a considerable part of the injected dose remained unabsorbed in the body cavity and/or blood stream of the tadpole. The whole body value of the administered dose was found by summing the values obtained for individual body parts. After 20 hours the amount of radioiodine de- creases gradually to 47 % and appears to main- tain itself at that level. This seems to indicate that extrathyroidal areas, which retain a consid- erable amount of the injected dose, also have a low turnover. These data agree, in general, with those provided for Rana pipiens tadpoles Text-fig. 3. Concentration of I131 in the various body tissues other than thyroid. Average determinations the same as in Text-fig. 1. 76 Zoologica: New York Zoological Society [43: 4: 1958] by Money, Lucas & Rawson (1955). Data obtained by Hunt & Dent (1957), who administered I131 by a short period of immer- sion in I131 solutions, indicate that after 2 days the total amount of iodine absorbed in Hyla versicolor tadpoles decreases from approxi- mately 50 % to 10 % whereas our values during the same period remain near 47 %. Dent & Hunt in 1952 pursued a study of the I131 distribution by radioautographing sections of various tissues. However, their study was not primarily quantitative. By actually measuring the fraction of the administered dose accumu- lated in various parts of the body by means of a well-type gamma ray scintillation counter, our study has allowed a quantitative description of the distribution of the radioiodine. It is interesting to note that the curves for the body tissues appear to fall into three sets of parallel curves (Text-fig. 3). The group com- posed of ventral white skin and dark skin of carcass, unexpectedly, shows higher values for the ventral white skin, which would be expected to contain much less tyrosine. Dent & Hunt (1952) found a denser radioautograph in the region of the skin. They attributed this to the binding of iodine by the tyrosine of the melanin pigment. The tail, with less skin and more mus- culo-skeletal tissue per unit weight, contained much lower iodine concentration, indicating that such tissue probablv does not serve as a storage area for iodine. Both viscera and coiled intestine continue to concentrate I131 even at a time when the other tissues seem to be reach- ing an equilibrium. Entero-hepatic recirculation is an important mode of excretion of thyroid hormones through the liver and intestine. The higher values for these organs may be a reflec- tion of such excretion, since the peaks occur at the same time as the thvroidal peak. This may be taken to indicate that the thyroid in the tad- pole does have a turnover, although a low one. Our experimental data indicate much indi- vidual variation, as also shown by both Hunt & Dent (1957) and Money, Lucas & Rawson (1955). Because each group of animals was kept in a large volume of water, the possibility of T131 reabsorption is negligible. Hunt & Dent (1957) also feel that the concentration of iodine in the water and the amount of food ingested by the tadpole do not effect the release of iodine. An extension of the experimental period beyond five days would have been desirable. Acknowledgements The authors of this paper wish to thank Pro- fessor Aubrey Gorbman for his guidance and criticism in the execution of this experiment. Summary 1. Per cent, uptake of injected I131 was calcu- lated for the whole body, thyroid and eyes of premetamorphic Rana clamitans. 2. Distribution of iodine was studied by com- parison of I131 concentration in the thyroid to the concentration of I131 in the other body tissues over a period of five days. 3. Total body per cent, uptake of I131 reaches a peak of 65 % at 20 hours and decreases to an equilibrium of 47 % at 72 hours. 4. The fraction of administered dose of I131 accumulated by the thyroid reaches a maxi- mum of approximately 2 % at 72 hours. Literature Cited Dent, J. N. & E. L. Hunt 1952. An autoradiographic study of iodine dis- tribution in larvae and metamorphosing specimens of Anura. Jour. Exper. Zool., 121: 79-97. Gorbman, A. & H. M. Evans 1941. Correlation of histological differentiation with beginning of function of developing thyroid gland of frog. Proc. Soc. Exper. Biol. & Med., 47: 103-6. Hunt, E. L. & J. N. Dent 1957. Iodine uptake and turnover in the frog tadpole. Physiol. Zool., 30: 87-91. Money, W. L., Virginia Lucas & R. W. Rawson 1955. The turnover of radioiodine by the Rana pipiens tadpole. .Tour. Exper. Zool., 128: 411-421. Saxen, Lauri, Erkki Saxen, Sulo Toivonen & Kauno Salimaki 1957. Quantitative investigation on the anterior pituitary-thyroid mechanism during frog metamorphosis. Endocrinology, 61: 35-44. Taylor, A. C. & J. J. Kollros 1946. Stages in the normal development of Rana pipiens larvae. Anat. Rec., 94: 7-23. 5 Oral Incubation in the Cichlid Fish Geophagus jurupari Heckel Melvin J. Reid & James W. Atz1 (Plate I) Historical Resume ALTHOUGH reproductive habits of the cichlid fishes assigned to the genus Geophagus were described as early as 1855, 1863 and 1865, our knowledge of them has remained, to this day, in a state of discon- certing confusion. In 1855 Castelnau described how his new species, Geophagus ( Chromys ) lapidifera, car- ried pebbles in its mouth to form a nest in which the eggs were laid. (Pellegrin, 1903). In 1862 the Reverend J. C. Fletcher and Sr. Henrique Antonii collected specimens of one or two species of Geophagus that had eggs or young in their mouths (Putnam, 1 863; Fletcher & Kidder, 1866). In 1865 Louis Agassiz also obtained a species of Geophagus in the mouth of which eggs and young in various stages of development were found (Agassiz, 1865), and within a year he had discovered additional species in a similar condition (Agassiz, 1866; Agassiz & Agassiz, 1868.) 2 The seeds of confusion, however, were already sown by these pioneers. Agassiz never identified or described his fishes in any detail, and some 14 bottles of unidentified Geophagus from the Thayer Expedition to Brazil, the one on which Professor Agassiz collected his fish, are stored at the Museum of Comparative Zoology. Similar limitations afflict most subsequent ob- servations on the group. Too often one cannot be certain of the identity of the fish concerned. Pellegrin (1903) recognized 17 species of Geo- phagus, while Regan (1906) accepted 12. Of these, three or four appear to have become 1 The junior author is Associate Curator, New York Aquarium, New York Zoological Society. 2 These five publications appear to be the first to mention oral incubation in the cichlid fishes. generally known, viz. Geophagus brasiliensis (Quoy & Gaimard), G. jurupari Heckel, G. surinamensis (Bloch) and perhaps G. gymno- genys Hensel, but one would be hard put to explain the different reproductive habits that have been ascribed to each of them except on the basis of misidentification of the form under observation. Without the background of a much needed taxonomic revision of the genus, as well as an attempt critically to identify each fish in question, it would be pointless to list the obser- vations of all the amateur and professional workers who have described their reproductive behavior. Our purposes will be served by a brief consensus, with mention of the most important papers and the most notable exceptions. At least ten different accounts of the breeding of Geophagus brasiliensis in captivity agree that the eggs and young are cared, for by both parents in typical cichlid fashion. Hensel (1870), how- ever, described how a specimen in nature took young into its mouth when disturbed and how these were found crowded into its mouth after the fish had been killed or stunned by a shot.3 Adloff (1922) similarly reported that he had seen the young of a freshly caught G. brasiliensis flee into the mouth of the parent several times. Von Ihering (1883, 1920) and Bruning (1931) stated that this species was a mouthbreeder, but their opinions appear to be based on the ob- servations of others. None of these authors mentioned the sex of the fish involved. Guim- araes (1930) says nothing about oral incubation in his description of the reproductive habits of this species, both in aquaria and in nature. Similarly, at least six aquarists have agreed 8 The fish was said to belong to Hensel’s n&wly de- scribed species, Geophagus scymnophilus sp. nov., which Pellegrin (1903) and Regan (1906) have both synonymized with G. brasiliensis. 77 78 Zoologica: New York Zoological Society [43: 5 that Geophagus gymnogenys reproduces itself in typical cichlid fashion, but Haseman (1911a, b), who declared that a subspecies of G. bra- siliensis was not a mouthbreeder, attempted to catch a female G. gymnogenys which was sur- rounded by small fish. These disappeared into her mouth and reappeared later. When the fish was finally caught, her mouth was full of young. On another occasion, a recently captured speci- men of his released young from its mouth. Although two aquarists have mentioned no signs of oral incubation in their breeding pairs of Geophagus surinamensis, Eigenmann (1912) collected a wild specimen carrying young, Beebe & Tee-Van (1922) two specimens similarly en- cumbered,4 and Puyo (1949) two females with eggs in their mouths. Eigenmann (1912) also found a specimen of G. jurupari that was shel- tering young, while Beebe & Tee-Van (1922) caught two fish of the same species that, follow- ing capture, spewed out 58 young between them. Another fish was found to have taken up some 60 young with which it had been placed in an aquarium five days previously, the fish apparent- ly being parent and offspring that had become separated at the time of capture.4 Professional ichthyologists and home aquarium fanciers are less at odds with regard to G. jurupari; Hartel (1936) andDvoskin (1955) have described how the female orally incubated both eggs and young in aquaria. Leitholf (1917), however, did not notice any mouthbreeding behavior in his breed- ing pair of this species. Among the less well known species of Geo- phagus, Eigenmann (1922) collected a female G. pellegrini carrying young. Briining (1918) described how a male G. acuticeps gave shelter to his offspring whenever danger threatened in an aquarium. Two other home aquarium reports on the latter species, however, do not mention mouthbreeding. Briining (1918) and Haseman (1911a) er- roneously believed that Geophagus shelters only the young in its mouth, never the eggs. As to the sex of the incubating parent, what little data there are implicate the female more often than the male, with a single observation that might well indicate that both members of a pair were simultaneously engaged in carrying young (Beebe & Tee-Van, 1922). The statements of Pellegrin (1908) and Miles (1947) that the male typically perfoms the nursing duties are therefore hard to justify. 4 We thank Drs. William Beebe and John Tee-Van for permitting us to use these unpublished data, which were gathered at the station of the Department of Tropical Research, New York Zoological Society, for- merly located at Kartabo, British Guiana. There appears to be no question about the identity of the fish of Eigenmann (1912), Beebe & Tee-Van (1922) and Puyo (1949), but it would be reassuring to check those of Haseman (1911a, b). Our opinion is that Geophagus juru- pari and G. surinamensis are mouthbreeders, while G. brasiliensis is not. About the other species we hesitate to commit ourselves. The need for further, more detailed, observations on all species of Geophagus is obvious. Present Observations The two fish whose behavior is the subject of this paper were kept in standing freshwater aquaria at the home of the senior author for 13 months before their first spawning. At that time the male was about five and one-half inches in total length, the female slightly more than five. About five months previously, the first signs of sexual dimorphism had appeared, accompanied by aggressive behavior on the part of the male, i Most notable were the more pointed extension ! of the posterior rays of the male’s dorsal fin and the much more conspicuous elongation of the anterior rays of his pelvics (Plate I). It was the latter outgrowth that seemed to indicate that the fish belonged to the species Geophagus acuticeps Heckel (Reid, 1956), but Dr. George S. Myers has recently examined the two fish and found that they belong to the species G. jurupari (Register Number, Stanford University, SU 49836). Ten spawnings have been recorded (Table 1 ) , but the bulk of the observations were made dur- ing three of them (Nos. I, III and IV). Details of methods of care and feeding are given in Reid (1956, 1957). Since the senior author was unable to make systematic observations through- out the day or at regular times on successive days, the data do not lend themselves to quanti- tative treatment. Nevertheless, sufficient time was devoted to observing the activities of the fish to reveal the general pattern of reproductive behavior as well as a number of interesting de- tails. Typical sequence of events.— Both male and female clean the surface on which the eggs are to be laid, within a few hours of spawning. The spawning act itself does not differ from that of many cichlids, the female laying many small batches of adhesive eggs, each batch being fer- tilized, in sequence, by the male. The eggs are guarded by both parents. Roughly 24 hours after laying, the eggs are picked up and orally incu- bated, by both parents or by the female alone. Although our observations cannot be considered conclusive on this point, we believe that the eggs 1958] Reid & Atz : Oral Incubation in Geophagus jurupari Heckel 79 o a o (J "3 O > cd 43 C d> p ” o c ,pH 00 t3 a — Qh^2 ° s o «* ^ C/5 C/5 ££ ti W) G CJ , Vh ^ a o a, ■5 o ffl 00 >< 00 o W S . Oh O Oh Z >5 ^ d) cd JG 43 ° m .5 u S (H ^ ~ X j* S- Uh ^ 00 Q 'e o s £ <3h o >' u. *Z2 cd u Dh o 43 a oo 73 c > •S S C Cd £ Oh g >, O-X) fli 00^ .9 2 » u a £ 3 _ ,2 •o E oo *2 £ ca ta 6 & o 55 Uh v s >4 ca Q "3 c o V) 00 ^ oo gr» 0> *n ^ u ’O ^ -a o 5 7d oo .9 o 0X1 g- oo O 0) cd O. O CQ iJ W-H Cd 4> o> 43 43 > w W I3 (3 (U feo 40- 3$. Uh ^ oo cd 73 O P 73 5 Sf 43 73 0) 1 « D-”0 d) 73 50 a o C/5 O 00 ”S « ■go^ | e^S G OO in 73 w o 0 3^0 O O d> > a> i-h oo cd _ °"S 5 2 o ^ P3 oo 0 1 o s 00 oo W CJ . o C 3 U CJ OJ •£L £ ^ - _ 00 « £ « « 'l “i Go's CJ 00 V - ° ca >, u OB G u to .5 B £ " Cl, b/j (D b£ £ 73 cd G £ O PJ cL G G bO C ‘5 £ a Oh c/d £ o JC c/3 *+3 Cd 3 O G 2 £ .fcj O Ph £ bX3 MH C G " 05 73 G , cd £ £ cd 3 6 * cd G ^ O 73 X) < 0? 50 3 ^ O cd JD 73 <3 CD G G O O X X "Tf- C1 cd « G G 2 O 00 73 G O cd 73 X s cd o O (N G >3 i cd Cd oo a) 73 ° rn S. ri G ■ cd CO z I Pipi ~ +* H SEE I I I ! I I I I E I £ E I I I I 3 CL u \ PipiPiPi ^ ^ Pi ^ HHHHHHl'-h I ~ I SSSSS £ £ £ I I I I VO«MO»0\0\OVOO' ir,n^invoeo'OM»o\a1o«r,nm't if,n>i»nop'Ooo\ Cl, < 3 a 3 o I I I I I I I 111111“ M I I SSS & 4> z L. . I s 33 3 CL Pi u u ~ H ~ E I E I I II I I I I HHHhf' II I I E ■ p i I i/-ir'oovooocoo\o\oo. XI aj CL C',rNmTfi/->r-ooLor'-ooo\©©*-H, rt £ ° IS t; « x3 .a a c« ~ 3 fc 3 3 5^T3 5 M ° u ^ I ^•2 « ° 2 s I XP«X1 O 3 L •3 Z1 a -a l. c ^ 03 ^ ^ cS S C^ i'Scx^ - - u « ^ £ 8 e-SS _ e< S co‘— Cl ^ e ^ 5 ^ 1/5 O (D 2 ^ c o c o^ C 4= 8 <3 QZ-s."^-S cj c S-fQ .a . aa_j-^H2Sc Bg3'°.0.2 ♦j <-> cw c/5 a ^ ^ 3«ig0 S.SP3.S ■of 2 o o J3Q 2 - "ui, w 3 os a ^ • •- *•§ 5?m 2 g f I-S3 P v d ^ O *-* c 5 rt-r w b£>^ <+h ■« 2'0og.S^°•> Sl . x,x>-~ a-*-* > . . y o o ^ =3 03 as ^ § 13 13 ^ >a.S w 3 3-2 d d . a c/5 c/5 cL,'a a a a c« Dh O- cd ~ •— C/5 C/5 *a >> - - T3 *J3 aq O- a- >a ^3 ^3 ,g2 n£%xxxz S « L S H Pi V c £2: ^ s x cr <3 a, J D-. o3 G ° e c ° • etc g T3 BS 2 « s .S u 2 >, Ofi“ c es « ^1 OJO J T3 2 e G xi a> O ^ t- a> K V w ^ co Oh C/J 3 G X> _ 0) D X -fi c w).S S CO 4-> CO o j3 -a i> J- C \G CO O Oh O CO 03 .~J i «= 2 •« c u. ° <“ H X W H h_ G s £ DhT3 ^ C G T-H CO G ° 5 l-s * >, CO 03 TD rH co *5 G dxj-T-T' .2 •> T3 *G Z* ■*“’ CO .0 73 a- ’I o-| G o •“ « C co E "c3 Ph > co •— < T3 « S co . G ; ■:i?y mmm . iKjij-.v-l Sill FllIsJI ' - ji _\r\<, | • \, « ^-> »IPII Illtlti iHlWH ASPECTS OF SOCIAL BEHAVIOR IN FIDDLER CRABS, WITH SPECIAL REFERENCE TO UCA MARACOANI (LATREILLE) - _ _ I - I ■ * _ — . J . 11 A Catalog of the Type Specimens of Fishes Formerly in the Collections of the Department of Tropical Research, New York Zoological Society1 Giles W. Mead2 During the past thirty-five years the De- partment of Tropical Research of the New York Zoological Society, under the direction of Dr. William Beebe, has contributed much to the study of fishes. The active field pro- gram resulted in large collections which were maintained and studied in the Department’s lab- oratory in the Zoological Park and in a long series of technical publications. In recent years, Dr. Beebe has donated parts of this collection to several institutions, where they continue to contribute to ichthyological research. The dis- tribution of the Department’s collection of fishes was completed in 1957 with the transfer of the remaining specimens to the U. S. National Mu- seum and the Natural History Museum, Stan- ford University. The type material resulting from the Department’s work is now in the collections of these two institutions and four others to which specimens were given prior to 1957. The list which follows has been prepared, at Dr. Beebe’s request, primarily to make known the present location of this type material. Its compilation was simplified by the catalogs and indices maintained by Dr. Beebe and through the cooperation of the curators of the several museums that now contain collections assembled by the Department of Tropical Research. The list is alphabetical, and all species are listed under the genus in which they were first described, regardless of subsequent transfer. Four species of deep-sea fishes were described by Dr. Beebe from sight observations from the bathysphere. These four, Bathysphaera Intacta (Beebe, 1932, Bull. New York Zool. Soc., 35 iContribution No. 990, Department of Tropical Re- search, New York Zoological Society. ^Ichthyological Laboratory, United States Fish and Wildlife Service, LT. S. National Museum, Washington 25, D. C. (5): 175), Bathyceratias trilychnus, Bathyem- bryx istiophasma and Bathysidus pentagrammus (Beebe, 1934, ibid., 37 (6): 190-192), have been omitted. The following abbreviations have been used: NYZS— New York Zoological Society. KOH— New York Zoological Society number assigned to specimens which were cleared in potassium hydroxide and stained with alizarine for osteological study. USNM— United States National Museum. SU— Natural History Museum, Stanford University. AMNH— American Museum of Natural History. CAS— California Academy of Sciences. Acanthemblemaria arborescens Beebe & Tee- Van, 1928, Zoologica (N.Y.), 10 (1): 244. Lamen- tin Reef, Port-au-Prince Bay, Haiti; Feb. 22, 1927. Type: (NYZS 6923) USNM 170566. Acanthemblemaria crocked Beebe & Tee-Van, 1938, ibid., 23 (3): 310. San Lucas Bay, Lower California, Mexico; April 2, 1936. Type: (NYZS 24824) SU 46497. Acanthemblemaria variegata Beebe & Tee-Van, 1928, ibid., 10 (1): 247. Lamentin Reef. Port-au- Prince Bay, Haiti; May 8, 1927. Type: (NYZS 7464) USNM 170569. Aceratias edentula Beebe, 1932, ibid., 13 (4): 102. 13 mi. south of Nonsuch, Bermuda; 1,000 fms.; Tune 2, 1931. Type: (NYZS 20571, misprinted as 20751 in text; KOH 871) USNM 170951. Ammodytes lucasanus Beebe & Tee-Van, 1938, ibid., 23 (3): 306. Cape San Lucas, Lower Cali- fornia, Mexico; April 25, 1936; from stomach of Euthynnus alletterata. Type: (NYZS 25249-A) SU 46501. Anomalopterus megalops Beebe, 1933, ibid., 13 (8): 159. 12 mi. south of Nonsuch, Bermuda; 700 fms.; July 10, 1929 (net 280). Type: (NYZS 11456) USNM 170957. Anchoviella longipinna Beebe & Tee-Van, 1928, ibid., 10 (1): 48. Bizoton, Haiti; April 1, 1927. Type: (NYZS 7460) USNM 170574. Arenichthys apterus Beebe & Tee-Van, 1938, ibid., 23 (3): 301. Arena Bank, Lower California, Mex- ico; 23° 29' 30" N„ 109° 25' 30" W.; April 20, 1936. Type: (NYZS 25361) SU 46498. 131 132 Zoologica: New York Zoological Society [43: 11 Aristostomias photodactylus Beebe, 1933, Copeia, 1933 (4) : 171. 10 mi. south of Nonsuch, Bermuda; 700 fms.; June 24, 1929. Type: (NYZS 10932) USNM 170931. Bathophilus altipinnis Beebe, 1933, op. cit., p. 162. 8 mi. south of Nonsuch, Bermuda; 800 fms.; June 24, 1929. Type: (NYZS 10885) USNM 170926. Bathytroctes drakei Beebe, 1929, Zoologica (N.Y.), 12 (1): 6. Hudson Gorge; 39° 15' N., 72° 00' W.; 800 fms.; July 7, 1928. Type: (NYZS 7690) USNM 170958. Caulolatilus guppyi Beebe & Tee-Van, 1937, ibid., 22 (1): 93. Port of Spain market, Trinidad; Dec. 16, 1936. Type: (NYZS 24737) USNM 170565. Cetomimus craneae Harry, 1952, ibid., 37 (1): 64. South of Nonsuch, Bermuda; 800 fms.; July 8, 1929. Type: (NYZS 11370) SU 17102. Cetomimus teevani Harry, 1952, op. cit., p. 61. South of Nonsuch, Bermuda; 700 fms.; Sept. 29, 1930. Type: (NYZS 19519) SU 17101. Chaenophryne crossotus Beebe, 1932, ibid., 13 (4): 83. 8 mi. south of Nonsuch, Bermuda; 500 fms.; June 15, 1931. Type: (NYZS 20809) USNM 170942. Chaenophryne draco Beebe, 1932, op. cit., p. 84. 10 mi. southeast of Nonsuch, Bermuda; 600 fms.; Aug. 15, 1931. Type: (NYZS 22396) USNM 170943. Chaenopsis deltarrhis Bohlke, 1957, Proc. Acad. Nat. Sci. Philadelphia, 109: 93. Gorgona Island, Colombia; 2° 57' 30" N., 78° 11' W.; 2-8 fms. Type: (NYZS 28672) SU 49250. Cherublemma lelepris Trotter, 1926, Zoologica (N.Y.), 8 (3): 119. Pacific off Colombia, 5° 03' N., 81° 08' W.; 140 fms. Type: (NYZS 5108) AMNH 8463. Chirostomias lucidimanus Beebe, 1932, ibid., 13 (4): 52. 10 mi. south of Nonsuch, Bermuda; 500 fms.; Aug. 10, 1931. Type: (NYZS 22200) USNM 170934. Cichlasoma haitiensis Tee-Van, 1935, ibid., 10 (2): 294. fitang Saumatre, Cul-de-Sac Plain. Haiti; March 15, 1927. Type: (NYZS 7302) USNM 170907. Cirrhitichthys corallicola Tee-Van, 1940, ibid., 25 (5) : 58. Gorgona Island, Pacific off Colombia; 2° 58' N., 78° 11' W.; March 30, 1938. Type: (NYZS 28710-A) SU 46504. Citharichthys gordae Beebe & Tee-Van, 1938, ibid., 23 (3): 302. Outer Gorda Bank, Lower Cali- fornia, Mexico; 60 fms.; April 23, 1936. Type (NYZS 25785) SU 46496. Corythoichthys bermudensis Beebe & Tee-Van, 1932; ibid., 13 (5): 113. Nonsuch Island, Bermuda; Aug. 18, 1930. Type (NYZS 9326) USNM 170913. Cremnobates argus Beebe & Tee-Van, 1928, ibid., 10 (1): 238. Lamentin Reef, Port-au-Prince Bay, Haiti. Type: (NYZS 7375) USNM 170904. Cyprinodon bondi Myers, 1935, ibid., 10 (3): 303. fitang Saumatre, Haiti; Feb. 20, 1933. Type: USNM 100960. Cypselurus antarei Beebe & Hollister, 1931, ibid., 12 (9): 83. 200 mi. north of Sombrero, B.W.I.; 21° 50' N., 63° 32' W.; June 30, 1932. Type: (Antares Exp. no. 6) USNM 170921. Dasyatis pacificus Beebe & Tee-Van, 1941, ibid., 26 (3) : 262. Port Parker, Costa Rica; Jan. 22, 1938. Type: (NYZS 26120) AMNH 15661, 15662, 15663 (three pieces of the body) and 15710 (lower jaw). Diabolidium arcturi Beebe, 1926, Bull. New York Zool. Soc., 29 (2): 80. Pacific; 4° 50' N., 87° W.; 514-900 fms. Type: (NYZS 6144; number on vial containing type: 6333) SU 46505. Dicrolene gregoryi Trotter, 1926, Zoologica (N.Y.), 8 (3): 116. Pacific; 4° 50' N., 87° W.; 844 fms.; May 31, 1925. Type: (NYZS 6063) AMNH 7511. Dixonina pacifica Beebe, 1942, ibid., 27 (8): 43. Port Culebra, Costa Rica; 10° 31' N., 85° 40' W.; Jan. 24, 1938. Type: (NYZS 26131) SU 46486. Dolichopteryx binocularis Beebe, 1932, ibid., 13 (4): 49. 14 mi. southeast of Nonsuch, Bermuda; 400 fms.; Aug. 4, 1931. Type: (NYZS 21867; KOH 960) USNM 170933. Dolopichthys gladisfenae Beebe, 1932, op. cit., p. 86. 6 mi. south of Nonsuch, Bermuda; 700 fms.; May 28, 1930. Type: (NYZS 15490) USNM 170944. Dolopichthys tentaculatus Beebe, 1932, op. cit., p. 88. 10 mi. southeast of Nonsuch, Bermuda; 600 fms.; Sept. 7, 1931. Type: (NYZS 23170) USNM 170945. Emblemaria micropes Beebe & Tee-Van, 1938, ibid., 23 (3) : 308. Inez Point, Santa Inez Bay, Gulf of California, Mexico; April 9, 1936. Type: (NYZS 24895) SU 46499. Eucinostomus mowbrayi Beebe & Tee-Van, 1932, ibid., 13 (5) : 115. Nonsuch Island, Bermuda; Sept. 30, 1930. Type: (NYZS 9328) USNM 170909. Eupomacentrus beebei Nichols, 1924, ibid., 5 (4): 63. Eden, Indefatigable Island, Galapagos; April 1, 1923. Type: AMNH 8270. Eupomacentrus rubridorsalis Beebe & Hollister, 1933, ibid., 12 (9) : 85. Chatham Bay, Union Island, Grenadines, B.W.I.; July 9, 1932. Type: (NYZS- Antares Exp. no. 97) USNM 170922. Eustomias satterleei Beebe, 1933, Copeia 1933 (4): 164. 8 mi. south of Nonsuch, Bermuda; 1,000 fms.; Sept. 10, 1929. Type: (NYZS 13457) USNM 170927. Eustomias schiffi Beebe, 1932, Zoologica (N.Y.), 13 (4): 54. 6 mi. south of Nonsuch, Bermuda; 600 fms.; May 29, 1930. Type: (NYZS 15653) USNM 170935. Exonautes marginatus Nichols & Breder, 1928, ibid., 8 (7): 429. Pacific east of the Galapagos; 2° 36'— 2° 8' N„ 85° l'-86° 31' W. Type: AMNH 9234. Exonautes nonsuchae Beebe & Tee-Van, 1932, ibid., 13 (5): 112. Near St. David’s Island, Ber- muda; May 15, 1929. Type: (NYZS 9983) USNM 170912. Gambusia beebei Myers, 1935, ibid. , 10 (3): 305. fitang de Miragoane, Haiti. Type: (NYZS 7168); missing. 1958] Mead: Type Specimens of Fishes 133 Gobiesox daedaleus boreus Briggs, 1955, Stan- ford Ichthyol. Bull., 6 ( 1 ) : 111. Conchaguita Island, Gulf of Fonseca, El Salvador; Dec. 22, 1937. Type: (NYZS 27550) SU 17392. Gobiesox stenocephalus Briggs, 1955, op. cit., p. 92. Puerto Parker, Costa Rica; Jan. 14, 1938. Type: (NYZS 27872) SU 17408. Gigantactis perlatus Beebe & Crane, 1947, Zoologica (N.Y.), 31 (4): 167. Pacific off Jicaron Island, Panama; 500 fms.; March 20, 1938. Type: (NYZS 28621) SU 46487. Gillellus quadrocinctus Beebe & Hollister, 1935, ibid., 19 (6): 222. (Also spelled quadrocintus, a lapsus). Union Island, Grenadines, B.W.I.; July 12, 1934. Type: (NYZS-Antares Exp. no. 180); missing. Gobiosoma chancei Beebe & Hollister, 1933, ibid., 12 (9): 87. St. George’s Bay, Grenada, B.W.I.; July 4, 1932. Type: (NYZS -Antares Exp. no. 22) USNM 170955. Gobiosoma macrodon Beebe & Tee-Van, 1928, ibid., 10 (1): 226. Lamentin Reef, Port-au-Prince Bay, Haiti. Type: (NYZS 7462) USNM 170896. Haplophryne hudsonius Beebe, 1929, ibid., 12 (2): 23. Hudson Gorge; 600 fms.; July 7, 1928. Type: (NYZS 7696; KOH 26) USNM 170954. Himantolophus azurlucens Beebe & Crane, 1947, ibid., 31 (4): 155. Pacific off Cape Mala, Panama; 7° N„ 79° 16' W.; 500 fms.; March 25, 1938. Type: (NYZS 28641) SU 46507. Hypleurochilus bermudensis Beebe & Tee-Van, 1933; ibid., 13 (7): 155. Marshall Island, Bermuda. Type: Harvard Mus. Comp. Zool. no. 33070. lridio bathyphilus Beebe & Tee-Van, 1932, ibid., 13 (5): 117. 1 mi. south of Nonsuch, Bermuda; 510 feet; Sept. 30, 1929. Type: (NYZS 9050) USNM 170910. Labrisomus albigenys Beebe & Tee-Van, 1928, ibid., 10 (1): 233. Lamentin Reef, Port-au-Prince Bay, Haiti; May 9, 1927. Type: (NYZS 7372) USNM 170899. Labrisomus haitiensis Beebe & Tee-Van. 1928. op. cit., p. 232. Bizoton, Port-au-Prince Bay, Haiti; March 15, 1927. Type: (NYZS 7170)' USNM 170903. Lampanyctus septilucis Beebe, 1932, ibid., 13 (4) : 68. 7 mi. south-southwest of Nonsuch, Ber- muda; 700 fms.; July 4, 1929. Type: (NYZS 14292-A) USNM 171199. Lampanyctus polyphotis Beebe, 1932, op. cit., p. 67. 5 mi. south of Nonsuch; 900 fms.; May 25, 1929. Type: (NYZS 10151) USNM 171200. Lamprotoxus angulifer Beebe, 1932, op. cit., p. 56. 15 mi. southeast of Nonsuch, Bermuda; 500 fms.; July 27, 1931. Type: (NYZS 21667) USNM 170936. Lasiognathus beebei Regan & Trewavas, 1932. Dana Rep. (1928-30 Exped.), 2: 90. Descr. based on Beebe, Bull. N. Y. Zool. Soc., 33 (2): 60 (fig- ure), without reference to specimen. Figure taken from a specimen which should stand as the type: (NYZS 9804; KOH 287) USNM 170956. Leptocephalus microphthalmus Beebe & Tee-Van, 1928, Zoologica (N.Y.), 10 (1) : 58. Port-au-Prince Bay, Haiti; March 19, 1927. Type: (NYZS 7080) USNM 170906. Leptophilypnus crocodilus Beebe & Tee-Van, 1928, op. cit., p. 219. Lamentin Reef, Port-au- Prince Bay, Haiti. Type: (NYZS 7467) USNM 170905. Leptostomias bermudensis Beebe, 1932, ibid., 13 (4): 59. IV2 mi. southeast of Nonsuch, Bermuda; 500 fms.; June 15, 1931. Type: (NYZS 20826) USNM 170937. Linophryne brevibarbata Beebe, 1932, op. cit., p. 94. 9 mi. southeast of Nonsuch, Bermuda; 900 fms.; July 16, 1929. Type: (NYZS 11656; KOH 973) USNM 170947. Linophryne quinqueramosus Beebe & Crane, 1947, ibid., 31 (4): 174. Pacific off Panama; 7° 24' N., 78° 35' W.; 500 fms.; April 4, 1938. Type: (NYZS 28709) SU 46506. Lipactis megalops Beebe, 1929, ibid., 12 (1): 19. Hudson Gorge; 1,000 fms.; July 28, 1925. Type: (NYZS 6633-A) missing since prior to 1932. Lophodolus lyra Beebe, 1932, ibid., 13 (4): 96. 10 mi. south of Nonsuch, Bermuda; 800 fms.; July 27, 1931. Type: (NYZS 21610) USNM 170949. Macromastax gymnos Beebe, 1933, ibid., 13 (8): 162. 8 mi. south of Nonsuch, Bermuda; 1,000 fms.; June 22, 1929. Type: (NYZS 10829) USNM 170960. Macrostomias calosoma Beebe, 1933, op. cit., p. 165. 12 mi. southeast of Nonsuch, Bermuda; 600 fms.; Sept. 15, 1930. Type: (NYZS 18781); miss- ing. Megalodoras irwini Eigenmann, 1925, Trans. American Phil. Soc., n.s., 22 (5) : 307-308. Beebe’s specimen, Indiana Univ. no. TRS 2567, was made the type of a new species which Eigenmann sup- pressed in proof when he found it to be a large representative of M. irwini, described on the fol- lowing page. It is evident from his account that he intended a specimen from Iquitos collected by Allen to be the type. Beebe’s fish is now in the California Academy of Sciences (CAS 20735). Melanocetus megalodontis Beebe & Crane, 1947, Zoologica (N.Y.), 31 (4): 152. Pacific 145 mi. north of Clarion Island; 500 fms.; May 17, 1936. Type: (NYZS 25791) SU 46488. Melanonus unipennis Beebe, ibid., 13 (4): 74. 10 mi. southeast of Nonsuch, Bermuda; 700 fms.; Aug. 15, 1931. Type: (NYZS 22397) USNM 170940. Melanostomias bulbosus Beebe, 1933, Copeia, 1933 (4): 166. 9 mi. southeast of Nonsuch, Ber- muda; 700 fms.; May 30, 1929. Type: (NYZS 10235) USNM 170928. Mnierpes macrocephalus catherinae Clark Hubbs, 1952, Stanford Ichthyol. Bull., 4 (2): 61. Piedra Blanca, Costa Rica; Feb. 4, 1938. Type: (NYZS 27178, misprinted as 28178) SU 46509. Mobula lucasana Beebe & Tee-Van, 1938, Zool- ogica (N.Y.), 23 (3): 299. San Lucas Bay, Lower California, Mexico. Type: (NYZS 24793) AMNH 15676 and 15675 (two pieces). 134 Zoologica: New York Zoological Society f43: 11: 1958] Neonesthes gnathoprora Cohen, 1956, ibid., 41 (2): 81. South of Nonsuch, Bermuda; 900 fms.; Aug. 17, 1929. Type: (NYZS 12501) SU 46381. Neonesthes nicholsi Beebe, 1933, Copeia, 1933 (4): 160. 9 mi. south of Nonsuch, Bermuda; 800 fms.; Sept. 1, 1930. Type: (NYZS 17529) USNM 170925. Ophioblennius ferox Beebe & Tee-Van, 1928, Zoologica (N.Y.), 10 (1): 242. Bizoton, Port-au- Prince Bay, Haiti; April 6, 1927. Type: (NYZS 7152) USNM 170901. Parabrotula dentiens Beebe, 1932, ibid., 13 (4) : 81.8 mi. southeast of Nonsuch, Bermuda; 800 fms.; June 12, 1930. Type: (NYZS 16110; published er- ror: NYZS 15882; KOH 556) USNM 170952. Paraclinus beebei Clark Hubbs, 1952, Stanford Ichthyol. Bull., 4 (2): 81. Piedra Bay, Costa Rica. Type: (NYZS 28152) SU 46512. Pherallodiscus varius Briggs, 1955, ibid., 6 (1): 131. Passavera Island, Chamela Bay, Jalisco, Mex- ico; Nov. 19, 1937. Type: SU 17807. Photichthys nonsuchae Beebe, 1932, Zoologica (N.Y.), 13 (4): 61. 7 mi. south-southwest of Non- such, Bermuda; 600 fms.; May 3, 1929. Type: (NYZS 9793; published error: NYZS 9973) USNM 170938. Photonectes bifilifer Beebe, 1933, Copeia, 1933 (4): 167. 9 mi. south of Nonsuch, Bermuda; 800 fms.; May 19, 1930. Type: (NYZS 15146) USNM 170929. Photonectes cornutus Beebe, 1933, op. cit., p. 169. 10 mi. south of Nonsuch, Bermuda; 600 fms.; Sept. 4, 1930. Type: (NYZS 17875) USNM 170930. Photostylus pycnopterus Beebe, 1933, Zoologica (N.Y.), 13 (8): 163. 9 mi. southeast of Nonsuch, Bermuda; 800 fms.; May 30, 1929. Type: (NYZS 10217) USNM 170959. Pomacentrus freemani Beebe & Tee-Van, 1928, ibid., 10 (1): 196. Sand Cay, Port-au-Prince Bay, Haiti; May 7, 1927. Type: (NYZS 7269) USNM 170898. Prionotus teaguei Briggs, 1956, Quart. Journ. Florida Acad. Sci., 19 (2-3): 101. 14 mi. southeast of Judas Point, Costa Rica; 9° 19' 32" N., 84° 29' 30" W.; March 1, 1938. Type: SU 46380. Psammobatus spinosissimus Beebe & Tee-Van, 1941, Zoologica (N.Y.), 26 (3) : 259. Pacific south of Cocos Island; 765 fms.; June 3, 1925. Type: (NYZS 6132) SU 46500. Station data given in text do not agree with Arcturus station 72, the pub- lished number, but with station 74. Pseudoscopelus stellatus Beebe, 1932, ibid., 13 (4): 75. 8 mi. southeast of Nonsuch, Bermuda; 300 fms.; July 7, 1931. Type: (NYZS 21555) USNM 170941. Quassiremus goslingi Beebe & Tee-Van, 1932, ibid., 13 (5): 110. Castle Roads, Bermuda; 30 feet; March 21, 1929. Type: (NYZS 8700) USNM 170563. Runula albolinea Nichols, 1924, ibid., 5 (4): 64. Indefatigable Island, Galapagos. Type: AMNH 8271. Rypticus bornoi Beebe & Tee-Van, 1928, ibid., 10 (1): 132. Lamentin Reef, Port-au-Prince Bay, Haiti; April 27, 1927. Type: (NYZS 7206) USNM 170572. Saccopharynx harrisoni Beebe, 1932, ibid., 13 (4): 63. 10 mi. southeast of Nonsuch, Bermuda; 900 fms.; June 11, 1931. Type: (NYZS 20802) USNM 170939. Scorpaenodes cortezi Beebe & Tee-Van, 1938, ibid., 23 (3): 304. Off San Jose Island, Gulf of California, Mexico; 24° 55' N., 1 10° 20' W.; April 8, 1936. Type: (NYZS 24889-A) SU 46503. Scorpaenodes russelli Beebe & Tee-Van, 1928, ibid., 10 (1): 189. Bizoton Reef, Port-au-Prince Bay, Haiti; 12 feet; April 27, 1927. Type: (NYZS 7207) USNM 170573. Somersia furcata Beebe & Tee-Van, 1934, Ameri- can Mus. Novitates, 730: 1. Hungry Bay, Bermuda; Nov. 12, 1933. Type: (NYZS 26165) USNM 170924. Stathrnonotus corallicola Beebe & Tee-Van, 1928, Zoologica (N.Y.), 10 (1): 249. Lamentin Reef, Port-au-Prince Bay, Haiti; April 22, 1927. Type: (NYZS 7463) USNM 170571. Stathrnonotus lugubris Bohlke, 1953, ibid., 38 (3): 145. Port Guatulco, Golfo de Tehuantepec, Mexico; 15° 43’ 30" N., 96° 08' W.; Dec. 3, 1937. Type: (NYZS 27236) SU 17748. Stomias fusus Beebe, 1929, ibid., 12 (1): 7. Hud- son Gorge; 600 fms.; Aug. 6, 1928. Type: (NYZS 7667) USNM 170953. Syngnathus mackayi nesiotes Herald, 1942, Stan- ford Ichthyol. Bull., 4 (2) : 128. Nonsuch, Bermuda. Type: (NYZS 8919) USNM 170915. Syngnathus pipulus Beebe & Tee-Van, 1932, Zool- ogica (N.Y.), 13 (5): 115. The Reach, Bermuda. Type: (NYZS 25152) USNM 170914. Tomicodon eos rhadinus Briggs, 1955, Stanford Ichthyol. Bull., 6 (1): 70. Tangola-Tangola Bay, Oaxaca, Mexico; Dec. 12, 1937. Type: (NYZS 27516) SU 18121. Trematorhynchus adipatus Beebe & Crane, 1947, Zoologica (N.Y.), 31 (4): 163. Pacific 71 mi. off Cape Corrientes, Colombia; 500 fms.; March 26, 1938. Type: (NYZS 28770) SU 46494. Trematorhynchus moderatus Beebe & Crane, 1947, op. cit., p. 164. Pacific, off Cape Corrientes. Colombia; 500 fms.; March 26, 1938. Type: (NYZS 28771) SU 46492. Trematorhynchus multilamellatus Beebe & Crane, 1947, op. cit., p. 165. 16 mi. southwest of Narbor- ough Island, Galapagos; 1,900 fms.; June 12, 1925. Type: (NYZS 6321) SU 46490. Trematorhynchus multiradiatus Beebe & Crane, 1947, op. cit., p. 166. Pacific, 11 mi. southwest of Jicaron Island, Panama; 500 fms.; March 20, 1938. Type: (NYZS 28773) SU 46491. Trematorhynchus paucilamellatus Beebe & Crane, 1947, op. cit., p. 166. Pacific, 20 mi. south of Cape Blanco, Costa Rica; 500 fms.; Feb. 7, 1938. Type: (NYZS 28250) SU 46489. Ultimostomias mirabilis Beebe, 1933, Copeia, 1933 (4) : 174. 12 mi. south of Nonsuch, Bermuda; 900 fms.; June 24, 1929. Type: (NYZS 10865) USNM 170932. Xenoceratias nudus Beebe & Crane, 1947, Zoolo- gica (N. Y.), 31 (4): 155. Pacific, 20 mi. south of Cape Blanco, Costa Rica; 500 fms.; Feb. 27, 1938. Type: (NYZS 28402) SU 46495. 12 The Influence of Environment on the Pigmentation of Histrio histrio (Linnaeus) C. M. Breder, Jr., & M. L. Campbell The American Museum of Natural History, New York (Plates I-III) Introduction The remarkable manner in which Histrio histrio (Linnaeus) disappears against a background of sargasso weed, with which it is so closely associated, has long been a mat- ter of comment by both biologists and casual observers. This condition is facilitated by the form of the fish and its pigmentation. While, as noted by Breder (1946), Histrio cannot be easily counted among those fishes which closely imitate specific parts of plants in the manner de- fined by him, its general “ragged” appearance enables it to become extremely inconspicuous against an equivalent appearance which the weed produces, although the latter is considerably different in detail. This effect, so far as form is involved, is greatly enhanced by the fish’s in- numerable small “tabs” of skin, assisted in a measure by the exserted rays of the pectoral fins. With pattern and coloration these features account for all the physical elements in the near “invisibility” of Histrio. The coloration and pat- tern, almost entirely caused by the arrangement and colors of the chromatophores, are the fea- tures to which this study has been chiefly ad- dressed, since it is this system which is basically responsible for the ability of these fishes to render themselves so well concealed. The general difficulty encountered in finding Histrio against a background of sargasso weed has been mentioned in many, often popular, ac- counts of such matters. See for example, Braam Houckgeest (1774), Ives (1889), Vignon (1931), Gordon (1938) and Breder (1949). Casual observations in the field indicate that not only do these fishes show this general resem- blance to Sargassum but that usually an indi- vidual matches the tan or yellowish color of the particular clump of weed on which it is found. While the range of coloration in these weeds is not great, there is sufficient variation to make such differences easily noticeable. Normally the fish are not found in other than living clumps of floating weed. Clumps which have been cast on the beach and refloated by a returning tide, after having been killed and dried in the interim, are much darker brown or sometimes nearly black. These have always been found to be de- void of Histrio. Thus these fishes are evidently not ordinarily called upon to respond other than to a limited range of color and pattern, as com- pared with the backgrounds against which most other color-matching fishes are exposed. For this reason, experiments were planned to deter- mine the limits of the range of chromatic changes possible in Histrio as compared with the range of background which they ordinarily encounter, and in reference to the kinds and arrangements of the chromatophores with which they were provided. The related various species of Antennarius, which are capable of much greater color and pattern changes, including brilliant reds, greens and yellows, are discussed comparatively with Histrio. Originally it had been planned to carry out an identical series of experiments on Anten- narius, but, as is developed in the concluding remarks of this paper, it became unnecessary to do this for present purposes. Most of the field observations mentioned herein were made while at The Lerner Marine Laboratory, on North Bimini in the Bahamas. All of the experimental studies were carried on in the laboratories of the Department of Fishes and Aquatic Biology of the American Museum of Natural History in New York. The living 135 136 Zoologica: New York Zoological Society [43: 12 material established there was collected at Bim- ini, some by Dr. Louis Krumholz and some by Dr. Vladimir Walters. Dr. Walters also kindly supplied certain field data pertinent to these studies. The necessary sea water was obtained through the kindness of Mr. C. W. Coates, Di- rector of the New York Aquarium, for which we are grateful. The description of the pigmen- tary systems involved are mostly paraphrased from notes prepared by Miss Priscilla Rasquin who was working at the time on the reproduc- tion and embryology of these same fishes, re- ported in Rasquin (1958). The plastic films employed in some of the experiments were sup- plied by the Visking Corporation, New York City, through the kindness of its representative, Mr. Ronald Basso, for which we are most appre- ciative. Experimental Backgrounds and Results All of the following experiments were carried out in seven aquaria, each of which measured 2 ft. long by 1 ft. by 1 ft. These were supplied with the running sea water in a small circulat- ing system set up especially for these experi- ments and described in detail by Breder (1957). The water used in the laboratory aquaria was obtained locally, most of it through the coopera- tion of the New York Aquarium at Coney Is- land. This was concentrated by evaporation to the approximate density of the water in which the fish were received, and held between 24° and 29° C., usually near 25°. The floors of these aquaria were covered about one-half inch deep in whitish calcareous sand from North Bimini. This served a practical maintenance purpose, as described by Breder (1957), and took on a reflected tint of the particular plastic shields employed. It was necessary to clean this sand periodically for both sanitary and chro- matic reasons. The illumination used was en- tirely artificial and consisted of fluorescent tubes, of the so-called “warm white” type which are generally suitable for small aquaria in that they permit the satisfactory growth of green plants. This light source was directly overhead. As Histrio customarily climbs about in sar- gasso weed, it was felt advisable to provide some sort of structure which would provide approxi- mately similar support for such activity. Most of the larger marine algae, especially Sargassum, are difficult to maintain satisfactorily in small aquaria and in any case would interfere with the artificial chromatic arrangements. Therefore ar- tificial supports of small glass rods 8V2 in. high were built in the form of a tripod with extended arms of lesser-sized rods, as is shown in several of the plates. These were colorless in themselves and sanitary and at most reflected or refracted the color of the surrounding plastic shields. The spontaneous growth of adherent algae made it necessary to clean them periodically. All the fishes were obtained from sargasso weed floating in the sea near Bimini. The small- est were shipped to New York by air express in the conventional plastic bags now in general use by aquarium dealers. Because of their well known voracious appetites, the fishes were kept separated by glass partitions in the tanks. They were fed daily with live Lebistes reticulatus Pet- ers, Astyanax mexicanus (Filippi) and Tilapia heudeloti (Bleeker) of sizes suited to the indi- vidual Histrio. Most of the individuals were induced to accept cut up pieces of Tilapia and Astyanax. Two sets of experiments were carried out, both in an effort to determine the extent of in- fluence of the colors seen by the fishes on their pigmentation. In the first set, sheets of colored plastic were placed against the outside of the four glass walls of the aquaria. Matching strips of plastic were placed inside each corner of the aquaria to mask the blackish aquarium cement used to hold the glass sides in place. This plastic, a vinyl chloride-acetate copolymer to which pig- ment had been added, was first thoroughly tested, in each of the colors used, for possible biological effects and found to be entirely satis- factory. Since the colored plastic sheets, Vs" stock, were somewhat translucent, transmitted light as well as that reflected from the inner surface of the sheets enhanced the chromatic values obtained. In the second set of experiments the same aquaria were used, modified in the following way. The protective shields of colored plastic were removed and the tripods were wound with colored strips of polyethylene sheeting of 0.1 mm. or less in thickness. These, too, had been tested for biological effects and found to be satisfactory. A lesser variety of colors was em- ployed for this series, because certain results in the first series had demonstrated some to be quite unnecessary. All other details were iden- tical with those described for the first series except that corner strips were not necessary in the black aquaria, since the corners were already black. The polyethylene fringes for the glass tripods were made as follows. Plastic strips 13A in. wide were folded lengthwise. Repeated cuts were made in these at right angles to the fold by small scissors. The cuts were made from the edges which had been pressed together in folding so that the cuts nearly reached the line of fold. One end of such a fringed strip was then knotted to the lower end of the tripod and wrapped up 1958] Breder & Campbell: Pigmentation of Histrio histrio 137 it spirally. Others were attached to the cross rods. The finished appearances of these struc- tures are shown in Plate I. It was found that the fishes made considerably more use of the tripods when so decorated and frequently nestled in this imitation sargasso weed, often disappear- ing as completely as they do when nestled in the natural plant. The purpose in the design of these two series of experiments was to make a distinction, if any existed, between a general over-all coloration of the environment as compared with the details of the weed in which the fish nestle. In a state of nature, all that these fish normally view is the more or less checkered sargasso weed, and its other closely matching attendant organisms in the bluish void of the lateral viewing of the open ocean. This is only interrupted by the com- paratively infrequent appearance of some pass- ing larger fish or smaller food object. In the first case it was the general over-all water color that was changed by the slightly translucent plastic shields, while in the second case the col- ors of the “weed” itself were modified. A schedule of the period of days in which the various individuals occupied the different aquaria is given in Table 1, covering both the first and second series. In all, 42 individuals were used in these experiments, 19 of which may be considered as controls, that is, fishes kept in clear glass aquaria and not exposed to specific colors as previously described. The 23 remaining fishes were exposed for varying times and under other differing conditions to these especially prepared colors. In addition to black and white, four colors were employed for this purpose: red, yellow, green and blue. There was one aquarium so shielded, by each of the six plastic shields above named. No attempt to de- fine these colors in terms of wave length has been made, since this refinement appeared to be quite unnecessary for the purposes of the proj- ect, and as it was well known in advance that these fishes could match a fair range of sargasso Table 1. Exposure of Histrio to Various Colors, in Days Fish number Aquaria Clear White Yellow Red Green Blue Black 1 54 <85 2 19* 83 3 20 4 66 5 13 6 183 7 14 8 178 9 48 80* 48*yellow 10 54 80* 48*yelIow 11 42 80* 48*yellow 12 128*blue 25 <29 13 127 15 3*black 101 <77 16 2* 18 19 5* 118 21 5* 62 23 5* 147*32f 26 5* 7 1 *82f 32 3* 48*1 14f 35 3* 54*82f 36 3* 101 37 3* 55*1 17f Note. Arrows indicate direction of change to an- other color, named where necessary. “f” indicates tripod has fringe of color indicated. In addition to the above 23 numbered individuals there were 19 controls as follows: 14, 17, 18, 20, 22, 24, 25, 27-31, 33, 34, 38-42. There is no indication of number of companions. if any, nor of shifts without color change, incident to operations in this tabulation. The fishes arrived and were placed in the aquaria indicated according to the following tabulation. Nos. 1 to 3, 3/22/56 Nos. 29 to 37, 12/25/56 Nos. 4 to 16, 5/1/56 Nos. 40 to 41, 5/27/57 Nos. 17 to 28, 12/6/56 No. 42, 6/23/57 138 Zoologica: New York Zoological Society [43: 12 weed hues. The fringed tripods were prepared in yellow, blue, black and clear, while red devel- oped spontaneously by the overgrowth of a red alga on a tripod fringed with clear plastic. Five fish were used in connection with these deco- rated tripods, most of which had been previously exposed to plastic shields of the same color. The days spent under each condition and the se- quence of change are given in Table 1. Other details are given in the text where they are pertinent. The missing numbers of individual fishes refer to the controls, which were main- tained in ordinary aquaria up to 193 days. The details of the experiments indicated in summary in Tables 1 and 2 are as follows, ar- ranged by color and followed by those concerned with changed colors. Black— The four fish exposed to black in the manner described all gave clear and definite re- actions. They all were very dark by the time they had been continuously exposed for 2 to 3 weeks. The resulting darkening for Fish No. 1 and Fish No. 26 is shown in Plate I after ex- posures of 30 and 103 days respectively. Yellow.— The ten fish exposed to yellow all showed an appropriate change in a more yellow over-all appearance and a lessening of pattern detail. This was, however, not nearly as pro- nounced as in the case of the black backgrounds, partly, at least, because the yellow backgrounds called for less of a change from the initial con- dition of the fishes than in the former. In Plate I, Fish No. 13 and Fish No. 37 are shown with plain and fringed tripods. These illustrations were made after the fish had been exposed for 38 and 85 days respectively. Red.— The differences between the unexposed fish and those exposed to red were slight and somewhat uncertain and could not be detected in a black-and-white photograph. Consequently no illustrations are given covering these three fishes. Green.— While there was no evident response to green, three fish which were simultaneously exposed to green for 79 days are shown in Plate II. These three fish. Nos. 9, 10 and 11, all showed different basic patterns when introduced into this single partitioned aquarium and main- tained them throughout the study. The possible significance of these patterns differs. Two other fish. Nos. 12 and 15, the latter having first been exposed to black, were also exposed to green, making a series of five fish. All performed in an essentially similar manner, including the black fish. No. 15, which, while lightening slightly, agreed with the rest in showing no direct response to green. The lightening was very limited, as would be expected because of the size Table 2. Sizes of Various Individual Histrio at Particular Times Fish number Size on arrival Size when photo- graphed Date when photo- graphed Figure number 1 4/20/56 1 26 18 65 3/19/57 2 13 6/7/56 3 37 25 77 3/19/57 4 9 7/18/56 5 10 7/18/56 5 11 7/18/56 5 2 5/2/56 6 14 5/2/56 7 28 15 48 1/4/57 8 29 43 65 2/26/57 10 Note. All measurements are given as millimeters in total length. Dates of measurements are approxi- mate. and age of the fish which had long passed its peak of maximum response to background col- ors. Blue.— Six fish. Nos. 3, 4, 5, 12, 21 and 32, were exposed to blue and showed absolutely no tendency to approach that color. It seemed that the fishes did not thrive nearly as well in the blue aquarium as in the others, several early deaths appearing which it was thought might have been in some obscure fashion connected with the color involved. White.— The four fish exposed to white, Nos. 2, 16, 19 and 36, all showed either a definite lightening or an evident great increase in bril- liance in the white spots made up of crowded leucophores, or both. There was no close ap- proach, however, to white in any real sense. In- deed, this would have been very difficult, for even if all the chromatophores could be reduced, it would require a tremendous overgrowth of leucophores to mask the underlying tissue col- ors. Plate II, fig. 6, shows Fish No. 2. Clear.— In addition to the two fish, Nos. 1 and 2, there were 19 “control” fish not listed in Table 1 which were kept in similar aquaria. The long series of fish kept in clear aquaria from 2 to 5 days were held there merely preparatory to trans- fer and the periods were too short to have any bearing on these experiments. All these fish changed less than those which were exposed to colored backgrounds, which gives support to the significance of the colors to the chromatic changes observed on the experimental fishes. Plate II, fig. 7, and Plate III, fig. 8, show Fish No. 14 as kept in an ordinary clear glass aquari- um and Fish No. 28 in association with clear fringe. 1958] Breder & Campbell: Pigmentation of Histrio histrio 139 The Pigmentary System and Its Limitations The following descriptions cover the kinds and distribution of the chromatophores and related structures as viewed on the living fishes through stereoscopic dissecting microscopes. These fea- tures cover the mechanism by which color and pattern changes are effected in Histrio and the cellular types present give an effective measure of the limitations present in their particular chromatophore system. The pigment cells present are of three kinds only, melanophores, xanthophores and leuco- phores. The xanthophores are deep yellow, near- ly approaching orange. The leucophores show no iridescence, all being of the milk white type. The typical condition found in Histrio, with little variation, is as follows. The most deeply pigmented patches in the dermal pattern seem to be composed entirely of melanophores and xanthophores. If there are any leucophores pres- ent in these areas they are evidently completely obscured by the overlaying of the more deeply colored cells. Both the melanophores and xan- thophores are small in size, but exceedingly numerous and very closely crowded together in the darker areas of the skin. The sharpness of the pattern shown by Histrio is accentuated by the abrupt ending of the more heavily pigmented areas, which have very sharp lines of demarca- tion between them and the lighter areas of the pattern. In the latter there are few, if any, strag- gling melanophores or xanthophores. Some leu- cophores are present in these clear areas, but they are sparse, to the point where the vasculari- zation of the dermis can be easily distinguished. The coloration of these lighter areas is based to a large extent on the nature of the underlying tissues, including principally a reddish suffusion of the blood, a pale yellowish from fat and a whitish tint from the muscles. The opaque white spots are composed entirely of tightly packed leucophores. All these features are displayed in the post dorsal area of skin shown in Plate III, fig. 9. Although this picture was taken of a freshly formalin-fixed fish, it is still typical of the living animal. In fishes which have become nearly entirely black, as Fish Nos. 1 and 26 of Plate I, figs. 1 and 2, there has clearly been an enormous in- crease in the number of melanophores. The greatly reduced clear areas contain straggling cells of both melanophores and xanthophores and the line of demarcation between the origi- nally dark and clear areas is not nearly as sharp as in the lighter-colored fishes. Some of the re- maining white spots contain xanthophores as well as leucophores while others remain as col- onies of pure leucophores. Fishes which have become more yellow, as Fish No. 37 of Plate I, fig. 4, show more clear spaces than the controls. Straggling xantho- phores are found in these spaces but no melano- phores. None of either were found in the white spots. The xanthophores are of a particularly brilliant orange hue. It seemed that fishes which had been kept in a red aquarium just barely suggested a slightly reddish hue, but this was too indefinite to be rigidly established. Fish No. 8, which had been kept in a red aquarium for 178 days, was such a case and its faint but seemingly reddish hue was evidently caused by its deeply colored xan- thophores, combined with many black melano- phores. The particular adjustment between them may indeed have been as far as the fish was able to go in the direction of matching a red background with its rather limited chromato- phore system. The yellow fish above mentioned has its intense yellow caused, not by paler xan- thophores, but by more of them and fewer melanophores than Fish No. 8. Xanthophores on the black fish. No. 1, are mostly covered by the very numerous melanophores, but what few were visible were definitely paler than on the other fishes. It should be noted in these connections that the presence of the white patches formed of massed leucophores sometimes gives a false sense of color, as they take on the hue of their surroundings, making a fish on yellow, for in- stance, appear to be yellower than it actually is. This effect helps enhance background matching in a purely passive way. Because of this condi- tion, caution was taken in the observations to make necessary allowances, often by viewing the fishes against a neutral gray background instead of the background of the colored con- tainers. While it is clear that the colors can and do change considerably in reference to the back- ground, differences in pattern, which are evi- denced in the smallest individuals obtained, do not vary to any noticeable extent except as certain markings may be obliterated by mask- ing, as in the very dark color phases. The basic design on each fish seems to be fixed for life. Whether this pattern is largely genetic or largely fixed at a very early stage, at sizes as yet not obtainable, is still unknown. As in all fishes showing marked ability to change their coloration, there are two elements in the response which must be distinguished in order to proceed with a satisfactory discussion. The rapid changes, sometimes almost instan- taneous, are all due to prompt reactions involv- ing the dispersal or concentration of the chroma- 140 Zoologica: New York Zoological Society [43: 12 tophore pigment granules. These, while striking in their speed and extent, are not as profound as those which take a longer time to appear, covering periods of weeks or months. The first kind are under the immediate control of the nervous and hormonal systems and the reac- tions are at speeds consistent with those systems. The slower type is based on the development of more color cells and thus involves a true morphological modification. The color changes here under consideration all refer to the second or morphological changes, as many days are involved in their development. Histrio is capable of making sudden changes in coloration but it is not, as a species, especially marked in this direction. When sudden change does occur, it is more apt to be associated with some influence other than background— as for example during breeding periods when radical changes in colora- tion sometimes appear for short periods. These changes are sometimes striking to the eye but are not nearly as fundamental as the great increase in the numbers of melanophores which an individual will develop against a black back- ground. Discussion It is shown in the preceding sections that Histrio responds rather slowly, but very defi- nitely, to colors in its surroundings. These re- sponses are evidently rather severely restricted to the colors found in sargasso weed, or perhaps with little extensions of them into blacks and yellows. This situation is evidently a matter of the physical limitations of the chromatophore system. Whether there is also a behavioral limita- tion involved, of course, cannot be verified by direct experimentation. Since, however, the not distantly related Antennarius can, in various of its species, show a much greater range of chro- matic adjustment and necessarily has a larger number of types of chromatophores, it would seem that Histrio, if not so restricted by its chromatophore system, should be able to do as well. This view is supported by the excellent matching effect which Histrio attains with its limited chromatic abilities within the environ- ment with which it is normally associated. Nearly all of the individuals of Histrio en- countered in the field are found in close associa- tion with Sargassum, both from personal ob- servation and the reports of collectors and the literature. The initial observation, which prompted the undertaking of the experiments herein described, was made in an aquarium in the Lerner Marine Laboratory. This tank, which had a light sand bottom, had been painted black on its back and two ends. Into it had been inad- vertently dropped a very small Histrio incident to some entirely unrelated operations, on No- vember 15, 1954. The only other contents of this 3 X 1 V2 X 1 V2 foot aquarium were the light sand bottom and a dozen or so small Sardinella. There was no Sargassum or other floating weed. The small Histrio usually was to be seen wedged in a corner near the surface. This fish, originally of the usual coloration and pattern, by Decem- ber 3 was a perfect match for the dead black background against which it had been living. Not only was the black background matched perfectly but there appeared on the sides of the fish a few small milk-white spots, which were the size and shape of a scattering of equally white volunteer calcareous growths on the black glass. The resemblance of the white spots to the growths may have been purely accidental, as these fishes normally display some milk-white spots which usually go unnoticed in their other- wise mottled pattern. However, as was noticed in the later experiments, the very smallest fishes made the most striking adjustments to back- ground, whereas the larger in no case did nearly so well. It is possible that the size and shape of the white spots on the fish above described were pattern adjustments, beyond the capabilities of the considerably larger fishes studied later. It may be that much greater adjustment to both colors and pattern of background is a capability of these fishes at sizes well below those ordi- narily obtainable. It would certainly be desirable to work from the transparent planktonic larvae just before they settle in the Sargassum in any attempt to analyze the potentialities of smaller fish. Such material is not obtainable by any pres- ently known means. Aside from this fish that turned black in an aquarium, there appears to be no previous record of black Histrio. Three years later the following observations were made near the laboratory. These constitute the only known occurrences of black Histrio in a state of nature and have a very distinct bearing on the present experiments. Dr. Vladimir Walters, working on other mat- ters, collected three such black Histrio on March 30, 1957, in one of the passages through the mangrove stands along the south side of South Bimini. The specimens were subsequently lost in shipment, but he described them as closely resembling the most thoroughly black fishes developed in the laboratory. Although many persons working out of the Lerner Marine Labo- ratory had collected in this same place and in many others basically similar to it, there have been no other reports of Histrio in these places whatever. In the present case, rotenone was used in collecting, which may be responsible for find- 1958] Breder <£ Campbell: Pigmentation of Histrio histrio 141 ing them at all. It is to be noted that especially heavy weather the preceding month had brought quantities of sargasso weed closer to the man- grove stands than usual. The presumed fate of these black Histrio, stranded in a mangrove swamp and living against a very black background, would seem to be obvious enough even if the precise reasons are not clear. In the light of aquarium observations, it would seem that such fish should be able to manage well enough, by taking on the habits of an Antennarius, which is something they may approximate in an aquarium without arti- ficial weed. That they are not very successful in mangrove swamps is evidenced by their nor- mal absence from such places. At least, it would appear that this is not based on any inability to match the mangrove background. It could be that there is some element in the mangrove association with which they are unable to cope. This could conceivably be the frequent abun- dance of large callinectids, so often found in such associations. If actual destruction of the fishes did not take place by some such means, it is conceivable that the fish might escape back to their normal environment, a seemingly un- likely event. The general reluctance of fishes to leave a background to which they are chro- matically adjusted would have to be over-ridden at some stage in any attempt to work out of a mangrove stand. There are no available step- by-step environmental niches through which they could make an easy transition from man- grove-black to sargasso-mottled, especially in view of the sluggishness of their chromatic ad- justments. The ability of Histrio to assume the behavior usually associated with Antennarius, as above noted, was nicely demonstrated by one that was placed in an aquarium where the sole fitting was a small shell, Melongena. This it associated with to the extent of spending much time huddled against it or even within the mouth of the shell. A typical posture of this fish is shown in Plate III, fig. 10. Not only were the postures reminis- cent of Antennarius but often the movements of the pectorals and the creeping about on the aquarium floor resembled the movements of that genus rather than Histrio. As might be expected, the darker markings of this fish took on a brown hue very close to that of the brown markings around the lip of the shell. The formal experiments agree well with the field observations and of course show that His- trio is well able to make chromatic adjustments involving mottled patterns from black through various brown and yellowish tans to a fair yel- low. Also because of the islands of leucophores, they are able to take on the appearance of over- laid whitish patches, not unlike various of the sessile epizootics on Sargassum. Actually this range goes little further than covering the range of possibilities of this somewhat variable weed but to which can be added any dominantly black environment such as a mangrove swamp. Other backgrounds do not yield such striking results. Green and blue, which are the backgrounds which one would expect to call for the presence of iridophores and combinations of chromato- phores into chromatosomes, yielded no recog- nizable response. Red, which might be expected to call for erythrophores, which too, are absent, yielded only questionable response of a slight sort. Actually what appeared to be a slight re- sponse was caused by a particular mixture of melanophores and xanthophores. This indeed may have been a tendency to produce a red color but one impossible of attaining any real success with the pigments present. White and clear, the latter in the form of plastic fringes, caused only a general lightening of the fish, i. e. a reduction in the number of melanophores and xantho- phores with perhaps some increase in the num- ber of leucophores. While this results in a better matching than red, green or blue, it is not in the same class at all with the effects induced by black or yellow. It is interesting in the above connections to consider the three fish shown in Plate II, fig. 5. Not only do they show marked differences in pattern but also at least equally marked differ- ences in amounts of pigmentation. As may be seen in the photograph, the fish at the left showed a comparatively heavily pigmented pattern of broad blotches while the central fish showed conspicuously less pigmentation and the right hand fish showed pigmentation comparable to that of the left hand fish but in a much finer pattern. It would seem almost that these fishes either reacted in no way at all to the green color and retained whatever tendency they had been impressed with before capture, or that they each reacted in a distinctly different manner to a common background which was neither very light nor very dark. It is evident in any case that this color green, to which so many fishes respond by matching, is evidently completely out of the chromatic reactivity of these fishes which respond so markedly to the yellow browns of the weeds which they normally inhabit. In these studies, several fish showed aberra- tions in their chromatic behavior that are not easily explained, the details of which are given in the preceding section. These would all appear to be due to some derangement in the complex system responsible for fishes responding with 142 Zoologica: New York Zoological Society [43: 12 their pigmentary effectors from colors thrown on the retina, as the mediating receptor. This is too poorly understood and too complex a matter to be discussed in present connections, except to suggest that these sometime reverse responses may be at the basis of the cause which impels a fish to change from a background- matching habit to one of background-opposing. Influences bearing on these two diametrically opposed reactions to background are further discussed by Breder & Rasquin (1955) and Breder (1955) in reference to other fishes. While it has been impossible to obtain suffi- cient material of Antennarius to perform a set of parallel experiments, observations on these fishes show them to have chromatic possibilities far exceeding that of Histrio. Barbour (1942) and Schultz ( 1957) indicated the highly variable coloration and pattern differences seen in these fishes. Individuals of Antennarius multiocellatus (Cuvier & Valenciennes) have been seen to vary through shades of red, vivid green, brilliant yel- lows, black and gray or brown mottlings. These are believed to represent only a few of the chro- matically possible arrangements of which this particular species is capable. Not only does this species, at least, show such background-match- ing proclivities, but may alternate it with back- ground-opposing coloration. The behavior of one individual in regard to this latter feature is discussed by Breder (1949). It is to be noted in reference to the above that this behavior in Antennarius was found to occur in environ- ments where a variety of other fishes also show the background-opposing response. An examination by Dr. Walters, of the pig- mentary system of a freshly fixed specimen of Antennarius ocellatus (Bloch & Schneider) from the Dry Tortugas, showed the following condi- tions. This fish, still pinkish in coloration, pos- sessed melanophores, erythrophores and leuco- phores, but no xanthophores or iridophores. A full study of the presences and absences of the various chromatophore types should be inter- esting and illuminating. These considerations lead to the idea that His- trio certainly descended from some form less restricted as to habitat, that it had an ancestor with greater chromatic range, perhaps similar to that found in Antennarius, and that therefore the condition presently found in Histrio is to be considered one of loss of certain pigmentary elements. Schultz (1957) considers Histrio and Antennarius as derived separately from what he calls “antennariid stock” rather than as having any closer relationship. That these retained pig- mentary elements permit the fish little ability to change color outside the range of those found in Sargassum certainly suggests a long-time asso- tion between the two. Also in reference to the Histrio in the mangrove swamp, it further sug- gests that such occurrences are probably very rare and are likely terminal wastage rather than some obscure but normal alternative open to these fishes in the ordinary run of their lives. Summary 1. Histrio has little ability to match back- grounds outside of the range of colors dis- played by Sargassum. 2. Black is well matched and yellow not quite so well while with other colors attempts to match are absent or negligible. 3. The only chromatophores present in the in- tegument of Histrio are melanophores, xan- thophores and leucophores, a condition which in itself is perhaps sufficient to explain the restricted chromatic ability of these fishes. 4. The related Antennarius is capable of much greater chromatic adjustment because of the presence of other types of chromatophores and is found to inhabit places which expose it to backgrounds which show a correspond- ingly greater range of colors. 5. It is inferred that Histrio descended from a chromatically more competent ancestor and the present chromatophore complement is one of loss and indicates a long association with Sargassum, in which these present ele- ments have sufficed. Bibliography Barbour, T. 1942. The northwestern Atlantic species of frog fishes. Proc. New England Zool. Club, 19: 21-40. Braam Houckgeest, A. E. van 1774. Bericht wegens den Lophius histrio. Verh. Holl. Maatsch. Wet., Haarlem, 15: 20-28. Breder, C. M., Jr. 1946. An analysis of the deceptive resemblances of fishes to plant parts, with critical re- marks on coloration, mimicry and adapta- tion. Bull Bingham Oceanogr. Coll., 10: (2), 1-49. 1949. On the relationship of social behavior to pigmentation in tropical shore fishes. Bull. Amer. Mus. Nat. Hist., 94: (2), 83-106. 1955. Special features of visibility reduction in flatfishes. Zoologica, 40: (8), 91-98. 1957. Miniature circulating systems for small laboratory aquaria. Zoologica, 42: (1), 1-9. 1958] Breder & Campbell: Pigmentation of Histrio histrio 143 Breder, C. M., Jr., & P. Rasquin 1955. Further notes on the pigmentary behavior of Chaetodipterus in reference to back- ground and water transparency. Zoologica, 40: (7), 85-90. Gordon, M. 1938. Animals of the Sargasso Sea merry-go- round. Nat. Hist., N. Y., 42: (1), 12-20. Ives, J. E. 1889. Mimicry of the environment in Ptero- phryne histrio. Proc. Acad. Nat. Sci. Phila.. 41: 344-345. Rasquin, P. 1958. Ovarian morphology and early embry- ology of the pediculates Antennarius and Histrio. Bull. Amer. Mus. Nat. Hist. 114: (4), 327-372. Schultz, L. P. 1957. The frogfishes of the family Antennariidae. Proc. U. S. Natl. Mus., 107: (3383), 45- 105. VlGNON, M. P. 1931. Le mimetisme chez les animaux marins. Terre et la vie, 1: 131-150. 144 Zoologica: New York Zoological Society [43: 12: 1958] EXPLANATION OF THE PLATES Note. In all photographs of living Histrio a tem- porary background of a neutral gray card was sub- stituted for colored plastic shields, except in Plate 2, fig. 5, in which the green plastic shield was left in place. In all but the excepted case this was necessary to make these fish sufficiently visible for photography. It is to be especially noted that in re- duction to monochrome the pictures lose much of the effective resemblance these fishes exhibit toward their colored backgrounds. Plate I Histrio after exposure to black and to yellow Fig. 1. Fish No. 1 after 30 days behind black shields. Fig. 2. Fish No. 26 after 82 days with black- fringed tripod. Fig. 3. Fish No. 13 after 38 days behind yellow shields. Fig. 4. Fish No. 37 after 85 days with yellow- fringed tripod. Plate II Histrio; various exposures Fig. 5. Fishes Nos. 9, 10 and 11 after 79 days behind green shields, showing basic pat- tern differences. Fig. 6. Fish No. 2 after 23 days behind white shields. Fig. 7. Fish No. 14 the day after arrival, for comparison with the experimental fishes. Plate III Histrio; chromatic and behavior details Fig. 8. Fish No. 28 after 85 days with clear- fringed tripod. Fig. 9. Magnified area at dorsal base on the right side of Fish No. 12 showing the main fea- tures of chromatophore cell types. Fig. 10. Fish No. 29 after 64 days’ association with a shell. BREDER & CAMPBELL PLATE I FIG. 1 FIG. 3 FIG. 2 FIG. 4 THE INFLUENCE OF ENVIRONMENT ON THE PIGMENTATION OF HISTRIO HISTRIO (LINNAEUS) BREDER & CAMPBELL PLATE II FIG. 5 FIG. 6 FIG. 7 THE INFLUENCE OF ENVIRONMENT ON THE PIGMENTATION OF HISTRIO HISTRIO (LINNAEUS) BREDER & CAMPBELL PLATE III FIG. 8 FIG. 9 FIG. 10 THE INFLUENCE OF ENVIRONMENT ON THE PIGMENTATION OF HISTRIO HISTRIO (LINNAEUS) 13 Studies on the Histology and Histopathology of the Rainbow Trout, Salmo gairdneri irideus. I. Hematology: Under Normal and Experimental Conditions of Inflammation1 Eva Lurie Weinreb Department of Zoology, University of Wisconsin, Madison (Plate I) MOST studies in comparative hematol- ogy have been limited to descriptions of the staining properties of fixed cells. The present study was made on living blood cells of the rainbow trout, Salmo gairdneri irideus, using phase contrast microscopy, to fur- ther elucidate physiological as well as mor- phological changes under experimental condi- tions. Such changes were compared with similar effects reported in more common laboratory animals. The circulatory system of the rainbow trout is sensitive to foreign stimuli and reflects the homeostasis of the animal. Changes in the blood picture were used, therefore, as criteria of sys- temic response to experimental conditions. Be- fore such changes could be evaluated, the nor- mal blood picture had to be determined and a standard established. Descriptive studies of teleost blood cells of various species have included perch (Yokoyama, 1947), carp and brook trout (Dombrowski, 1953) and salmon (Watson et al., 1956). De- tailed reports on rainbow trout have not been found. The response of leukocytes to various irritants has been described for many laboratory animals including dogfish shark (Reznikoff & Reznikoff, 1934), turtle (Charipper & Davis, 1932; Ryerson, 1943), perch (Yokoyama, 1947), mice and rats (Harlow & Selye, 1937). The influence of adrenal cortical hormones and ACTH on circulating leukocytes in mice, rats and rabbits has also been described (Dougherty & White, 1944; Palmer et al., 1951). The rela- 1A revised portion of the thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy at the University of Wisconsin. tion among leukocytes, ACTH and adrenal cor- tex was reviewed by Sayers ( 1950) . Although the blood-forming centers in fish differ from those found in mammals, and the individual cells vary betwen the classes, the author is retaining such terms as myeloid, myelo- cyte and lymphoid in order to be consistent with literature on fish hematology and to describe the various stages in granulocyte maturation in the fish as compared to similar well-defined stages in mammals. The prefix myelo- in this text is used not in reference to bone marrow, but to granulocytic as distinct from lymphoid elements. The author is grateful to Dr. Nellie M. Bilstad for her guidance, to Dr. John C. Neess for assist- ance with statistical methods, to Dr. Stanley Weinreb for preparation of the photomicro- graphs and critical reading of the manuscript, and to Mr. Henry Eichhorn, Jr., for technical assistance. This work has been supported by funds made available by the Wisconsin Alumni Research Foundation. Normal Blood Picture Materials and Methods The animals used for all studies were from the same hatching group (October 9, 1953). Fish were delivered from the Nevins State Fish Hatchery, Madison, Wisconsin, to the Univer- sity of Wisconsin Lake Laboratory and main- tained in oxygenated water at 12-12.5°C. for several days so they might become acclimated to the experimental tanks. The trout, of both sexes, averaged 15 cm. fork length and 50 gm. weight. The age ranged from 12 months at the start of this study to 21 months at its completion. 145 146 Zoologica: New York Zoological Society [43: 13 Since the variability of cell counts is great, each experimental group consisted of animals kept under the same conditions and of the same age and approximate size. After severing the tail, blood samples were collected from the haemal vessels, using no anticoagulants, and each animal was autopsied, with particular attention being paid to the hemopoietic structures. All trout used were in normal condition as far as could be determined. Living blood cells were studied and counted by phase contrast microscopy, using a dark con- trast-medium oil immersion objective. Wright’s stained smears were used for comparison (Plate I). The former method was found to afford greater accuracy of identification and precision in counting, the artefacts of staining and loss of fragile heterophils being minimal. Total red and white cell counts were made (Yokoyama, 1947), in order to determine the total number variability in different animals under normal and experimental conditions. Since the erythro- cyte count did not exhibit significant change, it was used as the basis for the differential leuko- cyte counts: all leukocytes within the range of 300 red blood cells were counted. The cells counted included heterophils, lymphocytes and thrombocytes; the number of eosinophils and basophils was negligible, and monocytes were absent. Results The criteria for cell identification in phase studies include size, shape, degree of motility, nuclear-cytoplasmic ratio, nuclear size and shape, chromatin pattern, cytoplasmic granula- tion and mitochondria. The mature erythrocyte is oval and nucleated, with abundant cytoplasm containing many very small, motile mitochondria. The smaller, round- er polychromatophilic erythrocyte differs in hav- ing less hemoglobin and a heavier nuclear mem- brane, similar to that of the lymphocyte. Lym- phocytes are the most abundant, being smaller and rounder than the red cell, often exhibiting pseudopod formation. The round or slightly indented nucleus is surrounded by a thin rim of cytoplasm with motile, rod-shaped mitochondria. The outstanding feature of the lymphocyte is the nuclear pattern; the chromatin forms a dis- tinct network of alternately dark and highly re- fractile, light areas. In the highly amoeboid heterophil, the par- tially lobed nucleus is in almost continuous roll- ing motion, while, after staining, the nucleus is lobed or ribbon-like. The cytoplasm, which exhibits continuous streaming and pseudopod formation, is filled with granules and filamentous mitochrondia. Cells comparable to myelocytes and metamyelocytes are not uncommon and may be distinguished from the mature heterophil by being less motile and round to oval in shape, with a more rounded nucleus and less granula- tion. The myelocyte is smaller and more rounded than either of the older cells. In both immature cell types the nucleus exhibits a folding motion with wrinkling of the surface, rather than the distinct roll of the heterophil. Although thrombocytes are usually in the blebbed state, intact cells are visible. Young cells are similar to lymphocytes, although smaller and rounder. Mature thrombocytes are elongated. In contrast to the lymphocyte, the cell may be identified by its indistinct chromatin network, paucity of cytoplasmic granules and barely mo- tile mitochondria. Occasional eosinophils and basophils are seen. Eosinophils are slightly larger than lymphocytes, with numerous, extremely refractile, spherical granules. The basophil, which is the largest leu- kocyte, has an eccentric spherical nucleus, in- distinct chromatin and a very prominent nucleo- lus. The large, round cytoplasmic granules are characterized by an internal, laminated or striated pattern. Macrophages, which are not usually found in a normal blood smear, are prominent in bac- terial and other infections. These cells are the largest in the circulation, being twice the size of the basophil. They contain abundant debris. Leukocyte counts were made on 10 normal rainbow trout. The number of fish leukocytes, particularly lymphocytes and thrombocytes, varies. In order to establish a normal standard, the mean ± standard error of mean (S.E.) and the range of mean, where range is the distance between the upper and lower 95% confidence limits, were determined for each cell type. The mean ± S.E. and range of mean for the hetero- phil, lymphocyte and thrombocyte are 2.3 ± 0.66 (0.98-3.62), 20.5 ± 2.31 (15.88-25.12) and 11.1 ± 2.33 (6.44-15.76), respectively. Effects of Turpentine, Cortisone and ACTH on Leukocytes Materials and Methods Trout were injected intraperitoneally under the pelvic fin with 0.2 cc./lOO gm. of turpentine, N.F. Controls were given injections of 0.6 cc./ 1 00 gm. of Ringer-Locke solution (0.65% ) . The animals were divided into two series; the first was sacrificed after 1, 3, 5, 7 and 24-hour inter- vals, the second after 1, 3, 5, 6, 7, 10, 24, 48, 60 and 72-hour intervals. Differential leukocyte counts were made after each time interval. Total 1958] Weinreb: Hematology of Rainbow Trout 147 red and white cell counts were also made 6 and 24 hours after turpentine injection. To study the effect of cortisone on leukocyte response, trout were divided into two series; each received 0.2 cc./lOO gm. of turpentine. Those in series 1 were given intraperitoneal injection of cortisone (Cortone Acetate, Merck and Co.) in saline suspension concurrently with the tur- pentine, while series 2 received cortisone 24 hours in advance of the turpentine. Animals were subdivided into dosage groups, receiving 0.3, 0.6 and 1.0 mg./cc., respectively. Controls were injected with 1.0 mg./cc. of cortisone with- out turpentine. Blood samples were taken 6 and 24 hours after turpentine injection as well as after the control cortisone injection. To determine whether administration of ACTH would stimulate adrenal tissue to secrete a hormone eliciting the same leukocyte response as cortisone, another group of trout was injected with ACTH (Corticotropin, ACTH, Armour Laboratories) in Ringer-Locke solution. Do- sages of 2 U. S. P. units (I. U.) were used and blood samples taken after 6 and 24 hours. Leukocyte counts were made and the 95% range of the mean in normal and experimental trout compared. Data were considered statis- tically significant when two ranges failed to overlap. Results Injections of Ringer-Locke solution resulted in no significant change in either heterophil or lymphocyte counts. The slight increase in hetero- phils encountered in a few animals was attrib- uted to handling and other unknown factors. Turpentine injection elicited a marked response. Total red cell counts did not differ significantly from normal, while total leukocyte counts were notably increased. Leukocyte counts of the first series, over a 24-hour period, are given in Table 1; those for series 2, over a 72-hour period, in Table 2. In response to injected turpentine, the hetero- phil count rose significantly within 5 hours, reached its peak at 6 hours and remained high for 60 hours. The lymphocyte count dropped significantly in 6 hours, and remained low for 72 hours. No significant change occurred in thrombocyte number in the first 24 hours, but decreases occurred at 48 and 60 hours. The early increase in heterophils is mainly due to release of myelocytes and metamyelocytes into the cir- culation. The mortality rate following turpentine injec- tion averaged 30%, the greatest loss occurring between the second and fourth post-injection hours. Leukocyte counts of trout following cortisone and concurrent injections of turpentine and cortisone are listed in Table 3. Counts taken after prior injections of cortisone are given in Table 4. Cortisone alone had little effect on the hetero- phil count; the change in lymphocyte count, however, was marked. Lymphopenia was noted 6 and 24 hours after injection, accompanied by thrombocytopenia. Cortisone, when given con- currently with turpentine, resulted in less hetero- philia than did turpentine alone, the increase being due to a preponderance of mature hetero- phils. The effect of cortisone was most apparent when injected in advance of the irritant. Al- though the largest dose yielded the maximum effect, little difference existed between dosage groups. After prior injection of cortisone, lymphopenia and thrombocytopenia were greater than that noted after concurrent injections. A comparison of mortality rates indicates that cortisone has a significant effect. Concurrent injections in series I maintained the average (30%) mortality rate. However, prior injections in series 2 average only 16%, a reduction of almost one-half. ACTH injection resulted in lymphopenia and thrombocytopenia, as did cortisone. In addition, heterophilia, due to increased mature cells, re- sulted. Leukocyte counts are given in Table 5. Discussion The blood response elicited in trout by tur- pentine is similar to that reported in other ani- mals, including the dogfish (Reznikoff & Rez- nikoff, 1934), turtle (Ryerson, 1943), chicken (Bradley, 1937) and perch (Yokoyama, 1947). Comparable effects were also reported in mice and rats after injections of adrenalin or formal- dehyde (Harlow & Selye, 1937). The response of the trout was not limited to the irritant, but wa~ also an expression of the shock reaction to the toxin. The high mortality seen in the first hours is attributed to this. The leukocyte response to pituitary and adrenal cor- tical hormones in trout is comparable to that noted in mammals. A decrease in leukocytosis (due to neutrophilia) was reported by Palmer et al. (1951) in turpentine-injected rats after administration of cortisone and ACTH, with the greater reduction following cortisone. The time relationship between administration of cortisone and turpentine was more important than cor- tisone dosage. Maximal inhibition of inflamma- tion was obtained after prior injection of cor- tisone, which permitted adequate time for ab- sorption. The effect on mortality also appears to 148 Zoologica: New York Zoological Society [43: 13 be dependent upon absorption sufficient to in- hibit shock. The heterophilia noted in trout after admin- istration of ACTH is similar to the neutrophilia elicited by ACTH in intact and adrenalecto- mized rats (Palmer et al., 1951) and in rats, mice and rabbits following ACTH or foreign protein (Dougherty & White, 1944). The neu- trophil or heterophil, therefore, is not under direct adrenal cortical control, but is subject to various influences. Lymphopenia following cortisone or ACTH is a more specific response. This reaction, re- ported absent after adrenalectomy or injection of other proteins (Dougherty & White, 1944), results from many unrelated stimuli, including turpentine, and is due to increased adrenal cor- tical activity initiated by ACTH. Decreased lymphocyte number with stress or hormone treatment in intact animals appears to follow adrenal cortical inhibition of the lymphoid organs. In rainbow trout, lymphopenia and throm- bocytopenia resulted from stress, cortisone and ACTH, while heterophilia followed stress and ACTH injection. Therefore, the mechanism of leukocyte control, as well as the physiological response of each cell type, in the trout is com- parable to that in the mammal. Since significant heterophilia was not caused by cortisone, it is inferred that granulocyte-forming centers are not under adrenal cortical control, as may be the case with lymphocytes and thrombocytes. In the trout, where granulocyte- and agranulocyte- forming tissue are located in the same hemo- poietic organs, multiple controls exist. Summary 1. Living blood cells from rainbow trout are described, and differential counts made, using phase contrast microscopy. In normal blood the predominant leukocyte is the lym- phocyte, the heterophil is scarce, and eosino- phils and basophils are seen only occasion- ally. 2. Changes in blood picture are used as criteria of systemic response to experimental condi- tions; the effects of turpentine, cortisone and ACTH on the leukocyte counts are deter- mined. 3. Turpentine produced a sterile inflammation, resulting in heterophilia, lymphopenia and thrombocytopenia. Blood counts, made over a 72-hour period, revealed marked hetero- philia at 6 hours due to release of myelocytes and metamyelocytes. 4. Cortisone, given concurrently and in advance of turpentine, reduced the inflammatory w .Its h t C o o T-H On d '—i V£> +1 »n NO + 1 ON Cl +1 •o NO +1 lO Cl in X +1 »n o (N Cl +1 NO Cl +1 O o d +1 o d +1 •n +1 o 1958] Weinreb: Hematology of Rainbow Trout 149 o u ts d o Z o o z £ o hJ -I o pH Q O 2 U4 Oh ei D O S aJ w > O C/D h* £ D O o U4 H u o p w hJ W p CQ < H rp PS CL. o W e-a H > L ON VO © q q q o oo oo ^ q q o o — I — ^ ri oo VO Cl "5J* P" q o o rn r' o 'd +1 r~ qvoqq VO © VO VD © P" © © ON p- +1 "3* © ON © © © q oo o rd r- cd < ON vd +1 rf; +1 rl © © © vd o © r-t m ON vd © vo vo r} ri CN © © +1 © © ON ri ri © © C (U 6 O CO o .52 as \ O cn • u O .9 'O G O & 03 d> ^ o\ 6 p vd r- +1 p p vd ^ y *- o> ^•59 6 § c pH Tt o rn +1 SO ri +1 poo rf tri as O VO O P o r-* d VO o Tf c4 Tt; oo\oo © — h m O +1 VO Tj- 00 o ON Tf Tt" + 1 pt^OO vc >d oo O (N VO o f^4 + 1 (N O O VO + 1 O O VO* ^f* J- ^3.Sq O o C O c « p o oc o O ^ 00 p d Ti- n 4.58-10.62 5.77 4.92 ± 0.94 3.04-6.8 1.92 3.64 ± 2.72 0-9.08 Table 4. Leukocyte Counts Following Injection of 0.2 cc./lOO gm. Turpentine and Cortisone Administered 24 Hours in Advance 1958] Weinreb: Hematology of Rainbow Trout 151 o w wo ri oo r- o NO wo n wo oo «o 1 o CO ri ri T-H d d +1 +1 +i +i +1 +1 +i X5 c wo oo On WO co wo H c3 ri CO £ d Tt — to O O o o o q os q o o o O O O o o o O ON NO HOW H- -h r~- &Q £ co co r- wo pi co rf t— H wo O rf wo to wo $ Jh d o o d wo o ON CO ON co -4 d NO CO CO NO CO r- oo Tf Pi © d ri wo T-H d t— H T-H t— • NO On T-H $ d CO Jn WO n ~+ to Os NO ON to wo NO Tf co d w o CO IS W OO OO ON CO CL, CO n to o r- Lh ri T-H T-H d ri r-H a> +1 +1 +1 +1 +i +i +1 CJ NO CO r- wo wo r- CO to s OO ON c o o o O O O O OO O O O O O O O o o o -H CO OS O ON NO O ^ O CO wo NO on re- *H -~ re c NO NO NO Tf rf 24 H c ,* CL, Lh u, 5 3 152 Zoologica : New York Zoological Society [43: 13 0) DC c ca fZ $ E o 1- ja H W +1 SC H O < tu O o o ^ o q u-i O ON Z P P- co D H U O Uj D W hJ o Si cl £ i-l W u 00 a ed cal $ in Os >o d q m o vo o o d fN O oo —1 d w j ra < H .a a, o SC W +1 c cs +1