QL 640.$ .vug HARVARD UNIVERSITY Library of the Museum of Comparative Zoology Vertebrate Ecology and Systematics A Tribute to Henry S. Fitch Edited By Richard A. Seigel Lawrence E. Hunt James L. Knight Luis Malaret Nancy L. Zuschlag The University of Kansas Museum of Natural History o \JkS UNIVERSITY OF KANSAS PUBLICATIONS MUSEUM OF NATURAL HISTORY Copies of publications may be purchased from the Publications Secretary, Museum of Natural History, University of Kansas, Law- rence, Kansas 66045. HARVARD UNIVERSITY H Library of the Museum of Comparative Zoology Front cover: The head of an adult Osage Copperhead (Agkistrodon contort rixphaeo- gaster) from Douglas County, Kansas. Drawing © 1984 by Linda Dryden. University of Kansas Museum of Natural History Special Publication No. 10 21 June 1984 Vertebrate Ecology and Systematics A Tribute to Henry S, Fitch Edited By Richard A. Seigel Lawrence E. Hunt James L. Knight Luis Malaret Nancy L. Zuschlag Museum of Natural History Department of Systematics and Ecology The University of Kansas Laurence. Kansas 66045 University of Kansas Laurence 1984 University of Kansas Publications Museum of Natural History Editor: Joseph T. Collins MU5, COMR ZOOL LIBRARY JUL U HARVARD UNIVERSITV Special Publication No. 10 pp. i-viii; 1-278; 79 figures 86 tables; 2 appendices Published 21 June 1984 Copyrighted 1984 By Museum of Natural History University of Kansas Lawrence, Kansas 66045 U.S.A. Printed By Alltn Press, Inc. Lawrence. Kansas 66044 ISBN: 89338-019-0 CONTENTS PART I. INTRODUCTION Henry S. Fitch in Perspective William E. Duellman 3 The Published Contributions of Henry S. Fitch Virginia R. Fitch 5 PART II. REPRODUCTIVE BIOLOGY AND POPULATION DYNAMICS Growth, Reproduction and Demography of the Racer, Coluber constrictor mormon, in Northern Utah William S. Brown and William S. Parker 13 Growth of Bullsnakes {Pituophis melanoleucus sayi) on a Sand Prairie in South Central Kansas Dwight R. Piatt 41 Communal Denning in Snakes Patrick T. Gregory 57 Parameters of Two Populations of Diamondback Terrapins (Malaclemys terrapin) on the Atlantic Coast of Florida Richard A. Seigel 77 An Ecological Study of the Cricket Frog, Acris crepitans Ray D. Burkett <• 89 Female Reproduction in an Arkansas Population of Rough Green Snakes (Opheodrys aes- tivus) Michael V. Plummer 105 Clutch Size in Iguana iguana in Central Panama A. Stanley Rand 115 Are Anuran Amphibians Heavy Metal Accumulators? Russell J. Hall and Bernard M. Mulhern 123 PART III. FEEDING AND BEHAVIOR Energetics of Sit-and-Wait and Widely-Searching Lizard Predators Robin M. Andrews 137 Feeding Behavior and Diet of the Eastern Coral Snake, Micrurus fu/vius Harry W. Greene 147 The Role of Chemoreception in the Prey Selection of Neonate Reptiles Pennie H. von Achen and James L. Rakestraw 163 Ecology of Small Fossorial Australian Snakes of the Genera Neelaps and Simoselaps (Ser- pentes, Elapidae) Richard Shine 173 Scaphwdontophis (Serpentes, Colubridae): Natural History and Test of a Mimicry-Related Hypothesis Robert W. Henderson 185 Dominance in Snakes Charles C. Carpenter 195 An Experimental Study of Variation in Habitat Selection and Occurrence of the Deermouse, Peromyscus maniculatus gracilis John H. Fitch 203 PART IV. SYSTEMATICS AND BIOGEOGRAPHY Herpetogcography in the Mazatlan-Durango Region of the Sierra Madre Occidental, Mexico Robert G. Webb 217 Systematic Review of the Percid Fish, Etheostoma lepidum Alice F. Echelle, Anthony A. Echelle, and Clark Hubbs 243 Anolis fttchi, a New Species of the Anolis aequatorialis Group from Ecuador and Colombia Ernest E. Williams and William E. Duellman 257 INDEX TO SCIENTIFIC NAMES 267 IV Preface This volume is the result of a symposium en- titled. "Perspectives in Fitchian Ecology," held on 9 August 1 980 in conjunction with the annual meetings of the Society for the Study of Am- phibians and Reptiles and the Herpetologists' League at Milwaukee, Wisconsin. The sympo- sium was organized to honor Dr. Henry S. Fitch on the occasion of his retirement in June 1980 after 32 years with the Department of System- atics and Ecology at the University of Kansas. Sixteen papers were presented in two sessions during the symposium and. aside from a few additions, the organizational format of this vol- ume closely follows that of the symposium. Manuscripts were submitted and accepted in late 1980 and 1981. but authors were given an op- portunity to update their contributions in early 1 983. In organizing the symposium we were sur- prised by the breadth of research conducted by the participants. Because of Fitch's influence on his past and present students and colleagues, this volume is not restricted to herpetological con- tributions. Thus, the topical emphasis of this vol- ume reflects Fitch's own research interests. The following is a breakdown by subject of the papers contained in this volume versus Fitch's pub- lished papers: ecology (this volume: 78%, Fitch: (73%); systematics and biogeography (17% vs. 19%): conservation (5% vs. 5%); and by taxo- nomic emphasis: squamates (this volume: 73%; Fitch: 62%); other amphibians and reptiles (14% vs. 7%); other vertebrates (13% vs. 23%). We wish to thank Max A. Nickerson of the Milwaukee Public Museum and Al Williams of the University of Wisconsin-Milwaukee and their respective staffs for logistical support in arrang- ing and conducting the symposium. A special note of thanks is extended to Virginia Fitch and other members of the Fitch family for assistance in the development of the symposium. The or- ganizational advice and encouragement of Wil- liam E. Duellman, Curator. Division of Herpe- tology; Philip S. Humphrey, Director, Museum of Natural History; and Richard F. Johnston. Chairman. Department of Systematics and Ecol- ogy, is greatly appreciated. Joseph T. Collins, Editor. Museum Publica- tions, deserves special recognition for his helpful advice and continued patience in answering our many questions concerning the development and execution of the symposium and this volume. The cheerful and patient assistance of Rose Etta Kurtz was invaluable. Finally, we are most grateful to the following persons for reviewing the manuscripts appearing in this volume: Robert D. Aldridge. Stevan J. Arnold, Reeve M. Bailey. Royce E. Ballinger. Thomas J. Berger. William S. Brown. Gordon M. Burghardt, Janalee P. Caldwell, Jonathan A. Campbell, David C. Cannatella. David K. Chisz- ar, Martha L. Crump, Arthur E. Dunham. Don- ald G. Dunlap. Henry S. Fitch. Darrell Frost. J. Whitfield Gibbons. Peter Gray, Harry W. Greene. Wendy Gorman. Harold Heatwole. James E. Huheey, John B. Iverson, Keith V. Kardong. Pe- ter Klopfer, Carl Lieb, Harvey B. Lillywhite. John D. Lynch. Richard Mayden. Roy W. Mc- Diarmid, Lawrence M. Page. William S. Parker, F. Harvey Pough. Rebecca A. Pyles, Steven M. Roble. Albert Schwartz. Richard Shine. Norman A. Slade. Linda Trueb, John Wiens. and Bernard Willard. Without the help of all these individuals this tribute to an outstanding biologist would not have been possible. Richard A. Seigel Lawrence F. Hunt James L. Knight Luis Malaret Nancy L. Zuschlag Lawrence. Kansas October 10. 1981 Contributors Robin M. Andrews. Department of Biology, Vir- ginia Polytechnic Institute and State Univer- sity, Blacksburg, Virginia 24061 William S. Brown, Department of Biology, Skid- more College, Saratoga Springs, New York 12866 Ray D. Burkett, Department of Biology, Shelby State Community College, P.O. Box 40568, Memphis, Tennessee 38104 Charles C. Carpenter, Department of Zoology, University of Oklahoma. Norman, Oklahoma 73019 William E. Duellman, Museum of Natural His- tory, University of Kansas. Lawrence, Kansas 66045 Alice F. Echelle. School of Biological Sciences, Oklahoma State University, Stillwater, Okla- homa 74078 Anthony A. Echelle, School of Biological Sci- ences, Oklahoma State University, Stillwater, Oklahoma 74078 John H. Fitch, Massachusetts Audubon Society, Lincoln, Massachusetts 01773 Virginia R. Fitch, University of Kansas Natural History Reservation, Lawrence, Kansas 66044 Harry W. Greene, Museum of Vertebrate Zo- ology, University of California, Berkeley, Cal- ifornia 94720 Patrick T. Gregory, Department of Biology, University of Victoria, Victoria, British Co- lumbia, V8W 2Y2 Russell J. Hall, U.S. Fish and Wildlife Service, Patuxent Wildlife Research Center, Laurel, Maryland 20811 Robert V\ . Henderson. Section of Vertebrate Zo- ology, Milwaukee Public Museum, Milwau- kee, Wisconsin 53233 Clark Hubbs, Department of Zoology, Univer- sity of Texas, Austin. Texas 78712 Bernard M. Mulhern. U.S. Fish and Wildlife Service, Patuxent Wildlife Research Center. Laurel, Maryland 208 1 1 William S. Parker, Department of Biological Sciences, Mississippi University for Women, Columbus, Mississippi 39701 Dwight R. Piatt, Department of Biology, Bethel College, North Newton, Kansas 671 17 Michael V. Plummer, Department of Biology, Harding University, Searcy, Arkansas 72143 James L. Rakestraw, Museum of Natural His- tory, University of Kansas, Lawrence. Kansas 66045 A. Stanley Rand. Smithsonian Tropical Re- search Institute, APO Miami. Florida 34002 Richard A. Seigel, Savannah River Ecology Lab- oratory, Drawer E, Aiken, South Carolina 29801 Richard Shine, School of Biological Sciences, University of Sydney, N.S.W. 2006 Australia Pennie H. von Achen, RD 2, Eudora, Kansas 66025 Robert G. Webb, Department of Biological Sci- ences, University of Texas, El Paso. Texas 79968 Ernest E. Williams, Museum of Comparative Zoology, Harvard University, Cambridge. Massachusetts 02138 ft i, . --.. i. c XS a t_ 00 O o X a Ih o o 6 o 2 H •3 c >> X) s Xj 00 c I 3 O c E E 00 00 in 00 .J; "5 O O a o X u o u a '-£ c o i. V'v Parti Introduction ■ Fig. 1. Henry S. Fitch in the field. Photograph by David M. Hillis. Vertebrate Ecology and Systematics— A Tribute to Hours S I it< h Edited b> R. A. Seigcl. L. E. Hunt. J I Knight. I Malaret and N. L. Zuschlag i I "t>84 Museum of Natural History, The University ot Kansas. Lawrence Henry S. Fitch in Perspective William E. Duellman Who is Henry Sheldon Fitch? This quiet, mod- est, unassuming man made his first entrance into the world of biologists by publishing on Oregon birds in the Condor in 1933. Yes, birds! Most of us think of Henry S. Fitch as a herpetologist. Yet, of his 1 50 published papers in the past 50 years, only about two-thirds of them deal with am- phibians and reptiles. Twenty others have been on mammals, 12 on birds, and others on spiders, molluscs, and plants. Most of us think of Henry S. Fitch as an ecol- ogist, but 25 of his papers are on systematics and include his classic work on alligator lizards pub- lished in 1934 and his highly perceptive study of western garter snakes published in 1940 (doc- toral dissertation at the University of California, Berkeley). His more recent systematic work has dealt with Middle American anoles— a field where most systematists have feared to tread. Fitch's best known works are on the natural history of reptiles. From his earliest papers on reptiles, he has provided extensive field obser- vations. In 1 948, he entered a "naturalist's heav- en"— the University of Kansas Natural History Reservation. There he began intensive studies on the biota of one square mile of deciduous hardwood forest — studies involving population densities, movements, food, growth rates, hiber- nation, and reproduction — all substantiated with massive quantities of data. Through his efforts this square mile is better known herpetologically than any other in the world. His studies on the natural history of reptiles are classics. Outstanding examples are the thor- ough study of the five-lined skink ( 1954) and the exhaustive study of the copperhead ( 1 960). More recently he has worked on the interactions of behavior and ecology, communities of anoles. and populations and conservation of iguanas. In addition to these systematic and ecological works. Fitch has provided us with important syntheses— reproductive cycles in lizards and snakes (1970) and sexual size differences in rep- tiles (1981). All of his works are characterized by careful and detailed studies on the existence of populations in nature. Vast quantities of such data combined with extensive laboratory and lit- erature research are reflected in his syntheses. These traits combined with dogged determina- tion to learn all there is to know about his sub- jects of study, his continued productivity, and his willingness to share his ideas, knowledge, and enthusiasm with students have assured him of a permanent place in the herpetological hall of fame. At the present time, many biologists com- monly are narrow specialists. Henry Fitch doesn't fit into a modern pigeon hole. He is a naturalist in the broadest sense of the word. His breadth of knowledge is matched by very few of his con- temporaries and scarcely imagined by most of his younger colleagues. An analogy can be drawn with the story of the hare and the tortoise, with Henry Fitch as the tortoise steadily plodding along his path of scientific endeavor, frequently being passed by various biological bandwagons, only to find them sometimes morassed or abandoned further down the road. He has avoided biopolitics. He has not been a vigorous proponent of controversial theories. Instead, he has continued to be a fine naturalist. But. his published works are among those com- monly cited in support of some theories or in the falsification of others. Thus, for half a century Henry S. Fitch has been a major contributor to our knowledge of the natural history of diverse kinds of animals. During this time he has introduced innumerable students to intensive field studies, has thought- fully guided the research of many graduate stu- dents, and has collaborated with a diversity of colleagues. A major factor in his remarkable and successful career has been a collaborator, assis- tant, caretaker and charming lady — Virginia R. Fitch. Few scientists can reflect on such a long and productive career, and yet upon officially retiring maintain such enthusiasm for an active research program. Henry Fitch's careful work on natural history is well worth emulating. Our knowledge of animals in nature would be far greater if there were many more biologists in the world who followed in the footsteps of Henry S. Fitch. Vertebrate Ecolog> and Systematics— A Tnbulc to Henr\ S I itch Edited by R. A. Seigcl. L. E. Hunt. J. I. Knight. L. Malarel and N. L, Zuschlag < 1984 Museum of Natural Hislorj I he 1 Iniversity of Kansas. Lawrence The Published Contributions of Henry S. Fitch Virginia R. Fitch Beginning with his first published paper in 1933. the writings of Henry S. Fitch have en- compassed a wide range of subjects and disci- plines, from reptilian ecology to bird behavior, from the economic relationships of rodents to an intensive study of spiders, and include such areas as taxonomy, life history', behavior, and repro- ductive biology. To date, he has produced 150 papers, all of which appear in the following list. Fitch's published works include as their subjects mammals (19 papers), birds (12), vertebrates in general (5), spiders (3). vegetation and habitats (4), and mollusks ( 1 ), as well as five book reviews, but papers on amphibians and reptiles ( 1 00) pre- dominate. His works are widely cited throughout scientific periodicals, and this list is presented both as a service to biologists and to document the impressive extent of the knowledge and breadth of interest of Henry S. Fitch. 1933. Bird notes from southwestern Oregon. Condor. 35:167-168 (with J. O. Steven- son). 1934. New alligator lizards from the Pacific Coast. Copeia. 1934:6-7. 1934. A shift of specific names in the genus Ger- rhonotus. Copeia, 1934:172-173. 1935. An abnormal pattern in a gopher snake. Copeia, 1935:144-146. 1935. Natural history of the alligator lizards. Trans. Acad. Sci. St. Louis. 29:1-38. 1936. Amphibians and reptiles of the Rogue River Basin, Oregon. Amer. Midi. Nat.. 17:634-652. 1938. Ranaboylii in Oregon. Copeia, 1938:148. 1938. An older name for Triturus similans Twitty. Copeia, 1938:148-149. 1938. A systematic account of the alligator liz- ards ( Gerrhonotus) in the western United States and lower California. Amer. Midi. Nat., 20:381-424. 1939. Desert reptiles in Lassen County. Cali- fornia. Herpetologica. 1:151-152. 1939. Leptodeira in northern California. Her- petologica. 1:152-153. 1940. A biogeographical study of the Ordi- noides artenkreis of garter snakes (genus Thamnophis). Univ. California Publ. Zool.. 44:1-150. 1940. A field study of the growth and behavior of the fence lizard. Univ. California Publ. Zool., 44:151-172. 1940. Some observations on horned owl nests. Condor. 42:73-75. 1941. The feeding habits of California garter snakes. California Fish and Game, 27:2- 32. 1941. Geographic variation in garter snakes of the species Thamnophis sirtalis in the Pa- cific Coast region of North America. Amer. Mid!. Nat.. 26:570-592. 1942. Interrelations of rodents and other wild- life of the range. Univ. California Agr. Exp. Sta. Bull.. 663:96-129 (with E. E. Horn). 1946. Observations on Cooper's hawk nesting and predation. California Fish and Game. 32:144-154 (with B. Glading and V. House). 1946. Feeding habits of the Pacific rattlesnake. Copeia. 1946:64-71 (with H. Twining). 1 946. Behavior and food habits of the red-tailed hawk. Condor. 48:205-237 (with F. Swenson and D. F. Tillotson). 1946. Trapping the California ground squirrel. Jour. Mammal.. 27:220-224 (with E. E. Horn). 1947. The California Ground Squirrel by J. M. Linsdale (Book review). Jour. Mamm.. 28: 191-192. 1947. A field study of a rattlesnake population. California Fish and Game. 33:103-123 (with B. Glading). 1947. Variation in the skinks (Reptilia: Lacer- tilia)oftheSkiltonianusgroup. Univ. Cal- ifornia Publ. Zool.. 48:169-220 (with T. L. Rodgers). 1947. Predation by owls in the Sierran foothills of California. Condor. 49:137-151. 1947. Ecology of a cottontail rabbit (Sylvilagus auduboni) population in central Califor- nia. California Fish and Game. 33:159- 184. 1947. Rattlesnakes on the range. Pacific Stock- man. 13(6):8-9 (with EC A. Wagnon). 1947. Rattlesnakes on western farm lands. Western Dairy Jour.. Sept.:23. 78-79 (with K. A. Wagnon). 1947. Ground squirrels mean destroyed forage. SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Western Livestock Jour.. Oct.:37, 109, 1955. 1 10. 1 12. 1948. Further remarks concerning Thamnophis ordinoidcs and its relatives. Copeia, 1948: 1955. 121-126. 1 948. Habits and economic relationships of the Tulare kangaroo rat. Jour. Mamm., 29:5- 1955. 35. 1948. Ecology of the California ground squirrel on grazing lands. Amer. Midi. Nat., 39: 1956. 513-596. 1948. A study of coyote relationships on cattle range. Jour. Wildlife Management, 12:73- 1956. 78. 1949. Sparrow adopts kingbirds. Auk, 66:368- 369. 1956. 1949. Outline for ecological life history studies of reptiles. Ecology, 30:520-532. 1956. 1949. Use of California annual-plant forage by range rodents. Ecology, 30:306-321 (with J. R. Bentley). 1949. Study of snake populations in central Cal- 1956. ifornia. Amer. Midi. Nat., 41:513-579. 1949. Road counts of snakes in western Loui- siana. Herpetologica, 5:87-90. 1950. A new style live-trap for small mammals. 1956. Jour. Mamm., 31:364-365. 1951. Remarks concerning the systematics of the collared lizard (Crotaphytus collaris), with a description of a new subspecies. Trans. 1956. Kansas Acad. Sci., 54:548-559 (with W. Tanner). 1951. A simplified type of funnel trap for rep- tiles. Herpetologica, 7:77-80. 19 56. 1952. The armadillo in the southeastern United States. Jour. Mamm., 33:21-37 (with P. Goodrum and C. Newman). 1952. The University of Kansas Natural His- 1957. tory Reservation. Univ. Kansas Mus. Nat. Hist. Misc. Publ., no. 4:1-38. 1952. (Book review) Ecological Animal Geog- raphy by Hesse, Allee and Schmidt. Wil- 1957. son Bull., 64. 1953. Ecology of the opossum on a natural area in northeastern Kansas. Univ. Kansas 1958. Publ., Mus. Nat. Hist., 7:309-338 (with L. L. Sandidge). 1953. (Book review) Natural Communities by 1958. L. R. Dice. Wilson Bull.. 65:121-123. 1954. Seasonal acceptance of bait by small mammals. Jour. Mamm., 35:39-47. 1954. Life history and ecology of the five-lined 1959. skink, Eumeces fasciatus. Univ. Kansas Publ., Mus. Nat. Hist., 8:1-156. 1959. Habits and adaptations of the Great Plains skink (Eumeces obsoletus). Ecol. Mono- gr., 25:59-83. Observations on the summer tanager in northeastern Kansas. Wilson Bull., 67:45- 54 (with V. R. Fitch). The coyote on a natural area in north- eastern Kansas. Trans. Kansas Acad. Sci., 58:211-221 (with R. L. Packard). A field study of the Kansas ant-eating frog, Gastrophryne olivacea. Univ. Kansas Publ., Mus. Nat. Hist., 8:275-306. An ecological study of the collared lizard (Crotaphytus collaris). Univ. Kansas Publ., Mus. Nat. Hist., 8:213-274. A ten-year old skink? Herpetologica, 12: 328. Early sexual maturity and longevity under natural conditions in the Great Plains nar- row-mouthed frog. Herpetologica, 12: 281-282. Temperature responses in free living am- phibians and reptiles of northeastern Kansas. Univ. Kansas Publ., Mus. Nat. Hist., 8:417-476. The forest habitat of the University of Kansas Natural History Reservation. Univ. Kansas Publ., Mus. Nat. Hist., 10: 77-127 (with R. L. McGregor). The molluscan record of succession on the University of Kansas Natural History Re- servation. Trans. Kansas Acad. Sci., 59: 442-454 (with D. H. Lokke). Ecological observations on the woodrat, Neotoma jloridana. Univ. Kansas Publ., Mus. Nat. Hist., 8:499-533 (with D. G. Rainey). Aspects of reproduction and development in the prairie vole (Microtus ochrogaster). Univ. Kansas Publ., Mus. Nat. Hist., 10: 129-161. Observations on hibernation and nests of the collared lizard. Crotaphytus collaris. Copeia, no. 4:305-307 (with J. M. Legler). Natural history of the six-lined racerun- ner (Cnemidophorus sexlineatus). Univ. Kansas Publ., Mus. Nat. Hist., 1 1:1 1-62. Home ranges, territories, and seasonal movements of vertebrates of the Natural History Reservation. Univ. Kansas Publ., Mus. Nat. Hist., 11:63-326. A patternless phase of the copperhead. Herpetologica, 15:21-24. Aspects of needed research on North VERTEBRATE ECOLOGY AND SYSTEMATICA American grasslands. Trans. Kansas Acad. 1 967. Sci.. 62:175-183 (with 5 other authors). 1 960. Criteria for determining sex and breeding maturity in snakes. Herpetologica, 16:49— 51. 1968. 1960. Autecology of the copperhead. Univ. Kansas Publ., Mus. Nat. Hist., 1 3:85-288. 1 969. 1960. (Book review). The Rusty Lizard, a Pop- ulation Study by W. Frank Blair. Copeia, 1960:386-387. 1961. Occurrence of the garter snake Thamno- phis sirtalis in the Great Plains and Rocky Mountains. Univ. Kansas Publ.. Mus. Nat. Hist., 13:289-308 (with T. P. Maslin). 1970. 196 1 . An older name for Thamnophis cyrtopsis (Kennicott). Copeia, 1961:112 (with W. W. Milstead). 1961. The snake as a source of living sperma- tozoa in the laboratory. Turtox News, 39: 247. 1961. Longevity and age-size groups in some 1970. common snakes. Pp. 396-414 in Verte- brate Speciation: A University of Texas symposium. Univ. Texas Press. 1970. 1963. Natural history of the racer Coluber con- strictor. Univ. Kansas Publ.. Mus. Nat. Hist.. 15:351-468. 1963. Observations on the Mississippi kite in southwestern Kansas. Univ. Kansas Publ.. 1 970. Mus. Nat. Hist.. 12:503-519. 1963. Natural history of the black rat snake (Elaphe o. obsoleta) in Kansas. Copeia. 1970. 1963:649-658. 1963. Spiders of the University of Kansas Nat- ural History' Reservation and Rockefeller Experimental Tract. Univ. Kansas Mus. 1971. Nat. Hist., Misc. Publ., no. 38:1-202. 1965. The University of Kansas Natural His- tory Reservation in 1965. Univ. Kansas Mus. Nat. Hist., Misc. Publ.. no. 42:1- 1971. 60. 1965. An ecological study of the garter snake, Thamnophis sirtalis. Univ. Kansas Publ.. 1971. Mus. Nat. Hist., 15:493-564. 1965. Breeding cycle in the ground skink, Ly- gosoma laterale. Univ. Kansas Publ., Mus. 1971. Nat. Hist., 15:565-575 (with H. W. Greene). 1966. Spiders from Meade County, Kansas. Trans. Kansas Acad. Sci.. 69:1 1-22 (with 1971. V. R. Fitch). 1967. Preliminary experiments on physical tol- erances of the eggs of lizards and snakes. 1971. Ecology, 48:160-165 (with A. V. Fitch). Ecological studies of lizards on the Uni- versity of Kansas Natural History Reser- vation. Pp. 30-44 in Lizard Ecology, a symposium. Univ. Missouri Press. Temperature and behavior of some equa- torial lizards. Herpetologica 24:35-38. Biotelemetric studies of small vertebrate behavior. Pp. 44-45 in Engineering Re- search, Center for Research, Inc.. Engi- neering Sci. Div., Univ. Kansas. Vol. Ill (1967-1968), E. D. Bevan, Ed. (with H. W. Shirer. W. K. Legler and D. D. Pip- pin.) Data acquisition systems for the study of vertebrate ecology. Pp. 19-20 in Re- search, Vol. IV, The Univ. Kansas, Cen- ter for Research, Inc.. P. McMillan. Ed. (with H. W. Shirer, K. Armitage. W. K. Legler, D. D. Pippitt. J. D. Pauley and J. F. Downhower.) Reproductive cycles in lizards and snakes. Univ. Kansas Mus. Nat. Hist., Misc. Publ., 52:1-247. Comparison from radiotracking of move- ments and denning habits of the raccoon, striped skunk, and opossum in north- eastern Kansas. Jour. Mamm., 5 1(3):49 1- 503 (with H. W. Shirer). A radiotelemetric study of spatial rela- tionships in the opossum. Amer. Midi. Nat. 84(1): 170-1 86 (with H. W. Shirer). Natural history of the milk snake (Lani- propeltis triangulum) in northeastern Kansas. Herpetologica 26(4):387-396 (with R. R. Fleet). Ecological notes on some common lizards of southern Mexico and Central America. The Southwestern Naturalist, 15:398-399 (with A. V. Fitch and C. W. Fitch). A radiotelemetric study of spatial rela- tionships in some common snakes. Co- peia. 1971:118-128 (with H. W. Shirer). (Book review). A Complete Field Guide to Nests in the United States by R. Head- strom. Jour. Wildlife Mgt., 35:188-189. A comparative analysis of aggressive dis- play in nine species of Costa Rican Anolis. Herpetologica, 27:271-288 (with A. A. Echelle and A. F. Echelle). A new anole from Costa Rica. Herpeto- logica. 27:354-362 (with A. A. Echelle and A. F. Echelle). Further observations on the demography of the Great Plains skink {Eumeces ob- SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY soletus). Trans. Kansas Acad. Sci.. 74:93- 98 (with R. J. Hall). 1972. Radio tracking of wild animals in their natural habitat. Pp. 35-36 in Research. Univ. Kansas Center for Research, Inc.. 1975. Vol. V, P. Nicholas, Ed. (with R. Hoff- mann, H. W. Shirer. L. A. Gold. R. L. Lattis, C. B. Rideout, and R. C. Waltner). 1972. Ecology of Anolis tropidolepis in Costa Ri- 1975. can cloud forest. Herpetologica, 28( 1 ): 1 0- 21. 1972. Variation in the Central American igua- 1976. nid lizard, Anolis cupreus, with the de- scription of a new subspecies. Occas. Pa- pers. Mus. Nat. Hist., Univ. Kansas, no. 8:1-20 (with A. A. Echelle and A. F. 1976. Echelle). 1972. Observations offish-eating and mainte- nance behavior in two species of Basilis- 1 976. ens. Copeia, 1972:387-389 (with A. A. Echelle and A. F. Echelle). 1973. Observations on the population ecology 1976. of the Central American iguanid lizard, Anolis cupreus. Caribbean Jour. Sci., 13(3- 4):215-229. 1973. Population structure and survivorship in 1976. some Costa Rican lizards. Occas. Papers, Mus. Nat. Hist., Univ. Kansas, no. 18:1- 41. 1973. A field study of Costa Rican lizards. Univ. 1976. Kansas Sci. Bull., 50(2):39-126. 1973. A new anole (Reptiha: Iguanidae) from southern Veracruz, Mexico. Jour. Herp. 1977. 7(2): 125-1 28 (with R. W. Henderson). 1973. Road counts of hawks in Kansas. Kansas Ornith. Soc. Bull., 24(4):33-35 (with H. A. Stephens and R. O. Bare). 1977. 1973. Yellow-billed cuckoo nesting at Univer- sity of Kansas Natural History Reserva- tion. Kansas Ornith. Soc. Bull., 24(2): 1 2- 15 (with P. von Achen). 1977. 1974. Observations on the food and nesting of the broad-winged hawk (Buteo platypte- rus) in northeastern Kansas. Condor. 76(3):331-333. 1974. Food habits of Basiliscus basiliscus in 1977. Costa Rica. Jour. Herp.. 8(3):260-262 (with R. R. Fleet). 1 975. A preliminary ecological study of the soft- shelled turtle, Trionyx muticus, in the 1978. Kansas River. Israel Jour. Zoology, 24: 28-42 (with M. V. Plummer). 1975. A comparative study of the structural and climatic habitats of Anolis sericeus (Rep- 1978. tilia: Iguanidae) and its syntopic conge- ners at four localities in southern Mexico. Herpetologica, 31:459-471 (with R. W. Henderson). A demographic study of the ringneck snake (Diadophis punctatus) in Kansas. Univ. Kansas Mus. Nat. Hist., Misc. Publ.. 62: 1-53. Sympatry and interrelationships in Costa Rican anoles. Occas. Papers, Univ. Kan- sas Mus. Nat. Hist., 40:1-60. A field study of the rock anoles (Reptilia, Lacertilia, Iguanidae) of southern Mexico. Jour. Herp., 10:303-311 (with R. W. Henderson). A new pholcid spider from northeastern Kansas. Bull. Kansas Ent. Soc. (with E. O. Maughan). Sexual size differences in the mainland anoles. Occas. Papers. Univ. Kansas Mus. Nat. Hist., no. 50:1-21. Field observations on rare or little known mainland anoles. Univ. Kansas Sci. Bull.. 51:91-128 (with A. F. Echelle and A. A. Echelle). A new anole (Reptilia: Iguanidae) from Great Corn Island. Caribbean Nicaragua. Contr. Biol. & Geol., Milwaukee Pub. Mus., no. 9:1-8 (with R. W. Henderson). Dragons for dinner. Wildlife Omnnibus, International Wildlife, 6(6): 1 7 (with R. W. Henderson). Age and sex differences in the ctenosaur. {Ctenosaura similis). Contr. Biol. & Geol., Milwaukee Pub. Mus., no. 1 1:1-1 1 (with R. W. Henderson). Age and sex differences, reproduction and conservation of Iguana iguana. Contr. Biol. & Geol., Milwaukee Pub. Mus., 13: 1-21 (with R. W. Henderson). Spatial relations and seasonality in the skinks, Eumeces fasciatus and Scincella laterale in northeastern Kansas. Herpe- tologica, 33:303-313 (with P. von Ach- en). Structure, movements and reproduction in three Costa Rican bat communites. Oc- cas. Papers Mus. Nat. Hist., Univ. Kan- sas, 69:1-28 (with R. K. LaVal). Inter- and intraspecific allometry in a dis- play organ: the dewlap of Anolis (Igua- nidae) species. Copeia, 1978(2):245-250 (with A. F. Echelle and A. A. Echelle). Behavioral evidence for species status of VERTEBRATE ECOLOGY AND SYSTEMATICS Anolis uniformis (Cope). Hcrpetologica 34(2):205-207 (with A. F. Echcllc and A. A. Echelle). 1978. Dragons: 25c/lb. Animal Kingdom, Feb./ March: 12-17 (with R. W. Henderson). 1978. A field study of the red-tailed hawk in eastern Kansas. Trans. Kansas Acad. Sci.. 8 1(1): 1-1 3 (with R. O. Bare). 1978. Sexual size differences in the genus Sce- loporus. Univ. Kansas Sci. Bull., 51(13): 441-461. 1978. Ecology and exploitation of Ctenosaura similis. Univ. Kansas Sci. Bull.. 51(15): 483-500 (with R. W. Henderson). 1 978. Two new anoles (Reptilia: Iguanidae) from Oaxaca, with comments on other Mexi- can species. Contr. Biol. & Geol.. Mil- waukee Pub. Mus.. 20:1-15. 1978. A field study of the prairie kingsnake (Lampropeltis calligaster). Trans. Kansas Acad. Sci.. 81:353-363. 1978. The plight of the iguana. LORE, Milwau- kee Pub. Mus., 28(3):2-9 (with R. W. Henderson). 1978. A 20-year record of succession on reseed- ed fields of tallgrass prairie on the Rocke- feller Experimental Tract. Univ. Kansas Mus. Nat. Hist.. Spec. Publ., 4: 1-1 5 (with E. R. Hall). 1979. Notes on the behavior and ecology of Ctenosaura similis (Reptilia: Iguanidae) at Belize City, Belize. Brenesia, 16:69-80 (with R. W. Henderson). 1980. Remarks concerning certain western gar- ter snakes of the Thamnophis elegans complex. Trans. Kansas Acad. Sci.. 83: 106-113. 1 980. Reproductive strategies of reptiles. Pp. 25- 31 in Reproductive Biology and Diseases of Captive Reptiles (J. B. Murphy and J. T. Collins, eds.), SSAR Cont. Herpetol.. 1:1-277. 1981. Sexual size differences in reptiles. Univ. Kansas Mus. Nat. Hist. Misc. Publ.. 70: 1-72. 1981. Coluber mormon, a species distinct from C. constrictor. Trans. Kansas Acad. Sci. 84:196-203 (with W. S. Brown and W. S. Parker). 1981. Thamnophis sirtalis. Cat. American Amph. Rept.. 270.1-270.4. 1983. 1983. 1982. Reproductive cycles in tropical reptiles. Occas. Papers. Mus. Nat. Hist.. Univ. of Kansas 96:1-53. 1982. Resources of a snake community in prai- rie-woodland habitat of northeastern Kansas. Pp. 83-97 in Herpetological Communities (N. J. Scott Jr., ed.). U.S. Fish and Wildlife Serv.. Wildl. Res. Rep. 13. 1983. Exploitation of iguanas in Central Amer- ica. Pp. 397^ 1 7 in Iguanas of the World: Their Behavior, Ecology, and Evolution (G. M. Burghardt and A. S. Rand, eds.) Noyes Press (with R. W. Henderson and D. M. Hilhs). Thamnophis elegans. Cat. American Amph. Rept., 320.1-320.4. Ctenosaura similis (Garrobo. Iguana Ne- gra. Ctenosaur). Pp. 394-396 in Costa Ri- can Natural History. (D. H. Janzen, ed.) Univ. Chicago Press (with J. Hackforth- Jones). 1983. Sphenomorphus cherriei (Escincela Par- da, Skink). Pp. 422-425 in Costa Rican Natural History. (D. H. Janzen, ed.) Univ. Chicago Press. 1983. Ecological succession in vegetation and small mammal population on a natural area of northeastern Kansas. Proc. Sev- enth North American Prairie Conf., Au- gust 1 980. Pp. 117-121 (with W. D. Ket- tle). In press. The Anolis dewlap: Interspecific vari- ability and morphological associations with habitat. Copeia (with D. M. Hillis). In press. Succession in small mammals on a nat- ural area in northeastern Kansas. Occas. Papers, Mus. Nat. Hist., Univ. Kansas (with V. R. Fitch and W. D. Kettle). In press. Geographic variation in clutch size and litter size in North American reptiles. Univ. Kansas Mus. Nat. Hist. Misc. Publ. In press. Ecological patterns of relative clutch mass in snakes. Oecologia (with R. A. Sei- gel). In press. Thamnophis couchi. Cat. American Amph. Rept. In press. Intergradation of Osage and broad- banded copperheads in Kansas. Trans. Kansas Acad. Sci. (with J. T. Collins). Part II Reproductive Biology and Population Dynamics Vertebrate t'cology and Systemalics— A Tribute to Henry S. Fiteh Edited by R. A. Seigcl. L. £ Hunt. J I Knight. L. Malaret and N. L. Zuschlag c 1984 Museum of Natural History. The University of Kansas. Lawrence Growth, Reproduction and Demography of the Racer, Coluber constrictor mormon^ in Northern Utah William S. Brown and William S. Parker Introduction Considerable interest has developed recently in comparative life history and demographic studies because the data point up a number of evolutionary strategies taken by separate inter- and intraspecific populations. To date, the data have been effective mostly in illustrating the se- lection and adaptive basis for the life histories of lizards, birds, and mammals among the ver- tebrates (Stearns 1976; Hutchinson 1978). Rare- ly have data on snakes perfused the general lit- erature even though a number of sound field studies of snake populations have been com- pleted (Blanchard et al. 1979; Branson and Baker 1974; Brown 1973; Carpenter 1952; Clark 1970, 1974; Clark and Fleet 1976; Feaver 1977; Fitch 1949, 1960, 1963, 1965, 1975; Gregory 1977; Hall 1969; Parker and Brown 1974, 1980; Piatt 1969; Prestt 1971; Spellerberg and Phelps 1977; Stewart 1968; Tinkle 1957, 1960; Viitanen 1967). Coluber constrictor (Serpentes, Colubridae) is known to occur from Guatemala to southern Canada (Conant 1975; Etheridge 1952; Stebbins 1966). The species is polytypic, with 10 de- scribed subspecies (Wilson 1970). In the United States, eight of the nine recognized geographic races occur east of the Rocky Mountains (Auf- fenberg 1955; Fitch 1963; Wilson 1970). C. c. mormon occurs west of the Continental Divide. This subspecies has been recorded in most of the states in the western third of the U.S. (Auffenberg 1955; Stebbins 1966; Wilson 1978:218.1). An extensive range hiatus in the Rocky Mountains, lack of intergradation, and differences in mor- phology and ecology between the midwestern subspecies, C. c. flaviventris, and C. c. mormon may warrant elevation of the latter taxon to species rank (Fitch, Brown and Parker 1981). Aside from several brief reports on various as- pects of the biology of C. c. mormon (see review of the literature in Fitch 1963), no comprehen- sive ecological study of this wide-ranging western form has been conducted. Fitch's (1963) study in Kansas of C. c. flaviventris is, to date, the most extensive ecological investigation of any popu- lation of Coluber constrictor. The present study focuses on the biology of the Western yellow-bellied racer. Coluber con- strictor mormon Baird and Girard, hereinafter called simply "racer." Our approach has been empirical and autecological and has concentrat- ed on one large population of this snake at a single locality in northern Utah over a four-year period. This paper treats growth, maturity, re- production, population structure, and demog- raphy of the racer. Widespread and abundant in North America, C. constrictor lends itself well to a study of its adaptive biology in several parts of its geographic range. Our attempt is to provide ecological comparisons of populations in Utah and Kansas. This study reveals different life his- tory strategies at the intraspecific level. Methods Snakes were captured in autumn 1 969 through spring 1973 at a communal denning area in a desert shrub habitat located 4 km W of Grants- ville, Tooele County, Utah (40°36'N, 1 12°32'W, elevation 1580 m), ca. 58 km WSW of Salt Lake City. This area is our primary study locality (area M) where all long-term mark and recapture field work was conducted. We recorded a total of 1 694 captures of 1046 racers at this site. Originally studied by Woodbury and his co- workers in the 1940's (Woodbury et al., 1951), the "main den" (den M) was later sampled in the mid-1960's by Hirth and King (1968) and again in the early 1970's by us. We discovered other actively used dens near den M; these were considered part of a discrete group which we called "M complex." A separate series of newly- discovered dens located 0.8 km to the south was designated "S complex" (Parker and Brown 1 973; Brown and Parker 1976a). The technique we used to capture snakes was to intercept them with a screen wire fence erected aound their hibernaculum. As the dens were sin- gle small rock piles located in fairly level terrain with sandy soils, it was possible to encircle each den completely. We sank steel reinforcing bars around a den, attached screening (ca. 95 cm high) to the stakes, and buried the base of the fence by 13 14 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY covering it with soil from a perimeter trench. Captures occurred almost daily in favorable weather as snakes attempted to enter a den in autumn and leave it in spring. The chronology of sampling Coluber and other snakes at area M is summarized in Parker and Brown ( 1 980). Our results pertain to the four- year period 1 969- 1 972 by sampling dens each autumn and spring from autumn 1969 through spring 1973. Data pre- sented for a given calendar year were derived from sampling in the autumn of that year (den M only) and the spring of the next (den M and other dens). Individuals were processed in the laboratory and most were released within 24 h after capture. Each snake was permanently marked by clipping ventral scutes (Brown and Parker 1 976b). Snout- vent length (SVL = distance from tip of snout to posterior edge of anal scute) and tail length to the nearest 0.5 cm (snakes > 1 year old) or to the nearest mm (hatchlings and juveniles) and live body weight to the nearest 0. 1 g (all snakes) were recorded for each individual at all captures. Re- productive condition of males was determined by obtaining cloacal smears and examining them microscopically for the presence of spermatozoa. Snakes in spring were released outside of their den fence, those in autumn were released inside. Snakes caught by hand on their summer range were released at the capture site. Other racers were collected from two nearby localities in northern Utah. Most snakes from these areas were sacrificed for food and repro- ductive data. These localities are designated as area SLC, vicinity of Salt Lake City, Salt Lake County, Utah; and area RB, Red Butte Canyon, 5 km E of Salt Lake City, Salt Lake County, Utah. Both areas SLC and RB provided data on clutch size. Female racers were marked and released in area RB and provided data on body weight changes. Some other females killed for exami- nation of reproductive tracts were from area M. These snakes included several casualties from our marked population and a few others taken > 2 km from the study dens and beyond the max- imum dispersal limits of racers from area M. In most years at area M hatching occurred around mid-August. Juveniles normally arrived at den M in early October at an average age of ca. 1.5 months. Winter dormancy lasted ca. 7 months (Oct. -Apr.) and the activity season ca. 5 months (May-Sept.) (Brown 1973; Parker and Brown 1980). As our sampling was in autumn and spring, snakes placed in a designated year class differed from their actual age by about 7 months. No growth occurred during hibernation so we assigned an equivalent age to autumn and spring-captured snakes as follows: hatchling (age 0),juvenile(1.5and8.5 months), 1 -year-old (13.5 and 20.5 months), 2-year-old (25.5 and 32.5 months), and so on. The simpler age designation in years corresponds to the number of full 5- month growing seasons which a snake had been through and facilitates analysis of age-specific aspects of the life history. Racers > 1 year old were sexed visually by the relatively thick (males) or thin (females) tail base. Juveniles lacked external sexual differences and those in 1972-1973 were sexed using a blunt probe to detect presence (males) or absence (fe- males) of hemipenial sacs. In earlier sample pe- riods juveniles were not sexed and numbers of male and female juveniles were apportioned as- suming a 1:1 sex ratio. Some of these juveniles were sexed later by recapturing them as marked 1 -year-olds after they had attained sufficient dis- cernible sexual dimorphism as yearlings. Assignment of males and females to specific age classes was based on size and growth of marked individuals. Sample means and 95% confidence intervals of length and weight were calculated for recaptured 1 -year-olds marked ini- tially as juveniles. Snakes in all sampling periods that compared closely to these values were as- signed as 1 -year-olds. Records for these initial juveniles and 1 -year-olds that were later recap- tured were then used to determine preliminary length and weight characteristics for 2- and 3-year- olds. Some individuals were thus followed from age 3 in 1969-1970 to age 6 in 1972-1973. By working in this step-wise procedure, many in- dividuals were aged through 6 years and a few through 7 years. Lacking prior captures made some error possible in assigning ages of 4 and 5 years to snakes early in the study, but our method of comparing sizes to known-age statistical val- ues was consistent and uniformly applied over all ages. We tended to be conservative in cases involving a size intermediate between two ages, e.g., if the snake was between the two- and three- year-old size, we designated it as a 2-year-old. Snakes too large for age determination, whether recaptured or not, were pooled as older adults (>6 years old). Yearly individual length and weight changes are based on successive spring or successive au- VERTEBRATE ECOLOGY AND SYSTEMATIC S 15 tumn captures. Annual growth increments thus include one intervening period of hibernation. Weight losses during winter dormancy did not differ significantly from year to year so both the spring-to-spring and autumn-to-autumn inter- vals used for determining annual growth rates are considered equivalent. Proportional annual increases or decreases in SVL or weight were calculated as the amount increased or decreased during the year divided by the initial size at the beginning of the year. For example, if a 1 -year- old male increased from 31.9 to 48.9 g (an ab- solute increase of 17.0 g/yr), the proportional increase would be 17.0 -5- 31.9 = 0.533/yr, or 53%. Survival rates were measured over two major periods in the annual cycle of Coluber at area M: (1) the winter period of hibernation and (2) the full year. Like growth rate calculations, annual survival rates include one intervening winter pe- riod and were calculated from spring-to-spring or autumn-to-autumn capture records. Population size estimates based on capture- recapture were calculated using the Jolly-Seber stochastic method following Caughley (1977) and Krebs (1978). Eight censuses at den M provided data for the Jolly-Seber analysis over three years (1970-1972). Snakes recaptured following their movement to a different den of M complex were included in the tabulations as were den M in- dividuals that were experimentally displaced from that den in autumn 1971 (cf. Brown and Parker 1976a). Thus, bias due to these factors was elim- inated. Population sizes were calculated sepa- rately for juveniles (both sexes combined) and for yearling and older (> 1 year inclusive) males and females. Statistical methods in this paper follow Sokal and Rohlf ( 1 969) and Woolf ( 1 968). Mean values are followed by ± one standard error of the mean (SE) with the extremes in parentheses. Results Sexual Dimorphism. — Weights of 73 male and 72 female snakes > 1 year old collected during the autumns of 1969-1972 were regressed on snout-vent lengths (Fig. 1). There was a highly significant (<0.01) greater in males (26.76 ± 0.11%. range 23.9- 28.9%, N = 73) than in females (25.07 ± 0.09%. range 23.1-27.0%. N = 72). Although statisti- cally significant, this distinction could not be used in visual sex determination. Size of Snakes of Known Age.— Snout-vent lengths and weights of 1236 Coluber of known 16 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 1. Sizes of Coluber constrictor mormon of known age, 1969-1972. Ages: H = hatchling (not sexed), J = juvenile, 1-7 = years (see text for method of designating age). Except for hatchlings, all measurements were recorded in spring at emergence from hibernation. Mean ± 1 SE, sample size and extremes in parentheses. Snout-vent length (cm) Weight (g) Age 83 99 <$<5 99 H 22.57 ± (19.7- 0.20(26) -23.8) 5.98 ± (4.1- 0.17(26) -7.7) J* 26.64 ± 0.48(11) (23.1-28.7) 27.17 ± 0.40(16) (24.3-29.7) 8.30 ± 0.33(11) (6.4-9.8) 8.78 ± 0.39(16) (6.4-12.2) 1 42.07 ± 0.19(141) (36.5-47.0) 43.01 ± 0.24(130) (35.5-48.0) 26.98 ± 0.34(141) (16.9-36.3) 28.36 ± 0.44(130) (16.1-39.9) 2 48.61 ± 0.13(149) (44.0-53.5) 52.41 ± 0.29(110) (43.5-59.0) 40.95 ± 0.35(149) (26.6-52.9) 51.59 ± 0.73(110) (35.0-69.3) 3 52.03 ± 0.13(116) (48.0-57.0) 57.49 ± 0.18(137) (47.5-63.5) 49.14 ± 0.44(116) (36.2-60.5) 66.25 ± 0.59(137) (43.4-79.2) 4 54.08 ± 0.16(77) (51.0-57.5) 59.95 ± 0.25(90) (53.0-65.0) 53.41 ± 0.54(77) (39.6-64.9) 71.38 ± 0.95(90) (49.8-96.2) 5 56.44 ± 0.15(67) (54.0-60.0) 62.03 ± 0.25(47) (55.5-66.0) 60.40 ± 0.63(67) (47.7-75.4) 79.47 ± 1.13(47) (65.1-93.1) 6 57.33 ± 0.30(21) (55.5-60.0) 63.25 ± 0.62(16) (56.5-66.5) 61.18 ± 1.40(21) (50.4-75.0) 83.95 ± 2.47(16) (68.9-103.8) 7 58.35 ± 0.37(7) (57.0-60.0) — 64.31 ± 2.22(7) (58.4-75.0) — * Additional 82 not sexed gives a total sample of N = 109 juveniles (sexes pooled: cf. Figs. 2 and 3). age up to 6 years in females and 7 years in males are shown in Table 1. A sample of 26 hatchling racers (age 0-1 day) was lab-reared in August 1972. Male and female juveniles that were sexed did not differ significantly in mean SVL (t = 0.8, 0.50 > P > 0.30) or weight (t = 0.9, 0.50 > P > 0.30). Sexes were pooled to obtain a larger sam- ple by including 82 juveniles not sexed in the first three years of the study. By the spring fol- lowing hatching (age ca. 8.5 months) the mean weight of 109 juveniles was 8.52 ± 0.14 (5.1- 1 2.4) g. At an age of 1 year there was a significant difference between males and females in SVL (/ = 3.0, P < 0.01) and weight (/ = 2.5, 0.02 > P> 0.01) (Table 1). Thereafter, a more rapid growth rate in females maintained a significant sexual disparity in size at all ages (cf. nonover- lapping 95% confidence limits in growth curves. Figs. 2 and 3). Asymptotic levels for length and weight were approached by the sixth year in both sexes. Female Age at Maturity.— The mean snout- vent length of 1 8 gravid females of unknown age from area M was 63.3 ± 1.25 (54.5-74.5) cm. Mean SVL ± 1 SD of 2-year-olds was 49.4-55.4 cm (Fig. 2), so most were probably immature. An estimate of the minimum size of sexual ma- turity (57 cm) was calculated by averaging SVL's of the smallest five gravid females and of the largest five 2-year-olds. In all years, the following proportions of known-age females >56.5 cm in spring were considered mature: 2-year-olds, 9 of 1 10 (8%); 3-year-olds, 105 of 137 (77%); 4-year- olds, 8 1 of 90 (90%); 5-year-olds, 46 of 47 (98%); 6-year-olds, 16 of 16 (100%). Male Age at Maturity. — Seminal fluid samples showed the presence of sperm in all of 21 males in autumn after the first full activity season at an age of ca. 1 3.5 months. Three yearlings having sperm in early October had been marked as ju- veniles a year earlier. In spring, all of 76 1 -year- olds and all of 98 2-year-olds tested had sperm. Positive samples in yearlings were obtained as early as 22 March at emergence and 22 Septem- ber at ingress. The smallest sexually mature male was 39 cm SVL and 17.8 g. None of the 195 one and 2-year-old males lacked sperm, and 250 of 26 1 older snakes (96%) had positive cloacal sam- ples. Weight Change during Hibernation. — Body weights of individual Coluber in spring following hibernation at den M were usually less than weights the previous autumn, indicating that most snakes had lost weight through the winter. A total of 333 autumn to spring records was obtained for snakes >1 year old. Of these, 315 (94.6%) VERTEBRATE ECOLOGY AND SYSTEMATICS 17 AGE ( yr ) Fig. 2. Growth in snout-vent length of Coluber constrictor mormon, 1969-1972. Data for hatchlings (H) and juveniles (J) include both 66 and 99 (sexes combined). Except for hatchlings. all records pertain to spring only. Horizontal lines = sample mean; solid rectangles = 95% confidence limits for population mean: open rectangles = ± 1 standard deviation (SD); vertical line = range. Means of 66 connected by dashed line, 99 by solid line. Sample sizes indicated above each bar diagram. involved a decrease in weight during the interval (Table 2). Proportions of weight loss records for 1 78 males (95.5%) and 1 55 females (93.5%) were similar, as were proportions for both sexes over four winters (1969-1970, 86.8%; 1970-1971, 92.5%; 1971-1972, 96.3%; 1972-1973, 97.2%). Females, averaging larger in size than males, lost significantly (/ = 3.5, P < 0.0 1 ) more weight than did males (Table 2). Eleven of 13 juveniles lost an average of 0.67 ±0.13 (0.3-1.6) g/snake. On a relative basis, juveniles lost 7.7% of their au- tumn body weights, not significantly more (F = 1.44, P > 0.05) than males (7.4%) and females (7.3%). Analysis of variance also showed that there were no significant between-years differ- ences in weight loss in males (absolute F = 1.15, P > 0.05; relative F= 0.30, P > 0.05) and in females (absolute F= 0.66, P > 0.05; relative F= 2.49, P> 0.05). Annual Age- specific Growth. — Absolute and relative yearly rates of increase in snout-vent length and weight are summarized in Tables 3- 6. As no recapture records were available to mea- sure growth in the season of hatching directly, growth calculations were based on differences be- tween means of hatchling (mid-August) and ju- venile (October) sizes. Young racers increased 16.0% in SVL and 36.3% in weight during this 1.5-month interval. Weight increase during the first year was rapid. Males increased an average of 225% and females 223% of initial juvenile weights (3.2-fold in- creases). One-year-old females nearly doubled their weight again in their second year (mean proportional increase 82%), achieving a growth rate 1.2 times greater than 1 -year-old males. By the time females reached an age of 3 years and most became sexually mature, they were 1.3 times heavier than an average 3-year-old male and 1 1 times heavier than the average hatchling. The 18 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY 100 80 a< 60 H I O UJ $ 40 20 H -i 1 r- 2 3 4 AGE ( yr ) T" 5 -r- 6 Fig. 3. Growth in weight of Coluber constrictor mormon, 1969-1972. Explanation and symbols as in Fig. 2. first full growing season was clearly the period of greatest rate of increase; thereafter growth rates declined steadily with age. Both absolute and relative growth rates in males were less than cor- responding rates in females at all ages. Unequal growth rates between years are in- dicated (Tables 3-6). We compared all age-spe- cific absolute rates of growth recorded in 1970 and 1971 against those in 1972. Five age inter- vals from 1-2 yr to 5-6 yr for each sex were tested. Significant (P < 0.05) between-years dif- ferences appeared in seven of 10 Mests of SVL increases and in eight of 10 Mests of weight in- creases. In particular. Coluber that were <5 years old grew significantly faster in 1970 and 1971 than in 1972. Annual Variation in Weight Changes. — To in- clude records of individuals of unknown age not analyzed above, proportions of all large snakes (males mostly >6 years old, females >4 years old) that increased in weight each year were com- pared to each year's rainfall (Fig. 4). A greater proportion of snakes gained weight during 1971 (85% of 142 records) than in either 1970 (70% of 46 records) or 1972 (44% of 162 records). Proportions of total annual rainfall in the 5-month activity period (May-Sept.) each year were 40% (1970). 39% (1971), and 19% (1972). Amounts of weight gained and lost are shown in Table 7. Individual weight gains were signif- icantly greater in 1970-1971 than in 1972 in males (t = 4.75, P < 0.00 1 ) and females (/ = 3.40, P < 0.0 1 ). Individual weight losses were not sig- nificantly greater in 1972 than in 1970-1971 in males (t = 0.30, 0.80 > P > 0.70) and females (t = 1.84, 0.10 > P > 0.05). Female Reproductive Cycle. — Females con- tained enlarged preovulatory oocytes in late April. May, and early June (Brown, unpubl. data). At other times of the year ovaries were small and contained no enlarging oocytes. Available evi- dence indicates production of a single clutch of eggs/2 per year in northern Utah. Clutch Size. — Clutch size was determined from VERTEBRATE ECOLOGY AND SYSTEMATICA 19 Table 2. Winter weight losses in Coluber constrictor mormon > 1 year old. Absolute loss is difference in body weight between last autumn capture and first spring capture at den M; relative loss is percentage of autumn weight. Mean ± 1 SE. sample size and extremes in parentheses. Year Absolute weight loss (g snake) Relali\e weight loss l".,i " 1969-1970 1970-1971 1971-1972 1972-1973 All snakes 4.04 ± 0.56(15) (0.7-8.6) 3.81 ± 0.34(39) (0.2-9.5) 4.72 ± 0.34(56) (0.3-12.2) 4.10 ± 0.36(60) (0.1-15.1) 4.23 ± 0.20(170) (0.1-15.1) 5.11 ± 0.75(18) (1.6-12.5) 6.09 ± 0.72(35) (0.1-20.7) 5.25 ± 0.33 (49) (0.9-10.4) 5.11 ± 0.57(43) (0.3-15.6) 5.39 ± 0.28(145) (0.1-20.7) 6.59 ± 0.80(15) (0.9-11.4) 7.25 ± 0.68(39) (0.3-20.1) 7.54 ± 0.42(56) (1.0-21.3) 7.56 ± 0.55(60) (0.2-21.9) 7.40 ± 0.29(170) (0.2-21.9) 9.17 ± 0.74(18) (3.2-14.3) 7.16 ± 0.63(35) (0.2-16.9) 6.61 ± 0.38(49) (1.7-14.2) 7.53 ± 0.61 (43) (0.5-18.6) 7.33 ± 0.29(145) (0.2-18.6) a sample of 43 reproductive females (Fig. 5). At area M the mean number of eggs/2 was 5.78 ± 0.24 (4-8), mode 5 (N = 18). For these females, a significant (r = 0.53. 0.05 > P > 0.01) linear correlation existed between body size and clutch size; SVL explained 28% of the variation in clutch size. Size of .Eggs. — Measurements of 54 eggs in nine clutches were recorded after oviposition in the laboratory. Eggs averaged 37.78 ± 0.75 (29.2- 54.3) mm in length. 18.00 ±0.14 (15.9-20.0) mm in width, and 7.80 ±0.17 (5.9-10.8) g. Data on egg size were not recorded in ten additional lab-deposited clutches. Eggs in three of these were weighed indirectly by dividing the female's ovi- positional weight loss by her clutch size. Mean egg weights calculated in this manner were 7.6. 7.8, and 9.8 g (overall, 8.4 g/egg, N = 18). Analysis of variance (model I) demonstrated a significant difference between clutch means of all three measurements of egg size (length F = 20.6, width F = 16.8, weight F = 38.2; P < 0.0 1 ). A model II analysis of the components of vari- ance showed that relatively more of the total variation occurred among clutches (73-86%) than among eggs within clutches ( 1 4-27%). The small- est clutch (mean weight 6.2 g/egg) differed from the largest (mean weight 10.6 g/egg) by a mean difference of 4.4 g/egg. There was no significant correlation between female size (SVL) and the mean weight of eggs in her clutch (r = 0.19. P > 0.05; N = 9). Incubation and Hatching. — Between 27 June and 1 July 1971 seven gravid females were col- lected in area RB. These females oviposited be- tween 8-25 July after 9-28 days in an environ- mental chamber maintained at 29°C. Three gravid females from area M collected between 27 June- 3 July oviposited in the laboratory between 12- 15 July. Hatching in the 1971 clutches occurred between 19-27 August, after a mean incubation period of 42.6 (41-44) days at 29°C. Nine area M females had enlarged ovarian oocytes between 3-7 June 1972. Four collected between 18-26 June oviposited in the laboratory between 26 June and 9 July. Eggs in three 1972 clutches hatched between 8-23 August after 44-45 days of incubation at 29°C. We followed three gravid females with im- planted radio transmitters at area M in 1972 (Brown and Parker 1976a). Two of these females oviposited on 21 and 23 June. Eggs of one clutch were excavated 36 days later and were lab-in- cubated at 29°C an additional 12-13 days; hatch- ing occurred on 11-12 August after 48 and 49 days incubation. At the second field site a hatch- ling was captured by fencing on 10 August. 50 days after oviposition. The third site was exca- vated on 6 August, 4 1 days after oviposition. and one freshly-hatched egg was recovered. In 1971. timing of reproduction between areas RB and M (located 65 km apart) was similar. If most females at area M had oviposited between 5-15 July 1971 and between 20-30 June 1972, with a probable natural incubation period of 45- 50 days, most hatching in the field around the communal dens occurred between 20-30 August 1971 and between 10-20 August 1972. Hatching Success. -In 1971 and 1972. 20 fe- males oviposited in the laboratory. A total of 20 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 3. Age-specific growth in snout-vent length of 199 66 Coluber constrictor mormon during 3 years. Mean ± 1 SE, sample size and extremes in parentheses. Aee Absolute increase (cm/year) interval 1970 + 1971 1972 1970-1972 J-l 15.85 ± 0.95(2) 13.14 ± 1.33(5) 13.91 ± 1.06(7) (14.9-16.8) (10.8-18.1) (10.8-18.1) 1-2 7.27 ± 0.35(37) 5.31 ± 0.37(16) 6.68 ± 0.29(53) (1.5-11.0) (2.5-8.0) (1.5-11.0) 2-3 4.33 ± 0.26 (43) 2.03 ± 0.19(16) 3.70 ± 0.24(59) (1.5-7.5) (0.5-3.0) (0.5-7.5) 3-4 2.69 ± 0.25(16) 1.55 ± 0.15(19) 2.07 ± 0.24(35) (0.0-6.5) (0.5-2.5) (0.0-6.5) 4-5 2.31 ± 0.50(8) 1.31 ± 0.18(21) 1.58 ± 0.20(29) (0.0-4.5) (0.0-3.0) (0.0-4.5) 5-6 1.50 ± 0.48(6) 1.05 ± 0.20(10) 1.22 ± 0.26(16) (0.0-3.5) (0.0-3.0) (0.0-3.5) 121 eggs were produced of which 105 (86.8%) were normal. Eggs that were improperly shelled and much smaller than usual were considered abnormal. Eight females that produced abnor- mal eggs had been maintained in captivity at 29°C for a mean of 17.5 (10-28) days prior to oviposition. Females that produced clutches of totally normal eggs were maintained for 1 1 .0 (3- 1 9) days prior to egg laying. Thus, the production of abnormal eggs may have been a function of the constant temperature and lengthy periods of captivity. The frequency of abnormal eggs pro- duced by these females is probably not repre- sentative of most gravid females in nature. Eggs laid by females from area M in 1972 were incubated at 29°C on soil from the locality where the animals were collected. Eggs were left undis- turbed in the same container in which the gravid female had been maintained prior to oviposition. Of 2 1 normal eggs produced by four females with completely normal clutches, 19 (90.5%) hatched. On 30 August 1971, we excavated a burrow sys- tem where a whipsnake {Masticophis taeniatus) had oviposited (Parker and Brown 1972). Fifteen Coluber eggs of a previous year were unearthed (one group of 10 eggs and one group of five). Of the 15 eggs, 14 (93.4%) had hatched successfully as indicated by egg slits made by the hatchlings. The combined sample (33 of 36 eggs hatched) indicates a hatching success of 92%. Weight Changes in Reproductive Females.— By recapturing females whose reproductive state Table 4. Age-specific growth in snout-vent length of 204 99 Coluber constrictor mormon during 3 years. Mean ± 1 SE, sample size and extremes in parentheses. Absolute increase (cm/year) interval 1970 + 1971 1972 1970-1972 j-i 16.56 ± 0.84(3) 14.90 ± 0.74(7) 15.40 ± 0.61 (10) (15.4-18.2) (13.0-17.7) (13.0-18.2) 1-2 10.27 ± 0.33(39) 9.10 ± 0.86(10) 10.30 ± 0.32(49) (6.5-15.5) (4.5-13.5) (4.5-15.5) 2-3 6.88 ± 0.35(38) 4.45 ± 0.47(19) 6.07 ± 0.32(57) (3.5-12.0) (1.0-10.0) (1.0-12.0) 3-4 3.62 ± 0.56(21) 1.92 ± 0.22(26) 2.68 ± 0.30(47) (0.5-10.5) (0.5-4.0) (0.5-10.5) 4-5 2.08 ± 0.37(13) 2.77 ± 0.43(13) 2.42 ± 0.29(26) (0.0-4.0) (1.0-5.5) (0.5-5.5) 5-6 2.17 ± 0.29(9) 0.75 ± 0.34(6) 1.60 ± 0.28(15) (1.0-3.5) (0.0-2.0) (0.0-3.5) VERTEBRATE ECOLOGY AND SYSTEMATICA 21 Table 3. Continued. Proportional increase scar 1970 + 1971 1972 1970-lv": .589 ± (.560 .178 ± (.032 .090 ± (.029 .053 ± (.000 .042 ± (.000 .027 ± (.000 029(2) -.618) 009 (37) -.293) 006 (43) -.160) 009(16) -.127) 009 (8) -.085) 009 (6) -.064) .502 ± (.404 .124 ± (.054 .041 ± (.010 .030 ± (.009 .024 ± (.000 .019 ± (.000 059(5) -.727) 010(16) -.187) 004(16) -.064) .003(19) -.057) .003(21) -.057) 005 (10) -.053) .527 ± .045(7) (.404-. 72 7) .162 ± .008(53) (.032-.293) .077 ± .005 (59) (.010-. 160) .040 ± .004(29) (.000-.085) .029 ± .004 (29) (.000-.085) .022 ± .005(16) (.000-.064) was determined, changes in weight prior to ovi- position were recorded. Ten females captured in spring at emergence and released between 26 April-20 May were later recaptured between 4- 26 June (Table 8). These animals were gravid as determined by palpation of enlarged ovarian oo- cytes or by subsequent oviposition in the field or laboratory. Recaptures occurred 1 7-47 (mean 3 1 ) days after release during which time these fe- males had increased by an average of 32.6% of their initial body weights. Absolute increases av- eraged 30.1 (14.6-48.8) g/snake during the pre- reproductive interval; the mean rate of weight gain was 0.97 ± 0.08 g/day. After ovipositing in the laboratory between 1- 14 July, parturient females were released in the field in July 1971 (3 22) and 1972 (2 22) (Table 9). One female oviposited in the laboratory both years. The four females were later recaptured in August and early September, 31-53 days after release. These spent females recovered an av- erage of 53% (24.7-71.1%) of their parturient weights (Table 9). Mean postreproductive weight recovery was 0.92 (0.45-1.38) g/day. For some area M females, additional weight records were obtained in a following year as they were again recaptured emerging from hiberna- tion. One female (No. 4. Table 9), weighed 1 14.0 g on 5 Sept. 1971 after reproduction. 100.5 g in spring 1972. and 109.2 g in spring 1973. Data showing very similar weights in the spring fol- lowing a known reproductive year as in the spring preceding that year are available for three fe- males in Table 8. Table 4. Continued. 1970 + 1971 Proportional increase year 19" .644 ± .041 (3) .550 ± .015(7) (.579-.719) (.448-.728) .247 ± .008(39) .202 ± .022(10) (.141-.356) (.095-.333) .135 ± .007(38) .085 ± .010(19) (.048-.261) (.018-.208) .063 ± .010(21) .033 ± .004 (26) (.017-.193) (.008-.071) .035 ± .007(13) .046 ± .007(13) (.008-.068) (.017-.091) .035 ± .005(9) .012 ± .005 (6) (.016-.056) (.000-.032) 19-D-I9-: .579 ± .032(10) (.448-.728J .238 ± .008 (49) (.095-.356) .119 ± .007(57) (.018-.261) .047 ± .005 (47) (.008-. 193) .041 ± .005(26) (.008-.091) .026 ± .005(15) (.000-.056) 22 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 5. Age-specific growth in weight of 179 6$ Coluber constrictor mormon during 3 years. Mean ± 1 SE. sample si/e and extremes in parentheses. Absolute increase (g/year) interval 1970 + 1971 1972 1970-1972 j-i 19.75 ± 1.15(2) 14.44 ± 1.29(5) 15.96 ± 1.35(7) (18.6-20.9) (12.0-18.9) (12.0-20.9) 1-2 16.30 ± 0.72(37) 10.48 ± 1.07(16) 14.54 ± 0.70(53) (3.2-23.3) (2.3-20.5) (2.3-23.3) 2-3 1 1.34 ± 0.58(43) 4.60 ± 0.97(13) 9.77 ± 0.63(56) (5.3-18.2) (0.2-12.1) (0.2-18.2) 3-4 8.09 ± 0.85(16) 2.88 ± 0.75(12) 5.86 ± 0.76(28) (1.7-17.1) (0.1-8.8) (0.1-17.1) 4-5 7.64 ± 1.29(8) 3.36 ± 0.82(15) 4.85 ± 0.81 (23) (1.5-12.8) (0.6-11.0) (0.6-12.8) 5-6 5.89 ± 1.07(9) 2.97 ± 0.15(3) 5.16 ± 0.88(12) (1.1-10.9) (2.7-3.2) (1.1-10.9) Frequency of Reproduction. — In 1971, five fe- males >57.0 cm SVL were captured in early summer between 15 June and 4 July. Of these, three were gravid. No enlarged oocytes were dis- cerned by palpation in two others, but they may have been gravid on the basis of post-emergence weight increases of 23.4% (1.11 g/day) in 1 4 days, and 34.2% (0.64 g/day) in 39 days, increase val- ues comparable to those for known gravid fe- males (Table 8). In 1972,21 females > 57.0 cm SVL were hand- captured in early summer between 2 and 28 June (nine records of those with a spring emergence capture are in Table 8). Of these, 18 were either gravid or recently parturient (bearing postero- lateral skin folds) when collected. One of the two apparently non-gravid females was a snake that had been followed by telemetry in August and early September 1971. Her apparent nonrepro- ductive state in 1972 may have been influenced by the transmitter in late 1971, perhaps by pre- venting normal feeding and weight recovery. Considering the combined sample (telemetry snake excluded) of 25 females captured in early summer, 22 (88%) were believed to be repro- ductively active. This evidence indicates egg lay- ing each year by a large proportion of mature females in the study area. Relative Clutch Mass. — In 12 female Coluber that produced clutches of normal eggs in the lab- oratory, RCM (Vitt and Congdon 1978) was cal- culated as the proportionate weight loss imme- Table 6. Age-specific growth in weight of 179 99 Coluber constrictor mormon. Mean ± 1 SE, sample size and extremes in parentheses. Age Absolute increase (g/year) interval 1970 + 1971 1972 1970-1972 J-l 20.13 ± 1.49(3) 17.30 ± 1.09(7) 18.15 ± 0.94(10) (18.4-23.1) (14.1-21.2) (14.1-23.1) 1-2 26.05 ± 0.93(39) 21.53 ± 2.21 (10) 25.12 ± 0.89(49) (15.8-38.0) (9.7-28.9) (9.7-38.0) 2-3 20.70 ± 1.05(38) 11.16 ± 1.50(18) 17.63 ± 1.04(56) (6.8-34.4) (3.1-28.9) (3.1-34.4) 3-4 16.53 ± 2.49(20) 6.38 ± 1.34(13) 12.53 ± 1.80(33) (1.0-42.3) (0.1-18.9) (0.1-42.3) 4-5 11.43 ± 1.28(13) 5.48 ± 1.55(8) 9.16 ± 1.16(21) (3.9-18.4) (0.2-15.1) (0.2-18.4) 5-6 9.41 ± 1.65(8) 6.65 ± 3.65(2) 8.86 ± 1.46(10) (4.5-18.1) (3.0-10.3) (3.0-18.1) VERTEBRATE ECOLOGY AND SYSTEMATICS 23 TABLE 5. Continued. Proportional increase- year 1970 + 1971 19" 1970-1972 2.153 ± (1.958- .647 ± (.137- .283 ± (.124- .171 ± (.003- .136 ± (.030- .099 ± (.019- .195(2) -2.348) .035(37) -1.081) .016(43) -.579) .021 (16) -.409) .022(8) -.223) .018(9) -.188) 1.744 ± (1.316- .384 ± (.076- .110 ± (.004- .058 ± (.002- .062 ± (.010- .046 ± (.038- .177(5) -2.363) .047(16) -.756) .025(13) -.314) .014(12) -.157) .016(15) -.232) .004(3) -.053) 1.882 - (1.316- .568 ± (.076- .243 ± (.004- .123 ± (.002- .088 ± (.010- .086 ± (.019- 147(7) 2.363) 033(53) 1.081) 017(56) .579) 017(28) .409) 015(23) .232) 015(12) .188) diately following egg laying (weight lost at oviposition/body weight prior to oviposition). Mean gravid weight of the 12 females was 111.6 ± 4.47 (82.3-137.2) g; mean parturient weight was 62.7 ± 2.73 (43. 1-77.0) g. Mean RCM was 43.8 ± 1.03% (37.9-49.2%). Females were weighed an average of 6 (1-13) days prior to oviposition during which time some weight loss would be expected through dehydration (al- though water was supplied ad libitum), so the measured relative weight loss due to oviposition was probably slightly higher than actual losses had weighing immediately preceded oviposition. However, there was no significant correlation be- tween weighing interval and percentage weight loss (r = 0.30. P > 0.05). Estimates of Population Size. — Total numbers of all individual racers captured at the various dens between 1969 and 1972 (Table 10) consti- tute a direct census which was influenced by (1) the snakes' fidelity to the several communal hibernacula and (2) the effectiveness of our en- circling fences in capturing them. We believe both possible sources of error were minimal, assuring a high reliability of our direct counts of individ- uals captured. Nonetheless, each year some un- marked snakes of all ages were caught (see "Age Structure"). Among all snakes one year old or older in 1971, 26-29% at dens M, 1 , and 5 were captures of unmarked individuals (Table 1 1). In 1972 at dens M and S3, new captures comprised 22% and 33% of the samples, respectively. Mark-recapture population estimates using the Jolly-Seber method resulted in population esti- mates for males and females > 1 year old only slightly higher than the actual number of snakes Table 6. Continued. Proportional increase >car 1970 + 1971 1972 19^0-1972 2.653 ± (.022 1.000 ± (.399 .442 ± (.159 .264 ± (.010 .160 ± (.049 .106 ± (.010 .384(3) -3.348) .038(39) -1.510) .028(38) -.873) .047 (20) -.789) .019(13) -.297) .027(8) -.247) 2.055 ± (1.156- .675 ± (.281- .219 ± (.052- .098 ± (.001- .078 ± (.003- .081 ± (.037- .252(7) -3.313) .076(10) -1.061) .037(18) -.754) .021 (13) -.285) .022 (8) -.219) .044 (2) -.124) 2.235 ± (1.156- .934 ± (.281- .370 ± (.052- .199 ± (.001- .129 ± (.003- .101 ± (.010- .219(10) -3.348) .038 (49) -1.510) .026 (56) -.873) .032 (33) -.789) .017(21) -.297) .022(10) -.247) 24 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY _ 40 ■ E o w 30 2 20 2 10 o 1.0 w .8 ct o ? .6 O .4 o CL o q: .2 - ^ 84 78 970 97 972 Fics. 4. Annual proportions of Coluber constrictor mormon that increased in weight in three successive years (1970-1972) compared to yearly rainfall. Upper histogram shows total annual rainfall (unshaded), May- Sept, total (stippled), and June-Aug. total (hatched) recorded at Grantsville, Utah. Weight change records (lower histogram) are for 1 79 <53 > 50.0 g (hatched bars) and 1 7 1 99 > 60.0 g (stippled bars) initial body weight; sample sizes above each bar. caught at den M in 1970 and 1971 (Table 12). The disparity was greater for juveniles, reflecting the greater difficulty of capturing them and their higher mortality rate. These factors lowered re- capture proportions and tended to raise the es- timated population of juveniles relatively more than the estimates for adults. The relatively low "difference factors" for older snakes indicated that the sampling technique effectively captured a high proportion of the adult population. Population Density. — Using maximum dis- persal distances recorded by Brown and Parker (1976a) (1.6 km from M complex, 1.8 km from S complex) as radii of circular areas, and assum- ing uniform dispersal in all directions from each den complex, areas occupied by the Coluber pop- ulations were 804 ha at M complex and 1017 ha at S complex. In autumn 1971 and spring 1972, when sampling was most complete, 528 Coluber weighing 29.728 kg were recorded at M complex, and in spring 1972, 271 Coluber weighing 15.795 kg were recorded at den S-3 in S complex (Table 10). Population and biomass densities at M and S complexes in 1971 were 0.66 and 0.27 snakes/ ha and 37 and 16 g/ha, respectively. Population census data were adjusted by calculated differ- ence factors (Table 12) to estimate total popu- lations. Adjusted population densities were 0.79 snakes/ha at M complex and 0.32 snakes/ha at S complex. Adjusted biomass densities were 39.8 g/ha at M complex and 16.7 g/ha at S complex (Table 13). The two den complexes are located ca. 875 m apart. Thus, a 600-ha region of overlap, encom- passing 60% of the S dispersal area and 75% of the M, could contain additive densities. The overlap densities were 0.78 snakes/ha and 39.6 g/ha. These are the most representative estimates of these parameters under the conditions of sam- pling and assumptions employed in the calcu- lations. Population Changes. — Population changes during our sampling are shown in Table 14. The racer population increased by 16.5% in 1970(den M), and by 16.7% (den M) and 18.9% (dens 1 and 5) in 1971. The population increases noted in 1 970 and 1 97 1 were not sustained during 1 972 when the populations declined by 22.2% (den M) and 20.3% (den S3). Sex Ratio. — For each den and sampling pe- riod, proportions of total numbers of males (822) and females (725) >1 year old were 0.531 and 0.469, respectively. In all but two sampling pe- riods, males outnumbered females (Table 10). Sex ratios were never significantly different from a 1 : 1 expectation as tested by chi-square for any den or sampling period. Sex ratio at birth was determined by eversion of hemipenes after injection for 18 lab-reared hatchlings randomly preserved in 1971. There were 9 males and 9 females in this sample. A sample of 1 7 juveniles in autumn 1972 and spring 1973 was sexed. There were 10 males and 7 fe- males in this sample (x 2 = 0.24, 0.70 > P > VERTEBRATE ECOLOGY AND SYSTEMATICS 25 Table 7. Annual absolute weight changes of 350 Coluber constrictor mormon in three successive years. Data are for 179 36 > 50.0 g and 171 29 > 60.0 g that gained or lost weight during a yearly interval. Mean ± 1 SE. sample size and extremes in parentheses. Increase (g, sr) Sex 1970 1971 1972 66 22 4.00 ± 0.77(14) (0.1-9.0) 13.53 ± 2.23(18) (1.0-31.5) 6.62 ± 0.45(64) (0.2-14.8) 14.26 ± 1.11 (56) (1.9-41.6) 2.98 ± 0.44(36) (0.3-9.3) 8.41 ± 1.20(35) (0.1-32.6) Decrease (g, >r) Sex |47() 1971 1972 66 22 6.44 ± 1.12(8) (2.9-10.9) 5.73 ± 1.17(6) (1.6-9.4) 1.76 ± 0.53(9) (0.3-5.5) 4.48 ± 1.35(13) (0.4-18.6) 4.28 ± 0.56(48) (0.1-19.7) 7.59 ± 0.88(43) (0.4-25.3) 0.50). Of 22 snakes marked as hatehlings (not sexed) and later recaptured as 1 -year-olds, there were 9 males and 13 females (x 2 = 0.41, 0.70 > P > 0.50). These combined samples indicate that sex ratio at hatching did not differ significantly from 1:1. Age Structure.— Ages of males and females at den M in each of four successive years and at other dens (dens 1 and 5 in M complex, den S3 in S complex) in two successive years are shown in Figs. 6 and 7. In samples following the initial sampling period at a den, proportions of marked snakes we recaptured were high, particularly if 1 -year-olds are excluded from the tabulations. Proportions of marked snakes >2 years of age averaged 80% in dens M, 1.5. and S3 between 1970-1972. The greatest proportion of "new" animals consisted of 1- and 2-year-old snakes. These ages accounted for 60.6-79.4% of all un- marked snakes during three years (Table 15). Pooled composite ratios of annual age groups for all dens (Fig. 8) show that in all years there were many immatures and young adults. Juve- niles through 5-year-olds comprised 6 1 .9-75.7% of all ages. Older individuals (all snakes aged 6+ years) represented 24.3-38.1% of the popu- lation in different years. Two distinct ages were evident among the younger classes. The first was the 1 -year-old group in 1969 which remained distinct as a cohort of 2-, 3-, and 4-year-olds through 1 972. The second was the juvenile age class in 1971 which showed up as a prominent cohort of 1 -year-olds in 1972. These 1 -year-old cohorts represented hatehlings produced in 1 968 and 1971, respectively. In 1 969. yearlings accounted for 27.4% of the population; in 1972 they made up 20.3% of the population (Fig. 8). The waves of age classes in the composite age distributions may correlate with yearly variation in climatic conditions and varying annual pro- portions of Coluber that increased in weight. A favorable year for growth (1971) apparently also was favorable for reproduction as reflected in the large 1 -year-old cohort in 1 972. Prior to our study, 1968 also appears to have been a favorable year as implied by the age structure showing a large proportion of 1 -year-olds in 1969. Juveniles re- cruited into the population in 1972 made up 4.6% of the population as compared to 1 1.2% in 1971, 7. 8% in 1970 and 7.9% in 1969. In contrast to 1970 and 1971, the poorest year for juvenile recruitment was 1972 when the proportion of juveniles was 4.6% (58% lower than in 1971). Survivorship.— Overwintering survival rates averaged 93.3% per winter for males (range 80- 98% per winter) and 93.4% per winter for females (range 89-98% per winter). Relatively little mor- tality (7% per winter) occurred during hiberna- tion in the four years of this study (Table 16). Annual age-specific survival rates of males (all ages >1 year) averaged .688 (1970). .810(1971), and .620 ( 1 972). Female annual survivorship (all ages >1 year) averaged .790 (1970), .782(1971), and .553 (1972). Annual survival rates for ju- veniles were .273 (1970), .233 (1971). and .226 (1972). Growth rates and age compositions also suggest that 1972 was an unfavorable year for 26 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY 90- 6 o80 I I- o z UJ 70 z > i I- Z> O z 60- 50 o o D A A A -IT 4 Li -10 co UJ o r - ^t— — i— — r 6 7 8 9 CLUTCH SIZE o 5 oo 10 Fig. 5. Clutch sizes-9 body size (SVL) relationship for Coluber constrictor mormon in 1971 and 1972. Circles represent laboratory oviposition records; triangles represent enlarged preovulatory oocytes. Solid symbols in scatter diagram (upper portion of figure) and corresponding shaded bars in histogram (lower portion of figure) are records from area M. Open circles and triangles (upper) and unshaded bars (lower) represent records from area RB, and squares area SLC. The regression line shows clutch size (Y) vs. snout-vent length (X) for area M females and is described by the equation Y = -0.56 + 0.10X. VERTEBRATE ECOLOGY AND SYSTEMATICS 27 Table 8. Prereproductive (late spring-early summer) weight increases in 10 gravid female Coluber constric- tor mormon at area M, 1971 and 1972. O = ovarian oocytes, E = oviducal eggs. Last two snakes were re- captured prior to being tracked by telemetry. Spring Summer Inter- \v< •lght increase Clutch weight u eight sal . Sl/C (g) (g) (days) g g das % 6, O 97.4 123.5 38 26.1 0.68 26.8 5. O 69.4 90.1 25 20.7 0.83 29.8 7. O 123.8 149.9 26 26.1 1.00 21.1 8, O 128.6 169.0 29 40.4 1.39 31.4 4. O 76.4 94.9 17 18.5 1.09 24.2 5, O 68.7 83.3 24 14.6 0.61 21.3 5.E a 111.3 152.1 36 40.8 1.13 36.7 7. E b 89.0 119.0 47 30.0 0.64 33.7 6, E<= 93.2 142.0 41 48.8 1.19 52.4 '?. E d 72.5 107.5 29 35.0 1.21 48.3 - 1 Oviposited in laboratory 3 July 1972, 1 5 days after recapture. b Oviposited in laboratory 26 June 1972,4 days after recapture. "Oviposited in field 24 June 1972. 16 days after release with transmitter (8 June). d Oviposited in field, date unknown: still gravid when last seen on 26 June 1972. 8 days after release with transmitter (18 June). Coluber compared to 1970 and 1971. Survivor- ship data for these two years were pooled for comparison to 1972 data (Table 17). Survivorship over all ages increased by 18% in males from 1 970 to 1971 and remained about the same in females between these two favorable years. In 1972. survivorship in males declined 20% from the combined 1970-1971 rate (.779). Female survivorship (all ages > 1 year) dropped 29% in 1972 from the combined 1970-1971 rate (.784). Age differences in survival rates between 1971 and 1972 indicate that a greater relative decline in 1972 (44%) occurred among 1 -year- olds (both sexes) and among snakes aged 6 years or older (26% decline in males. 39% in females: Fig. 9). Young to middle-aged individuals (ages 2-5) sulfered a lower relative reduction in sur- vival from 1971 to 1972 (4-28% in males. 20- 28% in females). Overall, males fared better than females in 1972 (Fig. 9). In the favorable years 1 970 and 1971. survivorship among all females of reproductive age (>3 years old) averaged .791 per year. In 1972. survivorship among all repro- ductive-aged females was .558. Our estimate of the survival rate of juveniles from the time they were captured at dens to an age of one year is 23%. We believe this value is based on sufficient data (22 1 -year-olds recap- tured of 94 juveniles released) to allow confi- dence in this estimate. Our life table (see below) utilizes an estimate of 17% survivorship during the first year (from egg to 1 -year-old). This sug- gests that the 1 .5-month interval between hatch- ing (mid-August) and arriving at the dens upon attaining juvenile age status (late September to early October) involved a 79% survival rate, and a survivorship of 72% from egg tojuvenile (Table 18). Applying our estimates of survivorship to the estimated total production of female eggs (302) by 106 reproductive-sized females captured at three dens (M, 1, 5) in 1971 yields an expected 221 female juveniles. We actually caught 32 fe- male juveniles at these dens in 1971. only 14% of the expected number. This could indicate one of three possibilities: (1) that our first-year sur- vivorship estimate is too high. (2) that the esti- mate of female productivity is too high, or (3) that a fairly large proportion of the annual crop Table 9. Postreproductive (late summer) weight increases in four parturient female Coluber constrictor mor- mon. Snake no. (area. \ear) Gravid Parturient Late summer Clutch weight weight weight Inters al si/e (g) (g) (g) (days) Weight increase g das 1. (RB. 1971) 2. (RB. 1971) 3. (RB. 1971) 3. (RB. 1972) 4. (M, 1972) 10 8 5 4 6 56.6 87.3 149.4 45 27.2 73.3 124.2 45 20.8 65.4 — 318 03.1 57.2 71.3 31 25.5 77.0 1 14.0 53 62.1 1.38 50.9 1.13 37.7* — 14.1 0.45 37.0 0.70 71.1 69.4 57.6* 24.7 48.1 * Increase that occurred in the 318-day interval between 1971 and 1972. 28 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 10. Numbers, live-weight biomass, and sex ratios of Coluber constrictor mormon captured at five hibernacula in M complex (dens M, 1,2, 3, 5) 1969-1972 and at one hibernaculum in S complex (den S3) in 1972. Number of snakes at den M is total different individuals for both autumn and spring sampling; for all other hibernacula totals are different individuals in spring only. Weights are for animals in spring unless only autumn capture was recorded. Males and females include all ages > 1 year old, juveniles (J) include both sexes < 1 year old (see text). ss ■■ JJ All snakes Total Total N cap- Propor- Weight N cap- Propor- Weight N cap- Weight N weight Yeai Den tured tion (kg) tured tion (kg) tured (kg) captured (kg) 1969 M 65 .508 3.667 63 .492 3.661 11 .091 139 7.419 1970 M 73 .507 3.709 71 .493 4.703 18 .138 162 8.550 1970 1 39 .534 2.055 34 .466 2.343 2 .016 75 4.414 1970 5 72 .518 3.727 67 .482 4.360 10 .081 149 8.168 1971 M 77 .484 4.318 82 .516 6.054 34 .281 193 10.653 1971 1 48 .533 2.767 42 .467 3.127 8 .067 98 5.961 1971 2 17 .654 0.760 9 .346 0.547 3 .025 29 1.332 1971 3 13 .448 0.697 16 .552 1.236 2 .014 31 1.947 1971 5 81 .526 4.439 73 .474 5.214 23 .182 177 9.835 1971 S3 137 .544 7.604 115 .456 8.011 19 .180 271 15.795 1972 M 75 .543 3.656 63 .457 3.852 10 .089 148 7.597 1972 S3 125 .581 6.170 90 .419 5.650 7 .067 222 11.887 of juveniles did not utilize the communal dens in their hatching year. We favor the third pos- sibility. Young snakes are perhaps not able to travel the distances necessary to get back to the dens for a variety of behavioral reasons. Many first-time users of the dens were yearlings (Table 15). The 302 female eggs should have resulted in 39 1 -year-olds the following year. Lacking den- fenced samples for dens 1 and 5 in spring 1973, however, we have no direct confirmation of the actual arrival of 1 -year-olds to any den except den M in 1972(14 yearlings taken). If we assume that the population had remained stable between 1 97 1 and 1 972, the 39 expected yearlings would have comprised 19.9% of all 196 females (the total number captured in 1971 ). This percentage compares closely to 1 7.7%, the actual mean pro- portion of 1 -year-olds to all females (6.1-35.3%, N = 10 den samples, all years). Thus, we feel that a 1 7% survivorship estimate from egg to yearling is accurate. Life Table. — Table 19 presents a life table for area M females. This life table is based on the following data and assumptions: (1) Sex ratio at hatching was 1:1. (2) Mean unadjusted age-specific fecundity for all younger females (5.0-5.8 eggs /2/yr, ages 2-6) was calculated from all body size (SVL) records in spring 1971 and an expected clutch size based on the appropriate regression. For all older females (ages >7), the mean un- adjusted fecundity (6.0-7.0 eggs/2/yr) was calculated by the same procedure in the body size range 66-75 cm. (3) Mean fecundity values were adjusted by as- Table 1 1. Proportions of marked Coluber constrictor mormon among recapture samples at four dens in three successive years. Total number of snakes captured in parentheses. Den >i -year-old > 2-year-old Year 66 M <5<5 '-■ 1970 M .520(73) .605(71) .603(63) .666(63) 1971 M .722(77) .743(82) .823(68) .789(76) 1971 1,5 .736(129) .713(115) .814(113) .832(95) 1972 M .773(75) .794(63) .887(62) .938 (49) 1972 S3 .648(125) .689(90) .788(99) .867(68) VERTEBRATE ECOLOGY AND SYSTEMATICA 29 Table 12. Comparison of sampled and estimated population sizes of Coluber constrictor mormon at den M in 1 970 and 1971. Difference factor is a proportion calculated by dividing the Jolly-Seber population es- timate by the actual number of snakes caught. Table 13. Population and biomass densities of Col- uber constrictor mormon in 1971. Difference factors used to adjust population sizes and total weights were calculated from data in Table 12. Population density Year Age Sex Number caught Jolly- Seber estimate Differ- ence factor Den com- plex Age sex Total number snakes captured Differ- ence factor Estimated total population size Snakes 1970 JJ > 1 year > 1 year <3 + 9 66 99 3 + 9 66 99 18 73 71 162 34 77 82 193 49.8 79.6 80.0 209.4 63.4 84.5 84.8 232.7 2.77 1.09 1.13 1.29 1.86 1.10 1.03 1.21 ha M S all all 528 271 1.21 1.21 639 328 0.79 0.32 All snakes Biomass density 1971 JJ > 1 year > 1 year All snakes Total live weight (kg) Differ- ence factor Estimated total weight (kg) Grams ha M JJ 66 0.569 12.981 16.178 1.86 1.10 1.03 1.058 14.279 16.663 99 all 29.728 32.000 39.8 suming no reproduction i at age 1 . 8% at age S JJ 0.180 1.86 0.335 2, 77% at age 3, and 90% at age 4 and all 66 7.604 1.10 8.364 subsequent ages; hatch sumed to be 92% at all ing success was as- ages. 99 all 8.011 15.795 1.03 8.251 16.950 16.7 (4) Estimated survivorship from age to 1 ( 1 7%) was based on the proportion of unmarked 1- year-old females (above). (5) Age-specific survival rates were based on all recapture records for the years 1 970 and 1971 combined. The life table (Table 19) calculated for Coluber indicates an increasing population (R (1 = 1.187). An approximate value of the mean generation time was calculated as T c = (2 xl x m x )/R = 6.89 years. The highest proportion (15.8%) of the net reproductive rate was contributed by 3-year-old females. Relative contributions to R () from the other adjacent age classes were 1 5.0% by 4-year- olds, 13.5% by 5-year-olds, and 10.8% by 6-year- olds. The life table for Coluber in Utah indicates a maximum longevity of some 1 5 years in the pop- ulation studied. That a few racers survive to this age is not unreasonable. Two males originally marked with metal jaw tags in autumn 1964 by Hirth (1966) survived through 1970 and 1971, periods of 6 and 7 years after marking. We es- timate these snakes were at least three years old when tagged (at a body size that would permit such tagging), so they reached minimum ages of 9 and 10 years. In autumn 1972 and spring 1973, 15 males were recaptured at den M that had been among an original group of 37 at least 3 years old when initially marked in 1969-1970. These animals were all 6 years old or older when last handled. At average growth rates calculated in this study both sexes of Coluber would reach maximum Table 14. Annual population changes of Coluber constrictor mormon in three successive years as mea- sured by total captures (all ages, both sexes) at four dens. Sampling periods: A = autumn, S = spring, num- ber = year. Rate of Year Den Sampling period Number of snakes captured increase yr Expo- Finite nential (A) (r) 1969 M A69, S70 139 — — 1970 M A70. S71 162 1.165 0.153 1971 M A71, S72 189 a 1.167 0.154 1971 1.5 S71 217" — — 1971 1, 5 S72 258 c 1.189 0.173 1972 M A72, S73 147 d 0.778 -0.251 1972 S3 S72 271 — — 1972 S3 S73 216 0.797 -0.227 a 193 (total)-4 (den shifts). b 224 (total)- 7 (den shifts). c 275 (total)— 1 7 (den shifts or displacement returns). d 148 (total)- 1 (den shift). 30 SPECIAL PUBLICATION- MUSEUM OF NATURAL HISTORY 66(0) FEMALE ^" — . 1 . k 6* 5 4 3 2 1 82(1) MALE a 1972 \''A I\\ '■'■'- '''^ j ' *.'■'■' " 99(3) :■ *— T 1 ■*•*•*•* 1*1 I 6 + 5 1 4 94(1) 3 X 2 ,«,• 1 :•:•:•:•:•:•:• i Q "7 1 J ■•■•■•'■•■•'•■•■;■••••••'•'< i 1 — 80(0) ■ T I I I I I * 6* c J 4 3 2 1 '.'•.■■': : ' 82(0) :•:•:•:•:•:•:•:•:■:] 1970 :•:•:•:•:•:•:•: 1 6* :S$:$x:i?:??S:?:W:i:¥ 5 .i/M, -"- 1 4 :•:• 71(3) 3 2 1 'X*X*X*X*X\ --*. »^ -l-fc. J xi'x i i J lO 20 30 AGE NUMBER OF SNAKES Fig. 6. Age and sex structure of Coluber constrictor mormon at den M in four successive years. For 1 970— 1 972 stippling = unmarked individuals and snakes not using den in previous year. For 1969 no stippling = individuals marked between 1964-1966 (Hirth, 1966); For other years no stippling = marked individuals (present study). Number in parentheses following num- ber of snakes indicates marked snakes which shifted to this den from another den and included among un- marked snakes. J = juvenile (sexed only in 1 972-1973); other numbers = age in years. sizes observed (ca. 75 cm SVL in females, 65 cm SVLin males) in about 10 years. Yearly variation in environmental favorability for growth and in- dividual variation in response to these condi- tions naturally would affect this estimate, but our size data indicate that larger racers may have been 10 years old or older. A body size distri- bution of large (>65 cm SVL) females of un- known age in 1971 was tabulated. Size range, number of individuals, percentage of total fe- males caught in 1971. and probable ages of these snakes were as follows: 65-66.5 cm. 17 (5.3%), 7-8 yr; 67-68.5 cm. 15 (4.7%), 9-10 yr. 69-70.5 cm, 15 (4.7%), 1 1-12 yr; 71-72.5 cm, 13 (4.1%), 13-14 yr; 73-75.0 cm. 4 (1.2%). > 1 5 yr. Two extremely robust individuals were recorded that far surpassed sizes even of snakes normally considered "large." One male measured 94.0 cm SVL and 153.6 g, and one female was 85.0 cm SVL and 194.2 g. These animals may have been unusually old individuals. Discussion Prior Studies at Utah Dens. — During our study of C. c. mormon in Utah, we investigated also two other colubrids, the desert striped whipsnake {Masticophis t. taeniatus) and the Great Basin gopher snake (Pituophis melanoleucus deserti- cola) (Parker and Brown 1980), as well as the Great Basin rattlesnake {Crotalus viridis lutosus; Parker and Brown 1974). The den M Coluber population was sampled for seven years in a 10- year period, from 1963-1965 (corresponding to spring 1964-1966 den fencings) by Hirth and King ( 1 968), and from 1 969-1972 (present study). Den M was not studied in the three-year period 1 966-1 968. Population growth, age structure, and overwintering survival rates may be compared with the earlier results of Hirth and King ( 1 968) and Hirth (1966). Increases apparently occurred in the Coluber population in 1964 (49 to 63 snakes) and 1965 (63 to 67 snakes) (Hirth and King 1968). In the three intervening years between studies, the pop- ulation grew from 67 to 139 individuals, an av- erage rate of increase of 26. 7% per year (/' = 0.182/ yr). Two favorable years of high individual growth, high survival rates, and high juvenile VERTEBRATE ECOLOGY AND SYSTEMATICA 31 FEMALE 95(2) ! i i SSSHSSSi ' ' ' * * ' v1,..i ' — 1 1 1 1 — ""^ 6* 5 4 3 2 1 J MALE 127(4; i himi i n 1 I S-3 1972 I » I l-l T T T T I T I f T T I I I I I 124 (o) m I n ■ V l ' lVr »l'«' . ' l '|-. i-H k n 1 ; ;; ■ ■ 6* 5 4 147 (0) ittmm n :::■ 3 2 1 •;v;v;«;«;«;*;»;»;»iV//iViWi' S-3 ■X\vXvX*X*X\v. - .v.\v J '>()<>))[" 1971 L^4_ 1 i i 130(7) i i ; w m — i — ■ — i •:•:•:*:•§ 6+ 5 4 3 2 1 145(7) »**«**■*» l>V«V»V«Il"lVr'l » »»» »»* ! & 5 1971 — — i — f"***"*'*'*'*'* , * , * , ***'*******»***»"' '*«"il; 107(3) 1 ' l * ■ ' ■ * ! [ ■ ' . ' ■ ■ ■ ■ ■ ■ ■■■'.V'.'. ' .V * K-X-Xv ■ •'•'■V i^ 6* 5 4 &#? 117 (4) 3 ? 1 '.'.'. '.'.'. 'A 1 . 1 . 1 . 1 . 1 . 1 . 1970 i i J LiixJ — 40 30 20 lO AGE 10 2 ° 3 ° NUMBER OF SNAKES 40 Fig. 7. Age and sex structure of Coluber constrictor mormon at three dens in two successive years. Lower two histograms show M complex dens 1 and 5 in 1970 and 1971 (spring 1971 and 1972 samples); upper two histograms show S complex den S3 in 1971 and 1972 (spring 1972 and 1973 samples). Symbols as in Fig. 6. recruitment ( 1 970 and 1971) resulted in contin- ued expansion of the population. An unfavorable year (1972) reversed these trends because of the adverse effects of a summer drought. Age structure in the Coluber population also may be compared with results of Hirth and King (1968). Their designation "hatchling" is equiv- alent to our juvenile age class, and their "juve- nile" group is equivalent to our 1 -year-old snakes based on average body weights calculated from their data (25.5-34.2 g/snake). Their proportions of "juveniles" were .245 ( 1 963). .222(1 964), and .493 (1965). These values compare to our pro- portions of .274 ( 1 969). .158(1 970), . 1 30 ( 1 97 1 ). and .203 (1972) for 1 -year-olds. Evidently, 1964 seems to have been a favorable vear for Coluber 32 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY FEMALE c (161) -i — i — i — i — i — i — i — i — i — i — " — i < ' ■" (380) -i — i — i — i — i — i — i — i — i — i — i — <~ (187) "I 1 1 1 1 1 1 1 1 1 1 1 1 ' ■ ' MALE 6+ 5 4 3 2 1 J 6+ 5 4 3 2 1 J 6* 5 4 3 2 1 J ■i 7•2^^^^^^^^x^ *•'•'•'^'^'•'•'■'■'^'^'■J'■'-*-'-'-** (209) ■ pyyKTgggg?W5 ' WW^T?TC??WWTCTCOTWB J (419) 971 r i i — i — i — i — < — > ' i ' (199) 1 '970 i — i — i — S — i — i — i — i — i — i — i — ' — i i ■ ■ i" 1 6+ fro 1 5 loo J 4 3 2 [ 1 J -ft —- . - . - . ■ . - . ■ ■ ■ ■ ■ ■ • ■ ' ■ • ■ • ■ • •x-v'/'.'. ■, ■.',',';'.'. '.'. ' . ' . ' . ' . ' . ' . ' .v a 1 I I 1 (71) 1969 -i — i — i — ? — r — » — i — i — ! — 8 10 12 14 16 16 14 12 10 8 6 4 2 AGE 2 4 6 PERCENTAGE OF SNAKES Fig. 8. Proportional age and sex structure of Coluber constrictor mormon at all hibernacula in four successive years (1969-1972). Number of individuals in parentheses. Age-specific proportions were calculated on the basis of the total number of snakes (both sexes) for each year. Stippling = 1 -year-old cohort in 1969 continuing as prominent older age groups in later years; hatching = juveniles in 1971 and 1 -year-olds in 1972. Annual total number of individuals reflects a varying number of dens sampled each year, not population changes. reproduction if nearly half of the population in est precipitation in the 1 3-year period 1960-1972 1 965 consisted of 1 -year-olds as the data of Hirth at the study area (Brown and Parker 1 982). and King (1968) imply. This is corroborated by Overwintering survivorship in 1964-1965 was rainfall data which show that 1 964 had the high- estimated to be 54% among 1 3 fall-released Col- Table 15. Proportions of new (unmarked) Coluber constrictor mormon of each age at four dens in three successive years. Data are for subsequent sampling periods following initial marking (sexes combined). Marked individuals (N = 35) that shifted from one den to another (thus "new" to that den) were excluded. Den No. unmarked snakes Age (; ears Year l 2 3 4 5 6 + 1970 1971 1972 M M. 1, 5 M, S3 63 109 102 .270 .404 .608 .381 .202 .186 .127 .174 .098 .083 .039 .079 .037 .029 .143 .101 .039 VERTEBRATE ECOLOGY AND SYSTEMATICS 33 Table 16. Overwintering survivorship of Coluber constrictor mormon > 1 year old estimated by recap- ture proportions at den M. Table 17. Annual survivorship of Coluber constric- tor mormon estimated by recapture proportions at three dens(M, 1,5) in 1970 and 1971 (years combined) and 99 N snakes released in autumn N sur- \ i\ ing to fol- lowing spring or later Sur- vival rate N snakes released in autumn N sur- viving to fol- lowing spring or later Sur- s isal rate Age 1970 + 1971 1972 N released N sur- \ iving to next age Sur- vival rate N released N sur- viving to next age Sur- vival rale Year 1 49 38 .776 $8 36 15 1969-1970 25 20 .800 29 27 .931 .417 1970-1971 52 49 .942 47 45 .957 2 62 49 .790 27 16 .593 1971-1972 67 66 .985 74 66 .892 3 23 18 .783 35 25 .714 1972-1973 66 61 .924 47 46 .979 4 11 8 .727 29 21 .724 Totals 210 196 .933 197 184 .934 5 6 + 16 92 13 71 .813 .772 15 66 10 42 .666 .636 uber > 1 year old (Hirth 1966). In contrast, our 1 2 53 54 39 44 .736 .815 29 23 29 9 19 .391 .655 data indicate that winter survival was 93% in 3 30 23 .767 54 31 .574 racers. Winter survivorship in adult Masticophis 4 16 14 .875 21 14 .666 was reported to be 65% and among juveni leand 5 6 + 13 75 10 59 .769 .787 15 48 9 23 .600 .479 adult Crotalus 66% (Hirth 1966). Our results for these species are 95% winter survival in adult JJ Masticophis (Parker and Brown 1980) and 96% in Crotalus (Parker and Brown 1974). Had the 41 10 .244 53 12 .226 sampling effort in Hirth's study continued into the next year, snakes not caught at spring emer- gence may well have been captured later. We made a particular effort to ensure high catch- ability of each den fence and searched the fenced dens continuously in favorable weather. Despite these precautions, some fall-released snakes were not recaptured until the next fall or spring when they were credited as having survived the pre- vious winter. For similar reasons, little can be concluded about the rates of increase of Coluber between 1 964 and 1 966 (Hirth and King 1968). There has been, however, a definite general trend of in- crease in the racer population over a 30-year period at area M (Parker and Brown 1973). Fac- tors involved in the population changes in this snake community, and the probable destruction of the communal dens resulting from burial by sand after an extensive range fire in 1974, are discussed in Brown and Parker (1982). Life History Strategies. — Several workers have generated new results from a demographic-evo- lutionary perspective of snakes (e.g.. Vial et al. 1977. using data of Fitch 1960) or have sum- marized the literature from this perspective (Turner 1977; Feaver 1977). We will attempt in this section, first, to focus on the reviews of Tur- ner (1977) and Feaver (1977) as they bear on other species of snakes for comparison with Col- uber constrictor mormon, and second, we will compare our data with those of Fitch (1963) on the midwestern subspecies, C. c. flaviventris. Survivorship in snakes is a parameter whose measurement is receiving increasing attention by field ecologists. Our data on annual survivorship of four species over three years (both sexes > 1 year old, number of individuals released in pa- rentheses) are as follows: Masticophis .862 ( 1 09). Crotalus .820 (44). Pituophis .750 (104), and Coluber .695 (892). Survivorship of Coluber in 1970 and 1971 was .781 (494). The other colu- brid species (Masticophis and Pituophis) did not seem to be as subject to between-year variation in survivorship as did Coluber. Whatever the factors may be in causing such interspecific dif- ferences in survival rates (discussed further in Brown and Parker 1982), all of these snakes had similar survival rates of around 80% per year in the more favorable years. Turner (1977) reviewed annual survivorship data of eleven other species of colubrids. These range from 0.30 to 0.75, with a mean survivor- 34 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY ^ .8 a* *" .7 I- 2 .6 a: id W .4 .3 • '/ O A^ / A' / / / / O— . • A A • -o / • -A_ --A X A A a'' / A 1-F ,972 - 1 1 1 i ' 12 3 4 5 6 AGE (yr) Fig. 9. Annual age-specific survivorship of Coluber constrictor mormon. Circles = 1971, triangles = 1972; solid symbols and dashed line = 33, open symbols and solid line = 22. ship of 0.50/yr. Feaver (1977) measured survi- vorship by mark-recapture proportions of Ner- odia sipedon over four years in Michigan and found the mean annual survival to be 0.35 in individuals > 1 year old. Apparently, the surviv- al rates we measured for the three species in Utah are among the highest known for colubrid snakes, and are more similar to the values reported for four species ofviperids (0.70-0.82; Turner 1977) and for one colubrid, Diadophis punctatus (0.75; Fitch 1975). One of the most difficult values to obtain in this study was an accurate estimate of survivor- ship in the first year of life. We have given our bases for arriving at an estimate of a 1 7% survival rate from egg to yearling. Fitch ( 1 963) estimated that survival through the egg stage was 50% in C. c. flaviventris, followed by a 31.3% survival from hatchling to yearling; these combined es- timates suggest a first-year survival rate of 1 5.6%, not greatly different from our estimate in Utah. Parker and Brown (1980) estimated first-year survival rates in Masticophis taeniatus and Pi- tuophis melanoleucus to be 8.3% and 20.0%, re- spectively. Feaver (1977) reported juvenile sur- vivorship rates of 19.3% in male and 23.5% in female Nerodia sipedon. It appears that juvenile and first-year mortality is considerably higher than adult mortality in these four species. This is not particularly surprising in view of the con- siderable number of potential predators of small- sized snakes in the habitats of some of these species (cf. Fitch 1963; Parker and Brown 1980). The resulting survivorship curves of the three Utah colubrids are type III curves (Deevey 1 947; cf. Fig. 1 1 in Brown and Parker 1982). Of the average annual mortality in C. c. mor- mon (around 21% per year in mature females), one third occurred during hibernation (7% per year). The remaining 14% per year mortality in females occurred during the summer. The sum- mer activity season may be divided into prere- productive and postreproductive periods. The prereproductive interval is approximately two months long, from emergence (early to mid-May) to egg laying (late June to mid-July). The post- reproductive period lasts approximately three months from the end of oviposition until ingress into hibernation (mid to late September). We have no data that measure the relative degree of summer mortality and how it is partitioned dur- ing the activity season. Additional risks likely Table 18. Estimates of the components of survivorship during the first year of life of Coluber constrictor mormon. Life history interval Egg to hatchling Hatchling to juvenile Juvenile to yearling Egg to yearling Approximate duration of interval 45 days (late June-mid Aug., same year) 45 days (mid Aug.-early Oct., same year) 345 days (Oct.-mid Sept., successive years) 450 days (late June-early Oct., successive years) Survivorship during interval 0.92 (.0204/day) 0.79(.0175/day) 0.23 (.0006/day) 0.17(.0004/day) VERTEBRATE ECOLOGY AND SYSTEMATICS 35 Table 19. Schedule of age-specific survivorship and fecundity of female Coluber constrictor mormon, x = age (years): P x = age-specific survival rate; l x = survi- vorship to age x: m v = number of female eggs produced each year by a female of age x; R„ = net reproductive rate. See text for assumptions and for adjustment fac- tors of m, schedule. Px lx Unad- justed ITU Adjusted m. I v m x 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 .170 .736 .815 .767 .875 .769 .787 .787 .787 .787 .787 .787 .787 .787 .787 .787 .000 .170 .125 .102 .078 .068 .053 .041 .033 .026 .020 .016 .013 .010 .008 .006 2.50 2.60 2.75 2.85 2.90 3.00 3.10 3.20 3.30 3.40 3.50 3.50 3.50 3.50 0.18 1.84 2.28 2.36 2.41 2.49 2.57 2.66 2.74 2.82 2.91 2.91 2.91 2.91 R = .023 .188 .178 .160 .128 .102 .085 .069 .055 .045 .038 .029 .023 .017 1.187 accrue to females during their increased move- ments in search of oviposition sites (Brown and Parker 1976a; Parker and Brown 1972. 1980) so it seems reasonable to suppose that a higher mor- tality may occur during the shorter prerepro- ductive phase than during the longer postrepro- ductive phase. In both a "good" and a "bad" year (1971 and 1972. respectively), three-year- old females had lower survival rates than either of the adjacent age classes. As most females ma- tured and presumably began reproduction for the first time at age 3, our data on higher mortality in 3-year-old females support this argument. To our knowledge, the only other attempt to measure the components of mortality in a snake population is that of Feaver ( 1977). Adult female N. sipedon suffered their heaviest losses (50% of the total annual mortality) in summer, adult males in spring (47% of the total). In each sex, mortality was higher in the season of most active repro- ductive behavior, i.e., spring mating activity in males, summer gestation and parturition in fe- males (Feaver 1977). Of the total annual mor- tality in N. sipedon, 32% occurred over the win- ter. This value is almost identical to our data (33% of the annual mortality was overwintering Table 20. Age-specific body size and fecundity in two populations of Coluber constrictor. Data are for C. < flaviventris in Kansas (Fitch 1963) and C. c. mormon in Utah (present study). H = hatchling, J = juvenile, numeral = years. Mean weight (g) •• K Mean clutch size Age Kansas Utah Kansas 1 tah Kansas I tali H 4.2 J 12.3 1 52.6 2 68.2 3 102.1 4 139.0 5 152.4 6 175.9 6.0 8.3 27.0 41.0 49.1 53.4 60.4 61.2 4.2 12.3 51.6 83.5 149.4 212.3 209.6 245.9 6.0 8.8 28.4 51.6 66.3 71.4 79.5 84.0 9.2 9.9 10.8 13.0 15.7 5.0 5.2 5.5 5.7 5.9 mortality in C. c. mormon). Feaver (1977) also reported a higher survivorship among A T . sipedon males (38%) than among females (30%) which is also the case for C. c. mormon (males 71%, fe- males 68% over all three years of our study). In Utah, the life history and demography of C. c. mormon may be summarized as follows: maturity in one year (males) and three years (fe- males), larger size and more rapid growth in fe- males than males, slight increase in fecundity with female size and age, large size of eggs rel- ative to female size, iteroparity. low juvenile sur- vivorship, and high adult survivorship. To de- velop insight into the possible adaptive aspects of the life history strategy of the Utah subspecies, it is helpful to compare it to the Kansas popu- lation of C. c. flaviventris studied over a number of years by Fitch (1963). Kansas racers (C. c. flaviventris) are larger than Utah racers (C. c. mormon) at all ages except at hatching (Table 20); therefore, racers in Kansas have a more rapid growth rate. Sexual maturity is reached in one year (males) and three years (females) in both populations. Racers in Kansas have an age-specific fecundity about twice as high as in Utah (Table 20), but Utah racers produce larger eggs and hatchlings and have a higher rel- ative clutch mass at all ages (Fig. 10). A consid- erable difference appears to exist in the age dis- tribution between the two populations, with Kansas racers constituting a younger-structured population than Utah racers (Fig. 11). The net fecundity distributions (age-specific contribution to the female replacement rate) of the two pop- ulations are similar, but Kansas females make a 36 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 21. Comparison of the major life history traits in two populations of Coluber constrictor. Where possible a measured value is given for each trait. Data from Fitch (1963) for Kansas (C. C. flaviventris) and present study for Utah (C. c. mormon). Life history parameter Kansas Utah Population density* Body size** Growth rate Reproduction Sexual maturity Clutch size Egg size Hatchling size Relative clutch massf Demography Age distribution Relative contribution to R by female of age x Juvenile survivorship Adult female survivorship Generation time higher (5.0/ha) larger (66 123 g) (25 155 g) faster 1 year (83) 3 years (22) larger (1 1.7 eggs) smaller (5.7 g) smaller (4.2 g) lower (.40) younger (72% 1-3 yr) (28% 4+ yr) higher over ages 2-6; peak at age 3 (19.3%) higher (31 %/yr) lower (62%/yr) shorter (5.1 yr) lower (0.8/ha) smaller (66 56 g) (92 69 g) slower 1 year (<3<5) 3 years (22) smaller (5.8 eggs) larger (7.8 g) larger (6.0 g) higher (.62) older (52% 1-3 yr) (48% 4+ yr) lower over ages 2-6; peak at age 3 (15.8%) lower (23%/yr) higher (79%/yr) longer (6.9 yr) * Value for Kansas from Turner (1977). ** Mean body weight, random samples > 1 year old C. c. flaviventris (N = 50 each sex; Fitch 1963) and C. c. mormon (136 <5<5, 114 22, den S3, spring 1972). t Mean clutch weight/mean body weight of non-gravid 22; mean RCM value for 5 ages (2-6+ years). somewhat higher relative contribution to R be- tween ages 2-6 years (Fig. 1 2). The distributions indicate that 3-year-old females contribute the most to R in each population. Utah racers have higher adult survival rates than do Kansas racers. Life tables developed for each population show that Utah racers have a somewhat longer esti- mated generation time, suggesting a less frequent turnover of the population. An overall summary of the major life history comparisons is presented in Table 21. It is apparent that there are several prominent reproductive and demographic differences be- tween racers in Kansas and Utah superimposed on a basic plan of biological similarities. Both subspecies of C. constrictor exhibit an identical growth pattern in which females mature later and grow larger than males. Feaver (1977) placed C. constrictor, Rhabdophis tigrinus, Thamnophis butleri, and Nerodia sipedon in this group as con- trasted with Crotalus viridis, Agkistrodon con- tortrix, and Elaphe quadrivirgata in which males grow to the larger size. There are several impor- tant reproductive and behavioral differences be- tween the two groups of snakes (cf. Shine 1978); generally the last group of species tends to show late maturity, high adult survivorship, small clutches, and large young as contrasted to the first species group which shows the opposite trends. Viewed at this level, one could apply a "K-selected" label to the second group and an "r-selected" label to the first. Whereas such a comparison may help to visualize the broad strategies, it is less capable of showing differences in an intraspecific comparison. C. c. flaviventris and C. c. mormon each seems to possess some "K" and some "r" attributes (cf. Pianka 1970; Stearns 1976). Our data on survivorship of C. c. mormon show that there were considerable between-years effects on survival in adult females and lesser effects in adult males when a dry year (1972) followed wetter, more favorable years (1970 and 1971). Juvenile survivorship, on the other hand, was not as strongly reduced in 1972 from 1970— 1971 levels. Under this regime (with adult mor- VERTEBRATE ECOLOGY AND SYSTEMATICS 37 3 4 5 AGE (YR) Fig. 10. Mean age-specific relative clutch mass in two populations of Coluber constrictor calculated as the clutch weight as a proportion of the mean non- gravid 9 body weight. Mean clutch weight was calcu- lated from clutch size and mean weight of eggs for C. c.flaviventris in Kansas (K; Fitch 1963) and C. c. mor- mon in Utah (U; present study). tality variable), a stable environment should fa- vor such traits as fewer young, longer life span, smaller reproductive effort, and slower devel- opment (Stearns 1976). It is not clear from the data available whether C. c. flaviventris has a more variable adult or juvenile survivorship and which environment, Kansas or Utah, is the more "stable." The Kansas habitat appears to be trophically more diverse. Insects (grasshoppers, crickets) are eaten by C. c. mormon almost ex- clusively, whereas C. c. flaviventris takes a mod- KANSAS n ' 242 1959 6* 5 4 3 2 1 -72 • 62 UTAH 19 69 n = 1.547 40 30 20 10 AGE 10 20 30 40 % of Sample Fig. 11. Age distributions in two populations of Coluber constrictor. Data are based on the proportion of snakes (both sexes combined, juveniles excluded) in six ages pooled over 4 years in Kansas (C. c. flaviven- tris: Fitch 1 963) and Utah (C c. mormon, present study). rr u_ o UJ < UJ o rr UJ a. 20 / / / / • ' \ / o K«— - Uo — 15 \ 8T' - / \, 10 / / • o 5 o i 1 1 1 1 2 3 4 5 6 AGE (yr) Fig. 12. Net fecundity distributions in two popu- lations of Coluber constrictor. Data plotted are l,m x values as a percentage of R contributed by females of each age. Data are based on a life table for C. c. fla- viventris in Kansas (K) developed by F. B. Turner (un- published) from data of Fitch ( 1 963) and on a life table for C. c. mormon in Utah (U: present study). erate proportion of small mammals, snakes, and lizards in addition to insects (Fitch 1963). No readily identifiable parameters of environmental predictability, resource limitation, or predation levels affecting racers at either locality are known. As Wilbur, Tinkle, and Collins (1974) have pointed out, until the precise role of these factors can be identified, the selective basis for the dif- ferences between the two taxa cannot be ex- plained by a simplistic dichotomous key of life history traits. Summary An intensive mark-recapture study of the Western yellow-bellied racer. Coluber constrictor mormon, was conducted over a 4-year period at two complexes of communal hibernacula in northern Utah. The objectives of the study were: ( 1 ) to gather detailed data on growth rates, sexual maturity, clutch size, population structure, and age-specific survivorship; and (2) to compare 38 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY these data with those of a population of C. c. flaviventris in Kansas studied by Fitch (1963). Racers were captured at their dens with en- circling screen fences each autumn and spring between 1969-1973. Snakes were measured, weighed, sexed, and permanently marked by scale clipping. The age of each individual was deter- mined by comparing its size to confidence in- tervals for length and weight of recaptured known- age snakes. A total of 1046 racers was captured 1694 times. Males became sexually mature at an age of <13.5 months. In females, 8% of 2-year-olds, 77% of 3-year-olds, and 90% of 4-year-olds were considered mature. Mean weight of hatchlings was 6.0 g, and juveniles 8.5 g. At an age of 1 year, females (.v = 28.4 g) weighed significantly more than males (x = 27.0 g) and females con- tinued to be significantly larger in both snout- vent length and weight at all ages. Body weight declined in 95% of the snakes over the winter; losses averaged 7.4% of initial autumn weight in both sexes. In 1971. a year of relatively high rainfall. 85% of racers gained weight over the summer, whereas in 1972. a dry year, only 44% gained weight. Age-specific growth rates were significantly higher in 1970 and 1971 than in 1972. Females produced a single clutch per year av- eraging 5.8 eggs. Eggs averaged 38 x 18 mm and 7.8 g. Mean clutch weight/female body weight ratio was 44%. Oviposition occurred in late June through early July: hatching occurred in mid to late August after an incubation period of 45-50 days. Hatching success was 92%. Sex ratio at hatching did not differ significantly from 1:1. Weight increases in prereproductive females in early summer averaged ca. 1 g/day as did post- reproductive weight recovery in late summer. Among randomly-collected females in early summer, 88% were gravid or parturient. In 1971, sampling at six dens yielded 528 rac- ers. The largest number recorded at a single den in one season (spring 1972) was 271 snakes. Us- ing maximum dispersal distances and assuming a uniform radial movement pattern from the dens, population density was 0.8 snakes/ha and biomass density was 40 g/ha. The population at den M increased from 139 to 189 individuals ( 1 8%/yr) over two successive favorable years (1970, 1971) and declined to 147 individuals (21% decrease) in an unfavorable year (1972). In all samples of snakes >1 year old, males comprised 53% and females 47% of the popu- lation. Age structure favored younger (<5 years) animals which comprised 62-76% of the popu- lation in different years. Large proportions of 1-year-olds in 1969 (27.4%) and juveniles in 1971 (11.1%) indicated that 1 968 and 1971 were years of high productivity. In contrast. 1 972 was a poor year for recruitment of juveniles (4.6% of the population). Overwintering survival rates averaged 93% in both sexes. Annual survivorship in juveniles was 23%. First year survival (egg to age 1) was esti- mated to be 1 7%. Adult survivorship in favor- able years was 78% in males and 79% in females. In an unfavorable year adult survivorship was 62% in males and 56% in females. Two other species of sympatric colubrid snakes in Utah had annual survival rates of around 80% per year. In contrast, literature reports for 1 1 species of col- ubrids indicate an average survivorship of ca. 50% per year. The 21% annual mortality in C. c. mormon may be partitioned into overwinter- ing (7%), prereproductive, and postreproductive mortality. We suggest that prereproductive mor- tality is higher in females from exposure to ad- ditional risks associated with egg laying. A life table for C. c. mormon calculated using the combined female survival rate in 1970 and 1971 showed a net reproductive rate (R ) of 1 . 1 87, a value indicating an increasing population. Three-year-old females contributed the highest proportion (15.8%) to R . Compared to the life history of C. c. flaviven- tris in Kansas. C. c. mormon in Utah is distinct in the following ways: ( 1 ) lower growth rates and smaller adult size; (2) lower age-specific fecun- dity; (3) larger eggs and hatchlings; (4) higher clutch weight/female body weight ratio; (5) lower juvenile survivorship and higher adult survivor- ship; and (6) older age distribution and longer generation time. These life history traits appear to fit some "r" and some *'K" strategies in each population. Without more detailed work on re- source levels, environmental stability, and pre- dation. we caution against simplistic interpre- tations in contrasting the two populations. Acknowledgments At the University of Utah our studies were supported by American Museum of Natural His- VERTEBRATE ECOLOGY AND SYSTEMATICS 39 tory (Theodore Roosevelt Memorial Fund) grants, a Biomedical Sciences Support Grant FR- 070902, a Graduate Research Fellowship to Brown, and an NDEA Fellowship to Parker. At Skidmore College, Denton W. Crocker, Chair- man, Biology Department, provided an oppor- tunity for portions of this work to be completed. Eric J. Weller, Dean of the Faculty at Skidmore College, allocated financial support to allow Brown's participation in the 1980 annual her- petology meetings and provided funds for manu- script preparation through a Mellon Foundation Grant for faculty development. For assisting us in numerous ways in the lab and in the field in Utah, we thank George C. Douglass, Richard J. Douglass. Thomas C. Juelson, Arthur C. King. John M. Legler, Grady W. Towns, and Robert M. Winokur. We thank Paul E. Feaver, Henry S. Fitch, Harold Heatwole, Richard Shine, Fred- erick B. Turner, Stephen C. Stearns, and an anon- ymous reviewer for critically reading the manu- script and suggesting improvements. Brown appreciates the assistance and support of his wife Betsy and children Amy, Lee, and Bonnie, and Parker similarly thanks his wife Beth. Elaine C. Rubenstein photographed some of the figures and Edie Brown competently typed the manuscript. Literature Cited AUFFENBERG, W. 1955. A reconsideration of the racer. Coluber con- strictor, in eastern United States. Tulane Stud. Zool., 2:89-155. Blanchard, F. N., Gilreath, M. R. and Blanchard, F. C. 1979. The eastern ring-neck snake (Diadophis punctatus edwardsii) in northern Michigan (Reptilia, Serpentes, Colubridae). J. Herpe- tol., 13:377-402. Branson, B. A. and Baker, E. C 1974. An ecological study of the queen snake. Re- gina septemvittata (Say) in Kentucky. Tu- lane Stud. Zool. Bot., 18:153-171. Brown, W. S. 1973. Ecology of the racer. Coluber constrictor mormon (Serpentes, Colubridae). in a cold temperate desert in northern Utah. Ph.D. Thesis. Univ. Utah, Salt Lake City. 208 p. Brown, W. S. and Parker, W. S. 1976a. Movement ecology of Coluber constrictor near communal hibernacula. Copeia, 1976: 225-242. 1976b. A ventral scale clipping system for perma- nently marking snakes (Reptilia, Serpentes). J. Herpetol.. 10:247-249. 1982. Niche dimensions and resource partitioning in a Great Basin desert snake communitv. Pp. 59-81. In Scott. N. J.. Jr. (Ed.). Herpe- tological Communities. U.S. Fish and Wild- life Service. Wildl. Res. Rep. 13. 239 p. Carpenter, C. C. 1952. Comparative ecology of the common garter snake (Thamnophis s. sirtalis), the ribbon snake ( Thamnophis s. sauritus), and Butler's garter snake (Thamnophis butleri) in mixed populations. Ecol. Monogr., 22:235-258. Caughley, G. 1977. Analysis of vertebrate populations. John Wi- ley & Sons. N.Y. 234 p. Clark, D. R.. Jr. 1970. Ecological study of the worm snake. Car- phophis vermis (Kennicott). Univ. Kansas Publ. Mus. Nat. Hist.. 19:85-194. 1974. The western ribbon snake {Thamnophis proximus): ecology of a Texas population. Herpetologica, 30:372-379. Clark, D. R.. Jr. and Fleet, R. R. 1976. The rough earth snake (Virginia striatula): ecology of a Texas population. Southwest. Nat., 20:467-478. Con ant, R. 1975. A field guide to reptiles and amphibians of eastern and central North America. 2nd ed. Houghton Mifflin Co.. Boston. Deevev, E. S. 1947. Life tables for natural populations of ani- mals. Q. Rev. Biol., 22:283-314. Etheridge, R. E. 1952. The southern range of the racer Coluber con- strictor stejnegerianus (Cope), with remarks on the Guatemalan species Coluber orten- burgeri Stuart. Copeia, 1952:189-190. Feaver, P. E. 1977. The demography of a Michigan population of Natrix sipedon with discussions of ophid- ian growth and reproduction. Ph.D. Thesis. Univ. Michigan. Ann Arbor. 131 p. Fitch, H. S. 1949. Study of snake populations in central Cali- fornia. Am. Midi. Nat., 41:513-579. 1 960. Autecology of the copperhead. Univ. Kansas Publ. Mus. Nat. Hist.. 13:85-288. 1963. Natural history of the racer. Coluber con- strictor. Univ. Kansas Publ. Mus. Nat. Hist., 15:351-468. 1965. An ecological study of the garter snake. Thamnophis sirtalis. Univ. Kansas Publ. Mus. Nat. Hist.. 15:493-564. 1975. A demographic study of the ringneck snake (Diadophis punctatus) in Kansas. Univ. Kansas Mus. Nat. Hist. Misc. Publ.. 62:1— 53. Fitch, H. S., Brown, W. S. and Parker, W. S. 198 1. Coluber mormon, a species distinct from C. constrictor. Trans. Kansas Acad. Sci., 84: 1 96- 203. Gregory, P. T. 1977. Life-history parameters of the red-sided 40 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY garter snake ( Thamnophis sirtalis parietalis) in an extreme environment, the Interlake re- gion of Manitoba. Natl. Mus. Canada Publ. Zool.. 13:1-44. Hall, R. J. 1969. Ecological observations on Graham's wa- tersnake {Regina grahami Baird and Gi- rard). Am. Midi. Nat.. 81:156-163. Hirth, H. F. 1 966. Weight changes and mortality of three species of snakes during hibernation. Herpetologica, 22:8-12. Hirth, H. F. and King, A. C. 1 968. Biomass densitites of snakes in the cold des- ert of Utah. Herpetologica, 24:333-335. Hl'TC hinson, G. E. 1 978. An introduction to population ecology. Yale Univ. Press, New Haven, Conn. 260 p. Krebs, C. J. 1978. Ecology: the experimental analysis of distri- bution and abundance. 2nd ed. Harper and Row, New York. 678 p. Parker, W. S. and Brown, W. S. 1972. Telemetric study of movements and ovi- position of two female Masticophis t. tae- niatus. Copeia, 1972:892-895. 1 973. Species composition and population changes in two complexes of snake hibernacula in northern Utah. Herpetologica, 29:319-326. 1 974. Mortality and weight changes of Great Basin rattlesnakes (Crotalus viridis) at a hibernac- ulum in northern Utah. Herpetologica, 30: 234-239. 1 980. Comparative ecology of two colubrid snakes, Masticophis t. taeniatus and Pituophis mel- anoleucus deserticola, in northern LHah. Milwaukee Public Mus. Publ. Biol. Geol. No. 7:1-104. PlANKA, E. R. 1970. On r and K selection. Am. Nat., 104:592- 597. Platt, D. R. 1969. Natural history of the hognose snakes Het- erodon platyrhinos and Heterodon nasicus. Univ. Kansas Publ. Mus. Nat. Hist.. 1 8:253- 420. Prestt, I. 1971. An ecological study of the viper, Viperabe- rus, in southern Britain. J. Zool. (London), 164:373-418. Shine, R. 1 978. Sexual size dimorphism and male combat in snakes. Oecologia (Bed.), 33:269-277. Sokal, R. R. and Rohlf, F. J. 1969. Biometry: the principles and practice of sta- tistics in biological research. W. H. Freeman Co., San Francisco. 776 p. Spellerberg, I. F. and Phelps, T. E. 1977. Biology, general ecology and behaviour of the snake, Coronella austriaca Laurenti. Biol. J. Linn. Soc. (London), 9:133-164. Stearns, S. C. 1976. Life history tactics: a review of the ideas. Q. Rev. Biol., 51:3-47. Stebbins, R. C. 1966. A field guide to western reptiles and am- phibians. Houghton Mifflin Co.. Boston. 279 p. Stewart, G. R. 1968. Some observations on the natural history of two Oregon garter snakes (genus Thamno- phis). J. Herpetol., 2:71-86. Tinkle, D. W. 1957. Ecology, maturation, and reproduction of Thamnophis sauritus proximus. Ecology, 38: 69-77. 1 960. A population of Opheodrys aestivus (Reptil- ia: Squamata). Copeia, 1960:29-34. Turner, F. B. 1977. The dynamics of populations of squamates, crocodilians and rhynchocephalians. Pp. 157-264. In Gans, C, and Tinkle, D. W. (Eds.), Biology of the Reptilia, Vol. 7. Aca- demic Press, New York. Vial, J. L., Berger, T. J. and McWilliams, W. T., Jr. 1977. Quantitative demography of copperheads, Agkistrodon contortrix (Serpentes, Viperi- dae). Res. Popul. Ecol., 18:223-234. Viitanen, P. 1967. Hibernation and seasonal movements of the viper, Vipera berus berus (L.), in southern Finland. Ann. Zool. Fenn., 4:472-546. Vitt, L. J. and Congdon, J. D. 1978. Body shape, reproductive effort, and relative clutch mass in lizards: resolution of a para- dox. Am. Nat., 112:595-608. Wilbur, H. M., Tinkle, D. W. and Collins, J. P. 1974. Environmental certainty, trophic level, and resource availability in life history evolu- tion. Am. Nat., 108:805-817. Wilson, L. D. 1970. The racer Coluber constrictor (Serpentes: Colubridae) in Louisiana and eastern Texas. Texas J. Sci., 22:67-85. 1978. Coluber constrictor. Cat. Amer. Amphib. Rept., 218.1-218.4. Woodbury, A. M. 1951. Introduction— a ten year study. Pp. 4-14. //; Woodbury. A. M., et al., Symposium: A Snake Den in Tooele County, Utah. Her- petologica, 7:1-52. Woolf, C. M. 1 968. Principles of biometry. D. Van Nostrand Co., Inc. Princeton, N.J. 359 p. Vertebrate Ecology and Systematica— A Tribute to Henry S Fitch Edited by R. A. Seigel. L. E. Hunt. J. L. Knight. L. Malaret and N. L. Zuschlag i 1 984 Museum of Natural History. The University of Kansas. Lawrence Growth of Bullsnakes (Pituophis melanoleucus sayi) on a Sand Prairie in South Central Kansas Dwight R. Platt Introduction Growth rates of snakes in natural populations have been studied for fifty years. Blanchard and Finster (1933) presented limited data on growth rates of recaptured garter snakes and water snakes. KJauber (1937) derived a growth curve for the southern pacific rattlesnake (Crotalus viridis hel- leri = C. v. oreganus) by analyzing a collection of preserved specimens and pointed out that the growth of captive snakes may be distorted. Sei- bert and Hagen ( 1 947) presented growth data for the plains garter snake (Thamnophis radix) and smooth green snake (Opheodrys vernalis) from a mark-recapture study of populations in Illinois. Henry S. Fitch (1949) was a pioneer in the study of free-living snake populations with his field work in central California. His analysis of growth in the northern pacific rattlesnake (Cro- talus viridis oreganus) has been widely cited. He and his students have provided many reports on growth rates of snakes in natural populations (Clark 1970, 1974; Clark and Fleet 1976; Fitch 1960, 1963a, 1963b. 1965, 1975; Fitch and Fleet 1970; Platt 1969). Other notable studies on growth rates in free-living populations of colu- brid snakes include those by Brown and Parker (1984), Carpenter (1 952), Feaver (1977), Fukada (1959, 1960, 1972, 1978), Heyrend and Call (1951), Imler (1945) and Parker and Brown (1980). Growth has previously been studied in the bullsnake (Pituophis melanoleucus sayi) in Nebraska by Imler (1945), in the pacific gopher snake (P. m. catenifer) in California by Fitch (1949) and in the great basin gopher snake (P. m. deserticola) by Parker and Brown (1980). Growth of several species of elapid and viperid snakes has been studied, including studies by Gibbons (1972), Heyrend and Call (1951). KJauber ( 1 956), Prestt( 1971), Shine (1978, 1980). Volsoe (1944) and Wharton (1966). Most of these investigations have indicated a high degree of individual variability in growth rates. Two methods have been used to study growth of snakes in natural populations: 1 ) summarizing growth records from marked, released and re- captured individuals; and 2) determining size at different ages, usually up to one year old, by an analysis of size frequencies in population sam- ples. My study of growth rates in snakes was part of a larger study of the ecology and population dynamics of sympatric species of snakes on the Sand Prairie Natural History Reservation in western Harvey County in south central Kansas. The objectives of the present study were: 1) to investigate the range and patterns of variability in growth rates and the effects of prey availability and of age and sex on growth; 2) to compare the growth rates and strategies of five species of snakes living in the same general environment: 3) to compare the two methods (above) of determin- ing growth rates. This paper reports results based on 709 cap- tures of 471 bullsnakes (Pituophis melanoleucus sayi) and on nine young which were hatched in the laboratory. Subsequent papers will describe other species studied and comparative aspects of the study. Methods This study used standard mark-recapture tech- niques with snakes trapped alive in drift fence traps (Fitch 1951, 1960; Platt 1969). Thirty to fifty trapping stations were used on a study area of 80 acres (32.4 hectares) in 1966. 1967 and early 1968. From late 1968 through 1974. traps at 100 to 120 stations were in operation on the study area and at up to 20 stations on adjacent pastures. A trapping station consisted of a low metal drift fence with a funnel trap fitted under each end. These traps set without bait intercepted movement of the animals. Most measurements were made in the labo- ratory. Snout-vent length (SVL) and tail length were measured to the nearest one millimeter with the snake stretched along a metal tape until it relaxed. Weights to the nearest 0.1 gram were measured on a triple beam balance. The snakes 41 42 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY were released within three to four days at the site of capture. Marking was accomplished by clipping or branding subcaudals or ventrals so that each snake was individually recognizable. Individual vari- ations in color pattern and scutellation were also recorded so that almost all recaptured snakes were individually identified with certainty. Sex- ing was accomplished by probing through the vent for the hemipenial sacs and was checked later by body proportions. Food records were obtained by palping fecal matter from the intes- tine and forcing stomach contents back up the gullet into the mouth for identification and then repalping into the stomach. Previous studies of growth have used snout- vent lengths, total lengths and/or weights as a measure of size. I used snout-vent length (SVL) because it is one of the least variable measures of size. Weight is more affected by the stage of the feeding cycle or the reproductive cycle while total length is affected by partial loss or differ- ential growth of the tail. Only measurements of live snakes were used in the analysis. Growth rates from recapture records were cal- culated by averaging growth increments during the period between captures for samples of re- captured snakes. Recapture records were used to calculate growth rates only if three weeks or more had elapsed since the previous capture. Although bullsnake eggs probably hatch on the study area in August, young snakes were not caught in traps until September. First-year snakes were defined in this study as those caught between September of their hatching year and the end of the next August. Records of recaptured first-year and old- er bullsnakes were readily distinguished by plot- ting the SVLs of snakes with respect to capture date. Growth rates were also calculated from the changes in mean SVL of population samples of identified age. Frequency distributions were cal- culated for the lengths of all bullsnakes caught in each two-week interval throughout the trap- ping season in each year. First-year snakes were readily identified in these frequency distributions and they did not overlap samples of older snakes in size until they had completed their first full year of growth. Mean snout-vent lengths were calculated for these first-year snakes captured in each month and growth rates were calculated from the means of these monthly samples. Snakes were trapped from late April or early May to the end of October or early November over a nine year period, 1 966-1 974. Growth was not continuous throughout the year. On my study area it was usually most rapid in early summer but occurred throughout the period of activity and trapping; probably little or no growth oc- curred during dormancy. Therefore the mean growth rates in this study were calculated using the 184 days from the first of May to the end of October as the growth season. Absolute growth rates were calculated as growth increment in SVL per month (30 days) excluding the period from 1 November to 30 April. Rela- tive growth rates were calculated as the growth increment per month per 100 mm median SVL. The median SVL was defined as the midpoint between the lengths at two successive captures or between the mean lengths of two successive population samples. Some studies (Carpenter 1952; Fukada 1959, 1960, 1978) have used the initial length at first capture to calculate relative growth rates. The median length is more similar to the length of the snake during the growing period. Relative growth rates calculated from median lengths are less affected by the duration of the period between captures. Mean values in this paper are usually accom- panied by one standard error. Homogeneity of variances was tested by an F test. Differences in the central tendency of different samples were tested by Student's / test for samples having sim- ilar variances and by the Mann-Whitney U test when the variances were heterogeneous (Cox 1980). Regression equations of weight on length were calculated by Bartlett's method (Simpson et al. 1960). Rodent populations, principal prey of bull- snakes, were sampled by the same drift fence traps used to capture snakes and by 100 baited small mammal live traps (constructed like traps described by Fitch 1950) set in a grid 1 50 meters on a side. Drift fence traps were operated con- tinuously from May through October while bait- ed live traps were operated for a few nights per month through the summer (May-August). Ro- dents caught in drift fence traps were recorded as number caught per 100 trap station days (TSD) while those caught in baited traps were recorded as number caught per 100 trap nights (TN). Study Area The Sand Prairie Natural History Reservation is 80 acres (32.4 hectares) of prairie on sand dunes VERTEBRATE ECOLOGY AND SYSTEMATIC S 43 Tabii 1. Rodents trapped on the Sand Prairie Nat- ural History Reservation in Kansas. Dnl'i fence traps Hailed traps Year No cil nap station days i rsD) No. rodents 100 TSD Medium Small sized sized speeies species No rodents 100 TN Medium Small si/ed sized species species No. of trap niehls (TN) 1967 1968 1969 1970 1971 1972 1973 1974 7317 9989 19.775 17.076 19,962 15.272 17.014 17,908 0.2 3.2 0.3 0.2 0.1 0.3 0.1 0.2 6.4 9.8 1.1 1.9 0.4 4.5 0.8 1.1 2369 1 161 1430 1088 1061 1518 2134 1747 1.3 3.4 2.0 1.3 1.5 1.8 0.6 managed as a natural area. Prior to its acquisition by Bethel College in 1 965. it was used as a pasture but was never cultivated. All snakes used in my analyses were captured on this study area or on immediately adjacent pastures. The Sand Prairie Reservation is in a band of wind-blown sand deposits, the Hutchinson Dune Tract of the Great Bend Lowland physiographic division (Frye and Leonard 1952;Schoewe 1949). The upland grass communities on the reserva- tion are dominated by little bluestem (Andropo- gon scoparius). Forbs and other genera of grasses ( Triplasis, Aristida and Panicum) also occur. The unliooded lowlands have dense tall grass com- munities dominated by switchgrass (Panicum virgatum), sand bluestem (Andropogon hallii), indiangrass (Sorghastrum avenaceum). eastern gammagrass ( Tripsacum dactyloides) and prairie cordgrass (Spartina pectinata). Thickets of chickasaw plum (Primus august ifolia) are com- mon on the uplands and buttonbush (Cephalan- thus occidentalis) and black willow (Sa/ix nigra) in the lowlands. The area is poorly drained and its low depressions between sand dunes are rel- atively wet, having ponds, shallow marshes or dry ground depending upon the amount of recent rainfall. A more complete description of the study area can be found in Piatt (1973. 1975). Results Prey Populations. — Prey of bullsnakes on the study area were predominantly rodents. Trap- ping success (Table 1 ) provides a rough measure of the size and activity of rodent populations. Medium-sized rodents, prairie voles (Microtus T\nn 2. Proportions of bullsnakes (Pituophis mel- anoleucus) containing recoverable food items in the stomach or residues in the intestines. Chi square tests were run on the differences in proportions of snakes containing food in successive years. N = number of snakes examined. Year Summer (Ma>-Aug.) Per- centage contain ing loud x" \utumn (Sept. -i K i i 1966 1967 1968 1969 1970 1971 1972 1973 1974 46 43 31 55 65 15 42 44 61% 72% 52% 58% 26% 33% 52% 68% 2.43 7 54** 0.66 27.94** 1.72 6.85** 4.46* Per- centage contain- N ing food 22 95% 29 76% 29 93% 72 56% 59 29% 27 22% 25 72% 50 48% 4.14* 4.71* 156.68** 17.61** 1.58 30.01** 14.28** * Significant at the 0.05 level. ** Significant at the 0.01 level. ochrogaster) and cotton rats (Sigmodon hispi- dus), were caught most readily in baited traps. Trapping success for these species was highest in 1968, the only year of the study in which signif- icant numbers of cotton rats were caught. After 1968 trapping success for Microtus and Sigmo- don declined, reaching a low point in 1973 and recovering somewhat in 1 974. Small rodents were predominantly western harvest mice (Reithro- dontomys megalotis). There were significant numbers of woods mice (Peromyscus leucopus) and a few plains pocket mice (Perognathus fla- vescens). Success in trapping small rodents was high through 1968 and again in 1970 and 1972, but was lowest in 1971. Other prey such as small cottontail rabbits (Sylvilagus lloridanus) and birds or bird eggs were also occasionally taken by bull- snakes but did not constitute an appreciable part of the diet. The percentage of captured bullsnakes that had recoverable food items in the stomach or scats in the intestine is a measure of feeding success (Table 2). The data on both trapping success and feeding success (Tables 1 and 2) indicate that food was readily available from 1966 through 1968, with highest availability in 1968. The pe- riod of greatest food scarcity was from autumn 1970 through 1971. Data from rodent trapping suggest low rodent populations in 1973. but the feeding rate of bullsnakes was twice as high in summer 1973 as in 1971. Growth Rates of Bullsnakes. — No significant 44 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 3. Absolute and relative growth rates of bull- snakes {Pituophis melanoleucus) in Kansas ( 1 966-1 974) determined from recapture records. N = number of useable recapture records. Mean followed by ± 1 stan- dard error. Table 4. Absolute and relative growth rates of bull- snakes (Pituophis melanoleucus) older than one year in Kansas determined from recapture records. The prob- able ages are based on size (see text). N = number of useable recapture records. Means followed by ± 1 stan- N Mean absolute growth rate (mm/30 days) Mean relative growth rate (%) Age Probable age N Mean absolute growth rate (mm/30 days) Mean relative growth rate First year 0-3 mo. 8-12 mo. >One year 100 26 74 64 56.5 ± 2.9 34.2 ± 3.5 64.4 ± 3.3 17.9 ± 2.0 9.4 ± 0.5 7.3 ± 0.7 10.2 ± 0.5 1.9 ± 0.2 (%) Second year Third year > Three years 34 12 15 26.9 ± 3.16 12.2 ± 2.66 6.0 ± 1.03 3.1 ± 0.39 1.2 ± 0.26 0.5 ± 0.08 differences between the sexes were discovered in the size of large snakes or in growth rates. Large males (> 800 mm SVL) averaged 991 ± 13.3 mm SVL (N = 105) and large females averaged 1014+ 13.4 mm (N = 91). Neither the vari- ances nor the means of these samples were sig- nificantly different (F = 1.14, P > . 10; t = 1.23, P > .10). Among large snakes from each of the nine years, four annual samples of males had higher mean lengths and five annual samples of females had higher means. In 1974, females (x SVL = 1036 ± 30.8 mm, N = 14) were signifi- cantly longer (/ = 3.565, P < .005) than males (x SVL =886 ± 25.0 mm, N = 1 0) but no other between-sex size differences were statistically sig- nificant. Parker and Brown (1980) reported that the largest female gopher snakes (P. m. deserticola) were 108 and 1 14 cm SVL, but males commonly exceeded 1 20 cm SVL. In P. m. sayi in Nebraska, Imler ( 1 945) reported that males averaged slight- ly longer than females but a female was the larg- est specimen measured. In my study in Kansas, the largest specimens had a slight preponderance of males. Of 17 snakes 1200 mm or more SVL, nine were males (1200-1420 mm SVL) and eight were females (1210-1300 mm SVL). The mean absolute growth rate calculated from recapture records for first-year males was 55.4 ± 4.3 mm/month (N = 49) and for first-year fe- males was 57.6 ± 4.0 mm/month (N = 57). Nei- ther the variances nor the means were signifi- cantly different (F = 1.08, P > .05;/ = 0.38, P > .50). Males > 1-year had a mean growth rate of 16.0 ± 2.2 mm/month (N = 30) and females > 1 -year had a mean growth rate of 19.7 ± 3.2 mm/ month (N = 34). The variance was significantly higher in the female sample (F = 2.20, P < .05) but the median lengths were not significantly dif- ferent (JJ= 0.81, P > .05). There was no signif- icant difference in average growth rate between the two sexes so samples of the two sexes were pooled. Growth rates were calculated from the pooled sample of recapture records (Table 3). The mean absolute growth rate for the first (post-hatching) autumn season (0-3 months of age) was only 53% and the relative growth rate only 72% of those during the succeeding summer (8-12 months of age). The mean growth rates also indicated an abrupt decrease in both absolute and relative growth after the first year. Growth rates calcu- lated from the recapture records of snakes > 1 year at initial capture were only 32% as great for absolute growth and 20% as great for relative growth as those for first-year snakes (Table 3). The abruptness of this change is somewhat ac- centuated by pooling the rates for all older snakes. It was not possible to separate age groups older than one year with any certainty on the graphs of recapture records. However, size and growth rates (calculated for arbitrary size groupings) were used in assigning snakes to an older "probable" age. Large bullsnakes recaptured within an Au- gust to August year were assigned to one of three groups: 700-800 mm, 850-950 mm, and > 1000 mm SVL during the previous August. The first group was most probably in its second year of growth, the second in its third year, and the third older than three years. The growth rates calcu- lated for these "probable" age groups from re- capture records are in Table 4. Mean snout-vent lengths were calculated for each monthly sample of first-year snakes in each year whenever sample size was greater than four (Table 5). Absolute and relative growth rates cal- VERTEBRATE ECOLOGY AND SYSTEMATICS 45 Table 5. Snout-vent lengths of bullsnakes (Pituophis melanoleucus) in successive months during their first year in Kansas. Young snakes were first caught in September and were dormant between October and May. Means were calculated only for sample sizes >4. Mean is followed by ± 1 standard error. Sample size is listed under the mean. The range of lengths is listed in parentheses. The year 1971-1972 was omitted because few first-year snakes were caught. Snout-venl lengths Year Sept. Oct. May June July Aug. 1966-1967 441 ± 13.0 10 (386-495) 453 ± 12.4 11 (365-506) 1 (638) 652 ± 10.8 18 (593-740) 747 ± 24.0 17 (607-860) 805 ± 17.6 10 (714-885) 1967-1968 429 ± 9.5 9 (404-465) 464 ±11.5 16 (406-566) 2 (503-529) 621 ± 16.8 15 (520-682) 3 (725-793) 3 (832-895) 1968-1969 1 (441) 488 ± 5.9 16 (375-560) 608 ± 16.0 7 (522-655) 3 (702-801) 842 ± 26.0 6 (808-911) 1969-1970 447 ± 4.4 44 (387-492) 470 ± 6.7 23 (425-545) 546 ± 8.5 8 (515-555) 650 ± 9.0 9 (612-710) 734 ± 10.8 11 (672-800) 4 (720-855) 1970-1971 439 ± 6.0 33 (363-508) 460 ± 6.6 18 (422-532) 528 ± 8.2 12 (474-565) 587 ± 7.9 15 (543-635) 3 (620-680) 714 ± 18.6 9 (660-822) 1972-1973 448 ± 12.5 11 (399-523) 447 ± 14.7 8 (414-542) 555 ± 15.5 10 (496-622) 650 ± 20.2 8 (547-719) 725 ± 14.4 6 (657-752) 1 (772) 1973-1974 445 ± 6.0 20 (392-495) 456 ± 7.0 29 (375-511) 515 ± 18.9 10 (397-587) 673 ± 23.9 5 (598-737) 749 ± 12.9 13 (662-816) 2 (777-814) Total sample 443 ± 2.8 128 464 ± 3.3 121 537 ± 8.2 43 630 ± 6.1 77 736 ± 14.3 49 788 ± 11.9 35 culated from the mean SVL and mean capture dates of the monthly samples are presented in Table 6. The composite sample from all years showed an absolute growth rate in the first autumn that was only 1 9% of the growth rate the succeeding summer (Table 6), an even greater difference than that shown by growth rates calculated from re- capture records. To further analyze growth, the juvenile snakes caught each autumn were grouped into two-week samples. Samples with five or more individuals were divided into thirds according to SVL. The mean SVL of the upper and lower thirds, the maximum and minimum SVL and the proportion of snakes with food in the stom- ach or scats in the intestines were calculated for each sample (Table 7). In many years there were some individuals that grew slowly and remained between 360 and 460 mm SVL in October. These slowly growing snakes were one factor in the low growth rates calculated for juvenile snakes in the first autumn. Mean growth rates were calculated for each year from recapture records of first-year and old- er snakes (Table 8). The mean growth rate for first-year snakes in 1971 was significantly lower than that for each of the other years (comparison of 1971 to 1967, / = 3.67, P < .005; 1971 to 1968, t = 6.16, P < .005; 1971 to 1970, t = 3.39. P < .01; 1971 to 1973, / = 3.56. P < .005; 1971 to 1974, / = 6.70, P < .005). Other differences in mean growth rates of first-year snakes in Table 8 were not statistically significant with the ex- ception that the growth rate for 1974 was sig- nificantly higher than that for 1967 (? = 2.16, P < .05). The differences in mean growth rates of older snakes were not significant because of the small sample sizes, but the between-year dif- ferences were similar to those of first-year snakes. The growth rates calculated from recapture rec- 46 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Tabi i 6. Absolute and relative growth rates of bullsnakes (Pituophis melanoleucus) during their first year in Kansas, determined from the data in Table 5. Values in parentheses are based on mean snout-vent lengths of small samples (<5). Absolute (mm growth rates month) Relative (mm/month growth rates /100 mm SVL) Year Autumn Sp-Oc Oc-May Summer May-Au 1st yr. S ii- Au Autumn Sp-( )c Oc-May Summer May-Au 1st yr. Sp-Au 1966-1967 18.0 — 72. 8 a 72.8 4.0 — 10.0 a 11.7 1967-1968 42.0 — (103.3) (81.1) 9.4 — (15.1) (12.7) 1968-1969 — — 121.0 a 88. 3 b — — 16.7- 13.3" 1969-1970 30.0 60.0 97.2' 72.4 C 6.6 11.8 15.2' 12.3' 1970-1971 26.2 49.8 65.6 55.0 5.8 10.1 10.6 9.5 1972-1973 90.0 86.4' 7 1 .0' 18.0 13.5' 12.1' 1973-1974 14.3 63.2 113.2 C 80.7' 3.2 13.0 17.9' 13.5' Total sample 25.2 62.6 86.6 70.4 5.6 12.5 13.1 11.4 ■ Because of inadequate samples in May, growth rates were calculated for the period June-August. b Because of inadequate samples in September, growth rates were calculated for the period October-August. ' Because of inadequate samples in August, growth rates were calculated for the period May-July or September- July. ords (Table 8) were much lower than those es- timated from the mean SVL of samples of first- year snakes (Table 6) but they varied in parallel fashion; the lowest growth rates by both methods of estimation occurred in 1971 and the highest in 1968, 1969, and 1974. Growth in weight is proportional to growth in length but more variable. Klauber (1956) found that the regression line of weight on length for rattlesnakes was of the form W = CL'\ where W is weight, L is length and C and P are constants characteristic for each population. Regression lines of this form were fitted to the data on weight in grams and SVL in meters of bullsnakes. The constants C and P for different samples of snakes and the projected weight at .8 meter SVL (ap- proximate SVL at one year of age) are presented in Table 9. From these regression equations a normal weight can be calculated for any SVL. Discussion Growth rates calculated from recaptures of first- year snakes (mean absolute growth rate was 56.5 mm/month; see Table 3) were consistently lower than growth rates calculated from changes in the mean length of monthly samples (mean absolute growth rate was 70.4 mm/month; see Table 6). The recapture records from the autumn of the hatching year may include a disproportionate number of small snakes. With the autumn rec- ords eliminated, the mean growth rate of recap- tured first year snakes was 64.4 mm/month, still somewhat lower than that calculated from mean lengths of monthly samples. This discrepancy may be partially due to a temporary decrease in growth after the capture experience. Fitch ( 1 949, 1975) and Clark ( 1 970) mentioned that recovery of normal growth after capture may take more than a month in rattlesnakes (Crotalus viridis), ringneck snakes (Diadophis punctatus) and worm snakes (Carphophis vermis). On the other hand, Fukada (1959) and Carpenter (1952) found no effect of capture on further growth of Natrix tigrina and garter snakes (Thamnophis). My re- sults indicated that capture did cause a short- term decrease in growth in many first-year bull- snakes, primarily limited to the first month after capture and to the smaller snakes (Table 10). Decreased growth following capture would prob- ably have a lesser effect on the growth rates of snakes > 1-year because most of the captures oc- curred over longer intervals. The decreased growth following capture was not due to depri- vation of food by interruption of feeding since the weights of recaptured male bullsnakes were not below normal (Table 1 1). Females were not tested because their weights were more variable in an annual cycle. Another cause of discrepancy in growth rates calculated by the two methods was differential mortality of less successful and more slowly growing snakes. Growth rates calculated from recaptures are based on a sample of all snakes in the population during the time interval, possibly biased in favor of those unsuccessful in feeding VERTEBRATE ECOLOGY AND SYSTEMATICS 47 Table 7. Snout-vent lengths and success of feeding for juvenile bullsnakes caught in half-month intervals in their first post-hatching autumn (September and October) in Kansas. From samples arranged in order of decreasing SVLs, mean and extreme SVLs are listed for the approximate upper and lower one-thirds of samples >4. Sample sizes for A and B are listed in C. The year 1971 was omitted because few juveniles were caught. Mean SVL Extreme SVL Half-moruh inters als Half-month intervals Year S 1-15 S 16-30 O 1-15 O 16-30 S 1-15 S 16-30 O 1-15 O 16-30 A. Lower l h of sample 1966 — 392 394 — 1967 — 402 409 432 1968 — — 440 458 1969 415 412 436 458 1970 405 415 440 — 1972 — 404 417 — 1973 — 414 396 448 B. Upper % of sample 1966 — 495 499 — 1967 — 450 504 530 1968 — — 515 515 1969 475 482 496 515 1970 454 485 490 — 1972 — 490 451 — 1973 — 471 479 501 394 386 365 435 — 389 406 414 — — 375 458 387 392 425 442 385 363 401 422 — 399 414 — — 392 375 408 472 495 506 500 — 466 512 566 — — 526 519 491 496 545 523 472 508 532 452 — 523 461 — — 495 510 511 C. Percent with food in digestive tract Half-month intervals s : 1-15 s 16-30 o 1-15 o 16-30 Year N % N % N % N "m 1966 3 100% 7 86% 7 100% 4 75% 1967 — 8 75% 7 57% 9 89% 1968 — — 12 92% 6 100% 1969 17 82% 27 26% 17 76% 6 33% 1970 14 14% 19 22% 14 43% 4 25% 1972 — 11 73% 5 60% — 1973 — 19 26% 15 47% 14 79% if they were more active in searching for food. Growth rates calculated from the mean lengths of snakes at identified ages are based on those snakes that survive over the growth interval. The absolute growth rate of first-year bullsnakes (70.4 mm/month) estimated from the mean lengths of snakes in September when first caught and in the following August was a good estimate of the growth of those first-year snakes that survive to one year of age. The absolute growth rate (56.5 mm/month) calculated from recapture records of first-year snakes is an estimate of growth in both survivors and non-survivors somewhat biased by decreased growth following capture. Differential mortality can also distort growth rates calculated from mean lengths in successive monthly samples and probably accentuated the differences in growth rates in the autumn and the following summer as estimated by both methods. Growth rates calculated by the two methods should be comparable for populations where growth is little affected by capture and there is little differential mortality based on size. There- fore growth rates calculated from recapture rec- ords were probably good measures of growth for older bullsnakes but less reliable for first-year snakes. Individual variability in growth of bullsnakes is high. The extreme SVLs in monthly samples of first-year snakes (Table 5) showed that indi- vidual snakes grew much slower and much faster than the mean growth rate. Although there was probably high mortality among the slow-growing snakes in the first year, some one year old snakes were <750 mm SVL while others were >850 mm SVL. 48 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY TABLE 8. Absolute growth rates ofbullsnak.es (Pitu- ophis melanoleucus) trapped more than once during a year in Kansas. N = number of useable recapture rec- ords. 1969 and 1972 were omitted because of the small number of recaptures. Means followed by ± 1 standard error. Age: first year summer only Age: >one year Year N Mean N Mean 1967 14 61.3 ± 6.5 — 1968 12 81.0 ± 7.3 6 31.8 ± 11.3 1970 10 66.2 ± 10.2 5 22.0 ±11.4 1971 16 35.1 ± 3.7 6 11.5 ± 4.4 1973 6 67.9 ± 11.7 — 1974 19 77.2 ± 5.0 6 15.6 ± 5.8 Nine young bullsnakes hatched in the labo- ratory on 1 9 August had a mean SVL at hatching of 363 ± 1.6 mm (415 mm total length) and a mean weight of 18.3 ± 0.39 gm. This is larger than the 334.4 mm SVL reported for hatchlings of the gopher snake (P. m. deserticola) (Parker and Brown 1980) and the 15 in. (381 mm) total length reported by Imler ( 1 945) for P. m. sayi in Nebraska. No information was obtained about the variability in size of hatchlings on my study area. Hatching dates on the study area are not known but young snakes were first captured in traps in September. Young snakes averaged 443 mm SVL in September (Table 5). Assuming that hatching occurred in mid-August, the growth rate for the first month would have been ca. 70 mm. On the basis of days in the period 1 May to 3 1 October, the growth rate calculated from the mean lengths of the October and May samples of first-year snakes was 62.6 mm/month (Table 6). These growth rates are more than twice as high as the growth rates calculated for autumn both from recapture records (Table 3) and from mean lengths of samples (Table 6). Growth is probably rapid in the first month after hatching but the estimate of the growth rate would be lower if the above estimated mean hatchling size was low and/or if hatching oc- curred earlier. The higher growth rate from Oc- tober to May could be the result of growth oc- curring during the dormant period. Snakes captured in their first autumn and subsequently caught the next summer grew 68.6 ± 4.28 mm/ month (N = 3 1 ). This is significantly higher (t = 2.62, P < .01) than the mean growth rate of 55.6 mm/month (N = 100) for all recaptured first-year snakes, but it is not significantly higher than the growth rate of 64.4 mm/month (N = 74) of first- year snakes during the summer. Although some growth may have occurred between the first of November and the first of May, evidence of any significant growth during this period is lacking. The higher growth rates calculated from the mean lengths of first-year snakes in October and May were probably caused by high mortality of the smaller poorly nourished snakes during the win- ter. In three of the four years with adequate sam- ples in both October and May, the difference between the minimum SVLs in October and May was greater than the difference between the max- imums (Table 5). Imler (1945) estimated that juvenile bulls- nakes grew approximately four inches ( 1 00 mm) in September and October following hatching. Table 9. Regression of weight in grams on SVL in meters of yearly samples of bullsnakes (Pituophis melano- leucus) in Kansas. The regression equation is of the form W = CL P . Mi lie Femaie N Regression constants Normal weight gm/0.8 m N Regression constants Normal weight Year C P C p gm/0.8 m 1967 38 247.7 2.7 135.6 38 274.5 2.8 147.0 1968 29 284.0 2.8 152.1 53 272.9 2.6 152.3 1969 56 275.7 3.1 138.0 49 271.4 3.1 135.9 1970 76 268.9 2.8 144.0 53 267.9 2.9 140.3 1971 63 240.4 2.7 131.6 24 230.7 2.6 129.2 1972 20 270.4 2.9 141.6 18 247.8 2.8 132.7 1973 48 261.3 2.8 140.0 44 290.5 3.0 148.8 1974 36 269.9 2.7 147.8 35 286.9 3.0 146.9 All years 384 265.6 2.8 142.2 325 274.1 2.9 143.5 VERTEBRATE ECOLOGY AND SYSTEMATICS 49 Table 10. Absolute growth rates (GR) of recaptured bullsnakes (Pituophis melanoleucus) in Kansas after different lengths of intervals between captures. N = useable recapture records. Means followed by ± 1 standard error. Interval be iween captures SVLat <30 days 30 to 60 days > 60 days capture, mm N Mean GR (mm month) N Mean GR (mm month) N Mean GR (mm month) <500 23 22.8 ± 5.6 16 53.0 ± 7.4 33 65.6 ± 4.0 500-599 11 28.6 ± 11.1 6 51.0 ± 8.1 10 52.7 ± 8.6 600-699 8 44.3 ± 10.7 14 53.2 ± 7.3 12 48.4 ± 5.7 700-799 14 27.0 ± 7.2 6 63.4 ± 17.2 12 39.9 ± 5.7 Parker and Brown (1980) reported no growth in juvenile gopher snakes (P. m. deserticola) in the first autumn. My results indicated that growth of juvenile snakes in autumn was variable. Some grew rapidly becoming 500-550 mm SVL in Oc- tober, or even in September in good years. Others grew more slowly and were 440 mm or less SVL in early October (Table 7). Growth of all juve- niles slowed in October and they added extra weight in good years. In 1968, the 16 juveniles caught in October averaged 41 gm, 20 per cent higher than normal weight (weight calculated from regression equation of weight on length). The low growth rates in autumn compared to summer for first-year bullsnakes (Tables 3 and 6) were partially due to high mortality of slow- growing snakes in late autumn and winter and partially due to an acceleration of growth in the summer. The most rapid growth of first-year bullsnakes was in June and July; growth declined toward late summer (Table 5). Imler (1945) reported growth of 14 in. (356 mm) during the first year for bullsnakes in Ne- braska, from an average length of 15 in. (381 mm) in September soon after hatching, to an average length of 29.5 in. (749 mm) at one year of age. These are probably total lengths. Parker and Brown (1980) indicated growth of ca. 280 mm for the first 1 3 months in gopher snakes (P. m. deserticola) in northern Utah. First-year bullsnakes on the Sand Prairie Reservation in Kansas had somewhat higher growth rates (422 mm/year) than these other two populations of the species. The absolute growth rate declined to approximately 40% of the first year rate in the second year, to 20% in the third year and to 10% or less after the third year (Table 4). The growth rate probably continued to decline in older snakes: individuals more than 1 100 mm SVL grew 5.2 mm/month (N = 9) and those more than 1200 mm SVL grew 0.7 mm/month (N = 3). Relative growth rates declined even more rapidly. The age at sexual maturity was not determined in this study. Growth rates showed a pattern of annual vari- ability which was related to prey availability. The highest rates occurred in the summers of 1968. 1 969, and 1974. The lowest rates occurred in the summer of 1971 (Tables 6 and 8). In both 1968 and 1969 all one-year old bullsnakes in samples collected in August were >800 mm SVL. In Au- gust 1971, some one-year old snakes had grown comparably to >800 mm SVL but others were <700 mm SVL (Table 5). These periods of high and low growth rates corresponded with the pe- riods of highest and lowest prey availability and feeding activity (Tables 1 and 2). The decrease in growth rates in 1971 was sta- tistically significant. Scarcity of prey in that year, when populations of both small and medium- sized rodents were low, adversely affected growth of bullsnakes. Periods of exceptional prey abun- dance had a less evident effect on growth of bull- snakes. Growth rates in 1968 were not appreci- Table 11. Comparisons of mean weights of male bullsnakes (Pituophis melanoleucus) recaptured within one growing season in Kansas to normal weights cal- culated from the regression equation of weight on SVL in samples with different intervals between captures. N = number of useable recapture records. Interval Mean Mean Normal between SVL weight weight captures N (mm) (gm) (gm) <21 days 19 634 79.4 74.2 21-30 days 13 649 81.5 79.1 31-60 days 16 730 114.2 110.0 > 60 days 9 835 168.0 160.3 50 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY II00-- I000-- ]> 900-- — 800+ to UJ _J > CO 700 600 500-- 400-- I f I A S FIRST ■i — I — I — I — I — t- -I— I — I — I — h S -H 1 1 1 1 H- M J J A S FOURTH MJJASO MJJA SECOND THIRD GROWING SEASON Fig. 1. Growth curve for bullsnakes (Pituophis melanoleucus sayi) in south central Kansas. Tick marks on the abcissa indicate the mid-point of each month. Vertical dashed lines represent the dormant period of six months. The solid growth line estimates the average growth of bullsnakes at different ages in nine years (1966-1974). Hollow circles represent mean lengths of snakes in monthly samples of first-year snakes (see text for age designations). Dashed lines represent mean growth rates in a poor year (1970-1971) and good years ( 1 968— 1969). ably higher than in 1969 and 1974 when rodent trapping was much less successful. Growth would not be expected to increase in proportion to the increase in prey populations if the snakes were already satiated. Also, at high feeding levels, a large proportion of the increased food intake probably goes into fat reserves rather than into increased growth in length. In 1968, weights of bullsnakes for any given length were higher than in any other year (Table 9). In 1969, although growth rates were relatively high (Table 6), feed- ing activity was much lower (Table 2) and weights were relatively low (Table 9). Young bullsnakes that accumulated extra fat reserves in the autumn of 1968 were able to maintain high growth rates in 1969 when rodent populations were probably only moderately high. The year 1974 had higher feeding activity (Table 2) and relatively high growth rates and weights (Tables 6, 8 and 9). Growth rates of juveniles in the autumn were variable. Low production, low survival and/or low growth of juveniles occurred in the autumn in 1970, 1971, and 1972. Feeding rates were also generally low (Table 7). A growth curve for bullsnakes for the years studied is presented in Fig. 1. The first year's growth was determined from the mean lengths of monthly samples while the growth curve be- VERTEBRATE ECOLOGY AND SYSTEMATICA 51 yond the first year was estimated from the growth rates of recaptured snakes. The lower dashed line represents the average growth in 1970-1971. a poor growth year, and the upper dashed line rep- resents growth determined from the combined samples for 1967-1968 and 1968-1969. good years. The data were inadequate to estimate the monthly variation in growth for snakes older than one year so the growth rate is applied uniformly through the active season. This growth curve in- dicates that bullsnakes were approximately 790 mm SVL at one year of age in August. 950 mm SVL at two years and 1030 mm SVL at three years. These estimates may be slightly low for the second and third years, since growth was probably more rapid before August than after August. If older snakes grew at the rate of 35- 40 mm per year (Table 4). a bullsnake 1 100 mm SVL would be five years old and one 1200 mm SVL would be seven to eight years old. This is similar to the estimate by Imler ( 1 945) that bull- snakes in Nebraska reach a total length of 49 inches (1245 mm; ca. 1100 mm SVL) in five to six years. Parker and Brown (1980) found that gopher snakes (P. m. desert kola) in Utah re- quired 1 8-20 years to reach 1 200 mm SVL. Fitch ( 1 949) estimated that P. m. catenifer in central California reached a SVL of more than 800 mm at two years of age. Growth of bullsnakes in cen- tral Kansas was comparable to that of bullsnakes in Nebraska but more rapid than populations of the same species in Utah and California. Different studies of growth rates in snakes can- not be precisely compared because of differences in methods used. But comparisons can be made on the basis of the general magnitude of growth in the first year of life. Information for some colubnd species on growth during the first year are presented in Table 12. The species are ar- ranged in order of increasing hatchling size. Growth rate of snakes in their first year is pos- itively related to both size of hatchlings and to normal adult size (Table 12). Hatchling size was either listed in the reports cited or was deter- mined from the growth curves reported and was rounded to the nearest five mm. Normal adult size was usually determined from length distri- butions as the mode of the snakes forming the largest size group in a population sample. The SVL was rounded to the closest 10 mm in small snakes, closest 50 mm in moderate-sized snakes and closest 100 mm in large snakes. If a fre- quency distribution was not included in a growth study, normal adult si/e was taken from the growth curve or from the author's statements about adult size. Other biological parameters modify the rela- tion between first-year growth rate and size: a) Geographic variation in growth rate probably was mediated through environmental limi- tation. The studies of growth of Thamnophis and Nerodia in Michigan with a short growing season reported lower first-year growth than studies of related forms in Kansas. The growth rates of Coluber constrictor mormon and Pi- tuophis melanoleucus deserticola in Utah, where ecosystem primary production is low- er, were much lower than those of related sub- species in Kansas. b) Taxonomic differences in growth are evident. Species of the genus Thamnophis have high growth rates in proportion to hatchling and adult size while the species of Elaphe and Lampropeltis studied appear to have low rel- ative growth rates. c) There is a relationship between first-year growth rate and age at sexual maturity. Fe- males of most of the species of moderate- sized snakes studied become sexually mature at one year of age. First-year growth amounts to at least half of the growth from hatchling to adult size. Females of large snakes, such as Elaphe, Pituophis, Coluber and Masticophis, usually take more than one year to mature and growth rates are not as high relative to size. Females of moderate-sized species with low growth rates, Lampropeltis triangulum in Kansas and Nerodia sipedon in Michigan, take two to three years to mature. Coluber constrictor mormon in Utah with lower growth rates is both smaller as an adult and takes longer to mature than the faster growing sub- species from Kansas. The smallest species of Elaphe studied. E. quadrivirgata, matures in one year but the larger species, E. climaco- phora and E. obsoleta, mature in three years and have relatively slower growth rates. First-year growth in viperid snakes ranged from 70-370 mm (10-45 mm/month) (Fitch 1949. 1 960; Gibbons 1 972: Klauber 1956: Prestt 1971; Volsoe 1944; Wharton 1966). Growth rates in elapid snakes in Australia up to 12 months of age ranged from 70 to 4 1 mm in a growth season 52 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 12. Growth increments (mm SVL) for colubrid snakes during the first year. Sex symbols are used to designate growth increments when authors reported different growth rates for males and females. An X designates growth rates of a pooled sample of the two sexes. The two numbers following each symbol are SVL of hatchling Millimeters growth in first year Species 50-100 100-150 150-200 200-250 I 'irginia striatula 2 110- -230 (8; 1) Carphophis vermis 9 i io- ns- -230(7; -260 2) Diadophis p. arnyi <5 9 1 lO- ll 5- -260(5; -290 2) Heterodon nasicus $ 150- -400 (6; 1) 9 150-450 Natrix tigrina* Thamnophis s. parietal is 6 170-550(6; 1) Nerodia sipedon <3 190-550 (5; 2) 9 190-750 Heterodon platyrhinos 3 195-500(6; 1) Thamnophis sauritus 5 200-450(5; 1) 9 200-550 Thamnophis s. sirtalis S 200-450(5; 1) 9 200-550 Lampropeltis triangulum X 210-600 (6; 2-3) Thamnophis proximus Coluber c. flaviventris Coluber c. mormon 6 225-550 (5; 3) 9 225-600 Elaphe quadrivirgata* 9 230-700 Masticophis taeniatus Elaphe climacophora* X 320-1600(6; 3) Elaphe obsoleta X 330-1200 (5; 3) Pituophis m. deserticola Pituophis m. sayi* Pituophis m. sayi * Values calculated by correction from data on total lengths. VERTEBRATE ECOLOGY AND SYSTEMATICS 53 Table 12. Continued. and normal adult SVL (see text). Listed in parentheses are the number of active or growing months in the year and the age in years of females at sexual maturity. The geographic location of the population studied is listed. Millimeters growth in first year 250-300 300-350 350^*00 400- -450 Authority Texas- Clark & Fleet 1976 Kansas— Clark 1970 Kansas— Fitch 1975 Kansas— Piatt 1969 S 165-500(6; 1) 9 175-600 Japan — Fukada 1959 9 170-650 Kansas— Fitch 1965 Michigan — Feaver 1977 S 195-600 Kansas— Piatt 1969 Michigan — Carpenter 1952 Michigan — Carpenter 1952 Kansas— Fitch & Fleet 1970 9 215-550(10; 1) Texas- Clark 1974 3 215- 9 215- -800 (6; 2) •900 Kansas— Fitch 1963b Utah- Brown & Parker 1984 S 240-850(6; 1) Japan — Fukada 1960 3 285-1000 (5; 3) 9 285-1000 Utah- Parker & Brown 1980 Japan — Fukada 1978 Kansas- Fitch 1963a 610 m Parker and Brown 1974 Preston 1964 males, non-gravid females. juveniles: 650 m 210-1600 m Macartney (pers. comm.) gravid females: 50 m 3-379 m Crotalus horridus males: 1400 m ?(1 only) Brown et al. 1982 females: 280 m 191-425 m Sistrurus catenatus <200 m Reinert and Kodrich 1982 munally in cavities in rock outcrops, sometimes associated with talus slopes (personal observa- tion); similar observations have been made for rattlesnakes elsewhere (Klauber 1972). At one site in the Chilcotin-Cariboo area of British Co- lumbia, the garter snakes Thamnophis elegans and T. sirtalis use a large cavity-riddled rocky mound which rises up from the surrounding grassland (Gregory, unpublished). Manitoba T. sirtalis hibernate in large sinkholes which are the result of slumps in the bedrock into subterranean cavities (Gregory 1 977a). Some snakes, however, den communally in impermanent structures such as prairie dog burrows (Klauber, 1972) or aban- doned ant mounds (Carpenter 1953; Lang 1971), but these dens usually have a lifetime of several years before they are no longer usable. Com- munal den sites are usually traditional in that they are occupied by snakes every year. They also frequently face south; thus, annual exposure to solar radiation is maximized. For other descriptions of communal hiber- nating sites of snakes, see reviews in Klauber (1972) and Parker and Brown (1973). VERTEBRATE ECOLOGY AND SYSTEMATICA 59 2. Relationship of den to summer habitat.— Snakes which hibernate singly or in very small groups may use sites within the summer range (Fitch and Glading 1 947; Naulleau 1 966). Com- munal dens of snakes, however, are frequently separated by fairly long distances from the sum- mer range, necessitating an annual migration back and forth between the two. Distance travelled ranges from a few hundred m to several km (Ta- ble 1): perhaps length of migration is inversely related to the availability of sites suitable for housing snakes in winter, but this idea has not been tested. In Coluber constrictor, mean dis- persal distance may be correlated with popula- tion density; presumably, in years with high numbers, individuals which disperse farther es- cape intraspecific competition for resources (Brown and Parker 1976). Some snakes which move between discrete denning and summer areas show a highly direc- tional form of dispersal (Gregory and Stewart 1975). while others do not (Parker 1976). Indi- vidual snakes often return to the same den or denning area year after year; measures of den fidelity of snakes in successive years are often in the 90-100% range (Fitch 1960; Viitanen 1967; Lang 1971; Brown and Parker 1976; Gregory 1977a. 1982; Parker and Brown 1980). Other authors have concluded that den fidelity is low (Noble and Clausen 1936); however, the defi- nition of what constitutes a den or denning area varies from study to study so that results are not necessarily comparable. In addition, distance be- tween neighboring dens may affect fidelity but is not always reported; Lang (1971) found lower fidelity and greater annual interchange between dens that were closer together. Nevertheless, a remarkable ability to home is shown by some species: Homing to specific den complexes less than 1000 m apart is almost 100% in Coluber constrictor \n Utah (Brown 1973; Brown and Par- ker 1976) and homing Thamnophis sirtalis in Manitoba must apparently pass close by other dens en route to their own dens each fall (Gregory and Stewart 1975). Other similar examples are given by Viitanen ( 1 967) and Lang (1971). The exact mechanisms used in homing are not known, but there is presumably selective value in returning to a den in which overwintering has been successful previously, even where other hi- bernating sites are abundant. This is important since high rates of mortality during hibernation have been reported for many snakes, especially the young (Bailey 1948; Carpenter 1953; Hirth 1 966a; Viitanen 1 967: Lang 1 969, 1971; Gregory 1977a, 1982; Parker and Brown 1980). Parker and Brown (1974, 1980) suggest that high mor- tality figures in such studies may be an artifact of handling and marking snakes, but the evidence in support of this contention is minimal. In cases where individuals generally return to the same den in successive years and where mat- ing usually occurs at the den site (see below), communal denning produces a large departure from panmixia. Over large areas, isolation of dif- ferent denning sites might ultimately be an im- portant contributing factor to differentiation within species (Gannon 1978). At the local level, however, populations at particular dens are probably never completely isolated demes. Den fidelity is rarely 100% so that some interchange occurs between dens. In addition, even in species which normally mate at the den. occasional mat- ing occurs away from the den when individuals from different hibernacula may come into con- tact (Gregory 1977a). Inter-den mating has been observed in some cases (Brown 1973; Brown and Parker 1976). Finally, there is no particular rea- son to believe that young snakes hibernating at a communal den for the first time necessarily use the same den as their parents, except when the young are born at the den. 3. Size and structure of denning popula- tions.— Sizes of overwintering aggregations of snakes have been reviewed by Klauber (1972). Parker and Brown (1973). and Gregory (1982). Most aggregations probably consist of much few- er than 100 individuals of all species combined. Some denning populations, however, may in- clude a few to several hundred individuals of a given species (Criddle 1 937; Viitanen 1 967; Lang 1969; Klauber 1972; Parker 1976). The largest denning populations known are those of Tham- nophis sirtalis in Manitoba, where numbers at one den fluctuated between about 4000 and 8000 in a four-year period (Gregory 1977a). Sampling the different size age groups in a snake population in proportion to their relative abundance is difficult because young snakes are smaller and often more secretive than adults. Nevertheless, it seems clear that young-of-year and/or juveniles are frequently absent (Viitanen 1967; Gregory 1977a: Sexton and Hunt 1980) or greatly underrepresented (Hirth et al. 1969; 60 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Klauber 1972; Parker and Brown 1973, 1980; Parker 1976; Brown and Parker 1976) at com- munal hibernacula. Prestt (1971), however, in contrast to Viitanen's ( 1 967) observations of the same species ( Vipera berus), found young hiber- nating with the adults. In some small species, the young also apparently hibernate with the adults (Noble and Clausen 1936; Lang 1971). Why young snakes often do not use the same dens as the adults is puzzling. Perhaps whether or not they do depends to some extent on the distance between the den and the summer hab- itat. If the young are born in the summer range a long way from the den, it may simply be too expensive energetically to make the journey if they can find suitable hibernacula closer by. This seems quite likely as smaller snakes can use sites which are inaccessible to the adults because of their size. This is probably the case for Tham- nophis sirtalis in Manitoba, where the adults, but no young-of-year, hibernate in limestone sinks several kilometres from the summer habitat (Gregory and Stewart 1975; Gregory 1977a). Young of this species are known to hibernate communally with two species of small snakes in ant mounds in nearby Minnesota (Lang 1971); this is presumably what occurs in the Manitoba summer habitat, where ant mounds are abun- dant. Young born or hatched closer to the den. on the other hand, may be more likely to hiber- nate with the adults (but see Viitanen 1967). Gravid females of many snake species show a tendency to aggregate in areas of localized shelter (Gregory 1975a); in some cases, this may occur at or near the den site (Viitanen 1967; Prestt 1971; Gregory, unpubl. obs.; see section on "Communal Denning and Mating Behavior in Thamnophis"). Other examples of gravid fe- males occasionally being found at dens are given by Preston (1964) and Galligan and Dunson (1979). Parker and Brown (1980) argue that young snakes hibernate elsewhere as a defense against intra- and interspecific predation by adults (in this case, Masticophis taeniatus and Coluber constrictor). The hypothesis that there is some disadvantage to young snakes in hibernating with the adults is supported by the observation that most young Masticophis using communal dens do not survive to age one and most one-year- olds at dens are not known to have used the dens the previous year (Parker and Brown 1980). However, young Thamnophis sirtalis hibernate with adults at a communal hibernaculum in Brit- ish Columbia despite occasional predation on them by adult T. elegans at the same den (Greg- ory, unpubl. obs.). Quantitative assessment of these ideas awaits further study, but it is clear that there is an important ontogenetic change in hibernation behavior in many species of snakes. 4. Fall and spring activity at dens. — Previous studies have revealed a great deal of variation in patterns and timing of entry into and emergence from hibernation of snakes at communal dens, including differences between species, sexes, and age/size groups at the same den (Viitanen 1967; Lang 1971; Prestt 1971; Brown 1973; Landreth 1973; Gregory 1974, 1977a, 1982; Brown and Parker 1976; Parker and Brown 1980). In most cases, fall and spring activity periods at dens span several days or weeks, but individual animals may be active above ground only briefly (e.g., Lang 1971). In other cases, however, individual snakes may remain active in the vicinity of the den for a large part of the fall and/or spring period (e.g., Viitanen 1967; Prestt 1971; Gregory 1974; Parker and Brown 1980), usually without feed- ing. The advantages of remaining active above ground at the den, rather than seeking cooler conditions below ground, are not clear, since such activity is energetically very expensive (Parker and Brown 1980). The significance of activity at dens is obvious in cases where snakes mate at the den site or nearby (Viitanen 1 967; Prestt 1971; Gregory 1974, 1977a; Bennion and Parker 1976; Parker and Brown 1980). Some species, how- ever, apparently mate away from the den (Brown 1973; Brown and Parker 1976; Parker and Brown 1980). Fall mating occurs in some snakes (Trap- ido 1940; Rahn 1942; Saint Girons 1957; Greg- ory 1977a), but in most cases is only an occa- sional phenomenon and less intense than in spring, which is the major breeding season for most temperate zone snakes. Prolonged activity at dens in fall is therefore generally not explained by mating behavior. Perhaps spring and fall activity is related in other ways to the reproductive cycle. For ex- ample, male Vipera berus bask at dens in fall to promote spermatogenesis, which will be com- pleted during basking the following spring (Vol- soe 1944). Vipera species, however, are different from all other temperate zone snakes, in which spermatogenesis is completed well before hiber- VERTEBRATE ECOLOGY AND SYSTEMATICS 61 nation (Aldridge 1979a). Thus, it is not obvious why individual male Masticophis taeniatus, which breed in spring, remain active at the den for up to 37 days before entering hibernation (Parker and Brown 1980). Females of some species of snakes undergo part of secondary vi- tellogenesis in fall (Aldridge 1 979b), but presence or absence of this pattern has not been correlated with fall activity or lack of it. In females of all temperate zone species, all or part of secondary vitellogenesis occurs in spring (Aldridge 1979b). If basking is important to this process, snakes in some cases may trade off the lack of food at the den site for the advantage of readily available shelter at times when cold weather could arise suddenly. We do not yet know enough about details of reproductive cycles (and factors af- fecting them) or fall and spring activity periods of most snakes to be able to correlate these fea- tures. Why Do Snakes Den Communally? Certain disadvantages appear to be inherent in the habit of communal hibernation. First, an- imals at dens in spring and fall may be very conspicuous because of their abundance, and may therefore attract predators. For example, crows take a fairly heavy toll of Thamnophis sirtalis at dens in Manitoba in early spring when the ground vegetation cover is sparse (Gregory 1977a). In- dividuals hibernating singly at isolated locations would be much less conspicuous. [Professional collectors for biological supply companies make even greater inroads in populations at these dens (Gregory 1977b). The problem of collection and/ or slaughter by humans at dens is also great for rattlesnakes, since these animals are often ac- tively persecuted (Klauber 1972; Galligan and Dunson 1979). However, human collection is a relatively recent phenomenon and cannot be considered a long-term selective force.] Another possible disadvantage of communal denning is related to the fact that the den and summer hab- itat may be quite far apart. In such cases, snakes have to migrate, often through unfavorable hab- itat, expending energy and possibly exposing themselves to a higher risk of predation. The question therefore arises as to why snakes den communally. Very small aggregations of snakes (see examples in Parker and Brown 1 973) may simply be fortuitous and irregular in oc- currence, but large aggregations probably have a different basis. There are at least three possible reasons, not mutually exclusive, for communal denning: 1. low availability of suitable hibernat- ing sites; 2. aggregation of snakes in hibernation to minimize losses of endogenously produced heat; 3. enhancement of mating success in the breeding season. A fourth possible advantage of communal denning is that it may lead to more efficient utilization of resources around the den; the area occupied by a dispersed population may change according to annual changes in snake population density and/or resource abundance (the "refuging" hypothesis, Parker and Brown 1980). However, it is not clear to me that com- munal denning is necessary for this system to operate and even if so, it is more likely to be a consequence of communal hibernation rather than a reason for its occurrence in the first place. Shortage of suitable hibernacula is undoubt- edly the main cause of communal overwintering in many cases. This argument has been used to explain winter aggregations of some lizards (Weintraub 1 968; Vitt 1974) and the rattlesnakes Crotalus viridis (Gannon 1978) and C. horridus (Brazaitis 1 980) and the occurrence of more than one species in large communal dens (e.g.. Car- penter 1953; Hirth el al 1969; Lang 1969). Smaller species of snakes may be less influenced by this factor than larger snakes since they are presumably capable of using cavities unavailable to the latter because of size. The problem of lim- ited availability of hibernacula is expected to be particularly serious in cold climates, where hi- bernation at considerable depth is critical for sur- vival. This correlates well with the observation that communal denning is an especially well de- veloped phenomenon at higher latitudes. Gan- non (1978) feels that availability of hibernacula is an important factor limiting the distribution of Crotalus viridis in southern Saskatchewan and Alberta. On the other hand, several authors have noted that there may be apparently usable hi- bernacula which go unused in any winter, even at high latitudes (Viitanen 1967; Lang 1971; Klauber 1972; Gregory 1 977a). Lang (1971) con- cluded that availability of ant mound hibernac- ula was therefore not a limiting factor on num- bers of three species of small snakes in Minnesota. This could be true, however, even if all hiber- nacula were used since there might still be space for more animals within individual hibernacula 62 SPECIAL PUBLICATION- MUSEUM OF NATURAL HISTORY (Parker and Brown 1973). In addition, what ap- pears to be a suitable den to the observer may not be seen as such by snakes; we need to know more about what qualities make a good hiber- nating site and to assess these qualities at poten- tial sites before we can reach a conclusion re- garding availability of hibernacula. The argument that snakes hibernate commun- ally so that they can aggregate below ground and reduce heat loss is difficult to support. It was put forward by White and Lasiewski (1971). with particular reference to rattlesnakes. In favor of this idea is the observation that rattlesnake dens blasted open in winter sometimes reveal large masses of animals (KJauber 1972); however, such behavior could be due to disturbance. Aleksiuk (1977) has also shown that Thamnophis sirtalis tend to huddle under cold conditions, but there is no evidence that this actually happens in the den during hibernation. Snakes hibernating in communal dens are frequently not in contact with one another, although small groups may be formed (Noble and Clausen 1936; Carpenter 1953; Lang 1971; Brown et al. 1974). and iso- lated individuals do not differ in body temper- ature from grouped individuals (Brown et al. 1 974). Use of energy reserves during hibernation per se is probably very low in most cases (Parker and Brown 1980), consistent with the observa- tion that hibernation usually takes place at a low temperature (e.g.. Brown et al. 1974; Sexton and Hunt 1980; Brown 1982; Gregory 1982), not a high one; use of energy reserves may be very high during activity at dens in fall and spring, but this is not taken into account in most studies, yielding considerable overestimates of the energetic cost of hibernation (Bartlett 1976; Parker and Brown 1980). Finally, an important physiological ad- aptation of many hibernating reptiles seems to be that metabolism is significantly depressed at low temperatures (e.g., Aleksiuk 1976; Johansen and Lykkeboe 1979; Gregory 1982). If this is interpreted as an energy-saving device during hi- bernation, then it is not surprising that snakes hibernate at fairly low temperatures, contrary to the predictions of White and Lasiewski (1971). Snakes hibernating with conspecifics at com- munal dens presumably have greater chances of finding mates in the mating season than they would have if they hibernated singly. This idea is difficult to test in the field, but it is obvious that mating opportunities at communal sites should be frequent simply because of the large numbers of snakes involved, especially for species which mate at or near the den in fall or spring. Even in species which do not mate right at the den site, communal hibernation may still en- hance mating chances since individuals dispers- ing from a small area should come into contact more often than when widely scattered (Parker and Brown 1 980). As in the case of the "refuging" hypothesis above, it may be argued that high probability of reproductive success is not a pri- mary reason for the occurrence of communal denning, but simply a secondary advantage of it. This problem is somewhat circular, however, since it is also possible that the prior evolution of early spring mating has resulted in selection for individuals that seek hibernacula used by conspecifics, or that the two have evolved jointly. We need studies which aim to unravel this ques- tion. In some cases, the advantages of communal denning in terms of mating extend beyond mere numbers. Once the mass overwintering habit is established, an opportunity is presented for mat- ing behaviors to evolve which take advantage of this situation. An example is provided by the garter snake, Thamnophis sirtalis. In this species, the different mating strategies of the two sexes seem to be reflected in significant differences in the dynamics of their behavior at the den during the breeding season. This example is examined in detail in the next section. Although the data analysis is largely a posteriori, its main function is to suggest testable hypotheses for further study and points for comparison with other commun- ally denning species which show different be- haviors. Communal Denning and Mating Behavior in Thamnophis The common garter snake, Thamnophis sir- talis, is the most widespread species of snake in North America. While this species does not den communally throughout its range, such behavior is well developed in the northern parts of its range. The study area in question here is in the Interlake region of Manitoba, near the northern limit of distribution of T. sirtalis. This region has a continental climate, with long cold winters and variable summers (Gregory 1 977a). Only four species of snakes occur in the study area, and T. sirtalis is by far the most abundant of these. Communal dens of T. sirtalis in the Interlake VERTEBRATE ECOLOGY AND SYSTEMATICS 63 are mainly limestone sinks, formed by the col- lapse of the ground surface into subterranean caves. The major den examined in this studs (Den 1 ) is a large, bowl-like depression about 20 m long x 12 m wide x 3 m deep; the bottom of the bowl is riddled with cavities leading under- ground. These dens occur on ridges between large marsh belts. Dens are abundant in such areas and are frequently less than 1 km apart. Popu- lations using dens may be very large. Den 1 is estimated to have housed as many as 8000 snakes during one winter, but population size fluctuates drastically from one year to another, apparently in response to variations in weather (Gregory 1977a). Den populations are exclusively adult (Fig. 1); it is not known where the young hiber- nate in this area. The summer habitat of these snakes is in the marshes between the ridges. In- dividuals may move as much as 18 km between den and summer range; migrations are unidirec- tional, with all animals moving south in summer despite the fact that suitable marshes are also found in other directions (Gregory and Stewart 1975). Despite these long migrations and the rel- ative closeness of dens to one another, homing success of individual snakes to the same den in successive years is about 96% (Gregory 1977a). The den is a central feature in the annual cycle of Interlake T. sirtalis. The hibernation period may be as long as six months; in addition, the fall and spring activity periods in the vicinity of the den may occupy up to 1 y h months each (Greg- ory 1977a). In extreme cases, therefore, individ- ual snakes may spend only three months away from the den during the year, and this is the only time in which feeding takes place (Gregory and Stewart 1975). Although occasional fall mating occurs, virtually all mating occurs at the den in spring after emergence (Gregory 1974, 1977a). Mating activity of males is apparently stimu- lated by the change from the cool conditions ex- perienced in hibernation to the warmer condi- tions above ground in spring (Aleksiuk and Gregory 1974; Hawley and Aleksiuk 1975; Crews and Gartska 1982; Gartska et al. 1982). Male courtship activity is directly related to temper- ature (Hawley and Aleksiuk 1975), but declines as the mating season progresses (Aleksiuk and Gregory 1974; Camazine et al. 1980). The tem- perature change associated with emergence from hibernation also stimulates sexual receptivity of females, but does not affect their attractivity to males (Licht and Bona-Gallo 1982). However, 300 380 460 540 620 700 780 860 SVL(mm) Fig. 1 . Size frequency distribution of T. sirtalis at Den 1 in fall 1972. Animals are grouped into intervals of 20 mm snout-vent length (SVL). n = sample size. Data from Gregory (1977a). Gartska et al. ( 1 982) indicate that females, which probably have an active role in mating, are more likely to mate when still cold from emergence; females which have warmed up may not be sex- ually receptive. In both sexes, sexual behavior is independent of gonadal activity (Crews and Gartska 1982; Gartska et al. 1982). The environmental problem faced by these snakes is that of a very short, and sometimes cold, active season. Time to reproduce and per- form other essential functions is therefore at a premium. Under such conditions, we should ex- pect the evolution of a mating system which maximizes the efficiency and success of mating for both sexes early in the season following hi- bernation. The two sexes have, in effect, different reproductive strategies: Males should mate as often as possible since this is their only means of increasing fitness; females, on the other hand, need mate only once per season (but see Gibson and Falls 1975 for evidence of multiple insem- ination of females, and discussion below) and should spend a minimum of time involved in mating activities per se, devoting instead more time to other activities critical to successful re- production. Differences in behavioral dynamics of the two sexes of Interlake T. sirtalis during the mating season appear to reflect these differ- ences in mating strategies. The relationships described above are difficult to test directly for lack of an appropriate control situation. A reasonable substitute for a true con- 64 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 2. Numbers of individuals of each sex of T. sirtalis caught at Den 1 in fall and spring of four over- wintering seasons (data from Gregory 1977a). Males Females 1969-1970 Fall 640 341 Spring 869 56 1970-1971 Fall 90 80 Spring 890 75 1971-1972 Fall 221 198 Spring 846 70 1972-1973 Fall 130 122 Spring 323 24 H (no difference in proportion of females in fall and spring of same overwintering season) rejected with P < .005 in each case (x 2 contingency table). trol, however, should be the behavior of the two sexes at dens in fall, when virtually no mating occurs. If the spring behavior is not specifically related to mating, we should expect the autumn behavior to be similar to it (or the reverse of it, chronologically). The first part of my analysis therefore is a comparison of spring and fall ac- tivity patterns of these snakes. I studied the population ecology of Interlake T. sirtalis from fall 1969 to spring 1973, a total of four overwintering seasons. The research was concentrated at Den 1 described above, but oc- casional samples taken at other dens confirmed that the results presented here are typical of den- ning populations in the study area. Samples of snakes were collected by hand at Den 1 on most days in each fall and spring in the period indi- cated above. Snakes were individually marked by clipping unique combinations of subcaudal scutes, measured (snout-vent length, SVL), and released, almost always on the day of capture. In each overwintering season, the ratio of number of individual males captured to the number of individual females captured was significantly dif- ferent in fall and spring when compared by x 2 contingency analysis (Table 2). Comparison of the four contingency table analyses by hetero- geneity-x 2 (Zar 1974) indicated that the data for all seasons except 1969-1970 could be pooled. In each case, the fall sex ratio was close to 1:1 (only fall 1969 was significantly different from 1:1, Gregory 1 977a), whereas the spring sex ratio was extremely biased towards males. This bias in the spring samples was presumably a result of the sampling method and would probably not have been observed if some other technique, such as fencing the den and capturing snakes leaving it, had been used. It seems very unlikely that it is due to any difference in winter mortality of males and females. The existence of this bias, however, is an important first piece of evidence that activity patterns of the snakes are different between fall and spring. An analysis of daily samples sheds more light on these two activity periods. In fall, the pro- portion of females remained high throughout most of the sampling period and was usually not significantly different from 0.5 (Fig. 2). In the data in Fig. 2, from fall 1971, the proportion of females in samples seems to decrease gradually towards the end of the sampling period, sug- gesting that perhaps females begin arriving at the dens and entering hibernation before males; however, data from other autumns do not con- firm this trend, but simply support the occur- rence of high proportions of females in daily samples. In contrast to autumn samples, the pro- portions of females in daily samples in spring were always low, frequently not statistically dif- ferent from zero (Fig. 2). In fall, individuals of both sexes were recaptured at similar, low rates (Fig. 3). The maximum number of captures of an individual during a given autumn was four for each sex, but maximum intervals between first and last captures were 43 days for males and 25 days for females. In spring, on the other hand, males were recaptured frequently, while females were rarely recaptured in the same spring (Fig. 3). Individual males were caught up to nine times in a given spring and the longest capture history was 36 days (Gregory 1974). No females were ever recaptured more than once in the same spring and the longest interval between first capture and recapture was four days (Gregory 1974). This difference between the sexes was particularly marked in spring 1971 (Fig. 4). These observations are the result of different patterns of emergence from hibernation and dis- persal from the den of males and females in spring. The proximate factors responsible for these dif- ferences are not known, but males begin emerg- ing in large numbers a few days before females and build up in numbers to a maximum in about mid-May, coinciding with the peak in mating activity, then decline in numbers (Gregory 1 974; Fig. 5). Female numbers presumably follow a similar curve since the proportion of females in VERTEBRATE ECOLOGY AND SYSTEMATICS 65 7 SEPT. 13 19 25 OCT. 13 19 25 1 .8 6 Q- 4 .2 o o 4 V J ki .. II Witt iln I * i i 25 1 APR. 13 19 MAY 25 31 6 JUNE Fig. 2. Proportion of females (P-99) in daily samples of T. sirtalis at Den 1 for fall 1971 (open circles) and spring 1972 (closed circles). Vertical lines are 95% confidence limits calculated on basis of binomial distribution. Daily sample sizes range from 3-55 for fall 1971 and 2-125 for spring 1972. daily samples does not vary greatly over the spring period (Fig. 2). In contrast to males, however, females apparently emerge throughout the spring period and spend little time at the den, dispersing very soon after emergence; road counts of dis- persing snakes indicate mostly females leaving early in the spring and increasing proportions of males leaving as the season progresses (Fig. 6). Females also emerge later in the day than do males, but they emerge progressively earlier as the season continues (Gartska et al. 1982). Except for the early part of spring when weath- er is sometimes cool, females are courted as soon as they emerge, or even while emerging ( Aleksiuk and Gregory 1974). Typically, many males si- multaneously court a single female, creating a writhing mating "ball'"' (Aleksiuk and Gregory 1974). Not surprisingly, the head-to-head ori- entation of male and female shown by many col- ubrid snakes is not required for successful court- ship in this species (Gillingham and Dickson 1980). Courtship and mating take several min- utes, but it is not usually possible to see which male manages to copulate with the female. This contrasts with the observations of others (e.g., Devine 1977), in which unsuccessful males leave before the successful male has finished copulat- ing. However, the numbers involved in mating activity in the Interlake are much larger than those reported elsewhere, obscuring actual cop- ulation. Mating almost always occurs on the ground, but males may follow females into low bushes and mate there (Gregory 1975b). Follow- ing copulation, the mating group breaks up rap- idly; the males seem to have no further interest in the female, which becomes unattractive for a day or more and even intolerant of further court- ship (Crews and Gartska 1982; Gartska et al. 1982), but turn to other emerging females in- stead. Devine ( 1 977) and Ross and Crews (1977) have shown that male garter snakes can distinguish between mature, non-mated and recently mated females and only court the former. The cues used are apparently pheromonal. The female attrac- tiveness pheromone is a non-volatile lipid, re- lated to vitellogenin, the precursor of yolk which is manufactured in the liver and circulates in the blood (Gartska and Crews 1981; Crews and Gartska 1982; Gartska et al. 1982). This pher- omone is presumably brought to the skin via a dermal vascular bed and is forced to the body surface through the thin skin between the dorsal and lateral scales. It is a contact pheromone. de- 66 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY 1.0 .8 .6 Q a u £ 1.0 .8 .6 .4 .2 SEPT. 13 19 25 OCT. 7 13 19 i 25 UL I flu 11 6 6 ii ii 25 1 APR. 7 13 19 MAY 25 31 6 JUNE Fig. 3. Proportion of recaptures from same season (P-recaps) in daily samples of T. sirtalis at Den 1 for fall 1971 (upper level) and spring 1972 (lower level). Open circles represent females and closed circles males. Vertical lines are 95% confidence limits calculated on basis of binomial distribution. Daily sample sizes range from 3- 30 (males) and 1-25 (females) for fall 1971, 2-1 10 (males) and 1-15 (females, plus some days with no captures) for spring 1972. tected by the male via the vomeronasal system, and may or may not be the same as the trailing pheromone, which allows species-specific trail- ing of females by males and has its most pro- nounced effect during the spring mating period (Ford 1978, 1981, 1982; Ford and Low 1982). In any case, males are not sensitive to the female attractiveness pheromone early in the season when mating opportunities are very low; how- ever, as the season progresses, females become slightly more abundant relative to males, and males become sensitive to the pheromone and more discriminating about potential mates (Gartska et a/. 1982). Mated females are unattractive to males be- cause of a male-inhibiting pheromone. Follow- ing copulation, a plug forms in the cloaca of the female (Devine 1975); this copulatory plug is apparently manufactured in the renal sex seg- ment of the male (Crews and Gartska 1 982). The male-inhibiting pheromone is probably made by the male at the same time as the plug (Ross and Crews 1977; Crews and Gartska 1982), although Devine (1977) suggests that the female produces VERTEBRATE ECOLOGY AND SYSTEMATICA 67 1.0r .6 Q a u . 1 < * 1 401- 451- 501- 551- 601- 651- 400 450 500 550 600 650 700 SVL (mm ) 701 Fig. 7. Proportion of potentially reproductive (P- repro.) female T. sirtalis from Interlake dens (spring and fall 1972 and spring 1973 combined) in different snout-vent length (SVL) groups. Vertical lines are 95% confidence limits calculated on basis of binomial dis- tribution. Samples sizes range from 3-35. Data from Gregory (1977a). basking in spring (Volsoe 1944; Nilson 1980), thus acquiring additional benefits by emerging early. On the other hand, males of species which do not mate in spring should not be expected to emerge before females; this appears to be the case for Crotalus viridis, which mates in late summer in British Columbia (Macartney, pers. comra.). Benefits accruing to females from this system are less obvious. One significant advantage is that reproductive females are almost certain to mate, although the extent, if any, to which fe- males are able to exercise any choice of mate is unknown. The majority of females at dens are capable of reproduction (Fig. 7; compare with Fig. 1), but a significant number are not. How- ever, probability of being reproductive is cor- related with body size (Fig. 7; Gregory 1977a) and larger females are courted with much greater frequency than are smaller females (Hawley and Aleksiuk 1976). Males therefore waste little en- ergy on non-reproductive females. Copulatory plugs, indicating recent mating, are seen much more frequently in females over 500 mm SVL than in smaller females, an observation consis- tent with the data in Fig. 7. Hawley and Aleksiuk ( 1 976) also provide evidence from laboratory ex- periments that the probability of mating for fe- males is correlated with body size. Almost all reproductive females leaving dens are mated whereas none of the non-reproductive females are (Table 3; Gregory 1977a). [The female attractiveness pheromone may provide the basis for selection of large mates by VERTEBRATE ECOLOGY AND SYSTEMATICA 69 males (Crews and Gartska 1982; Gartska el al. 1982). Since the pheromone is related to vitel- logenin and larger females produce more yolk because they have bigger broods (Gregory 1 977a), larger females may be more attractive because they produce more pheromone (Crews and Gart- ska 1 982; Gartska el al. 1982). Males might even choose mates on the basis of a previous year's reproductive output (Crews and Gartska 1982) since lipid may be stored in the skin of females (Gartska and Crews 1981; Crews and Gartska 1982; Gartska el al. 1982). In any case, males can select the potentially most fecund mate.] A second advantage to females of this pattern of activity is that they are mated almost im- mediately upon emergence, reducing their time of exposure to predators (Crews and Gartska 1981) and allowing them to disperse quickly to the summer habitat and begin feeding. They thus spend a minimum of time active without feeding. This is important because summers are very short in this area and gravid females do not feed in advanced stages of gestation (Gregory and Stew- art 1975). Late spring and early summer may therefore be an important time of year for re- producing females to balance their energy bud- gets. Males are not under such energetic con- straints as reproductive females and may obtain additional benefits from being near shelter at the den if cold weather strikes in spring. Females apparently trade off this advantage for the others mentioned above. It is presumably also advan- tageous for non-reproductive (usually smaller) females to leave the den soon after emergence since they would then extend their feeding season and might reach a larger size by the end of sum- mer; larger females tend to produce bigger broods (Gregory 1977a). The observations and conclusions reported here are probably not unique to communal dens of garter snakes in the Interlake. Partly to answer this question, in 1979 I began monitoring activ- ity patterns of T. sirialis and T. elegans at a communal den in the Chilcotin-Cariboo region of British Columbia, also an area with long, cold winters. Data for only the first year and a half of the study are presented here, but some trends are apparent. Unlike the Interlake dens, this den is occupied in the summer by gravid females (Figs. 8 and 9); they apparently give birth there and the young remain at the den for their first winter. Few adult snakes are seen at the den in fall; per- Table 3. Distribution of mated and non-mated fe- male T. sirialis in reproductive and non-reproductive categories. Samples from summer habitat and from roads in vicinity of dens in April and May 1972; re- productive and mating status determined by dissection (data from Gregory 1977a). Mated Non-maled Potentially reproductive Non-reproductive 20 4 7 H„ (no difference in proportion of mated females in reproductive and non-reproductive categories) rejected with P < .001 (x 2 contingency table). haps the fall activity period is very brief in this case. In other respects, however, activity at this den seems basically similar to that described for T. sirialis in the Interlake. Spring collections from this den are again heavily biased in favor of males in both species (Figs. 8 and 9). Apparently, 77. sirialis leaves the den earlier in spring than 77. elegans; it probably also emerges earlier. In 77. elegans, as in Interlake 77. sirialis, the females that are caught in spring are usually not seen again at the den in the same season, whereas males are recaptured frequently over the spring period. Mating of both species occurs at the den in spring, although mating balls are seen much less frequently than in the Interlake. More often, evidence of spring mating is obtained from oc- casional females found with copulatory plugs in their cloacas. The spring activity pattern of Interlake 77. sir- talis is therefore probably typical of communally denning garter snakes. If so. it may be an im- portant part of the suite of adaptations allowing garter snakes to be so successful in the rigorous environments which limit the northern distri- bution of most other North American reptiles. Questions Although various aspects of communal den- ning in snakes have been studied in some detail, there remain many gaps in our knowledge of this phenomenon. Many of the questions which need to be answered are interrelated and include the following: What are the important physical fea- tures of suitable hibernating sites? How do dis- persion and abundance of suitable hibernating sites in a given area affect the distance snakes migrate between hibernacula and summer range? 70 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY in nj c CO o 6 10 o o o In ff7yi-n ■^ — a 22-23/4/30 n=60 _ 29/4-1/ty80 n=17 _ 2-3/5/79 n=16 _ 8/ty80 n=6 _ 15-16yty79 n*15 _ 22/5/80 n=2 __ 5-6/5/79 n=1 _ 18-W^BO n=1 _ V7/79 n-1 __ 25/V79 n=12 I i I i L ■ ■ i ■ \ =10 Snakes 200 300 400 500 600 700 800 SVL(mm) Fig. 8. Size frequency distribution of T. sirtalis at den in Chilcotin-Cariboo region for various times of year (1979 and 1980 combined). Animals are grouped into intervals of 25 mm snout-vent length (SVL). Dates given are day/month/year; n = sample size. Data above each line represent males (open areas) and unsexed juveniles (hatched areas); data below line represent females (dark areas represent obviously gravid females). Why do some snakes disperse in a particular di- rection, especially if suitable habitat is also avail- able in other directions? Why are young snakes often poorly represented at communal dens and where do they hibernate? Where, in relation to the den, are young snakes born or hatched, and does this influence the likelihood of them using the same den as the adults? How do individual snakes find their way back to the same den over long distances year after year, especially where several dens are present in the same general area? Are new dens occasionally colonized (or old ones recolonized following a disturbance) and if so, how and at what rate? What is the extent of ge- netic isolation among populations at dens in a given area and how is new genetic material in- troduced to a den? Why are snakes sometimes active at dens for long periods of time in fall and/ or spring without feeding or mating? Several of these questions are discussed by Parker and Brown (1980), who also suggest possible ap- proaches to some of them. Underlying all of this is the question of why snakes den communally. Complex, apparently co-ordinated patterns of emergence and mating, such as that shown by Interlake Thamnophis sir- talis, can probably function only in a communal denning situation. An important hypothesis therefore is that snakes which hibernate com- munally have a reproductive advantage over those which hibernate as isolated individuals. This hypothesis should be testable. The ideal way to make such a test would be to make direct comparisons of communal and non-communal hibernators within the same population, but I know of no examples in which both types occur. Comparison of the same species in widely sep- arated parts of its range is somewhat risky be- cause the environmental pressures may differ markedly in the two locations. A more useful VERTEBRATE ECOLOGY AND SYSTEMATICS 71 20- 20 CO o P7W. Ll g*q 22-23/V80 n=23 29/4 -1/5/80 n= 50 2-3/5/79 n=88 6/5/80 n=18 15 -1^5/79 n-56 22/5/80 n=19 28/5/79 n=9 5-6/E/79 n=17 _ 18-19/^80 n-10 11/7/79 n-U -=^- ■ ■»-»• t . i - - 25/S/79 n.13 ]»10 Snakes 100 200 300 400 500 600 700 SVL(mm) Fig. 9. Size frequency distribution of T. elegans at den in Chilcotin-Cariboo region for various times of year (1979 and 1980 combined). Symbols as in Fig. 8. approach might be to compare the ecology of communally and non-communally hibernating species within the same region; if the species in- volved are similar in abundance, size, temper- ature tolerance, etc., it might at least be possible to eliminate the alternative hypothesis that com- munal hibernation results from a shortage of overwintering sites. However, different species may use different strategies to solve the same environmental problem (Wilbur ct al. 1974; Stearns 1976) and different species may hiber- nate communally for different reasons. In fact, the hypotheses put forward in this paper to ac- count for communal denning are not mutually exclusive and may be difficult to separate by ob- servation. The most fruitful approach to this problem therefore is probably through carefully planned manipulative experiments on specific cases. I am now designing such studies. Summary In northern regions, where winter may be sev- eral months long, many snake species hibernate communally in large aggregations of up to a few thousand individuals in extreme cases. Com- munal hibernacula are usually permanent struc- tures, often used annuallv bv the same individ- 72 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY uals, and are sometimes a considerable distance from the summer habitat. Denning populations frequently consist mainly of adults and the snakes may be active at dens for some time each spring and fall. Communal hibernation probably re- flects low availability of overwintering sites in many cases. Another advantage, however, may be that individuals hibernating with conspecifics have enhanced chances of successfully mating early in the active season, an important adap- tation where summers are short. Analysis of ac- tivity of garter snakes (Thamnophis sirtalis) in spring at dens in Manitoba indicates that the two sexes have different behavior patterns consistent with their different reproductive strategies (i.e., males mate more than once per season, females probably once only). Males emerge in fairly large numbers early in spring, whereas females emerge in smaller numbers throughout the spring. Most reproductive females mate immediately upon emergence and then disperse to the summer hab- itat, thereby presumably maximizing the length of their summer activity period. Males, in con- trast, remain at the den for longer periods in spring and continue mating. Thus, emergence patterns are co-ordinated in such a way that mat- ing opportunities are maximized for all individ- uals in the population. Such a system can operate only in the context of communal denning. In fall, when mating is rare, the two sexes do not behave differently from one another. Preliminary data from a garter snake (T. sirtalis and T. elegans) den in British Columbia suggest similar behav- iors. Other communally denning snakes do not necessarily show these kinds of behavior pat- terns, but this does not negate the presumed ad- vantages of such behaviors. It should be possible to design experiments which test hypotheses aris- ing from the garter snake model. Acknowledgments I thank Henry Fitch for stimulating me to study snake ecology through his many papers on the natural history of various species of snakes. I also thank Linda Gregory for her enthusiastic criti- cism of my work and Malcolm Macartney, Anna Roberts, and Neil Dawe for their interesting dis- cussions about garter snake dens and mating be- havior. Malcolm Macartney also allowed me ac- cess to his preliminary data on Crotalus viridis. Pat Konkin (UVic Statistics Laboratory) wrote the computer program for calculating confidence limits for proportions and Gary Caine assisted in preparing figures. Funds for computing were provided by a UVic Faculty Research Grant and other funds by an operating grant from the Nat- ural Sciences and Engineering Research Council of Canada. The manuscript was typed by Barbara Waito. Literature Cited Aldridge, R. D. 1979a. Seasonal spermatogenesis in sympatric Cro- talus viridis and Arizona elegans in New Mexico. J. Herpetol., 13:187-192. 1 979b. Female reproductive cycles of the snakes Ar- izona elegans and Crotalus viridis. Herpe- tologica, 35:256-261. Aleksiuk, M. 1976. Metabolic and behavioural adjustments to temperature change in the red-sided garter snake ( Thamnophis sirtalis parietalis): an in- tegrated approach. J. Thermal Biol., 1:153- 156. 1977. Cold-induced aggregative behaviour in the red-sided garter snake (Thamnophis sirtalis parietalis). Herpetologica. 33:98-101. Aleksiuk, M. and Gregory, P. T. 1974. Regulation of seasonal mating behavior in Thamnophis sirtalis parietalis. Copeia, 1974: 681-689. Ataev, Ch. 1974. Characteristics of hibernation of the Cauca- sian agama in the Kopet-Dag Range. Soviet J. Ecology, 5:159-161. Bailey, R. M. 1948. Winter mortality in the snake Storeria de- kayi. Copeia, 1948:215. Bartlett, P. 1976. Winter energy requirements of Sceloporus occidentalis in the Mojave desert. Comp. Biochem. Physiol., 55A: 179-181. Bennion, R. S. and Parker, W. S. 1976. Field observations on courtship and aggres- sive behavior in desert striped whipsnakes, Masticophis t. taeniatus. Herpetologica, 32: 30-35. Brazaitis, P. 1980. Denizens of a secret underworld. Animal Kingdom, 82:14-18. Brown, W. S. 1973. Ecology of the racer. Coluber constrictor mormon (Serpentes, Colubridae), in a cold temperate desert in northern Utah. Ph.D. Dissertation, Univ. of Utah, Salt Lake City. Brown, W. S. 1982. Overwintering body temperatures of timber rattlesnakes (Crotalus homdus) in north- eastern New York. J. Herpetol., 1 6: 1 45-1 50. Brown, W. S. and Parker, W. S. 1976. Movement ecology of Coluber constrictor VERTEBRATE ECOLOGY AND SYSTEMATICS 73 near communal hibernacula. Copeia, 1976: 225-242. Brown, W. S., Parker, W. S. and Elder, J. A. 1974. Thermal and spatial relationships of two species of colubrid snakes during hiberna- tion. Herpetologica, 30:32-38. Brown, W. S.. Pyle, D. W., Greene, K. R. and Fried- LAENDER, J. B. 1982. Movements and temperature relationships of timber rattlesnakes (Crotalus horridus) in northeastern New York. J. Herpetol., 16:151- 161. Camazine, B., Gartska, W., Tokarz, R. and Crews, D. 1980. Effects of castration and androgen replace- ment on male courtship behavior in the red- sided garter snake ( Thamnophis sirtalis par- ietalis). Horm. Behav., 14:358-372. Carpenter, C. C. 1953. A study of hibernacula and hibernating as- sociations of snakes and amphibians in Michigan. Ecology, 34:74-80. Crews, D. 1975. Psychobiology of reptilian reproduction. Science, 189:1059- 1065. Crews, D. and Gartska, W. R. 1982. The ecological physiology of a garter snake. Sci. Am., 247(5):159-168. Criddle, S. 1937. Snakes from an ant hill. Copeia, 1937:142. Devine, M. C. 1975. Copulatory plugs in snakes: enforced chas- tity. Science, 187:844-845. 1977. Copulatory plugs, restricted mating oppor- tunities and reproductive competition among male garter snakes. Nature, 267:345-346. Duguy, R. 1963. Biologie de la latence hivernale chez Vipera aspis L. Vie et Milieu, 14:31 1-443. Fitch, H. S. 1 960. Autecology of the copperhead. Univ. Kansas Publ. Mus. Nat. Hist., 13:85-288. 1963a. Natural history of the black rat snake (Elaphe o. obsoleta) in Kansas. Copeia, 1963:649- 658. 1963b. Natural history of the racer Coluber con- strictor. Univ. Kansas Publ. Mus. Nat. Hist., 15:351-468. 1965. An ecological study of the garter snake, Thamnophis sirtalis. Univ. Kansas Publ. Mus. Nat. Hist., 15:493-564. Fitch, H. S. and Glading, B. 1947. A field study of a rattlesnake population. Calif. Fish and Game, 33:103-123. FitzSimons, V. F. M. 1 962. Snakes of southern Africa. Purnell and Sons, Capetown, S.A. Ford, N. B. 1978. Evidence for species specificity of phero- mone trails in two sympatric garter snakes, Thamnophis. Herp. Review, 9:10-11. 1981. Seasonality of pheromone trailing behavior in two species of garter snake, Thamnophis (Colubndae). Southwest. Nat., 26:385-388. 1 982. Species specificity of sex pheromone trails of sympatric and allopatric garter snakes (Thamnophis). Copeia, 1982:10-13. Ford, N. B. and Low, J. R. 1 982. A biological mechanism for determining di- rection of pheromone trails having low vol- atility. Am. Zool., 22:852. (abstr.). Galligan, J. H. and Di nson, W. A. 1979. Biology and status of timber rattlesnake (Crotalus horridus) populations in Pennsyl- vania. Biol. Conserv., 15:13-58. Gannon, V. 1978. Factors limiting the distribution of the prai- rie rattlesnake. Blue Jay, 36:142-144. Gartska, W. R., Camazine, B. and Crews, D. 1 982. Interactions of behavior and physiology dur- ing the annual reproductive cycle of the red- sided garter snake (Thamnophis sirtalis par- ietalis). Herpetologica, 38:104-123. Gartska, W. R. and Crews, D. 1981. Female sex pheromone in the skin and cir- culation of a garter snake. Science, 2 1 4:681- 683. Gatten, R. E., Jr. 1978. Aerobic metabolism in snapping turtles, Chelydra serpentina, after thermal acclima- tion. Comp. Biochem. Physiol., 61A:325- 337. Gibson, A. R. and Falls, J. B. 1975. Evidence for multiple insemination in the common garter snake, Thamnophis sirtalis. Can. J. Zool., 53:1362-1368. Gillingham, J. C. and Dickinson, J. A. 1980. Postural orientation during courtship in the eastern garter snake, Thamnophis s. sirtalis. Behav. Neur. Biol., 14:358-372. Gregory, P. T. 1 974. Patterns of spring emergence of the red-sided garter snake ( Thamnophis sirtalis parietalis) in the Interlake region of Manitoba. Can. J. Zool., 52:1063-1069. 1975a. Aggregations of gravid snakes in Manitoba, Canada. Copeia, 1975:185-186. 1975b. Arboreal mating behavior in the red-sided garter snake. Can. Field-Nat., 89:461-462. 1977a. Life-history of the red-sided garter snake (Thamnophis sirtalis parietalis) in an ex- treme environment, the Interlake region of Manitoba. Publ. Zool. No. 13, Nat. Mus. Canada. 1977b. Rare and threatened snake species of Can- ada. Pp. 122-126. In Mosquin.T. and Such- al, C. (eds.), Canada's Threatened Species and Habitats. Special Publ. No. 6, Canadian Nature Federation. Ottawa. 1 982. Reptilian hibernation. Pp. 53-1 54. In Gans. C. and Pough. F. H. (eds.). Biology of the Reptilia. Volume 13, Physiology D. Aca- demic Press, London. Gregory, P. T. and Stewart, K. W. 1975. Long-distance dispersal and feeding strategy of the red-sided garter snake (Thamnophis sirtalis parietalis) in the Interlake of Mani- toba. Can. J. Zool., 53:238-245. 74 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Hawley, A. W. L. and Aleksiuk, M. 1975. Thermal regulation of spring mating behav- iour in the red-sided garter snake {Tham- nophis sirtalis parietalis). Can. J. Zool., 53: 768-776. 1976. Sexual receptivity in the female red-sided garter snake ( Thamnophis sirtalis parietalis). Copeia, 1976:401-404. Hirth. H. F. 1 966a. Weight changes and mortality of three species of snakes during hibernation. Herpetologica, 22:8-12. 1 966b. The ability of two species of snakes to return to a hibernaculum after displacement. Southwest Nat., 1 1:49-53. Hirth. H. F., Pendleton, R. C, King, A. C. and Downard, T. R. 1969. Dispersal of snakes from a hibernaculum in northwestern Utah. Ecology, 50:332-339. Johansen, K. and Lykkeboe, G. 1979. Thermal acclimation of aerobic metabolism and d-Hb binding in the snake, Yipera be- rus. J. Comp. Physiol., 130B:293-300. Jolly, G. M. 1965. Explicit estimates from capture-recapture data with both death and immigration — sto- chastic model. Biometrika, 52:225-245. Kinghorn, J. R. 1 964. The snakes of Australia. Rev. ed. Angus and Robertson, Sydney, Australia. Klauber, L. M. 1 972. Rattlesnakes. 2nd ed. (2 vols.). Univ. of Cal- ifornia Press, Berkeley. Landreth, H. F. 1973. Orientation and behavior of the rattlesnake, Crotalus atrox. Copeia, 1973:26-31. Lang, J. W. 1 969. Hibernation and movements of Storeria oc- cipito-maculata in northern Minnesota. J. Herpetol., 3:196-197. (abstr.). 1971. Overwintering of three species of snakes in northwestern Minnesota. M.S. Thesis, Univ. of North Dakota, Grand Forks. Licht, P. and Bona-Gallo, A. 1982. Dependence of vitellogenesis on low temper- atures and mating in the garter snake, Tham- nophis sirtalis parietalis. Am. Zool., 22:859. (abstr.). Mayhew, W. W. 1965. Hibernation in the horned lizard, Phryno- soma m'calli. Comp. Biochem. Physiol., 16: 103-119. MOBERLY, W. R. 1963. Hibernation in the desert iguana, Dipsosau- rus dorsalis. Physiol. Zool., 36:152-160. Naulleau, G. 1966. Etude complementaire de Factivite de 17- pera aspis dans la nature. Vie et Milieu, 1 7: 461-509. NlLSON, G. 1 980. Male reproductive cycle of the European ad- der, Vipera berus, and its relation to annual activity periods. Copeia, 1980:729-737. Noble, G. K. and Clausen, H. J. 1936. The aggregation behavior of Storeria dekayi and other snakes, with especial reference to the sense organs involved. Ecol. Monogr., 6: 269-316. Parker, W. S. 1 976. Population estimates, age structure and den- ning habits of whipsnakes, Masticophis t. taeniatus, in a northern Utah Atriplex-Sar- cobatus community. Herpetologica, 32:53- 57. Parker, W. S. and Brown, W. S. 1973. Species composition and population changes in two complexes of snake hibernacula in northern Utah. Herpetologica, 29:319-326. 1 974. Mortality and weight changes of Great Basin rattlesnakes {Crotalus viridis) at a hibernac- ulum in northern Utah. Herpetologica, 30: 234-239. 1 980. Comparative ecology of two colubrid snakes, Masticophis t. taeniatus and Pituophis mel- anoleucus deserticola, in northern Utah. Publ. Biol. Geol. No. 7, Milwaukee Pub. Mus. Patterson, J. W. and Davies, P. M. C. 1978. Energy expenditure and metabolic adapta- tion during winter dormancy in the lizard Lacerta vivipara. Jacquin. J. Thermal Biol., 3:183-186. Preston, W. B. 1 964. The importance of the facial pit of the north- ern Pacific rattlesnake {Crotalus viridis ore- ganus) under natural conditions in southern British Columbia. M.Sc. Thesis, Univ. of British Columbia, Vancouver. Prestt, I. 1971. An ecological study of the viper, I 'ipera be- rus, in southern Britain. J. Zool., Lond., 164: 373-418. Rahn, H. 1942. The reproductive cycle of the prairie rattler, Crotalus viridis. Copeia, 1942:233-240. Reinert, H. K. and Kodrich, W. R. 1982. Movements and habitat utilization by the massasauga, Sistrurus catenatus catenatus. J. Herpetol., 16:162-171. Ross, P., Jr. and Crews, D. 1977. Influence of the seminal plug on mating be- haviour in the garter snake. Nature, 267:344- 345. Saint Girons, H. 1957. Le cycle sexuel chez Vipera aspis (L.) dans l'ouest de la France. Bull. Biol, de France Belgique, 91:284-350. Sexton, O. J. and Hunt, S. R. 1980. Temperature relationships and movements of snakes {Elaphe obsoleta, Coluber constric- tor) in a cave hibernaculum. Herpetologica, 36:20-26. Shine, R. 1 979. Activity patterns in Australian elapid snakes (Squamata: Serpentes: Elapidae). Herpeto- logica, 35:1-1 1. VERTEBRATE ECOLOGY AND SYSTEMATICA 75 Stearns, S. C. 1976. Life-history tactics: a review of the ideas. Quart. Rev. Biol., 51:3-47. Trapido, H. 1 940. Mating time and sperm viability in Storeria. Copeia. 1940:107-109. VllTANEN, P. 1967. Hibernation and seasonal movements of the viper, I 'ipera bents (L.). in southern Finland. Ann. Zool. Fenn.. 4:472-546. Vitt. L. J. 1 974. Winter aggregations, size classes and relative tail breaks in the tree lizard, Urosaurus or- nat its (Sauria: Iguanidae). Herpetologica. 30: 182-183. VOLS0E, H. 1 944. Structure and seasonal variation of the male reproductive organs of Vipera bents (L.). Spolia. Zool. Mus. Hauniensis, 5:1-157. Weintraub, J. D. 1968. Winter behavior of the granite spiny lizard. Sceloponts orcutti, Stejneger. Copeia, 1968: 708-712. White, F. N. and Lasiewski, R. C. 1971. Rattlesnake denning: theoretical consider- ations on winter temperatures. J. Theor. Biol., 30:553-557. Wiklund, C. and Fagerstrom, T. 1977. Whvdo males emerge before females? Oeco- logia, 31:153-158. Wilbi k. H. M., Tinkle, D. W. and Con ins. J. P. 1974. Environmental certainty, trophic level, and resource availability in life history evolu- tion. Am. Nat.. 108:805-817. Woodbury, A. M. and Hardy, R. 1 940. The dens and behavior of the desert tortoise. Science. 92:529. Woodbury. A. M., Vet as, B., Julian, G.. Gi issmeyer, H. R.. Heyrend, F. L.. Call. A.. Smart. E. W. and Sanders, R. T. 1951. Symposium: a snake den in Tooele County. Utah. Herpetologica. 7:1-52. Zar, J. H. 1974. Biostatistical analysis. Prentice-Hall. Engle- wood Cliffs. N.J. Vertebrate Ecology and Systematics— A Tribute to Henry S. Fitch Edited by R. A. Sergei, L. E. Hunt, J. L. Knight. L. Malaret and N. L. Zuschlag ! 1984 Museum of Natural History. The University of Kansas. Lawrence Parameters of Two Populations of Diamondback Terrapins (Malaclemys terrapin) on the Atlantic Coast of Florida RlC HARD A. SEIGEL Introduction The diamondback terrapin, Malaclemys ter- rapin has several attributes which would seem to make it an interesting subject for ecological study. These include a unique habitat for che- lomans (brackish water), an extremely wide lin- ear range (Massachusetts-Texas), and the dis- tinction of being the "most celebrated of American turtles," a reflection of the popularity of this turtle as a gourmet food item in the early 20th century (Conant 1975). Despite these fea- tures, our knowledge of the life history of this species remains surprisingly limited. Studies of terrapins in the wild have dealt mainly with re- production (Finneran 1948; Reid 1955; Burger and Montevecchi 1975; Montevecchi and Burger 1975; Burger 1976a, 1976b, 1977; Auger and Giovannone 1979; Seigel 1980b. 1980c). epizoic fouling (Jackson and Ross 1971; Ross and Jack- son 1972; Jackson et al. 1973), mortality (Seigel 1978, 1980a) and hibernation (Lawler and Mu- sick 1972; Yearicks et al. 1981). Data on the population biology of Malaclemys are few, es- pecially under natural conditions. Cagle (1952) reported growth rates and age at maturity for Louisiana intergrades (M. t. pileata x littoralis), and Hurd et al. (1979) described the size struc- ture and population size of M. t. terrapin from Delaware. Most data on population biology are based on captives (Hildebrand 1929, 1932; Allen and Littleford 1955), and must be viewed with caution due to the unnatural conditions under which the turtles were maintained (Carr 1952; Burnley 1969). From 1977 to 1979 I studied the life history and ecology of the Florida east coast terrapin. M. t. tequesta, at the Merritt Island National Wildlife Refuge, Brevard County, Florida. This paper presents data on the growth rates, popu- lation structure, and age at maturity for two pop- ulations of Malaclemys under natural condi- tions. Materials and Methods The Merritt Island refuge consists of three large, brackish water lagoons, each surrounded by a series of canals and ditches which are perma- nently filled with water. A more detailed descrip- tion of the area is presented elsewhere (Seigel 1979). For the purposes of this study, terrapins were collected primarily from the northern ends of two lagoons, known locally as the Indian and Banana rivers (Fig. 1). Indian River turtles were collected by deploying small mesh (maximum diameter = 6 cm) gill nets along a narrow canal bordering a dike road. Two nets were set per- pendicular to the shoreline to block off a 100 m section of the canal. Turtles moving up and down the canal became entangled in the nets and were removed within two hours of capture. Turtles from the Banana River were collected by walking surveys around a small man-made spoil island. Turtles were captured by hand while they basked along the shoreline, or while they swam and fed in the clear waters surrounding the island. The following straight-line measurements were recorded to the nearest 0. 1 cm using vernier cal- ipers; carapace (CL) and plastron (PL) length, length of the right abdominal scute, and medial length of visible abdominal annuli. Wet body weight was recorded to the nearest 1 g with a spring balance. All turtles were given an indi- vidual mark (Ernst et al. 1974) and released at point of capture. Plastral annuli have been used to estimate the growth rate and age of several species of turtles, using a variety of techniques (see Graham 1979 for review). In my study, age was estimated using the method of Sexton (1959). Growth was esti- mated using Sergeev's ( 1 937) formula of L,/L 2 = C,/C 2 , where C, represents the annuli length, C 2 the abdominal scute length, L, the plastron length when the annuli was formed, and L 2 the current plastron length. Since large, female Malaclemys >16 cm PL often lacked one or more annuli. they were excluded from this analysis. Statistical tests follow Ott (1977). Means are followed by ± one standard deviation. Results and Discussion Growth and Sexual Maturity. — One hundred thirteen Malaclemys were examined from the 77 78 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY 80° 40' Fig. 1. Merritt Island National Wildlife Refuge. Shaded areas represent lagoonal waters. Indian River study site is shown by cross-hatching. Banana River study site by cross. VERTEBRATE ECOLOGY AND SYSTEMATICA 79 E o X 12- 8 - H 2 3 —r- 4 O cr h- 20 i < _J CL 16 12- 8 H T 3 ■~r- 4 5 nr 6 AGE Fig. 2. Relationship between age and plastron length in 53 female and 13 male Malaclemys from the Indian River. Vertical bars represent sample range. Indian River and 44 from the Banana River. Fifty-three of the Indian River turtles bore dis- tinct growth annuli, but heavy shell damage from barnacles (Seigel 1983) obliterated most annuli on terrapins from the Banana River. Ontogenetic change in the relative size of the abdominal scute, such as that noted by Moll and Legler ( 1 97 1 ) for tropical Pseudemys scripta, was minor in this study ( < 2%), so no correction factor was needed. Fig. 2 shows the relationship between age and plastron size. The wide variability in size within a particular age class observed in Malaclemys frequently occurs in other turtles (Gibbons 1 968: Ernst 1971. 1975. 1977: Plummer 1977b). Growth of the sexes is relatively constant and similar for the first two years of life, but begins to diverge after age three, when male growth rates decline, but females continue to grow at a steady rate. The curve for both sexes shows a marked decline in growth as sexual maturity is reached (see below). Fig. 3 shows the relationship be- tween percent growth/year and plastron size. Most 80 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY 100 -| £ 80 > LL! h- < DC o DC CD 60 - \- 40 - 20 - <> 4) II I) II T -L .L _L <> I 3.5 - 1 1 1 1 1 1 1 5.5 7.5 9.5 11.5 PLASTRON LENGTH (CM) 13.5 15.5 Fig. 3. Relationship between growth rate (%/year) and plastron length for Indian River Malaclemys. Vertical bars represent sample range. rapid growth occurs at PL 3-3.9 cm, followed by a sharp decrease, and then a more gradual decline in growth to <5%/year in mature individuals. This pattern is similar to that of most other fresh- water turtles, especially the genera Chrysemys and Pseudemys (see Bury 1979 for review). Lim- ited data from turtles recaptured after six months or more support the above growth estimates. Two mature females of 13.8 and 14.6 cm PL grew at annual rates of 5.4% and 2.9% respectively. Six mature females of > 1 5.0 cm PL grew at a mean annual rate of 2.2% (range = 0-7.1%). Based on these values, the largest female in the Indian Riv- er population (PL = 17.7 cm) would be approx- imately 15 years old. Longevity in this popula- tion is estimated to be about 20 years. Fig. 4 compares the PL/age relationships of Malaclemys from different parts of the range. Florida Malaclemys grow at a slightly faster rate than terrapins from North Carolina or Louisiana (Cagle 1952). Although Florida Malaclemys are larger at hatching than turtles from the other populations (Seigel 1980c), this difference in ini- tial size is insufficient to account for the differ- ences in Fig. 4. Gibbons ( 1 967) showed that even local populations of Chrysemys picta varied widely in growth rates because of differences in local feeding habits and food quality. Most data suggest that the feeding habits of Malaclemys are relatively similar throughout its range (Cagle 1952; Wood 1976; Hurd et al. 1979; R. Seigel, pers. obs.; but see Cochran 1976), with no com- parable dramatic differences such as Gibbons ( 1 967) noted. It therefore seems unlikely that the differences in growth rates seen in Fig. 4 are due to differences in local feeding habits. However, the North Carolina turtles were captives, and were fed fish as supplements to their normal food VERTEBRATE ECOLOGY AND SYSTEMATICS 81 14 12-| ? o ■E10 o z LLt _l Z 8- o cc l- (/) < c 14 3 4 AGE 5 J J Fig. 4. Comparison of growth rate of Malaclemys from different parts of the range (sexes combined). See text for data sources. (mollusks), so their growth may have been some- what affected. The differences in Fig. 4 may re- flect the longer activity and growing season of M. I. tequesta, which at Merritt Island is active from mid-February to late November (Seigel, unpub. data), whereas North Carolina captives were only active from May to October (Hilde- brand 1932). No data on the activity season of Louisiana terrapins are available, but from a cli- matic viewpoint, it is probably more similar to Florida than North Carolina. The smallest female showing evidence of sex- ual maturity (oviducal eggs or corpora lutea) was 13.5 cm PL, and all females > 14.0 cm PL were mature. Fig. 2 shows that most females reach 13.5-14.0 cm by age four, but that some may not attain maturity until age five. The smallest male considered mature (based on secondary sexual characteristics and enlarged testes) was 9. 1 cm PL, and all males >9.5 cm PL were consid- ered mature. According to Fig. 2, males may reach this size as early as the second year of life, but most probably do not mature until age three. Hildebrand (1932) suggested that sexual matu- rity in Malaclemys was related to size rather than age, and my results support this idea. Table 1 shows the size and age at maturity for Malacle- mys from different parts of the range. Size at maturity is rather uniform for both sexes, while age at maturity is more variable. Bury (1979) summarized the data on growth and sexual maturity for freshwater, mainly north- temperate turtles, and made the following con- clusions: 1) males often mature earlier and at a smaller size than females; 2) growth is most rapid before maturity is reached; 3) in temperate re- gions, individuals in southern populations ma- ture earlier than northern conspecifics: 4) sexual 82 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 1. Size and age at sexual maturity for male and female Malaclemys terrapin from different parts of the range. Subspecies (locality ) 8 si/e (plastron length I and age (yrs) at maturity $ size (plastron length) and age (yrs) at maturity Authority terrapin (New Jersey) 13.2cm/?yrs — Montevecchi and Burger (1975) terrapin^ (North Carolina) 13.7 cm/7 yrs 9.0 cm/5 yrs Hildebrand(1932) pileata x littoralis (Louisiana) 16.0 cm/6 yrs 9.9 cm/3 yrs Cagle(1952) tequesta (Honda) 13.5 cm/4-5 yrs 9.5 cm/2-3 yrs This study Captive population. maturity is usually related to attaining a certain size rather than a certain age. The data from subtropical Malaclemys appear to conform closely to these patterns. Sexual Dimorphism. — The sexes of M. t. te- questa are highly dimorphic, both for external characters and body size. Male terrapins have smaller, narrower heads, darker carapacial mark- ings, and larger carapacial keels than females. In addition, adult males have long, thick tails, with the vent posterior to the margin of the carapace, while adult females have short, narrow tails with the vent beneath the overhanging carapace. The mean PL and weight of 1 1 3 Indian River females was 15.4 ± 1 .00 cm, and 886 ± 1 93 g. The same measurements for 13 Indian River males were 10.4 ± 0.69 cm, and 283 ± 50.9 g. Using the terminology of Fitch (1981). M. t. tequesta has a FMR (female to male ratio) for length and weight of 148 and 313, respectively. Thus, while females are ca. 1.5 times male length, they are > 3 times male weight. Fitch (1981) reported that other turtle species showed similar variation in FMR between length and weight. Berry and Shine (1980) and Fitch (1981) re- viewed sexual size dimorphism in turtles and other reptiles, and related such dimorphism to intrasexual competition and divergent reproduc- tive strategies. The greater body size of female M. t. tequesta probably evolved as a means to increase reproductive potential, since clutch size is positively correlated with body size in Mala- clemys (Montevecchi and Burger 1975; Seigel 1980c). Large males on the other hand, would gain no reproductive advantage over smaller males, because male combat for mates is appar- ently absent in Malaclemys (Seigel 1 980b). How- ever, small body size permits males to mature at a younger age than females, possibly increasing lifetime reproductive potential. Population Structure. — The size class structure of Indian and Banana river females, and that of Indian River males are shown in Figs. 5 and 6. Too few Banana River males were captured for analysis. The age structure of Indian River ter- rapins is shown in Fig. 7. There is a noticeable lack of small or immature individuals in both populations. The under-representation of small terrapins is probably the result of sampling error which favored adults. Small Malaclemys (<9 cm PL) were rarely seen at either study site, and it seems likely that behavioral differences exist be- tween adults and juveniles that reduce the prob- ability of capturing small terrapins. Hurd et al. (1979) advanced a similar hypothesis to explain the lack of small individuals in a Delaware pop- ulation of M. t. terrapin, but also noted that local habitat destruction may have caused "cata- strophic mortality" among young terrapins. No evidence for such habitat destruction exists at Merritt Island (at least recently), and I suggest that sampling error, rather then heavy mortality, is sufficient explanation for the lack of immature individuals. The population size structures of Indian and Banana river females are significantly different (Mann-Whitney L'test, P < 0.01), with the most striking differences reflected in the 17.0 cm and larger size classes (Fig. 5). Thirty-one percent of the female turtles from the Banana River were > 1 7.0 cm PL. compared to only 3% of the Indian River terrapins in the same size range. Conceiv- ably, these differences could be artifacts of the different sampling techniques used, but this seems VERTEBRATE ECOLOGY AND SYSTEMATICS 83 20-i 10 - \° o\ > O 2 LU D O LU DC BANANA RIVER N =44 r^ io - INDIAN RIVER N = 113 . ._, I I i 10 - 1 1 — T , 1 0- T 13.5 13.7 14.2 14.7 15.2 15.7 SIZE (CM) 16.2 16.7 17.2 175 Fig. 5. Size structure of female Malaclemys from the Indian and Banana rivers 1977-1979. Size distributions did not differ between sampling periods (Mann-Whitney U test, P > . 10). so data are combined for presentation. unlikely, since both techniques were effective in capturing large individuals and it was in the larg- er size classes that most of the differences be- tween the populations were found. Gibbons (1967) found that local populations of C. picta may vary widely in size structure, reflecting dif- ferences in feeding habits, food quality and growth rates. I was unable to determine if Indian and Banana river Malaclemys grew at different rates, but qualitative analysis of stomach contents from females from the two sites showed no striking differences (Seigel, unpub.), and it seems im- probable that feeding habits varied sufficiently to account for the differences in size structure seen in Fig. 5. An alternative hypothesis is dif- ferential mortality. Elsewhere (Seigel 1980a) I have shown that Indian River females are subject to raccoon predation during the nesting season. No comparable predation was observed along the Banana River. Each time an Indian River female nests, there is a small but definite risk of encountering a predator and being eliminated from the local population. Because M. t. tequesta may nest up to three times/year (Seigel 1980c), the probability of Indian River females surviving many nesting seasons may be low. This may re- sult in a lower proportion of females surviving to reach the larger (i.e. older) size classes that might otherwise occur in the absence of preda- tion. The under-representation of larger females in the Indian River when compared to the Ba- nana River might be a reflection of the higher mortality rate among mature Indian River fe- males. Sex Ratio.— The ratio of females/males for Merritt Island Malaclemys is shown in Table 2. Ratios varied seasonally, but were always sig- nificantly different from unity (binomial distribu- tion, P < 0.05). The samples showing the least skewed ratios (March-April) were taken during 84 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY >o 40 - s INDIAN RIVEF 1 5 30- z LU O 20- LU DC LL 10 - i i i 1 i 1 9.2 9.7 10.2 10.7 SIZE (CM) 11.2 >ll.7 Fig. 6. Size structure of male Malaclemys from the Indian River, 1977-1979. Differences in size structure between sampling periods were insignificant (Mann- Whitney (7 test, P > .10), so combined data are shown. the spring mating season (Seigel 1 980b), at a time when males reach their greatest representation in local populations. The 5:1 ratio taken at this time probably represents the best estimate of the "true" sex ratio in this population. Hildebrand (1932), Hurd et al. (1979), and Yearicks et al. (1981) have also reported female-biased sex ra- tios in Malaclemys, although Cagle (1952) noted the opposite trend in offshore populations from Louisiana. The skewed sex ratios in Merritt Island Ma- laclemys are not the result of sampling bias or investigator error as suggested by Gibbons ( 1970) for other studies where the sex ratio was not 1:1. The different sampling techniques used, as well as the number of populations examined, probably precluded most bias due to sampling error, and the distinct sexual dimorphism in Ma- laclemys (see above) greatly reduced the chances of mistaking juvenile males for females, as Gib- bons (1970) suggested. Bury (1979) found that of 39 studies of sex ratio in freshwater turtles, skewed ratios were reported in only 13 (33%). Assuming no bias due to sampling error, cases of skewed sex ratio such as that of M. t. tequesta are in the minority. Such skewed ratios may be the result of differential juvenile mortality (which has been neither confirmed or rejected), or tem- perature-modified sex determination during in- cubation, for which extensive evidence exists (see Bull and Vogt 1979; Yntema 1 979 and references >- o 50 C 40 " UJ 30 - o lu 20 U_ I AGE (YRS) Fig. 7. Age structure of 12 male and 107 female Malaclemys from the Indian River, based on age-size relationships calculated from growth annuli. Males >5 years old, and females >7 years old, could not be ac- curately aged, so they are included with the oldest classes. Males are represented by shaded bars, females by open bars. therein). If skewed sex ratios among adult turtles are a reflection of incubation temperatures, it remains to be determined if such ratios occur fortuitously (depending on the nest site choosen by the female) or if these ratios have been se- lected for by such factors as resource abundance, as was suggested by Nichols and Chabreck ( 1980) for Alligator mississipiensis. Population Size. — Schnabel population size estimates and 95% confidence limits (Overton, 1969) for the Indian and Banana rivers were 404.7 (95% C.L. = 182.8-790.5) and 212.5 (95% C.L. = 58.7-627.3), respectively. The wide vari- ability in the confidence limits are probably a result of two factors: 1 ) a low recapture rate which did not exceed 50% until the last sampling pe- riod, and 2) the apparent ability of Malaclemys to freely move into and out of the study area for short time periods (due to the lack of natural Table 2. Sex ratio (females/males) of Merritt Island Malaclemys during different sampling periods. Study site Sampling period Sex ratio N Banana River February-November 9:1 47 Indian River February-November 10:1 99 Indian River March-April 5:1 26 VERTEBRATE ECOLOGY AND SYSTEMATICS 85 barriers restraining movement). Plummer ( 1 977a) found that temporary movements of Trionyx muticus out of his Kansas study site greatly in- creased the variability of his population size es- timates. Although Malaclemys at Merritt Island showed relatively long-term (ca. 18 months) fi- delity to a particular area (Seigel 1979), it is prob- able that short-term movements took place at both study sites, so the above estimates may be somewhat biased. These population estimates and the size limits of the two sampling areas were used to construct density estimates. The Indian River sampling area covered 2.27 acres, yielding a density of 178.3 individuals/acre; the Banana River sam- pling area was 1.62 acres, yielding a density of 131.1 individuals/acre. These figures are some- what higher than most reports of density in fresh- water turtle populations (Bury 1 979), but are not as high as the 239 individuals/acre reported by Ernst (1971) for C. picta. Biomass estimates, based on the above figures and wet body weight were 390.0 kg/ha for the Indian River, and 355 kg/ha for the Banana River. Both the density and biomass estimates may be somewhat inflated as a result of a) an arbitrary and possibly unreal- istically low estimate of the population bound- aries, and/or b) the tendency of Malaclemys to form large aggregations during the breeding sea- son (Seigel 1980b). However, it seems clear that Merritt Island Malaclemys may attain a consid- erable density and biomass in local areas, at least during certain times of the year. Summary The growth rates, age at maturity, population size and population structure of the Florida east coast terrapin, Malaclemys terrapin tequesta were studied from 1977 to 1979 at the Merritt Island National Wildlife Refuge, Brevard County, Flor- ida. Data from two areas (Indian and Banana rivers) are presented. Growth was most rapid immediately after hatching, declining to <5%/ year in mature turtles. Females matured at a plas- tron length of 13.5-14.0 cm, at an age of 4-5 years. Male terrapins reached maturity at a plas- tron length of 9.0-9.5 cm, at an age of 2-3 years. Female terrapins attain a much larger body size than do males, with a mean FMR (female to male size ratio) of 148 for length and 313 for weight. Such dimorphism probably reflects divergent re- productive strategies between the sexes; females benefit from large body size via increased repro- ductive potential, whereas males attain only a small body size, but reach maturity earlier than females. The two study populations differed sig- nificantly in size structure, with the Banana Riv- er population having relatively more individuals in the larger size classes. This may reflect higher mortality among Indian River females. The sex ratios of both populations were significantly dif- ferent from 1:1, with females outnumbering males by at least 5:1. Schnabel population size esti- mates for the Indian and Banana rivers were 404.7 and 212.5, respectively, and it appears that Malaclemys may attain a considerable density and biomass in local areas. Acknowledgments Assistance in the field was provided by E. Scott Clark, Timothy R. Claybaugh, John D. Galluzzo, Mary T. Mendonca, Boyd Thompson, and Sher- ry Williams. Thanks go to the U.S. Fish and Wildlife Service for making field sites accessible, and for logistical support. I also thank Gunther Schlager for statistical advice. The critical re- views of William E. Duellman, Carl H. Ernst, Henry S. Fitch, and an anonymous reviewer im- proved this manuscript. Particular appreciation goes to my wife, Nadia, for constant help both in the field and during preparation of the manu- script, and to the late James D. Anderson for advice and encouragement. This research was supported by NASA contract NASI 0-8986, to L. M. Ehrhart. Literature Cited Allen, J. F. and Littleford, R. A. 1955. Observations on the feeding habits and growth of immature diamondback terrapins. Herpetologica, 11:77-80. Auger, P. J. and Giovannone. P. 1 979. On the fringe of existence. Diamondback ter- rapins at Sandv Neck. Cape Natur.. 8:44- 58. Berry. J. F. and Shine, R. 1980. Sexual size dimorphism and sexual selection in turtles (Order Testudines). Oecologia, 44: 185-191. Bull, J. J. and Vogt, R. C 1979. Temperature-dependent sex determination in turtles. Science, 206:1 186-1 188. Burger. J. 1976a. Behavior of hatchling diamondback terra- pins (Malaclemys terrapin) in the field. Co- peia. 1976:742-748. 86 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY 1976b. Temperature relationships in nests of the northern diamondbaek terrapin. Malacle- mys terrapin terrapin. Herpetologica, 32:4 1 2- 418. 1977. Determinants of hatehing success in dia- mondbaek terrapin, Malaclemys terrapin. Amer. Midi. Natur.. 97:444-464. Burger, J. and Montevecchi, W. A. 1975. Nest site selection in the terrapin Malacle- mys terrapin. Copeia. 1975:1 13-1 19. Burnley, J. M. 1969. Diamondbaek terrapin. Int. Turtle and Tor- toise Soc. J.. 3:32-34. Bury. R. B. 1979. Population ecology of freshwater turtles. Pp. 571-602. In Harless, M. and Morlock, H. (eds.), Turtles, Perspectives and Research. Wiley-Interscience, New York. Cagle. F. R. 1952. A Louisiana terrapin population (Malacle- mys). Copeia, 1952:74-76. Carr, A. F. 1952. Handbook of turtles. Comstock Publ. As- soc, Ithaca. 542 p. Cochran, J. D. 1978. A note on the behavior of the diamondbaek terrapin, Malaclemys t. terrapin (SchoepfF) in Maryland. Bull. Md. Herpetol. Soc, 14: 100. CONANT, R. 1975. A field guide to reptiles and amphibians of eastern and central North America. Hough- ton Mifflin Co., Boston. 429 p. Ernst, C. H. 1971a. Growth of the painted turtle, Chrysemys pic- ta. in southeastern Pennsylvania. Herpe- tologica, 27:135-141. 1971b. Population dynamics and activity cycles of Chrysemys picta in southeastern Pennsyl- vania. J. Herpetol., 5:151-160. 1975. Growth of the spotted turtle, Clemmvs gut- tata. J. Herpetol., 9:313-318. 1977. Biological notes on the bog turtle, Clemmys muhlenbergii. Herpetologica, 33:241-246. Ernst, C. H., Barbour, R. W. and Hershev, M. F. 1 974. A new coding system for hardshelled turtles. Trans. Kentucky Acad. Sci., 35:27-28. Finneran, L. C. 1948. Diamond-back terrapin in Connecticut. Co- peia, 1948:138. Fitch, H. S. 1981. Sexual size differences in reptiles. Misc. Publ., Mus. Natur. Hist., Univ. of Kansas 70: 1-72. Gibbons, J. W. 1967. Variation in growth rates in three popula- tions of the painted turtle, Chrysemys picta. Herpetologica. 23:296-303. 1 968. Population structure and survivorship in the painted turtle, Chrvsemvs picta. Copeia, 1968:260-268. 1970. Sex ratios in turtles. Res. Popul. Ecol., 12: 252-254. Graham, T. E. 1979. Life history techniques. Pp. 73-95. In Har- less, M. and Morlock. H. (eds.). Turtles, Per- spectives and Research. Wiley-Interscience, New York. HlLDEBRAND, S. F. 1929. Review of experiments on artificial culture of diamond-back terrapin. Bull. U.S. Bur. Fish., 45:25-70. 1932. Growth of diamondbaek terrapins size at- tained, sex ratio and longevity. Zoologica, 9: 551-563. Hurd, L. E., Smedes, G. W. and Dean, T. A. 1979. A ecological study of a natural population of diamondbaek terrapins (Malaclemys t. terrapin) in a Delaware salt marsh. Estuaries, 2:28-33. Jackson, C. G., Jr. and Ross, A. 1971. Molluscan fouling of the ornate diamond- back terrapin, Malaclemys terrapin macro- spilota. Herpetologica, 27:341-344. Jackson, C. G., Jr., Ross, A. and Kennedy, G. 1973. Epifaunal invertebrates of the ornate dia- mondbaek terrapin, Malaclemys terrapin macrospilota. Amer. Midi. Natur., 89:495- 497. Lawler, A. R. and Musick, J. A. 1972. Sand beach hibernation by a northern dia- mondbaek terrapin, Malaclemys terrapin terrapin (SchoepfF). Copeia, 1972:389-390. Moll, E. O. and Legler, J. M. 1971. The life history of a neotropical slider turtle, Pseudemys scripta (SchoepfF). in Panama. Bull. Los Angeles Co. Mus. Natur. Hist. 1 1: 1-102. Montevecchi, W. A. and Burger. J. 1975. Aspects of the reproductive biology of the northern diamondbaek terrapin. Malacle- mys terrapin terrapin. Amer. Mildl. Natur., 94:166-175. Nichols, J. D. and Chabreck, R. H. 1980. On the variability of alligator sex ratios. Amer. Natur.. 116:125-137. Ott, L. 1977. An introduction to statistical methods and data analysis. Duxbury Press. Belmont. 730 p. Overton, W. S. 1969. Estimating the numbers of animals in wild- life populations. Pp. 403-455. //; Giles, R. (ed.). Wildlife Management Techniques. Wildlife Soc Washington. Plummer, M. V. 1 977a. Activity, habitat and population structure in the turtle, Trionyx muticus. Copeia. 1977: 431-440. 1977b. Reproduction and growth in the turtle, Tri- onyx muticus. Copeia, 1977:441-447. Reid, G. K. 1955. Reproduction and development in the northern diamondbaek terrapin, Malacle- mys terrapin terrapin. Copeia, 1955:310-31 1. VERTEBRATE ECOLOGY AND SYSTEMATICA 87 Ross, A. and Jackson, C. G.. Jr. 1972. Barnacle fouling of the ornate diamondback terrapin, Malaclemys terrapin macrospilota. Crustaceana, 22: 203-205. Seigel, R. A. 1 978. Simultaneous mortality in the diamondback terrapin. Malaclemys terrapin tequesta Schwartz. Bull. N.Y. Herpetol. Soc, 14:31- 32. 1979. Reproductive biology of the diamondback terrapin. Malaclemys terrapin tequesta. Master's thesis, Univ. of Central Florida, Orlando. 40 p. 1980a. Predation by raccoons on diamondback ter- rapins, Malaclemys terrapin tequesta. J. Herpetol., 14:87-89. 1980b. Courtship and mating behavior of the dia- mondback terrapin, Malaclemys terrapin te- questa. J. Herpetol., 14:420-421. 1980c. Nesting habits of diamondback terrapins (Malaclemys terrapin) on the Atlantic Coast of Florida. Trans. Kansas Acad. Sci., 83:239- 246. 1983. Occurrence and effects of barnacle infesta- tions on diamondback terrapins (Malacle- mys terrapin). Amer. Midi. Natur.. 109:34- 39. Sergeev, A. 1937. Some materials to the problem of reptilian post-embryonic growth. Zool. J. Moscow, 1 6: 723-735. Sexton, O. J. 1959. A method for estimating the age of painted turtles for use in demographic studies. Ecol- ogy. 40:716-718. Wood, R. C. 1976. $25 per egg. N.J. Outdoors. 3:14-15. 26. Yearicks, E. F., Wood, R. C. and Johnson, W. S. 1981. Hibernation of the northern diamondback terrapin, Malaclemys terrapin terrapin. Es- tuaries, 4:78-80. Yntema, C. L. 1979. Temperature levels and periods of sex de- termination during incubation of eggs of Chelydra serpentina. J. Morphol., 159:17- 28. Vertebrate Ecology and Systematics— A Tribute to Henry S. Filch Edited by R A. Seigel. L. E. Hunt. J. L. Knight. L. Malaret and N. L. Zuschlag < 1484 Museum of Natural History. The University of Kansas. Lawrence An Ecological Study of the Cricket Frog, Acris crepitans Ray D. Burkett Introduction The cricket frog, Acris crepitans, is a useful subject for population studies since it is generally abundant throughout most of the year and tends to form separate and distinct populations. It oc- curs in a variety of habitats, such as along lakes, pond, rivers, streams and occasionally tempo- rary ponds or rain pools and even relatively dry stretches of intermittent streams. Most Acris ag- gregate on relatively level, bare areas at the water's edge, avoiding steep, vegetation-covered slopes in most instances. Cricket frogs venture into water away from the shore line only when mats of algae are present on the surface. Earlier knowledge of the ecology of Acris was based mainly on short notes summarized by Wright and Wnght (1942). More recent studies include those by Turner (1960b) and Ferguson el al. (1965) on Acris gryllus\ and those by Py- burn (1958, 1961a, 1961b), Blair (1961), Fer- guson et al. (1967), Bayless (1969b), Labanick ( 1 976), and Johnson and Christiansen ( 1 976), on A. crepitans. Nevo (1973a, 1973b) has studied both species and Bayless (1969a) studied sym- patric populations of both species. Some comparisons were made with popula- tions of Acris from other locations, but the main objective of this study was to determine if there were any differences in the ecology of popula- tions of cricket frogs in different habitats sepa- rated by only a few kilometers. Description of Study Areas Populations of Acris were studied in and near Lawrence, Douglas County, Kansas, in the fall of 1 96 1 , and from fall 1 963 through spring 1 966. Two populations were studied intensively by capture/recapture and toeclipping: one in a wooded pond and stream at the University of Kansas Natural History Reservation (KUNHR). about 1 1 km northeast of Lawrence, and the oth- er in an open reservoir and 1 1 rectangular ponds at the University Fish Laboratory (FL) on the southwestern part of the campus. The Kansas River lies between the two populations as a pos- sible barrier to gene flow. The pond at the Reservation was created by the construction of an earthen dam impounding water on its northeastern side. Water overflowing the pond drains down a stream to the southwest and into a small creek that empties into the Kan- sas River about four km east of Lawrence. The pond has a maximum circumference of about 435 m. The northeastern end of the pond is shal- low and swampy with numerous willows (Salix) along its edge. Honey locust (Gleditsia triacan- thos) borders much of the northern edge and northwestern edge of the dam. The southeastern end of the pond is almost always shaded by large oaks (Quercus velutina), elms (Ulmus ameri- cana) and ash (Fraxinus americanus). Much of the northwestern edge of the pond and dam are bordered by small trees, shrubs, herbs and grass- es. Algae are common in the pond in a zone from about 0.3 to 0.9 m from shore. For a detailed description of the Reservation see Fitch (1952, 1965) and Fitch and McGregor (1956). At the Fish Lab the reservoir is on a south- facing slope and the 1 1 ponds are located about 90 m south of the reservoir. Each pond is drained through pipes that empty into a small stream south of the ponds. The stream continues south until it reaches the Wakarusa River, which enters the Kansas River about 1 1 km east of Lawrence. The reservoir fluctuates considerably in depth since water is used to fill the ponds. The maxi- mum circumference of the reservoir during my study was about 365 m, and the minimum 230 m. The only trees around the reservoir are small saplings of Populus and Salix, which occur in about equal numbers. Methods A total of 2492 frogs were captured at the Nat- ural History Reservation, and 1077 were cap- tured at the Fish Laboratory. Owing to the large numbers of individuals that were sometimes present, frogs were marked serially rather than individually. Areas (not exceeding 100 m long and 4.5 m wide) were marked off at each locality. All individuals captured in each area were given a unique mark for that area and date. Frogs cap- tured during the initial sampling period and sub- 89 90 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 1. Estimated reproductive output for. 4cris cre- pitans in two populations in northeastern Kansas. See text for explanation. Location Year Approx. no. in pop. in Ma\ \ppro\ no. of females Estimated no. of eggs laid (200-275 per female) KUNHR Pond 1964 90 27 5400-7425 1965 318 95 19,000-26,125 Stream 1966 100 30 6000-8250 Fish Lab Ponds 1965 68 32 6400-8800 Reservoir 1965 47 22 4400-6050 1966 24 11 2256-3025 sequently recaptured, were given an individual mark on the third recapture. Frogs found un- marked at the second or subsequent sampling periods were given a unique mark. The place of capture, precise to the nearest meter, was re- corded whenever frogs were given individual markings and during subsequent recaptures. Whenever new groups were being marked, they were captured in an area of pond-margin up to 6 m in length, then released in the center of the area. Results and Discussion Breeding Season The breeding season of Acris varies somewhat with geography and weather. Periods of calling near Austin, Texas have been reported as early as 30 January and as late as 10 September (Blair 1961). Smith (1961) stated that in Illinois calling occurs from late April until late summer; in Iowa, Johnson and Christiansen (1976) reported call- ing from mid-May until late July. In this study, large choruses were heard as early as late April in 1965 and mid-May in 1964 and 1966, and as late as 23 June 1964. Calling ceased by the last week in July. Each year calling began in the day- time; but later, as temperature increased, calling also occurred at night. Observation of gravid and post-partum fe- males indicates that most spawning occurred from late May to early July. Eggs became fully devel- oped by about mid-April, and almost all females remained gravid until the end of May. Post-par- tum females first appeared in the populations on 6 June 1964, 31 May 1965 (KUNHR) and 29 June 1965 (FL). In the 13 July 1965 sample (KUNHR), some females were fully gravid, but most had only a few eggs remaining in one or both oviducts. One female containing only a few eggs was found as late as 3 August, and it appears that females may either retain a few eggs, or mate at least twice during the breeding season, laying a portion of their eggs each time they mate. In Kansas, however, there is no second breeding involving young-of-the-year as reported by Py- burn in Texas (Blair 1961). Metamorphosed young first appeared about 1 July, and incompletely metamorphosed frogs still retaining tails were observed as late as 29 Sep- tember. Larval development generally requires from five to ten weeks in northeastern Kansas (Burkett 1969), and transformation into frogs takes about two days (Wright and Wright 1942). In northern Texas, I have found motile sperm in young males in early September. In the Kansas populations sperm were grouped in clumps in late September, but by early October nearly all males had well-formed sperm. A few males also develop chin spotting in the fall, but the vocal pouch is formed in late March or April, preced- ing the breeding season. These findings are in contrast to Brenner's ( 1 969) conclusions that Ac- ris must overwinter before attaining sexual ma- turity. Sex Ratio and Reproductive Potential Tadpoles and juveniles of Acris are easy to sex if dissected, as the testes are well formed and black. The sex ratio in newly metamorphosed Acris was found to be about four females per male, but in frogs over three months old, males usually were predominant. From September un- til the following July, males averaged between Fig. 1. Growth rate of two populations of Acris crepitans, illustrated by combined samples for each month. Range, mean, one standard deviation, and two standard errors are indicated for each sample. Numbers above or below each line indicate sample size. April to July samples are subdivided to indicate males, left; entire sample, center, and females, right. VERTEBRATE ECOLOGY AND SYSTEMATICS 91 35 361 288 54 148 JO 30h U) c -J 25 890 142 220 C (D ^ 20 +-» =J o " 151 47 ||. II ii 10 L 274 65 J I69 n II ' J J 1 32 33ri i \ 91 u ir J- 127 J L J L JASON MAM J J KUNHR 35 141 T 97 56 JZ 30 -4-> U) c 45 50 50(100) 41-45 47 46(97.9) 36-40 30 29(96.7) 31-35 16 2(12.5) <31 19 0(0) Table 2. Opheodrys aestivus: Presence of cloacal sperm in females (>35.0 cm SVL) at different times of the year. ovaries throughout the sampling period (Fig. 1). Follicles measuring 1-3 mm proliferate in post- reproductive females in July and August and in- crease in size to a maximum 5 mm by October. Two snakes collected in mid-February had a mean of 9 follicles measuring 1-3 mm and 9.5 follicles measuring 3-5 mm. The February data are sim- ilar to the September-October data and indicate that very little follicular activity occurs over win- ter. Vitellogenesis resumes in spring. Rapid yolk- ing and enlargement to approximately 1 5-25 mm occur in May and oviducal eggs are present from late May to early July. In 8 pre-ovulatory females which contained enlarged (>12 mm) follicles the mean number in the left and right ovaries was 2.6 ± .26 and 3.6 ± .38, respectively. Fourteen of 23 post-ovu- latory females (60. 1%) had a disparity in the cor- responding number of corpora lutea and ovi- ducal eggs on each side. In 1 snakes extrauterine transfer of ova involved only 1 ovum and in 2 snakes movement of 2 ova was involved. A sig- nificantly greater number of ova moved from the left ovary to the right oviduct than vice versa (x 2 = 7.14, P < .01). Reciprocal transfers, of course, could not be detected. The average num- ber of oviducal eggs in the left and right oviducts was 2.2 ± .15 and 4.1 ± .23, respectively. It appears that the right oviduct, which is longer than the left, receives more ova because of great- er ova production by the right ovary and because of greater extrauterine transfer from the left ovary. Oviposition dates in the laboratory from fe- males collected 14 June- 14 July 1979 were 1-7 July (N= 11); 8-15 July (N = 18); 16-23 July (N = 11). Oviposition in 1978 apparently oc- curred earlier (Table 3) or perhaps laboratory stress caused a delay in oviposition in 1979. Snakes collected in 1979 seemed to be delayed in follicular enlargement and ovulation as de- termined by palpation. Body fat (Fig. 1 , Table 4) is the greatest in early Period Snakes (N) Snakes (N) with cloacal sperm (%) Apr-May Jun-Aug Sep-Oct 10 16 17 8' (80.0) 2 2 (12.5) (0) 1 Spermatozoa extremely abundant. 2 Only 3 spermatozoa detected in 2 samples from each of 2 snakes collected in June. spring and late fall and least in June. Two Feb- ruary females averaged 8.3 (range = 5.3-11.4). These data indicate that fat is not depleted in overwintering snakes. However, the significant reduction of fat during secondary vitellogenesis and ovulation from late May to June (/ = 9.27; P < .001) suggests that energy needed for these processes is derived from fat reserves. Although the pattern of fat cycling resembles that in ju- veniles (Table 4), an abrupt significant decrease in fat is not evident. Additionally, mature fe- males have more fat overall than do juveniles (r = 3.17, P < .01). Clutch Size. — The number of eggs in a clutch was determined by counting the eggs actually deposited (42 snakes), by counting oviducal eggs (28 snakes), or by counting enlarged (> 12 mm) ova (8 snakes). Clutch size in 77 snakes was 6.1 ± .21 (range 3-10). Estimates of clutch size by these methods are subject to some error. En- larged, atretic follicles (N = 4) were present in the ovaries of 3 snakes which had oviducal eggs. Additionally, 1 1 unfertilized, unshelled ova were deposited singly and separate from clutches among the 40 laboratory snakes. The relative number of enlarged follicles that do not termi- nate as a part of an actual clutch, therefore, ap- pears to be small. This fact along with the fact that only 8 of 77 clutch size estimates were based solely upon number of enlarged ova probably has not biased the clutch size estimate greatly. Clutch size averaged 6.3 ± .59 eggs based on number of enlarged follicles which is not significantly dif- ferent from the total estimate (/ = .20; P > .50). No evidence was seen which indicated that more than 1 clutch was produced in a given year. Incubation and Hatching. — Of 180 eggs laid and incubated in the laboratory, 1 6 1 (89.4%) were fertile. One hundred forty-five (90.0%) of the fer- tile eggs hatched. Of 19 infertile eggs, 18 (94.7%) VERTEBRATE ECOLOGY AND SYSTEMATICA 107 CO Ld _l CJ 10 b 2 tr UJ DD iioh < !±! 6h 2 I APR- 15 MAY JUL r -I — i i i— II OE 16-31 MAY 16 w=u* AUG +' I I I JUN I SEP- io oct HI2 HP I I I 12 O 8 2 II OE 3 II OE DIAMETER OF FOLLICLES (mm) C7> I- >- O O CO -8 _ < Fig. 1. Opheodrys aestivus: Number of various sized follicles and amount of body fat in mature different times of the year. The horizontal lines of the dicegrams are means; vertical lines are ranges: delimit 95% confidence limits. OE = oviducal eggs. females at rectangles were either partially or totally unshelled. For totally shelled eggs the modal embryonic stage (Zehr 1 962) at oviposition was 25 (N = 37, range 21-27). The range of stages in a single clutch of 7 eggs was 25-26. For partially shelled eggs or those that were inviable at or soon after ovipo- sition (determined by the rapid loss of tonicity and growth of mold) the mode was 18 (range = 14-19; N = 10). Statistics relating to egg size are given in Table 5. Incubation ranged 36-43 days and averaged Table 3. Opheodrys aestivus: Number of snakes (>35.0 cm SVL) determined to be gravid by palpation in the field during various times in 1978. Period Snakes (N) No. gravid (%) 5-20 Jun 39 39(100) 22-30 Jun 32 20(62.5) 2-7 Jul 29 11 (37.9) 12-31 Jul 50 0(0) 39.2 ± .12 (N = 142). Hatching occurred 9-30 August in the laboratory. Hatchlings were seen in the field on 31 August and on 10 September in 1978. Statistics relating to hatchling size are sum- marized in Table 5. Of 141 hatchlings. 67 (47.5%) were males. Sex ratio is not significantly different from 1:1 ( x 2 = .35; P > .50). Male hatchlings weigh about the same as females (1.37 ± .023, Table 4. Opheodrys aestivus: Percent body fat of ju- venile females and adult females. Data are x ± 1 SE (N) fat (g)/body weight (g) x 100. Period Juvenile Adult Apr- 15 May 4.1 ± 1.17(5) 7.2 ± .62(14) 16-31 May 1.7 ± .11 (2) 7.9 ± .57(16) Jun 2.1 ± .49 (9) 1.9 ± .20(30) Jul 2.3 ± .34 (9) 2.3 ± .38(16) Aug 2.4 ± .54 (7) 5.1 ± .57(6) Sep-Oct 3.2 ± .42 (6) 8.6 ± .77(9) 108 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY i.o- X Ll) >■ Q O CD X UJ X o I- _J ° 35 40 45 50 SNOUT-VENT LENGTH (cm) Fig. 2. Opheodrys aestivus: The relationship of clutch weight/post-reproductive body weight and snout-vent length for snakes collected in 1979. The regression equation is Y = .00 IX + .588 (r = .024, P > .75). 1.36 ± 0.23 g; / = .01, P > .90) and have about the same SVL (13.9 ± .10, 14.0 ± .10 cm; / = .02, P > .90). Reproductive Effort. — Reproductive effort of a female is that organism's total investment in a current act of reproduction (Pianka 1976). In snakes reproductive effort has been crudely es- timated using the ratio of clutch weight to non- reproductive female weight (C/B) (Clark 1970; Fitch 1975; Pianka and Parker 1975; Shine 1977). Because in O. aestivus there is no parental care (egg brooding or oviducal retention) most of the reproductive investment should be contained in the egg itself and therefore the ratio C/B should be representative of reproductive effort (but see Vitt and Congdon 1978). The risks involved in transporting the enlarged ova and eggs in the maternal body are assumed to be negligible. In O. aestivus C/B averages .64 and does not change with body size (Fig. 2). Less than .06% of the variation in C/B is explained by body size. Larger snakes produce both larger eggs and larger clutch- es (Fig. 3). There is a possible trend toward small- er eggs with increasing clutch size (Fig. 4) al- though there is great variation (r 2 = 1.7%) and the regression is not significant. Larger eggs pro- duce significantly larger hatchlings (Fig. 5). Discussion Opheodrys aestivus appears to have a typical female reproductive cycle for a temperate ovipa- rous snake. From the limited data available for a comparison of geographic variation in repro- ductive attributes, other reports appear to con- form with this population. In southern Louisiana Tinkle ( 1 960) and in Illinois Morris (1982) found similar results in O. aestivus with regard to size at sexual maturity, the ovarian cycle, and repro- ductive potential. Apparently, mating is limited to spring in this population although fall mating may occur in other populations (Richmond 1956). Table 5. Opheodrys aestivus: Egg and hatchling statistics. All data are expressed as x ± 1 SE (N); range. Wgt. (g) Max. width (cm) Max. length (cm) SVL (cm) Shelled eggs Hatchlings 1.62 ± .015(190); 1.17-2.26 1.37 ± .016(144); .82-1.76 9.9 ± .04(190); 8.4-11.9 24.8 ± .23(190); 16.2-34.2 13.9 ± .07(144); 10.7-16.1 VERTEBRATE ECOLOGY AND SYSTEMATICS 109 10- LiJ 8 M CO 5 6 h- _l o 4 r- X C5 UJ 5J < LxJ 2 - 1.8 - .6- w 1.4 1.2- _ •* • •• - ® • • • • • - • • •• - • •••• • •••••^X' • • - • •• • •• •J^-^»»#« • • • - • •• - • ^- ••••••• • • - 1 • •• 1 1 1 1 1 • - ® • • " • - • • • ^ — - • • • ^**"* • - • • • " ^ • • - • • • - • • - ' _i — i i i , • — i — i 35 40 45 50 SNOUT- VENT LENGTH (cm) Fig. 3. Opheodrys aestivus: A. The relationship of clutch size and snout-vent length. The regression equation is Y = .23X - 4.1 (r = .58, P < .001). B. The relationship of egg weight and snout-vent length. The regression equation is Y = .02X + .61 (r = .46, P < .05). McCauley (1945) observed mating behavior on 18 May of a captive male O. aestivus directed toward a female O. vernalis. The data on egg and hatchling size, clutch size, and dates of ovipo- sition and hatching of several anecdotal reports (Conant 1938; Conant and Downs 1940; Mc- Cauley 1945; Curtis 1950; Guidry 1953; Car- penter 1958; Sabath and Worthington 1959; Smith 1961; Anderson 1965; Webb 1 970; Mount 1975; Morris 1982; and others summarized in Wright and Wright 1957) generally are similar to the present report. The greatest differences are in time of oviposition and in incubation time. Since the timing of these probably is dependent upon phenological events and temperature, re- spectively, it is not surprising that they are so variable. The extremely wide variation in ovi- position dates (17 June-28 August) and subse- quent hatching in southeastern Texas (Guidry 1953) suggests the possibility of multiple clutch- ing. Seemingly, enlargement of follicles to the 5 mm stage occurs throughout the activity season and could be described as primary vitellogenesis (sensu Aldridge 1979). The rapid enlargement from the 5 mm follicle to ovulatory size is con- fined to late April and May and may be described as secondary' vitellogenesis (sensu Aldridge 1 979). 110 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY • _ 1.9 - • • u> • • h- • X S2 1.7 • • • • UJ £ • • • • CD • CD • ■ w 1.5 - • • ^__g.^ z • • < • UJ • s 1.3 • • • • i i i 1 • | 6 8 CLUTCH SIZE 10 Fig. 4. Opheodrvs aestivus: The relationship of egg weight and clutch size. The regression equation is Y .014X + 1.65 (r= -.13. P < .50). Aldridge (1979) described Type I secondary vi- tellogenesis in which rapid yolking of ova is con- fined to spring and Type II secondary vitello- genesis in which rapid yolking begins in late summer and fall, becomes dormant in winter, and resumes in spring. He classified O. aestivus as having a Type II pattern based upon Tinkle ( 1 960). My interpretation of Tinkle ( 1 960) is that of a Type I pattern, as is that of the present study. Although not mentioned by Aldridge, I suspect that intraspecific vitellogenic patterns might vary geographically. Differential production of ova by left and right ovaries and extrauterine transfer of ova are prob- ably simple consequences of space limitations. Shine (1977) stated that it is difficult to see how extrauterine transfer would affect clutch size and embryonic survivorship and therefore would be of doubtful selective importance. However, be- cause in snakes the right oviduct is longer than the left, production of ova by the left ovary be- yond that which the left oviduct can accom- modate probably would result in greater follic- ular atresia were it not for differential extrauterine transfer to the right oviduct. The decision to stage the embryos at ovipo- sition was prompted by the data of Blanchard (1933) for O. vernalis, who found incubation to vary from 4 to 23 days. This species is a possible example in the evolutionary transition from ovi- parity to viviparity (Packard et al. 1977) with a relatively short incubation period and a presum- ably wide range in the degree of embryonic de- velopment at oviposition. Opheodrvs aestivus, however, has little variation in embryonic de- velopment at oviposition and has an incubation period similar to many other oviparous colu- brids (Fitch 1970). Determination of oviposition dates in the field by palpation assumes that each mature female breeds annually. This apparently is the case as all 39 females palped 5-20 June in 1978 were gravid (Table 3). all 40 mature females collected for the laboratory oviposited in 1979, and all 1 25 females autopsied April-July (1977-1979) were in breeding condition or had evidences of recent oviposition. The annual ovarian cycle is corre- lated with an annual fat cycle which presumably provides much of the energy needed for repro- duction. Energy, in the form of stored fat reserves or as an outcome of foraging success, has been implicated as the major factor in the control of VERTEBRATE ECOLOGY AND SYSTEMATICS 11 1.4 16 1.8 EGG WEIGHT (g) 2.0 Fig. 5. Opheodrvs aestivus: The relationship of hatchling weight and egg weight. The regression equation is Y = .79X + .115 (r= .85. P < .001). frequency of female reproduction in several snakes (summarized in Wharton 1966; Gibbons 1972). Production of equal numbers of male and fe- male hatchlings of similar size and weight is in accordance with Fisher's sex ratio theory and is the usual situation in snakes (Shine and Bull 1977). Because metabolism decreases with body weight in snakes (Galvao et al. 1965) propor- tionally more energy may be available for repro- duction. The risks involved when time and en- ergy are allocated to reproduction may decrease survivorship and therefore the expectation of fu- ture progeny (reproductive value). Therefore, a younger snake with a higher expectation of future progeny might be expected to devote less time and energy to reproduction than an older snake which has less expectation (Pianka and Parker 1975; Pianka 1976). Tests of this hypothesis in snakes have shown diverse results. In Carpho- phis vermis (Clark 1 970) C/B increases with body size(=age). In Diadophis punctatus (Fitch 1975), Masticophis taeniatus (Pianka and Parker 1975), and O. aestivus (present study) C/B remains con- stant with body size. In Notechis scutatus and Pseudechis porphyriacus (Shine 1977) C/B de- creases with body size. Pianka and Parker ( 1 975) and Pianka ( 1976) suggested that correlations be- tween reproductive effort and reproductive value might be greater in multiple-brooded species than in single-brooded species where proximal factors such as resource availability might have a greater effect. In all of the above studies the snakes were single-brooded. However, in a study of annual reproductive variation in O. aestivus (Plummer 1983) it was shown that C/B and other repro- ductive attributes did not vary between years in which snakes stored greatly different quantities of lipids. Even if reproductive effort remains con- stant with age (as in O. aestivus), the absolute energy allocated to reproduction actually in- creases. The increased energy available in O. aes- tivus is reflected in the production of larger eggs and larger clutch sizes (Fig. 3). Fecundity in snakes is often related to body size (Fitch 1970; Shine 1977; Aldridge 1 979; present study). Shine ( 1 978) found that in about 66% of species (including O. 112 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY aestivus) females attain a larger body size than males. Shine suggested that one reason for this disparity was that selection has favored large body sizes in the females because of greater fecundity. Another reason for increased body size might be that larger snakes produce larger eggs which pro- duce larger hatchlings (Fig. 5). In general, larger hatchlings should enjoy higher survivorship and be better competitors (Pianka 1976). In the lizard Sceloporus undulatus (Ferguson and Bohlen 1978) larger hatchlings from late broods enjoy greater survivorship than do smaller hatchlings, but larger hatchlings from early broods have survi- vorship similar to smaller hatchlings. Although to my knowledge there are no comparable data for snakes, if female fitness was increased by pro- ducing larger eggs then selection should favor either larger parental body size or decreased clutch sizes (Pianka 1976). Smith and Fretwell ( 1974), Pianka (1976), and Stewart ( 1979) discuss models which predict that with a constant reproductive effort, an increased female size may result in either larger clutches or larger sized eggs. These models assume a neg- ative correlation between clutch size and egg weight. In O. aestivus there is no statistical re- lationship between clutch size and egg weight (Fig. 4). Although the correlation between SVL and egg size is not strong (/' = .46), it appears that in this population correlates of female body size are selection for increased clutch size as well as for increased egg size. Summary Various aspects of female reproduction in Opheodrys aestivus were examined by specimen autopsy and from the study of living snakes in the field and in the laboratory. These snakes ma- ture at 36-40 cm SVL and breed annually there- after. Ovarian follicles measuring 1-5 mm in di- ameter are present in mature snakes throughout the year. Rapid yolking of follicles occurs in the spring and ovulation begins in late May. Extra- uterine transfer of ova is common. One clutch is produced per year (.v = 6. 1 eggs). Coelomic fat bodies cycle annually and presumably provide energy for vitellogenesis and ovulation. Ovipo- sition occurs in late June and July. Ninety per- cent of the eggs laid were fertile and 90% of fertile eggs hatched in the laboratory. At oviposition the modal embryonic stage (Zehr 1962) was 25 (range 21-27). Incubation averaged 39 days. The sex ratio of hatchlings was not significantly dif- ferent from 1:1. Male and female hatchlings are similar in length and weight. Reproductive effort (jc = .64) did not change with body size. Larger females produce both larger clutches and larger eggs. Larger eggs produce larger hatchlings. Acknowledgments I thank several students who were involved in various aspects of this project. They are T. M. Baker, F. W. Brown, D. B. Farrar, M. W. Pat- terson, D. E. Sanders, and M. White. W. B. Rob- erson assisted in the laboratory. R. A. Aldridge, J. S. Jacob, R. Shine and an anonymous reviewer made helpful suggestions regarding the manu- script. M. Groves and J. Huckeba willingly typed the numerous versions of the manuscript. I owe much to H. S. Fitch who stimulated and refined my interests in living organisms under natural conditions. To him this paper, and this entire volume, is rightfully dedicated. This study was supported in part by grants from Sigma Xi and Harding University. Literature Cited Aldridge, R. D. 1 979. Female reproductive cycles of the snakes Ar- izona elegans and Crotalus viridus. Herpe- tologica, 35:256-261. Anderson, P. 1965. The Reptiles of Missouri. Univ. Missouri Press. 330 p. Blanchard, F. N. 1933. Eggs and young of the smooth green snake, Liopeltis vernalis (Harlan). Papers Michigan Acad. Sci., Arts, Letters, 17:493-508. Carpenter, C. C. 1958. Reproduction, young, eggs and food of Okla- homa snakes. Herpetologica, 14:1 13-1 15. Clark, D. R. 1970. Age-specific "reproductive effort"' in the worm snake Carphophis vermis (Kennicott). Trans. Kansas Acad. Sci., 73:20-24. Conant, R. 1938. The Reptiles of Ohio. Amer. Mid. Nat., 20: 1-200. Conant, R. 1975. A field guide to reptiles and amphibians of eastern and central North America. 2nd ed. Houghton Mifflin Co., 429 p. Conant, R. and Downs, A., Jr. 1940. Miscellaneous notes on the eggs and young of reptiles. Zoologica, 25:33-48. VERTEBRATE ECOLOGY AND SYSTEMATICS 113 Curtis, L. 1950. A case of twin hatching in the rough green snake. Copeia, 1950:232. Ferguson, G. W. and Bohlen, C. H. 1978. Demographic analysis: a tool for the study of natural selection of behavioral traits. In Greenberg, N. and MacLean, P. D. (eds.). Behavior and Neurology of Lizards. NIMH. DHEW Publ. No. 77-491. Fitch, H. S. 1970. Reproductive cycles in lizards and snakes. Univ. Kansas, Mus. Nat. Hist., Misc. Publ., 52:1-247. 1975. A demographic study of the ringneck snake {Diadophis punctatus) in Kansas. Univ. Kansas, Mus. Nat. Hist., Misc. Publ., 62:1- 53. Galvao, P. E., Tarasantchi, J. and Guertzen- STEIN, P. 1965. Heat production of tropical snakes in rela- tionship to body weight and body surface. Am. J. Physiol., 209:501-506. Gibbons, J. W. 1972. Reproduction, growth and sexual dimor- phism in the canebrake rattlesnake (Cwtalus horridus atricaudatus). Copeia, 1972:222- 227. Guidrv, E. V. 1953. Herpetological notes from Southeastern Texas. Herpetologica, 9:49-56. McCauley, R. H., Jr. 1945. The Reptiles of Maryland and the District of Columbia. Hagerstown. Maryland, 194 p. Morris, Michael A. 1982. Activity, reproduction, and growth of Opheodrys aestivus in Illinois (Serpentes: Colubridae). Nat. Hist. Misc., 214:1-1 1. Mount. R. H. 1975. The Reptiles and Amphibians of Alabama. Auburn Univ. Agric. Exp. Station, 347 p. Packard, G. C, Tracy. C. R. and Roth, J. J. 1977. The physiological ecology of reptilian eggs and embryos, and the evolution of viviparity within the class Reptilia. Biol. Rev.. 52:71- 105. PlANKA, E. R. 1976. Natural selection of optimal reproductive tactics. Am. Zool.. 16:775-784. Pianka, E. R. and Parker, W. S. 1975. Age-specific reproductive tactics. Amer. Nat., 109:453-464. Plummer, M. V. 1981. Habitat utilization, diet and movements of a temperate arboreal snake (Opheodrys aes- tivus). J. Herpetol., 15:425-432. 1983. Annual variation in stored lipids and repro- duction in green snakes (Opheodrys aesti- vus). Copeia. 1983:741-745. Richmond, N. D. 1956. Autumn mating of the rough green snake. Herpetologica, 12:325. Sabath, M. and Worthington, R. 1959. Eggs and young of certain Texas reptiles. Herpetologica, 15:31-32. Schoener, T. W. 1977. Competition and the niche. In Gans. C. and Tinkle, D. W. (eds.). Biology of the Reptilia. Academic Press, pp. 370. Shine, R. 1977. Reproduction in Australian elapid snakes II. Female reproductive cycles. Aust. J. Zool., 25:655-666. 1978. Sexual size dimorphism and male combat in snakes. Oecologia, 33:269-277. Shine, R. and Bull, J. J. 1977. Skewed sex ratios in snakes. Copeia. 1977: 228-234. Smith, C. C. and Fretwell, S. D. 1 974. The optimal balance between size and num- ber of offspring. Am. Nat., 108:499-506. Smith. P. W. 1961. Amphibians and Reptiles of Illinois. Bull. Illinois Nat. Hist. Survey, 28:1-298. Stewart, J. R. 1979. The balance between number and size of young in the live bearing lizard Gerrhonotus coeruleus. Herpetologica, 35:342-350. Tinkle, D. W. 1960. A population of Opheodrys aestivus (Reptil- ia: Squamata). Copeia, 1960:29-34. Turner, F. B. 1977. The dynamics of populations of squamates, crocodilians. and rhynchocephalians. Pp. 157-264. In Gans, C. and Tinkle, D. W. (eds.). Biology of the Reptilia. Academic Press. Vitt, L. J. and Congdon, J. D. 1978. Body shape, reproductive effort, and relative clutch mass in lizards: Resolution of a par- adox. Amer. Nat. 112:595-608. Webb, R. G. 1970. Reptiles of Oklahoma. Univ. Oklahoma Press, 370 p. Wharton, C. H. 1966. Reproduction and growth in the cotton- mouth Agkistrodon piscivorus Lacepede. of Cedar Keys, Florida. Copeia, 1 966: 149-1 6 1 . Wright, A. H. and Wright, A. A. 1957. Handbook of Snakes. Vol. I. Comstock Publ.. 564 p. Zehr. D. R. 1962. Stages in the normal development of the common garter snake, Thamnophis sirtalis sirtalis. Copeia, 1962:322-329. Vertebrate Ecologs and Systematics — A Tribute to Henry S. Fiteh Edited by R. A. Sergei. L. E Hunt. J. L. Knight. L. Malaret and N. L. Zuschlag c 1984 Museum of Natural History- The l ! niversil\ of Kansas. Lawrence Clutch Size in Iguana iguana in Central Panama A. Stanley Rand Introduction Iguana iguana, laying up to 6 dozen eggs in a single clutch, is one of the most prolific lizards in the new world. Prized as food and heavily hunted in many parts of its range, its conserva- tion and the possibilities for sustained yield har- vesting have been discussed (Fitch et al. 1983). Though reproductive potential is important in any understanding of population dynamics, only one detailed study of the number of eggs which female iguanas produce has been published. Fitch and Henderson (1977) for Nicaragua. The pres- ent paper describes the size and weight of clutch- es produced by female iguanas and their relation to female size in Panama. The clutch size and reproductive investment in Iguana iguana is compared with that described for other lizards. Materials and Methods Clutch size data were collected from females caught during the nesting season, late January to early March, between 1968 and 1980 in the vi- cinity of Panama City and Gamboa, Republic of Panama. Some females were caught and allowed to nest in a large outdoor enclosure, others were killed or found as fresh road kills and the eggs removed from the oviducts. The sample was not randomly selected from the population, rather, because of my interest in the relationship be- tween female size and clutch size, the few females deliberately shot were selected because they were very large or very small. The following mea- surements were taken on 30 females: snout-vent length (SVL), female wet weight without eggs, clutch wet weight before significant hydration and clutch size (number of eggs). In some cases clutch volume, clutch dry weight (oven dried at 105°C) were also measured. Regressions were compared using covariance (Snedecor 1956). Results The results from examination of females with oviductal eggs are given in Table 1 . Sizes of these animals were not distributed normally nor do they suggest that the population can be divided into age classes. Even though animals were col- lected to emphasize the extremes, only one fe- male was below 300 mm SVL. Because sampling was not random, ranges are probably more ac- curate representations of the population than are means. Female weight is closely correlated with SVL (N = 30, /• = 0.86. P < .001) (Fig. 1) par- ticularly if logs of both are plotted (N = 30, r = 0.91, P < .001). Number of eggs per clutch ranged from 9 to 71 (N = 30, mean = 40.6) and was closely pos- itively correlated with female size. The correla- tion of egg number with SVL (N = 30, r = 0.78, P < .001) (Fig. 2) is about equal to that with female weight (N = 30, r = 0.79. P < .001). A better predictor of the number of eggs that a female will lay, and one that can be used in the field, is her weight before she has laid her eggs (i.e., her own body weight plus the weight of her clutch) (N = 28. r = 0.88, P < .001) (Fig. 3). The weight of 28 clutches ranged from 84 to 1086 g (mean = 538 g) and is closely correlated with female size (N = 28, r = 0.83. P < .001). Mean egg weight per clutch (clutch weight/num- ber of eggs) ranged from 9.3 to 16.0 g (N = 28, mean = 13.1). Eggs within a clutch appear quite uniform in size. Larger females tended to lay larger eggs but the correlation of mean egg weight to SVL, though significant (N = 28, r= 0.55, P < .01), is not as high as the correlations already cited. The single very small female with her very- small eggs contributes greatly to this correlation; if she is excluded the correlation is lower (N = 27. /• = 0.35. .01 < P < .05). The water content of the eggs varied little, 57- 67% (N = 11, mean = 62.3%). There was no sig- nificant correlation of water content either with egg weight or female size. The percentage that the clutch contributed to the combined weight of female and clutch (rel- ative clutch mass of Vitt and Congdon 1978) ranged from 19.7 to 39.9% (N = 28. mean = 30.3%). It shows a weak positive correlation with female SVL (N = 28. r = 0.30. .01 < P < .05) which depends heavily on the single small fe- 115 116 SPECIAL PUBLICATION- MUSEUM OF NATURAL HISTORY Table 1. Clutch size in Panamanian Iguana iguana. Female Clutch Mean egg Relative Expenditure Percent SVL Number water in in mm Wt. in g of eggs weight in g weight in g clutch mass' per egg 2 clutch 224 289 9 83.7 9.3 22.5 2.5 62.8 304 843 26 371.0 14.3 30.6 1.2 61.9 310 814 26 315.6 12.1 27.9 1.1 65.8 316 910 26 319.0 12.3 26.0 1.0 61.8 320 774 25 320 994 24 293.1 12.2 22.8 0.9 57.5 330 1042 35 433.3 12.4 29.4 0.8 61.9 331 1136 40 476.1 11.9 29.5 0.7 64.2 340 907 30 407.8 13.6 31.0 1.0 340 1114 46 558.6 12.1 33.4 0.7 340 883 37 340 836 30 415.3 13.8 33.2 1.1 350 1219 29 464.9 16.0 27.6 1.0 350 1141 48 635.4 13.2 35.8 0.7 350 1019 35 462.7 13.2 31.2 0.9 350 1246 53 614.5 11.6 33.0 0.6 353 1322 56 693.0 12.4 34.4 0.6 67.0 359 1422 46 551.1 12.0 27.9 0.6 59.9 360 886 42 462.0 11.0 34.3 0.8 363 1341 42 568.1 13.5 29.8 0.7 62.0 370 902 46 585.6 12.7 39.4 0.9 380 1299 24 319.0 13.3 19.7 0.8 390 1369 28 445.1 15.9 24.5 0.9 390 1937 54 662.4 12.3 25.5 0.5 390 1639 56 742.7 13.3 31.2 0.6 397 1784 59 724.1 12.3 28.9 0.5 410 2306 69 926.1 13.4 28.7 0.4 410 1639 56 742.8 13.3 31.2 0.6 410 1721 49 717.2 14.6 29.4 0.6 430 1636 71 1085.8 15.3 39.9 0.6 N= 30 30 30 28 28 28 28 11 x = 354.2 1212.3 40.6 538.4 13.1 30.3 0.7 62.4 1 Percent relative clutch mass = clutch weight/(female weight + clutch weight) x 2 Percent expenditure per egg = relative clutch mass/number of eggs x 100. 100. male; if she is excluded r falls to 0.14 (P < .05). The percentage that a single egg represented of the combined weights of female and clutch (ex- penditure per egg) varied from 0.4 to 2.4% (N = 29, mean = 0.7%). It also shows a correlation with female weight (N = 29, r = 0.37, P < .01). Discussion and Conclusions In any lizard an upper limit to the possible volume of its clutch is set by the space available within the female for eggs. Two sorts of obser- vations suggest that most gravid female iguanas have reached this limit. First, most have their abdomen distended and on dissection are found to have their body cavity packed with eggs with no extra space, even for food. Second, most points in Fig. 2, relating clutch weight to female length, lie close to the regression line, and there are only a few females that have a smaller clutch weight than their length predicts. Since most females produce as large a clutch as there is space for. it seems probable that clutch size usually is not limited by food availability. Iguana clutches in Nicaragua were measured by Fitch and Henderson (1977). Table 2 shows that, comparing the population in Nicaragua with that in Panama, the slopes of the various regres- sion lines are not significantly different (P < .05), but that the elevations of some of them are. Com- pared with those in Panama, female iguanas in Nicaragua of the same size have lower body weights and lay lighter clutches. They also lay fewer and smaller eggs. These differences balance VERTEBRATE ECOLOGY AND SYSTEMATICS 117 2000 1000 — 400 - J_ ±. ± X X v x X x x x x x X x x L 250 300 350 SNOUT VENT LENGTH, IN MM WO 450 Fig. 1. The relationship between weight and snout-vent length in Panamanian iguanas. so that females of the same weight in two pop- ulations lay about the same number of eggs but have a lower total clutch weight in Nicaragua. The relative clutch mass and the investment per egg is not significantly different between the two populations. No other equally extensive data on Iguana iguana clutch sizes has been published. Hirth ( 1 963) reported SVL and clutch size for 7 females from Tortuguero, Costa Rica (Fig. 2). These clutch sizes are intermediate between those of Pana- manian and Nicaraguan females of similar sizes. This suggests the possibility of a geographical trend in reproductive strategies that would be worth exploring. However, it does not appear to continue into South America. Hoogmoed ( 1973) reports clutch sizes from 24 to 57 in Surinam. Muller (1972) reports clutches from 14 to 70 at Santa Marta, Colombia, and egg weights aver- aging 13.0 g (12.4-14.0). Detailed comparison with South American populations awaits more data. Wiewandt (1983) has compared reproductive patterns among iguanine lizards. He distin- guished three groups of genera on ecological grounds: 1) those in mainland deserts of tem- perate North America (Dipsosaurus and Sau- romalus); 2) those on dry subtropical islands (Cy- clura); and 3) those in mainland tropical areas (Iguana and Ctenosaura). The mainland tropical group grows the most rapidly, matures earliest, and has the largest clutch sizes and the lowest ratio of egg weight to female weight. In Iguana iguana, though its eggs are small relative to female size, the weight of its total clutch, relative to that of the female, is about the same as it is for the three other iguanine species for which Wiewandt gives data (Sauromalus obesus, Cyclura carinata and Cyclura coronuta stejnegeri). The marine iguana of the Galapagos was not classified in his scheme but is extreme within the iguanines in having very few, very- large eggs with a high investment per offspring but a low investment per clutch (Carpenter 1 966). Wiewandt attributes the reproductive pattern in Iguana iguana and Ctenosaura similis to the relatively high predation pressure on young liz- ards in these species. Tinkle et al. (1970) have reviewed the reproductive strategies of a wide taxonomic and geographical representation of lizards. That survey included few iguanines. or other large tropical herbivorous lizards and it is 118 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY 80 60 CO U3 5/|0 LLl 00 20" + X X X X X X * X X X + . XX , X + B B t X , + x B + 1 + + + X + ♦ + | * + < X X v + . * X X 200 250 300 350 SNOUT VENT LENGTH IN 400 450 Fig. 2. The relationship between clutch size and snout-vent length in iguanas from Panama (X); Nicaragua ( + ) (Fitch and Henderson 1977), and Tortuguero. Costa Rica (0) (Hirth 1963). interesting to compare our data with their results. Tinkle et al. recognize two different reproductive strategies among lizards: those that mature dur- ing their first year and lay more than one clutch per season, and those that delay maturation for more than a year and lay only a single clutch per year. Iguana iguana, and most iguanines (see Wiewandt for possible exceptions) clearly fall into 80 60 GO CD CD °<4G en UJ OQ 20 x x X X XX X *\ X X _ — I • 1 » 1 1 * 1 1 ( » 1 1 1 1 1 200 1000 2000 3000 FEMALE WEIGHT + CLUTCH WEIGHT IN GRAMS Fig. 3. The relationship, in iguanas from Panama, between clutch size and the weight of the female plus the weight of her clutch. VERTEBRATE ECOLOGY AND SYSTEMATICS 119 HO.. 30-- oo CD £20-- 10- ■ ++ *t + +4 + + + lol — ' ' ' — 260 — ' "" SNOUT VENT LENGTH AT MATURITY, IN 300 Fig. 4. The relationship between clutch size and minimum snout-vent length at maturity. + = single brooded iguanas (Tinkle et al. 1970), 1 = J 'vblyrhynchus cristatus, (Carpenter 1966), 2 = Sauromalus obesus, 3 = Cten- osaura similis, 4 = Cyclura cahnata, 5 = Cyclura coronuta stejnegeri (2-5, Wiewandt 1983), 6 = Iguana iguana from Panama, 7 = Iguana iguana from Nicaragua (Fitch and Henderson 1977). the second group. In contrast to the largely trop- ical iguanines, most of Tinkle et al.'s single brooded species are temperate in distribution. Fig. 4 plots the relationship between minimum size at first reproduction and mean clutch size for the species that Tinkle et al. included in their single brooded group as well as the data for Igua- na iguana from Panama and Nicaragua and 4 other iguanines. As Wiewandt noted, Ambly- rhynchus cristatus, Cyclura cahnata and C. co- ronuta stejnegeri and Sauromalus obesus have few eggs for their size. Fig. 4 shows that Cteno- saura similis and Iguana iguana have clutches close to but still slightly below those predicted from the smaller lizards. We do not have enough data to plot other iguanines but those for which we do have some data (Brachylophus faciatus, Sauromalus varius and S. hispidus, Conolophus subcristatus, and Iguana delicatissima) all seem to have clutch sizes below those oflguana iguana and Ctenosaura similis and those of the smaller, single brooded lizards. Number of eggs per clutch is an important parameter in a reproductive strategy. It is not, however, a very good index of reproductive effort (Tinkle and Hadley 1975; Vitt and Congdon 1978); in part, because a female iguana expends a great deal of energy in traveling to a nest site, digging a nest burrow, defending it, filling the burrow and returning to her home range (Rand and Rand 1976). Even for the clutch itself, clutch mass or calorific content is a better measure of reproductive effort than is number of eggs. We do not have calorific data for iguanas but Ballin- ger and Clark (1973) and Vitt (1978) have shown that calorific content per unit weight is quite con- stant for the eggs of a variety of numbers of liz- ards. Vitt (1978) has shown that the ratio of ca- lorific content of clutch to calorific content of female is similar to the ratio of wet weight of clutch to wet weight of female. That our ratio of dry to wet weight of iguana eggs lies within the range that Vitt reported for other lizards is sup- port for our assumption that iguana eggs are probably not too different from other lizards in calorific content per unit wet weight. For Iguana iguana, in Panama, wet clutch weight averages 30.3% of the total wet weight of female and clutch (relative clutch mass): this is close to the mean of 27.7% that Vitt and Cong- don (1978, Table 2) give for 17 much smaller North American iguanid lizards. Not surprising- 120 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 2. A comparison between clutch sizes of Iguana iguana from Panama and Nicaragua. Panama N r a b X = Snout-vent length (mm) 30 .86** -1872 8.708 Y = female weight (g) X = log snout-vent length 30 QJ** -9.62 2.841 Y = log female weight X = snout-vent length 30 .78** -59.4 .282 Y = number of eggs X = snout-vent length 28 .83** -936 4.141 Y = clutch weight X = snout-vent length 28 .55** 6.34 .0186 Y = mean egg weight X = snout-vent length 28 .30* 18.2 .033 Y = relative clutch mass' X = snout-vent length 28 .84** 3.59 -.008 Y = expenditure per egg 2 X = female weight 30 .80** 6.10 .0284 Y = number of eggs X = female weight 28 .82** 35.8 .4054 Y = clutch weight X = female + clutch weight 28 .89** 1.70 .0222 Y = number of eggs 1 Relative clutch mass = clutch weigh t/( female weight + clutch weight). 2 Expenditure per egg = relative clutch mass/number of eggs. 3 Panama minus Nicaragua. * = significant at the 5% level. ** = significant at the 1% level. ly, Iguana iguana, which is an arboreal, highly cryptic herbivore, has a relative clutch mass above that for active foragers and more like that of the sit-and-wait predators. Because Iguana iguana clutch size is so much larger, the investment per offspring, relative to female size, is much smaller than in the other iguanids that Vitt and Congdon report. That the relative clutch mass is larger in large females than small ones in both Panama and Nicaragua (Table 2) suggests that Iguana iguana behaves in accord with the general prediction by Williams (1966) that, within a species, relative reproductive effort should increase with age be- cause as an animal ages there is less cost to future reproduction from a high effort at the present age. Pianka and Parker (1975) develop this idea and discuss the conditions under which an ani- mal should modify its current reproductive effort in order to maximize its total lifetime reproduc- tive value, and discuss some examples both of reptiles which increase their relative reproduc- tive investment with age and those that do not seem to do so. In these iguanas, the greatest dif- ference in relative reproductive effort is between very small females and all the rest. Differences between moderate and large females are much less strong. Until we have some estimates of growth and mortality for iguanas at different ages we cannot interpret the selective pressures that have produced the observed patterns and eval- uate the relevance of Williams' suggestion and Pianka and Parker's model to the evolution of iguana reproductive biology. Certainly other fac- tors such as body shape and the ability to acquire or process resources may change with female size and influence relative clutch mass. Summary In Iguana iguana from Panama the number of eggs and the weight of a clutch laid by a female VERTEBRATE ECOLOGY AND SYSTEMATICS 121 Table 2. Continued. Nicaragua F vali jes comparing Differences in N r a b Slopes Elevations adjusted means 5 24 93** -1490 7.113 1.79 9.49** 164 24 .96** -10.4 2.945 0.12 14.59** .159 24 .90** -41.5 .2170 1.75 4.17* 4.48 24 90** -724 3.158 2.39 19.38** 127 24 .69** -1.39 .0349 2.95 25.02** 2.15 24 .45* 9.92 .049 0.31 3.87 24 .89** 3.19 -.007 0.77 0.44 24 .81** 8.20 .0256 0.24 0.10 24 .85** -16.1 .3901 0.04 4.07* 67.5 24 .82** 8.01 .019 0.78 0.59 are closely correlated with both her length and her body weight. The latter two variables are also closely correlated. Egg weight, relative clutch mass, and expenditure per egg all correlate pos- itively with female size. Iguanas from Nicaragua (Fitch and Henderson 1977) show similar relationships between these variables; though females from Nicaragua tend to be lighter and to lay fewer and smaller eggs than do females of equal length in Panama. Rel- ative clutch mass and relative egg size are about the same in the two populations. Compared with other lizards. Iguana iguana has a clutch size which is only a little below that predicted for its size at reproductive maturity from the relationship between these variables seen in other, much smaller lizards. The clutch weight, as a percentage of female size is also not very different. Among its close relatives Ctenosaura similis is like Iguana iguana but Cyclura, Sau- romalus, Amblyrhynchus, and Conolophus seem to have fewer eggs than predicted on the basis of size at maturity. Acknowledgments I would like to thank Bob Henderson and Hen- ry Fitch for sending me their raw data on clutch- es of iguanas in Nicaragua used in their 1977 paper. Mike Ryan and Kathy Troyer criticized the manuscript usefully. Literature Cited Ballinger, R. E. and Clark, D. R., Jr. 1973. Energy content of lizard eggs and the mea- surement of reproductive effort. J. Herpetol., 7:129-132. Carpenter, C. 1 966. The marine iguana of the Galapagos Islands, its behavior and ecology. Proc. Calif. Acad. Sci., 34:329-376. Fitch, H. S. and Henderson, R. W. 1977. Age and sex differences, reproduction and conservation of Iguana iguana. Contribu- tions in Biology and Geology, Milwaukee Public Museum, No. 13. Fitch, H. S., Henderson, R. W. and Hillis. D. M. 1983. Exploitation of iguanas in Central America. In Burghardt, G. M. and Rand, A. S. (eds.). Iguanas of the World: Behavior. Ecology, and 122 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Conservation. Garland STPM Press. New York. Hikih. H. F. 1963. Some aspects of the natural history of Igua- na iguana on a tropical strand. Ecology. 44(3): 613-615. HOOGMOED, M. S. 1973. Notes on the herpetofauna of Surinam. IV. The lizards and amphisbaenians of Surinam. W. Junk. The Hague. I-IX: 14-19 pp. MlLLER, H. 1972. Ukologische und ethologische studien an Iguana iguana L. (reptilia: Iguanidae) in Ko- lumbien. Zoologische Bertrage N. F. 1 8: 109— 131. Pianka, E. R. and Parker. W. S. 1 975. Age-specific reproductive tactics. Amer. Nat.. 109:453-464. Rand. W. M. and Rand, A. S. 1 976. Agonistic behavior in nesting iguanas: a sto- chastic analysis of dispute settlement dom- inated by the minimization of energy cost. Z. Tierpsychol. 40:279-299. Snedecor, G. W. 1956. Statistical methods. Iowa State Univ. Press, Ames. Iowa. pp. 1-534. Tinkle. D. W., Wilbur, H. M. and Tillev, S. G. 1970. Evolutionary strategies in lizard reproduc- tion. Evolution, 24(l):55-74. Tinkle, D. W. and Hadlev, N. F. 1975. Lizard reproductive effort: caloric estimates and comments on its evolution. Ecology, 56: 427-434. Vim, L. J. 1978. Caloric content of lizard and snake (Reptilia) eggs and bodies and the conversion of weight to caloric data. J. Herpetol., 12:65-72. Vitt, L. J. and Congdon, J. D. 1978. Body shape, reproductive effort, and relative clutch mass in lizards: resolution of a para- dox. Amer. Natur.. 1 12(985):595-608. WlEWANDT, T. A. 1 983. Evolution of nesting patterns in iguanine liz- ards. In Burghardt, G. M. and Rand, A. S. (eds.). Iguanas of the World: Behavior, Ecol- ogy and Conservation. Garland STPM Press, New York. Williams, G. C. 1966. Adaptation and natural selection. Princeton University Press, Princeton, New Jersey, pp. 1-307. Vertebrate Ecology ami Systematics— A Tribute to Henry S Fitch Edited by R. A. Seigel, L. E. Hum. I I .. Knight. I Malaret and N. L. Zuschlag 1984 Museum of" Natural Histoiv I he I 'inversus of Kansas. Lawrence Are Anuran Amphibians Heavy Metal Accumulators? Russell J. Hall and Bernard M. Mulhern Introduction Concern about heavy metals in the environ- ment has increased recently, partly as a result of increased awareness of their potential effects, and also because of the prospect of expanded use of fossil fuels in processes that release metals. From time to time amphibians have been examined as possible indicators of contamination by heavy metals. Their habitats, abundance, and ease of sampling have made them convenient subjects for such purposes. There have been indications that amphibians may be unusual in their ability to accumulate metals. A survey of the copper content of the livers of a wide range of vertebrate species (Beck 1956) indicated extremely high levels (up to 1 640 ppm) in Bufo marinus; average copper concen- trations in livers were generally much lower in other species, although one species of marine fish had higher average concentrations. Surprisingly high concentrations of lead in the livers of some frogs from a remote and apparently uncontam- inated area were reported by Schroeder and Tip- ton (1968). Gale el al. (1973) found up to 1590 ppm (dry weight) of lead in tadpoles from a con- taminated area and these results suggest that they have a much greater ability to concentrate en- vironmental lead than do the other species sam- pled. The iron content of one sample of Rana catesbeiana tadpoles analyzed in our own labo- ratory reached the startling level of 19.000 ppm (dry weight). It is the purpose of this paper to present data from our own work at the Patuxent Wildlife Re- search Center (PWRC) and information from the literature in order to ascertain whether the con- centrations of metals in amphibians fall outside the normal range of variation of other animals. This paper will assemble data which may bear on the questions of whether amphibians are par- ticularly susceptible to heavy metal pollution, whether they can accumulate levels which may be hazardous to their predators, and whether they can be of value as monitors of heavy metal con- tamination. Methods Adult amphibians collected at the Patuxent Wildlife Research Center were generally ob- tained from the Island Marshes, artificial habi- tats built for waterfowl management, or they were picked up on service roads on wet nights. Larval amphibians and fish were captured by seine or dip net from Harding Spring or Mabbott ponds: both are shallow, moderate-sized artificial ponds surrounded by wooded areas. The research cen- ter is not known to be contaminated by heavy metal residues. Iron is naturally abundant in the soil and groundwater. Possible alteration of Harding Spring Pond by runoff from a nearby landfill has led us to undertake a program mon- itoring organochlorine and heavy metal levels in certain animals found in different areas of the center. A sample of 10 leopard frogs (Rana pip- iens) obtained from the National Fish and Wild- life Health Laboratory, Madison. Wisconsin was also analyzed for heavy metal levels. Tissue samples were homogenized in a blender and a 5 g portion was weighed into a crucible for heavy metals analysis. A separate 5 g portion was weighed into a round-bottom flask to deter- mine mercury levels. Digestion for mercury anal- ysis used the method described by Monk (1961). Mercury was determined by cold vapor atomic absorption spectrophotometry using the method of Hatch and Ott (1968) with a Coleman model MAS-50 mercury analyzer. The lower limit of reportable residues was 0.02 ppm. The sample used to determine other metals was dried in an oven and then charred in a muffle furnace where the temperature was raised to 550°C at the rate of 100°/hr and left overnight. The cooled ash was dissolved over a hot plate in approximately 2 ml of concentrated nitric acid and 1 ml of concen- trated hydrochloric acid, transferred to a 50 ml polypropylene centrifuge tube, and diluted with dionized water. Analysis was by flame atomic absorption spectrophotometry with a Perkin-El- mer model 703 equipped with a deuterium arc background corrector, an AS-50 autosampler. and a PRS-10 printer. The lower limit of reportable 123 24 SPECIAL PUBLICATION -MUSEUM OF NATURAL HISTORY Table 1. Cadmium in amphibians. Area Tissue PPM N Sample Wel weigh! Dry weight Reference Tadpoles (sp.) Pb contaminated Whole body 1.4-3.0 8 pools Gale et al. 1973 Uncontaminated Whole body 1.1 1 pool Gale et al. 1973 Tadpoles (R. catesbeiana) Uncontaminated Whole body 0.16-0.24 1.2-2.0 5 pools PWRC (R. clamitans) Uncontaminated Whole body 0.10-0.19 1.0-1.8 2 pools Toads (Bufo Uncontaminated 2 Whole body 0.15-4.0 0.72-26 4 PWRC spp.) Livers 0.08-0.13 0.64-1.1 2 Kidneys 1.9 6.1 1 pool (pooled) Carcasses 0.19-7.3 0.68-25 12 Frogs (Rana Uncontaminated Whole body 0.10-0.36 11 PWRC spp.) Adult amphibians Contaminated by (Triturus, Zn mill Rana, downwind Whole body 10.7 1 Dmowski and Pelobates, Karolewski Bufo) 1979 protected area Whole body 1.3-14.4 8 Dmowski and Karolewski 1979 Uncontaminated Whole body 0.3-1.4 11 Dmowski and Karolewski 1979 1 Unpublished data, Patuxent Wildlife Research Center. : Some of these animals were captured on or near highways; these local sources of contamination may explain the high levels of cadmium found in some samples. residues was 0.1 ppm. Recoveries of all metals from fortified chicken livers ranged from 8 1 to 1 1 6%. Residues were not corrected for percent recovery. All concentrations are on a whole body, wet weight basis unless stated otherwise. Tadpoles analyzed at Patuxent Wildlife Research Center were prepared in pools of at least 10 individuals each; variability among pools was relatively low. Most reports in the literature do not have esti- mates of variability. Indications that variability may be extreme, however, have led us to use ranges rather than means for comparative pur- poses. Results and Discussion Metal Residues Reported in Amphibians Arsenic — Concentrations of arsenic in livers of adult Rana, Bufo, and Bombina collected in an apparently uncontaminated area (Byrne et al. 1975) averaged 0.164 ppm; they averaged 0.079 ppm in an area known to be contaminated with mercury. These concentrations are within nor- mal levels for most organisms (Schroeder and Balassa 1966) and are considerably lower than levels found in many freshwater animals (Wa- gemann et al. 1978). Cadmium. — Levels in lead-contaminated areas (Gale et al. 1973) generally exceeded those re- corded in apparently unpolluted areas. Cad- mium concentrations in tadpoles (Table 1) ana- lyzed in our laboratory are about an order of magnitude higher than the concentrations found in a variety offish species in New York (Lovett et al. 1972) and approximately twice the levels found in fish from the same areas. Concentra- tions of cadmium in livers and kidneys of toads and carcasses of frogs (Table 1) generally fall within the ranges reported for other animal ma- terials by Fleischer et al. (1974). Carcass and whole-body concentrations in toads (up to 7.3 ppm wet weight) exceeded levels reported in oth- VERTEBRATE ECOLOGY AND SYSTEMATICA 125 Table 2. Copper in amphibians. PPM Sample Area Tissue Wet weight Dp. ueight Reference Tadpoles (sp.) Uncontaminated Pb contaminated Whole body Whole body 8 17-48 1 pool 8 pools Gale el at. 1973 Tadpoles (sp.) Pb contaminated Whole body 169 1 pool Jennett et al. 1977 Tadpoles (Rana catesbeiana) Uncontaminated Whole body 1.4- -3.2 12-19 5 pools PWRC R. clamitans Uncontaminated Whole body 0.93- -1.2 8.8-12 2 pools Adult {Rana temporaria) Uncontaminated Whole body 53.1- -207.1 159.9-845. I 2 106 Pasanen and Koskela 1974 Adult (Rana, Bufo, Bombina, Triturus spp.) Hg contaminated Liver 4.0- -318.9 13.8-1099.7 24 Bryne et al. 1975 Adult (Bufo, Hyla, Lim- nodynastes) Uncontaminated Liver 10-1640 21 Beck 1956 Adult (Bufo marinus) Uncontaminated Liver 367-2091 6 Goldhscher et al. 1970 Adult (Bufo) Uncontaminated Whole body 2.1- -5.0 9-20 4 PWRC Adult (Rana) Uncontaminated Whole body 1.2- -3.5 9.0 11 PWRC Adult (Rana catesbeiana) Uncontaminated Blood Plasma Liver Fat Eggs 0.5- 0.3- -5.9 -1.5 10-29 0.4-1.4 0.19-0.26 40 1 1 pools 40 40 t Sarata 1938 1 Unpublished data, Patuxent Wildlife Research Center. 2 Means; highest single observation was 1367 ppm. er terrestrial animals from uncontaminated areas (Table 1). Cobalt. — Bullfrog tadpoles had up to 12.760 pCi/g (dry weight) of cobalt-60 after treatment of an experimental pond (Brungs 1963). The level in tadpoles before treatment was 12 pCi/g. Amounts in the flesh of treated adult bullfrogs ranged up to 105 pCi/g, or less than 1% of the maximum levels in the larvae. Amounts of co- balt-60 taken up by the tadpoles exceeded those in various species of fish and invertebrates. Tadpoles collected from a presumably uncon- taminated area were analyzed in our laboratory; they averaged 1.92 ppm of cobalt in R. cates- beiana tadpoles and 0.38 ppm in R. clamitans tadpoles. A mixed species group of adult anurans averaged 2. 1 ppm (range 0.73-3.2). Higher levels were found in the carcasses when kidneys and carcass remainders were analyzed separately. Copper. —Copper levels may be strikingly high in adult anurans, but they are highly variable, even in contaminated areas (Table 2). Copper in the livers of Bufo marinus is sequestered in 1\ - sosomes where it is unable to damage the he- patocytes (Goldfischer et al. 1970). These au- thors stated that similar levels in adult humans (not contained in lysosomes) would result in se- rious or fatal damage to the liver. This associa- tion with lysosomes explains how toads can tol- erate high copper burdens, but does not explain how they accumulate such concentrations in ap- parently uncontaminated habitats (Table 2). Liv- er levels of copper up to 3000 ppm have been reported in sheep on a diet with copper-molyb- denum imbalances and exceedingly high liver levels of copper have been found in Bedlington Terrier dogs with a hereditary and fatal liver dis- order (National Research Council 1 977). The high levels of copper found in some amphibians may likewise indicate unusual dietarv or metabolic 126 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Tabi f 3. Lead in amphibians. Area Tissue PPM N Sample Wet weight basis Dry weight basis Reference Tadpoles (sp.) Uncontaminated Pb contaminated Whole body Whole body 28 36-1590 1 pool 8 pools Gale et al. 1973 Tadpoles (sp.) Pb contaminated Whole body Eviscerated 4139 93-213 1 pool 2 pools Jennett et al. 1977 Tadpoles (R. cates- beiana) {R. clami- Uncontaminated Uncontaminated Whole body Whole body 2.5-3.2 1.4-1.5 14-23 14-15 5 pools 2 pools PWRC tans) Frogs (Rana spp.) Toads (Biifo spp.) Uncontaminated Uncontaminated Whole body Carcass Kidneys Livers 0.88-3.2 21 4.9 2.1-2.6 1.6-11 0.73-75 15 19-21 11 3 pools 1 pool 2 PWRC PWRC Adult frogs {Acris, Biifo) Frogs (Rana spp.) Uncontaminated Pb contaminated Uncontaminated Whole body Whole body Livers Kidneys 3.5 2 3.7-3.0 2 1.0-31.3 1.3-10.2 14 31 Rolfe et al. 1977 Schroeder and Tipton 1968 Adult amphibians ( Triturus, Rana, Pelobates, Bufo) Contaminated by Zn mill downwind protected area Uncontaminated Whole body Whole body Whole body 461 46-202 1 8 1 1 Dmowski and Karolewski 1979 1 Unpublished data, Patuxent Wildlife Research Center. 2 Means. influences. Levels of copper in tadpoles are with- in the normal range of concentrations reported in various animal materials by National Re- search Council (1977). Chromium. — Tadpoles collected on the Pa- tuxent Wildlife Research Center averaged 3.8 ppm (R. catesbeiana) and 1.6 ppm (R. clami- tans). Chromium concentrations in adult B. americanus, B. woodhousei, R. sphenocephala ranged from 1.8 to 5.4 ppm. A single individual had 56 ppm, but it also had an unusually high concentration of nickel and the sample may have been contaminated during homogenization. Ten R. pipiens averaged 0.48 ±0.12 ppm. Pooled samples of B. woodhousei carcasses and kidneys had 0.66 and 0.95 chromium. The levels of chro- mium reported in amphibians are higher than the representative levels given by Schroeder et al. (1962b) for a variety of animals. Iron. — Singh ( 1978) studied seasonal variation in the iron content of the serum of Rana tigrina. Averages were 0.99 ppm for males and 0.93 ppm for females. Levels tended to be highest during the breeding season. Concentrations were similar to those in humans, greater than those in fish, but less than those of laying fowl. A pooled sam- ple of several species of adult frogs analyzed at Patuxent averaged whole body levels of 90 ppm. Kidneys (180 ppm) contained more than car- casses (21 ppm). Tadpoles accumulated large body burdens of iron; R. catesbeiana larvae av- eraged 2600 ppm and R. clamitans 500 ppm. There is little published information about iron in wild animals, but these levels are several or- VERTEBRATE ECOLOGY AND SYSTEMATICA 127 Table 4. Mercury in amphibians. PPM (wet Sample Area Tissue weight basis) N Reference Amphibian spawn Uncontaminated — 0.0002-0.012 6 Byrne ct al. 1975 Hg contaminated — 0.365 1 Newt larvae Hg contaminated Whole body 0.18.0.22 2 Byrne ct al. 1975 Tadpoles (sp.) Hg contaminated Whole body 0.41.0.49 2 Byrne ct al. 1975 Tadpoles (R. catesbeiana) Uncontaminated Whole body 0.05-0.10 5 pools PWRC (R. clamitans) Uncontaminated Whole body 0.04-0.10 2 pools PWRC Frogs (Rana spp.) Uncontaminated Whole body <0.01-0.14 11 PWRC Toads (Bufo spp.) Uncontaminated Whole body 0.04-0.14 4 PWRC Adult amphibians Uncontaminated Muscle 0.04-0.48 19 Bvrne ct al. (Bufo, Rana, 1975 Bombina, Triturus) Adult amphibians Hg contaminated Muscle 1.39-2.85 9 Byrne ct al. (Bufo, 1975 salamander) Frogs {Rana spp.) Hg contaminated Carcass <0. 10-0. 18 4 Dustman ct al. 1972 Liver 0.28-0.74 4 Frogs {Rana, Bufo, Uncontaminated Liver 0.07-2.3 16 Bvrne ct al. Bombina) 1975 Hg contaminated Liver 1.2-16.2 8 Unpublished data, Patuxent Wildlife Research Center. ders of magnitude higher than levels reports in humans or in foods of animal origin (National Research Council 1979). Lead. — High levels of lead in tadpoles from contaminated waters of Missouri's New Lead Belt are reported by Gale et al. (1973) and Jennett ct al.(\911). But Rolfe etal. (1977) could not detect elevated lead levels in adult anurans from an area near a highway which presumably received lead from automotive exhausts. Despite this negative evidence, it is thought that elevated lead levels in some of the PWRC samples (Table 3) might have resulted from the proximity of the collec- tion sites to highways. The range of concentra- tions found in adult frogs falls within the range reported in humans (Schroeder and Tipton 1968) and levels seem to be close to those found in small mammals (Getz et al. 1977). The data in- dicate little tendency for lead to be preferentially stored in liver, kidney, or carcass. Other authors have not found lead to be concentrated in par- ticular tissues (Schroeder and Tipton 1 968; Getz et al. 1977). When lead contaminated earth- worms were fed to Xenopus laevis, animals con- suming up to 170 Mg/day accumulated up to 81 ppm (dry weight) in the kidneys and 3 1 ppm in the liver (Ireland 1977); relatively greater amounts of lead are stored in soft tissues in Xen- opus than in Peromyscus. Mercury. — Reports of mercury in amphibians are summarized in Table 4. An extensive study by Byrne et al. (1975) surveyed mercury levels in a number of species of amphibians in both contaminated and uncontaminated areas. Dust- man et al. (1970) reported much higher concen- trations of mercury in predatory birds and fish than in amphibians from a contaminated area. Higher concentrations of mercury have been re- ported in amphibians collected from other con- taminated areas but they do not seem to con- centrate mercury as much as some other species. Nickel. — Levels of nickel of 2.7 ppm for R. catesbeiana and 0.9 ppm for R. clamitans were found in tadpoles from the Patuxent Wildife Re- search Center. Specimens of adult anurans ana- lyzed ranged between 0.9 and 2.9 ppm. with a 128 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 5. Zinc in amphibians. PPM Wet weight Dry weight Sample Area Tissue basis basis N Reference Tadpoles (sp.) Uncontaminated Whole body 62 1 pool Gale et at. 1973 Pb contaminated Whole body 160-1090 8 pools Tadpoles (sp.) ?b contaminated Whole body Carcass (without GI tract) 2808 240-256 1 pool 2 pools Jennett etal. 1977 Tadpoles (R. cates- Uncontaminated Whole body 9.7-15 60-86 5 pools PWRC beiana) (R. clami- Uncontaminated Whole body 3.7-6.0 33-59 2 pools tans) Adult amphibians Uncontaminated Liver 11.0-28.3 38-98 2 20 Byrne et al. (Bufo, Rana, 1975 Bombina, Hg contaminated Liver 20.0-33.4 69-1 15 2 4 Triturus) Adult frogs Uncontaminated Liver 12.2-24.6 45.0-90.7 106 Pasanen and (Rana Koskela temporaria) 1974 Adult toads Uncontaminated Whole body 25-94 85-460 4 PWRC (Bufo spp.) Carcass (pool of 10) 84 300 1 pool Kidneys 39 120 1 pool (pool of 10) Adult frogs Uncontaminated Whole body 6.2-31 130 11 PWRC (Rana spp.) Adult amphibians Contaminated by Dmowski and (Triturus, Zn mill Karolewski Rana, downwind Whole body 534 1 1979 Pelobates, protected area Whole body 56-206 8 Bufo) Uncontaminated Whole body 104-301 11 1 Unpublished data, Patuxent Wildlife Research Center. 2 Calculated on the basis of average percent moisture given by the authors. single specimen containing 27 ppm (see account on chromium). Ten leopard frogs (R. pipiens) averaged 1.39 ±0.13 ppm. Pooled kidneys and carcasses of B. woodhousei each had 0.76 ppm nickel. Levels of nickel in wild species are poorly known, but the levels reported here seem to be within representative levels reported by Schroe- der etal (1962a). Zinc— Levels of zinc in amphibians (Table 5) show little variation between tadpoles and adults or between background and metal-contaminated areas except in an area directly contaminated by zinc (Dmowski and Karolewski 1 979) and in tad- pole samples from lead-contaminated areas (Gale et al. 1973; Jennett et al. 1977) which show def- inite evidence of elevated levels. Zinc concen- trations in whole tadpoles were about 10 times those in fish and in eviscerated tadpoles were about double those in fish (Jennett et al. 1977). Zn-65 reached approximately 60,000 pCi/g (dry weight) in bullfrog tadpoles inhabiting an artif- ically contaminated pond (Brungs 1963), a level close to that found in carp and higher than in mollusks or bluegills. Levels in adult bullfrogs VERTEBRATE ECOLOGY AND SYSTEMATICS 129 Table 6. Other elements in amphibia. Sample Area Tissue PPM N Element Wet weight Dry weight Reference Bromine Adult Uncontam- Livers 2.6-15.1 3 Bvrne et al. amphibians inated Hg contam- inated Livers 1.35-1.59 2 1975 Calcium Adult R. temporaria Uncontam- inated Livers 48.9-67.3 176.0-259.6' 106 Pasanen and Koskela 1974 Iodine Adult Uncontam- Livers 0.076-0.145 3 Bvrne et al. amphibians inated Hg contam- inated Livers 0.049-0.092 2 1975 Magne- sium Tadpoles R. cates- beiana Uncontam- inated Whole body 58-160 5 pools PWRC 2 R. clami- Uncontam- Whole 14-29 2 pools tans Adult R. inated Uncontam- body Livers 81.0-170.4 315.1-663.8 1 106 Pasanen and tempdraria inated Koskela 1974 Man- ganese Tadpoles R. cates- beiana Uncontam- inated Whole body 14-42 5 pools PWRC R. clarm- Uncontam- Whole 1.1 2 pools tans Tadpoles inated Uncontam- body Whole 710 1 pool Gale el al. (sp.) inated Pb contam- inated body Whole body 262-5650 8 pools 1973 Selen- Adult Uncontam- Livers 0.069-0.077 2 Bvrne el al. ium amphibians inated Hg contam- inated Livers 3.2-4.7 2 "1975 1 Means, 2 Unpublished data, Patuxent Wildlife Research Center. ( — 4000 pCi/g) were much higher than those in turtles but less than half the levels found in cray- fish. Other Elements. — Bromine, calcium, iodine, magnesium, manganese, and selenium have been reported in amphibian tissues (Table 6) but either the potential threat posed by the materials seems remote or there is little comparative information. Manganese was present at higher levels in tad- poles than in crayfish or small fish from waters contaminated with lead (Gale et al. 1973). Se- lenium levels also may be higher in areas con- taminated by mercury than in areas not contam- inated by mercury (Byrne et al. 1975). Other Radionuclides.— There is some infor- mation on radionuclides in addition to those (Co- 60 and Zn-65) discussed earlier. Radioactive ce- sium has been reported in a single mixed species sample of amphibians in the food of green herons by Domby et al. (1977). The sample contained 0.2 pCi/g (wet weight) of Cs- 1 34 and Cs- 1 37 and was from an area polluted by radioactive wastes. Bullfrog tadpoles which originally averaged 33 pCi/g (dry weight) of cesium- 137 contained 21,000 pCi/g after the experimental addition of radionuclides (Brungs 1963). Adult bullfrogs contained 426 pCi/g. Brungs also found that tad- poles accumulated 7923 pCi/g of strontium-85 but that adult frogs accumulated only 521 pCi/ g in their flesh. Both Cs-137 and Sr-85 were ac- cumulated to higher levels in tadpoles than in other animals but were accumulated less by adult frogs than by most aquatic animals. Rana tem- poraria tadpoles take up more yttrium-90 (a breakdown product) than they do the parent strontium-90 material from treated watei (Lucas 130 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY and Pickering 1958). Pendleton and Hanson (1958) looked at cesium- 137 uptake in a variety of organisms following addition of 6 pCi/ml of the radionuclide to the water of a concrete-lined pond. They analyzed concentration factors after approximately 90 days when levels of Cs-137 in the water had apparently stabilized. They found high concentrations in bullfrog tadpoles, with the bulk of the material stored in the gut fraction. Still higher levels were found in spadefoot toad (Scaphiopus hammondi) tadpoles and adult bull- frogs. Comparison of Anuran Amphibians with Other Animals Brungs (1963) published a number of useful comparisons of the abilities of aquatic animals to take up radionuclides. The highest recorded tissue levels of Co-60, Zn-65, Sr-85, and Cs-137 were all recorded in tadpoles. Somewhat lower levels were found in pelecypods (Co-60, Zn-65, Sr-85) and gastropods (Cs-137). Concentrations in bluegill sunfish and carp tended to be much lower except for Zn-65 and Sr-85 which tended to accumulate in bone. One possible explanation for the high body burdens in tadpoles is their relatively large gut capacity and the chance that a large part of the metals recorded was in the gut cavity and had not actually been assimilated. Separate analyses of gut and the remainder of the carcass confirmed the presence of high levels in the gut fraction, but, with the exception of Sr- 85, body remainders still had greater accumu- lations that most other animals. Also of interest is the fact that the highest levels of radioactivity in tadpoles occurred relatively soon after exposure; other species usually took longer to reach maximum levels and they main- tained high levels longer than did the tadpoles. Brungs suggested that the high levels of radio- nuclides recorded were the result of the vertical distribution of the contaminants in the experi- mental ponds and the tendency of tadpoles to feed on fine sediments. Shortly after addition to the aquatic system, the radionuclides become at- tached to fine particles and settle to the bottom. Tadpoles consume them there and accumulate high levels before various processes had distrib- uted the contaminants more generally through- out the the system. Support for the assertion that feeding habits rather than physiological factors produced the high levels observed in tadpoles is seen in Brungs' data on adult bullfrogs; they ac- cumulated much lower levels of all the metals than did tadpoles, and less than detritivores such as crayfish. Relatively high concentrations of Cs- 1 37 were found in tadpoles in ponds experimentally dosed by Pendleton and Hanson (1958), but the levels were lower than those reported in sunfish, shrimp, and adult frogs. These comparisons were based on data collected some months after the addition of Cs-137 to the system. The authors stated that tadpoles are among organisms which take up the metal rapidly, accumulating it faster than do adult fish, frogs, or seed plants, but Pendleton and Hanson (1958) did not present specific data on the speed of uptake by tadpoles. The apparent differences between these results and those of Brungs are due to the different time spans be- tween dosing and observation; both the relative amounts of Cs- 1 37 in tadpoles compared to oth- er animals and its absolute concentrations de- clined as the time after dosing approached 80 days (Brungs 1963). Most of the lead, zinc and copper in tadpoles from a lead-contaminated area were in the gut (Jennett et al. 1977). However, concentrations in the rest of the body tended to be higher than those in fish from the same waters. These results support the idea that the uptake of the metals is through the diet. Comparison of tadpole gut and contents with those of bass and bluegills indicates an approximate 10-fold greater concentration of the three metals in the amphibian samples, also supporting Brungs' (1963) contention that feed- ing habits produce the higher levels in amphib- ians. Getz et al. (1977) compared lead in different freshwater animals in urban and rural areas. They pointed out that lead levels were higher when the animals (fish and invertebrates) were more closely associated with silt substrata; analyses showed that the uppermost layers of sediment were high- est in lead. Getz et al. (1977) concluded that physical contact with silt and the direct ingestion of lead in silt and detritus were important in uptake. They believed that food chain concen- tration did not occur. Pooled samples of tadpoles of two species, and fish collected from two nearby ponds, are com- pared in Fig. 1 . These results do not closely cor- respond to those metal levels reported by Gale et al. (1973) and Jennett et al. (1977) nor the radionuclides documented by Brungs ( 1963) be- VERTEBRATE ECOLOGY AND SYSTEMATICA 131 3i 2 H Pb M B G 2- 1- Cu 300i 200 100- Mg M P B G M P B G o- Cr 4- 2- n n 8 0.2 0.1- M B G Cd 30 n 20- 10- Mn 3000- 2000- 1000- M P B G o.io- 0.05- Hg □ M P B G 2.0- 1.0- — Ni M M M B G O 40- | 30H 20-" 10-1 Zn EL M B G 2 H M P Co n n M Umbra pygmaea P tepomis g/bbosus B Rana cafesbe/ana G Rana clamitans Fig. 1. Levels of metals detected in pooled samples offish and amphibians collected from two ponds at the Patuxent Wildlife Research Center. Altogether, mudminnows ( Umbra pygmaea) were represented by four pools, bullfrogs (R. catesbeiana) by five pools, green frogs (R. clamitans) by two pools, and pumpkinseed sunfish (Lepomis gibbosus) by seven pools. Pools of the first three species ranged from five to 50 individuals each: sunfish were of various sizes and each sample was comprised of from one to several individuals. Arithmetic means are shown. cause the concentrations in tadpoles are not con- sistently higher than those in fish. Levels in bull- frog tadpoles are higher than those in fish in 6 of the 1 1 metals, but green frog tadpoles contain levels lower than the fish in all but one of the metals. Zinc shows a trend opposite to that seen in the other studies and. with the exception of iron and manganese, the differences between 132 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY groups of animals are much less than those re- ported earlier. The concentrations of metals re- ported in Fig. 1 seem to show some real differ- ences, but they indicate that conditions favoring the uptake of specific metals do not always result in the greatest uptake occurring in tadpoles. Presumably the availability of metals to the different animals and their potential for uptake are influenced by the habits of the animals (see Steinwascher 1978, 1979) and the distribution of the metals within the environment. Distri- bution of metals in the ponds seems to differ from that in the systems examined by other au- thors, perhaps because our areas were essentially uncontaminated and had stable levels of most of the metals rather than a single treatment (Brungs 1 963) or a continuous (Jennett et al. 1977) influx of contaminants. The result would be a greater dispersion of the metals and less tendency for tadpoles to accumulate them. This apparent ten- dency for tadpoles to selectively take up contam- inants which have only recently entered an aquatic system, or which enter on a more or less contin- uous basis, would seem to make them good in- dicators of environmental contamination. them excellent indicators of contaminated en- vironments. Metals transported into an aquatic ecosystem would first collect in sediments where tadpoles could accumulate them, as has been sug- gested in the case of lead (Getz et al. 1977; Jen- nett et al. 1977). Residual metals in uncontam- inated areas, or those which have been in the ecosystem for some time, should tend to become more widely dispersed (Brungs 1 963) and to pro- duce patterns similar to those seen in samples analyzed in our laboratory. Thus because of their apparent tendency to selectively accumulate those metals adsorbed to surface sediments, it might be possible to use tadpoles .to identify ongoing contamination. Acknowledgments H. M. Ohlendorf and C. Brand collected some of the samples. D. Brown and P. McDonald helped with preparation of the manuscript. Drafts of the manuscript were reviewed by W. N. Beyer, E. H. Dustman, and J. C. Lewis. Literature Cited Conclusions 1 ) Adult amphibians of certain species can ac- cumulate extremely high levels of copper in the liver. It seems likely that dietary imbalances or metabolic factors, rather than high environmen- tal levels, result in this accumulation. It has been shown that some anurans are protected from these high copper levels, but individuals with such ac- cumulations may be toxic to their predators. There is little evidence that adult amphibians can concentrate other metals to a greater extent than other vertebrates. 2) Tadpoles accumulate high levels of certain metals, including lead, zinc, copper, cobalt, ce- sium, strontium, iron, and manganese, because of their contact with them in sediments and sus- pended particles. There is extensive literature, not reviewed here, on the toxic effects of metals on amphibians and other aquatic vertebrates which indicates that these organisms are suscep- tible to poisoning by metals. Doubtless their un- usual powers of accumulation can sometimes re- sult in metals in tissues reaching toxic levels. 3) The apparent tendency for tadpoles to pick up metals from surface sediments might make Bec k, A. B. 1956. The copper content of the liver and blood of some vertebrates. Aust. J. Zool., 4:1-18. Brungs. W. A. 1963. The relative distribution of multiple radio- nuclides in a freshwater pond. Ph.D. Thesis. Ohio State Univ., Columbus. Ohio: 97 p. Byrne. A. R., Kosta, L. and Stegnar, P. 1975. The occurrence of mercury in amphibia. En- viron. Lett., 8:147-155. Dmowski, K. and Karolewski, M. A. 1979. Cumulation of zinc, cadmium and lead in invertebrates and in some vertebrates ac- cording to the degree of an area contami- nation. Ekologia Polska, 27:333-349. Domby, A. H., Paine, D. and McFarlane, R. W. 1 977. Radiocesium dynamics in herons inhabiting a contaminated reservoir system. Health Physics, 33:523-532. Dustman, E. H., Stickel, L. F. and Elder, J. B. 1 970. Mercury in wild animals: Lake St. Clair 1 970. Pp. 46-52. In Hartung. R. and Dinman, B. D. (eds.). Environmental Mercury Contam- ination. Ann Arbor Sci. Publ., Ann Arbor, Michigan. Fleischer, M., Sarofim, A. F., Fassett, D. W., Hammond, P., Shacklette, H. T., Nisbet, I. C. T. and Epstein, S. 1974. Environmental impact of cadmium: A re- view by the panel on hazardous trace sub- stances. Environ. Hlth. Perspectives. Exp. Issue No. 7:253-323. VERTEBRATE ECOLOGY AND SYSTEMATICS 133 Gale, N. L., Wixson, B. G.. Hardie, M. G. and Jennett, J. C. 1973. Aquatic organisms and heavy metals in Mis- souri's New Lead Belt. Water Resour. Bull.. 9:673-688. Getz, L. L., Haney, A. Q., Larimore, R. W.. Mc- Nurney, J. W., Leland. H. V.. Price, P. W.. Rolfe, G. L., Wortman, R. L., Hudson, J. L., Solomon, R. L. and Reinbold, K. A. 1977. Transport and distribution in a watershed ecosystem. Ch. 6 In Boggess, W. R. (ed.). Lead in the environment. National Science Foundation RANN Program NSF/RA 770214. Goldfisc her, S., Schiller, B. and Sternwieb, I. 1970. Copper in hepatocyte lysosomes of the toad, Bufo marinus L. Nature, 228:172-173. Hatch, W. R. and Ott. W. L. 1968. Determination of sub-microgram quantities of mercury by atomic absorption spectro- photometry-. Anal. Chem., 40:2085-2087. Ireland, M. P. 1977. Lead retention in toads Xenopus laevis fed increasing levels of lead-contaminated earthworms. Environ. Pollut., 2:85-92. Jennett. J. C, Wixson, B. G.. Lowsley, I. H.. PURUSHOTHAMAN, K... BOLTER, E., HEM- PHILL, D. D., Gale, N. L. and Tranter, W. H. 1977. Transport and distribution from mining, milling, and smelting operations in a forest ecosystem. Ch. 7. In W. R. Boggess (ed.), Lead in the Environment. National Science Foundation RANN Program NSF/RA 770214. LOVETT. R. J., GUTENMANN, W. H., PAKKALA, I. S.. Youngs, W. D.. Lisk. D. J.. Burdick. G. E. and Harris, E. J. 1972. A survey of the total cadmium content of 406 fish from 49 New York State freshwa- ters. J. Fish. Res. Brd. Canada, 29:1283- 1290. Lucas, J. W. and Pickering, D. C. 1958. Direct absorption of dissolved strontium-90 and yttrium-90 by tadpoles of Rana tem- porary. Nature, 182:1242-1243. Monk, H. E. 1961. Recommended methods of analysis of pes- ticide residues in food stuffs. Report by the Joint Mercury Residues Panel Anal., 86:608- 614. National Resear( h Coun< ii . 1977. Copper. National Academy of Sciences, Washington, DC. 115 p. 1979. Iron. University Park Press. Baltimore. MD. 248 p. Pasanen, S. and Koskela, P. 1974. Seasonal changes in calcium, magnesium, copper and zinc content of the liver of the common frog, Rana temporaria L. Comp. Biochem. Physiol., 48A:27-36. Pendleton, R. C. and Hanson, W. C. 1958. Absorption of Cesium-137 by components of an aquatic community. Proc. 2 nd United Nations Conf. on Peaceful Uses of Atomic Energy, 18:419-422. Rolfe, G. L., Haney, A. and Reinbold, K. A. 1977. Environmental contamination by lead and other heavy metals, Vol. II. Ecosystem anal- ysis. Final Rept. National Science Founda- tion RANN Program, Inst, for Envtl. Stud. Univ. Illinois Urbana-Champaign. Sarata, U. 1938. Studies in the biochemistry of copper. XXX. Seasonal changes in the amount and distri- bution of copper in tissues of the cultivated bullfrog. Japan. J. Med. Soc, 4:65-69. Schroeder, H. A. and Balassa, J.J. 1966. Abnormal trace elements in man: Arsenic. J. Chron. Dis., 19:85-106. Schroeder, H. A., Balassa, J. J. and Tipton, I. H. 1962a. Abnormal trace elements in man: nickel. J. Chron. Dis., 15:51-65. 1962b. Abnormal trace elements in man: chromi- um. J. Chron. Dis.. 15:941-964. Schroeder, H. A. and Tipton, I. H. 1968. The human body burden of lead. Arch. En- vir. Hlth., 17:965-978. Singh, K. 1978. Serum iron level of the common Indian frog Rana tigrina Daud. Experientia. 34:433-434. Steinwascher, K. 1978. The effect of coprophagy on the growth of Rana caiesbeiana tadpoles. Copeia. 1978: 130-134. 1979. Competitive interactions among tadpoles: Responses to resource level. Ecology. 60: 1172-1183. Wagemann, R., Snow, N. B.. Rosenberg, D. M. and Lutz, A. 1 978. Arsenic in sediments, water and aquatic bio- ta in lakes from the vicinity of Yellowknife. Northwest Territories, Canada. Arch. En- viron. Contam. Toxicol., 7:169-191. Part III Feeding and Behavior Vertebrate Ecolog> and Systematics— A Tribute to Henrj S Fitch Edited by R. A. Seigel, L E Hunt. .1 I Knight. I. Malarct and N. L. Zuschlag • IV84 Museum of Natural Historv The University of Kansas. Lawrence Energetics of Sit-and-Wait and Widely-Searching Lizard Predators Robin M. Andrews Introduction The foraging tactics of insectivorous lizards, like those of most other predators, appear to be dichotomous (Pianka 1978). In North America, iguanid lizards exemplify the "sit-and-wait" tac- tic in which prey are sought passively from a fixed perch site. Sight of a moving prey item elicits ambush or pursuit. Teiid lizards, on the other hand, exemplify the "widely-searching" tactic in which prey are sought actively while the lizard moves through the habitat. These two tac- tics may represent a fundamental means of par- titioning the food niche (Pianka et al. 1979). Each tactic apparently gives maximal foraging effi- ciency (in time or energy units) under conditions of varying prey abundance (Norberg 1977) or structural configuration of the habitat (Stamps 1977). The sit-and-wait and the widely-searching tac- tics are each associated with an "adaptive syn- drome" of predator characteristics (Eckhardt 1979). In addition to characteristics strictly re- lated to foraging, the adaptive syndromes of ig- uanid and teiid lizards differ markedly in several ways. Iguanids have more stereotyped responses to novel items in their environment than do teiids (Regal 1978). Iguanids are strongly territorial, teiids lack home range defense (Stamps 1977). Iguanids escape predators by cryptic behavior and, once discovered, by the use of known routes to hiding places. In contrast, teiids rely on rapid flight to escape their predators (Vitt and Congdon 1978; Schall and Pianka 1980). Clutch size per unit body weight is higher for iguanids than for teiids (Vitt and Congdon 1978). Although many aspects of the adaptive syn- dromes of sit-and-wait and widely-searching predators have been described, the energetic costs and benefits of each tactic are unknown. For ex- ample, the low searching costs of the sit-and-wait tactic are often associated with relatively low pre- ferred body temperatures. Sceloporine and an- oline iguanids have preferred body temperatures of 35°C or less even in well insolated environ- ments (Blair 1960; McGinnis 1966; Andrews, unpublished data; Huey and Webster 1 976; Ben- nett and Gorman 1979). In contrast, (macro) tends such as Cnemidophorus not only have high searching costs (Bennett and Gleeson 1979). but their foraging tactic is associated with preferred body temperatures of about 40°C (Asplund 1 970; Schall 1977; Bennett and Gorman 1979). Thus, for many iguanids (notable exceptions are desert lizards such as Holbrookia and Callisaums), the metabolic cost of foraging is low compared to that of teiids not only because of the low levels of activity associated with the sit-and-wait tactic but because of low activity temperatures. The major objectives of this study were to an- swer two questions: 1 ) What are the relative en- ergy intakes of lizards using sit-and-wait and widely-searching tactics when both forage in the same habitat? 2) Does the proportional alloca- tion of assimilated energy to production and me- tabolism differ for lizards using the sit-and-wait tactic and the widely-searching tactic? Lizard Subjects and Field Sites Field studies were conducted in the Chiricahua Mountains of Arizona. The lizard subjects were Sceloporus jarrovi (viviparous, Goldberg 1971) and Cnemidophorus exsanguis (parthenogenetic. Cole and Townsend 1977). These species are an ideal pair for comparative studies of feeding be- havior and energetics. First, they are broadly sympatric in oak-pine-juniper woodland. Sec- ond, they are of similar size; females of both species reach a maximum weight of about 20 g. Third, their ecology is comparatively well known (Simon 1975; Congdon 1977; Schall 1977; Ruby 1977; Ballinger 1979). Observations were made from 1 1 July to 8 August 1979. At this time, the fat reserves of both species are increasing rapidly (Goldberg 1972; Schall 1978). Since energy stored by female lizards prior to winter inactivity contributes di- rectly to the development of offspring or eggs that will be produced the following spring (Hahn and Tinkle 1965; Gaffney and Fitzpatrick 1973), en- ergy available for fat storage is directly related to the reproductive effort of both S. jarrovi and C. exsanguis. Moreover, adult S. jarrovi females. 137 138 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 1. Prey items used in laboratory feeding experiments, their dry weights, ash contents, and proportional representation in the feeding regimes of the lizard subjects. Prey taxa Drv wt. (%) Ash (%) S jarrovi C. exsanguii Coleoptera Tenebrio molitor, adult Dermestes caninus, adult Dermestes caninus, larvae Phyllophaga sp., adult Chaulognathus pennsylvanicus, adult Lepidoptera (various moths) Orthoptera Blattella germanica. adult male 59 61 34.0 5.1 38.4 4.4 43.4 2.5 28.6 6.1 25.3 4.9 34.0 4.4 37 16 31.0 6.7 4 23 at least, allocate large amounts of energy to growth during the summer and fall (Tinkle and Hadley 1975). Since body size is related to clutch size (Vitt and Congdon 1978; Balltnger 1979), the energy available for growth is directly related to reproductive effort. Observations were made at two sites. Site A was located in North Fork Canyon immediately west of the Southwestern Research Station (SWRS) at 1650 m elevation. The study area extended for 3 km along North Fork Creek (and a gravel road). Both S. jarrovi and C. exsanguis were found in the open woodlands along the creek and in the more sparsely wooded hillsides. How- ever, individuals of both species were most abun- dant near the creek and most observations were made there. Site B was located in Cave Creek Canyon 7 km SW of Portal. The site was bisected by the South Fork of Cave Creek. Although S. jarrovi was abundant, C. exsanguis was not found in this well shaded riparian forest. Methods Faecal and Urinary Production by Free- ranging Lizards Lizards were captured by hand or by noosing. They were placed individually in plastic bags and taken to the SWRS where they were held in a screened open-air laboratory for 48 h before re- lease at the place of capture. Scats produced dur- ing this period were collected, their faecal and urinary portions were separated, dried at 65°C for 48 h, and then weighed to the nearest 0. 1 mg. Mean maximum and minimum temperatures in the laboratory during the study were 32°C and 16°C, respectively. Mean maximum and mini- mum temperatures outside of the laboratory' (SWRS weather records) were 3 1 . 5°C and 1 1 .5°C, respectively. Relation Between Faecal and Urinary Production and Food Intake In order to estimate the food intake of free- ranging lizards, the relation between faecal and urinary production and food intake was deter- mined in my laboratory at Virginia Polytechnic Institute and State University (VPI and SU) in June and July 1980. Lizards were individually caged in 20 gal. aquaria. Temperatures during the experiments were set to simulate field con- ditions in July. Night (12 h) temperatures were 1 6°C and day temperatures were 26°C. Heat lamps suspended over one end of each cage allowed the lizards to thermoregulate normally from 0900 to 1 500 h. Lizards were fed a pre-weighed quantity of live insects each morning for 3 days. The ratio of live to dry weight for each prey type or species used was determined for representative individ- uals (Table 1 ). About noon at the end of the 3 d feeding period, lizards were placed in plastic bags. The urinary and faecal material produced in the following 48 h were treated the same as that pro- duced by field-collected lizards. Prey items used for the feeding experiments were selected on the basis of what lizards were eating in the field (see Results). One-half of the faecal material produced by lizards collected on Site A was used to determine prey taxa and size. Each faecal pellet was softened in a detergent- water solution and gently teased apart. Prey taxa were identified from their chitinous remains and VERTEBRATE ECOLOGY AND SYSTEMATICS 139 their lengths were estimated roughly at 5 mm intervals. A major assumption of this method of esti- mating food intake is that lizards are active every day and that they defecate regularly. From ob- servations made near my site B, Simon and Mid- dendorf ( 1976) found that the percent of adult S. jarrovi active every day was 75% in July and 100% in August. Thus, the assumption of daily activity is probably valid for S. jarrovi but has not been tested for C. exsanguis. Observations on 5. jarrovi and C. exsanguis maintained in large cages under simulated field conditions sug- gest that defecation occurs at least every morning following the attainment of preferred body tem- peratures (see also Cowles and Bogert 1944). Ash contents of faeces, urinary wastes, and the various prey types used in the laboratory exper- iments were measured by heating samples for 1 h at 550°C in an ashing oven. The mass of all materials is presented as ash-free dry weight. Activity Periods and Body Temperatures Any lizard seen was considered active. Because 5. jarrovi individuals were readily found during all daylight hours. I assumed that their activity period potentially spanned 10-12 h. In contrast. C. exsanguis individuals were encountered most frequently in the morning. To define the activity period of C. exsanguis, a series of 30-minute censuses was conducted on 3 and 4 August. All individuals encountered while I slowly walked about 2 km through site A were counted. Body (cloacal) temperatures (T b ) were mea- sured immediately after capture with a Schult- heis quick-reading thermometer. Temperatures of lizards which avoided capture for more than a minute were not taken to avoid bias. Shaded air temperatures were taken at 1 m and at 1 cm above the place where the lizard was first seen. Results Food Intake by Free- ranging S. jarrovi and C. exsanguis Various species of beetles made up about 60% of the insects eaten by lizards in the feeding ex- periments and moths and cockroaches made up the other 40% (Table 1 ). This particular feeding regime was similar to the natural diets of the two Table 2. Prey items of Sceloporus jarrovi and of Cne- midophorus exsanguis in July-August 1979. Propor- tion of total prey is given for each species followed by modal length category in parentheses. Pre\ laxa S \UTTOVi unguis Coleoptera (adults) Formicidae Lepidoptera (adults) Hymenoptera (adults) Araneida Orthoptera Miscellaneous* .306(5-10) .518 (<5) 0.0 .082(5-10) .023 (<5) .023(10-15) .047 (<5) .189(5-10) .500(5-10) .122(10-15) 0.0 .067 (<5) .078(10-15) .044(5-10) 90 * S. jarrovi: 4 Homoptera-Hemiptera; C. exsanguis: 2 Homoptera-Hemiptera. 1 mantid. 1 Chilopoda. lizard species (Table 2). Judging by both fre- quency and size, beetles were probably the most important component of the diets of both S. jar- rovi and C. exsanguis. Orthoptera were probably the second most important component of the diet of S. jarrovi and Lepidoptera were probably the second most important component of the diet of C. exsanguis. Ants were not used in the feeding experiments although they comprised about half of the items eaten by both species in the field. Because of their small size (bulk) their contribution to total energy intake was probably low. Using stomach contents to evaluate the diet of C. exsanguis in New Mexico. Medica (1967) also found the major items (by volume) to be beetles and Lepidoptera. with Hymenoptera (mostly ants) to be relatively unimportant. Food intake of field-collected S. jarrovi and C. exsanguis females was estimated as I, -= F*CFF-'*W- 0.83 and I„ = U*CFU-'*W-°- 83 where I, and I u are the respective estimates of food intake based on faecal and urinary produc- tion. F and U are faecal and urinary production (mg dry wt) during the 48 h of confinement, re- spectively. CFF and CFU are the factors which convert F and U to food intake for faecal and urinary production, respectively, and W 3 is live body weight in g raised to a power of 0.83 to adjust for weight specific metabolic rates (Ben- 140 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 3. Daily food intake (I, and I u ) of S. jarrovi and C. exsanguis based on production of faecal and urinary material (sec text for details). Spec ies and site In I, ± SE (mg'g~°-"-d" In I„ ± SE (mg-g-°" d ■') S. jarrovi— A S. jarrovi— B C. exsangi 0.05), mean values were used to estimate food intake. Respective mean values ( ± SE) for S. jarrovi and C. exsanguis were CFF = 0.56 ± 0.053 and 0.24 ± 0.024 and CFU = 0.2 1 ± 0.027 and 0. 1 9 ± 0.027. Natural log transformations were used to normalize the I, and the I u data for statistical analyses. Some female-sized male S. jarrovi were included in the analyses (5 of 9 and 9 of 29 individuals on Sites A and B, respectively) since the faecal production of these males did not differ from females on either site (P > 0.05, two-tailed /-tests). Al- though individuals were captured at various times during the day (see below), regression analyses indicated that time of capture was not related to food intake (P > 0.05). Differences in I, and I u among the C. exsanguis females and the two populations of S. jarrovi (Table 3, Fig. 1) were statistically significant (P < 0.05, analysis of variance). A posteriori tests showed that C. exsanguis females had a signifi- cantly (P < 0.05) greater I, and I u than both the S. jarrovi populations, and that the S. jarrovi populations did not differ from one another for either I, or I u (P > 0.05, Duncan's multiple range tests). Therefore, in subsequent analyses the data for S. jarrovi females have been combined. The two estimates of food intake for C. exsanguis differed by only 4% on a In scale (Table 3) and by only 1 1% on an arithmetic scale. Since the correlation between U and I 3 for S. jarrovi was not statistically significant, I u was not deter- mined. Activity Periods and Thermoregulation Scleoporus jarrovi and C. exsanguis differed considerably in the apparent length of their ac- tivity periods. I made observations from about 0800 to 1730 h with comparable times spent in the field in the morning and in the afternoon. The number of S. jarrovi individuals observed in the morning and the afternoon was very sim- ilar. In contrast, C. exsanguis individuals were active primarily in the morning; only 3 of 37 individuals collected were caught in the after- noon. The census data also indicated that peak activity was in the morning (Table 4). Body temperatures of S. jarrovi were depen- dent on weather conditions (Figs. 2 and 3). On site A where temperatures were measured under sunny conditions, S. jarrovi individuals main- tained relatively constant T h s (Mean ± SE = 34.2 ± 0.36°C). In contrast, on site B about one- half of temperature measurements were taken under overcast or intermittently cloudy condi- tions. At these times, T b s averaged 31.1 ± 0.6 PC. During sunny conditions T b s averaged 35.8 ± 0.34°C. Body temperatures of C. exsanguis were in- dependent of ambient temperatures (Fig. 1), av- eraging 40.0 ± 0.3 PC. The one individual with a T b of 34°C had probably just emerged from a burrow. Discussion During the July-August study period, 5. jar- rovi females, using sit-and-wait tactics, had a sig- nificantly lower intake of food than did C. exsan- guis females which were using widely-searching VERTEBRATE ECOLOGY AND SYSTEMATICA 141 CO 00 o E 100i 60 20 A A A • A A a* A A 4 a A* A A A. * 10 15 20 W eg) Fig. 1. Food intake based on faecal production (I r ) of field-collected Sceloporus jarrovi (Site A. filled circles, and Site B. filled triangles) and Cnemidophorus e.xsanguis (Site A, open triangles) as a function of body weight. tactics. The result is particularly interesting in that S. jarrovi was active more than twice as long as C. e.xsanguis. Moreover, since lizards were foraging in the same habitat, individuals of both species potentially had the same kinds and abun- dances of prey available to them. Thus, the widely-searching tactic appears to be more effi- cient both in terms of time spent and energy acquired. In order to compare the energy that S. jarrovi and C. e.xsanguis females have available for pro- duction, I, and I u were partitioned into their ma- jor components as I = R + P + FU where I is food intake. R is metabolism. P is production, and FU is the combined faecal and urinary wastes. The energy value of food intake was determined as I times 5800. the mean caloric value for a variety of insects (Griffiths 1977). Digestion and assimilation efficiencies of small insectivorous lizards are quite similar (Harwood 1978; Johnson and Lillywhite 1979). Therefore. FU was estimated as 20% of I for both S. jarrovi and C. e.xsanguis (Johnson and Lillywhite 1979; Andrews and Asato 1977). The parameters used to estimate field metabolism of a 12 g S. jarrovi female are from Table 3 and Appendix A. Table 1. of Congdon (1977). Since his Ash Spring site and my sites were located within a few km of one another, I have used his July-August deter- minations directly. The field metabolism of C. e.xsanguis was estimated from metabolic data collected on Cnemidophorus murinus, a West In- dian species. Metabolic rates of C. murinus were determined under standard conditions for both resting individuals and for individuals moving Table 4. Numbers of Cnemidophorus individuals seen during 30 min censuses conducted on 3 and 4 August 1979. Shaded air temperatures 1 m above ground are shown for the time the census was began. Both days were sunny. Census period T(°C) N 800-830 23 2 930-1000 27 6 1030-1100 27 6 1100-1130 27 5 1200-1230 29 2 1400-1430 28 1 1700-1750 29 142 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY O o LU DC Z) h- < DC LU Q_ LU 45 35 25 A A A A ZA A AA A A A A & v> • • 10 11 T ~ 4 6. 25 - 010,

S Fitch Edited h\ K V Scigel. L. E Hunt. J L Knight. L. Malaret and N. L. Zuschlag c 1984 Museum of Natural Hislon. The I'nncrsin of Kansas, Lawrence Feeding Behavior and Diet of the Eastern Coral Snake, Micrurus fulvius Harry W. Grefne Introduction Snakes are prominent predators in many ter- restrial, aquatic, and tropical marine commu- nities, and exhibit some unusual morphological and behavioral modifications for this role. They rely heavily on chemical senses for locating food (Burghardt 1970; Chiszar and Scudder 1980) and their usual method of locomotion (lateral un- dulation) is energetically more efficient than te- trapody (Chodrow and Taylor 1973). Perhaps most importantly, these "limbless tetrapods" possess an extremely flexible jaw apparatus that permits the ingestion of large prey items without the assistance of limbs or mastication (Gans 1961). Although many species swallow prey alive and struggling, others immobilize it by constric- tion, venom injection, or a combination of these methods (Gans 1978; Greene and Burghardt 1978; Kardong 1980). It is now clear that venom delivery systems comprise at least three grades of structural com- plexity and that these have evolved indepen- dently in several lineages of snakes (see Gans and Gans 1978; Savitzky 1978; 1980; Kardong 1980; Cadle. in press, for extensive discussion and re- views). Opisthoglyphs (many species of colu- brids) possess enlarged, grooved teeth on the pos- terior ends of otherwide normal, elongate, toothed maxillae. Proteroglyphs (elapids and hydro- phiids) have one or two enlarged, canaliculate, anterior teeth on each short, nonmobile or slight- ly mobile maxilla. Solenoglyphs (viperids and atractaspids) have a single, very elongate hollow fang on each highly movable maxillary bone. Studies on several solenoglyphs of the family Vi- peridae show that these snakes often strike and release prey, then relocate it before swallowing (e.g., Klauber 1956; Duellemeijer 1962; Nalleau 1966; Minton 1969; Kardong 1975; Chiszar and Scudder 1 980). Although there are isolated notes on the feeding behavior of opisthoglyphs and proteroglyphs in the literature (e.g.. Armitage 1965, Lambins 1967, for African elapids). the only extensive accounts are for certain sea snakes (Voris et al. 1978; Radcliffe and Chiszar 1980). With few exceptions (e.g.. Shine 1977, for ela- pids; Voris et al. 1978. for hydrophiids), we also know very little about the dietary ecology of pro- teroglyphs. This general lack of descriptive stud- ies on proteroglyphs, particularly terrestrial forms, hampers broader considerations of functional morphology, adaptive radiation, and commu- nity structure in snakes (cf. Arnold 1972; Rabb and Marx 1973; Kardong 1980; Savitzky 1980; Greene. MS). In the present paper I provide a description of feeding behavior, an ecological characterization of the food habits, and a discussion of factors affecting diet composition in a venomous coral snake. Micrurus fulvius. This species occurs in the southeastern United States and northeastern Mexico, in habitats ranging from subtropical swamps and lowland forests to semiarid scrub (Wright and Wright 1957). It is a northern rep- resentative of an essentially Neotropical radia- tion of the cosmopolitan front-fanged family Elapidae(Roze 1967; see Savitzky 1978. and Ca- dle and Sarich 198 1 for contrasting views on the relationships of coral snakes). Eastern coral snakes have been found crawling on the surface and in or under rocks, logs, stumps, litter, and burrows (Wright and Wright 1957; Gentry and Smith 1968). There is perhaps seasonal and geographic variation in diel activity, but these snakes are predominantly diurnal (cf. Neill 1957; Wright and Wright 1957; Jackson and Franz 1981). An average adult is ca. 50-85 cm long and weighs 20-55 g. Wright and Wright ( 1957), Shaw (1971). Campbell (1973). Greene ( 1 973a. 1 973b). Quinn (1979). and Jackson and Franz (1981) summa- rized some aspects of the biology of this species. Methods Behavioral Observations. — Sixty-five com- plete feeding sequences on live and dead prey by- four captive coral snakes were observed (one fe- male, three males; total lengths 52.5-85.0 cm; from Dallas. Hidalgo, and Nacogdoches Coun- ties, Texas). The snakes were individually housed in glass terraria that measured 32 x 32 x 62 cm 147 148 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY or 27 x 32 x 52 cm. Each cage had a gravel sub- strate covered with leaf litter, a water bowl, and at least one large piece of bark for cover. Water was sprinkled over the leaves two or three times each week. The snakes were kept in a dark room that usually had a temperature of 22-24°C, but occasionally rose to 30°C. A 100 W bulb on top of the perforated metal cover of each tank raised the temperature at one end to ca. 24-26°C for 10 hr each day. Observations were timed with a stop watch and recorded on audio tape or with a 35 mm camera and electronic flash. Captive coral snakes were offered live or dead prey, as available, of the following species: Anolis carolinensis, Eumeces tetragrammus, E. fascia- tits. Scincella lateralis, Carphophis amoenus, Coluber constrictor, Diadophis punctatus, Elaphe obsoleta, Heterodon platyrhinos, Nerodia ery- throgaster, N. rhombifera, Opheodrys aestivus, Sonora semiannulata, Storeria dekayi, TantiUa gracilis, T. nigriceps, Thamnophis proximus, Tropidoclonion lineatum, and Virginia striatula. Live prey was released in a cage as far from the coral snake as possible. Dead prey was held with forceps ca. 20 cm from an active snake and jig- gled to simulate prey movements; if there was no response, the prey was moved closer until it was seized. Trail Following.— \ used a modified version of the arena used by Gehlbach et al. (1971), con- sisting of an 80 x 80 cm piece of white duck cloth (28 strands/cm 2 ) in a plastic swimming pool. An octagonal trail lane with segments 20 cm on an outer side and 1 cm wide was marked on the cloth with small, faint broken lines of indelible ink. Prior to an experiment a potential prey item was restricted to the trail lane by a portable 8 cm high cardboard alley and allowed to crawl around for one or two circuits. Then the prey animal and the cardboard alley were removed. Next a coral snake was confined in the center of the arena for three minutes in a bottomless 1- gal plastic jar. The snake was released by lifting the jar, and its behavior observed under a 60 W red light positioned so that the arena was very dimly lit. The cloth arenas were machine washed, rinsed, and dried after each test. Diet Studies. — Museum specimens were opened with a ventral incision and the orienta- tion of each prey item in the gut was recorded. The identity and approximate total length (TL) of each item was determined if possible, often on the basis of a tail or a tail and posterior portion of a body, by comparisons with published in- formation and intact reference specimens. Ad- ditional records were obtained from conversa- tions or correspondence with collectors and from the literature (Matthes I860; Hay 1893; Mitchell 1903; Strecker 1908; Schmidt 1932; Loveridge 1938, 1944; Klauber 1946; Ruick 1948; Minton 1949; Clark 1949; Telford 1952; Curtis 1952; Highton 1956; Martin 1958; Kennedy 1964; Myers 1965; Neill 1968; Chance 1970; Malloy 1971; Fisher 1973; Jackson and Franz 1981). Snout-vent (SV), tail, and head lengths of pre- served coral snakes were measured when possi- ble. Because many museum specimens had dam- aged heads, SV was used for comparisons with prey TL. I estimated the weights of common prey items from published statements and from live measurements of four Scincella lateralis, one Leptotyphlops dulcis, three Storeria dekayi, three Tantilla sp., six Tropidoclonion lineatum, and eight I 'irginia striatula. The average total lengths of all snakes in east Texas were taken as the midpoints of the ranges for adults given in Con- ant (1975). In a few cases I weighed preserved coral snakes and intact prey after blotting them on paper towels. I evaluated geographic variation in food habits by grouping records for Texas in four subsam- ples: "east Texas" (mixed deciduous and pine forests), "north central Texas" (tall grass-prairie- forest ecotone), "central Texas" (forested hill country of the Edwards Plateau and the extreme eastern edge of the Chihuahuan Desert), and "south Texas" (semiarid thorn scrub and sub- tropical forest, see Gould 1969, for vegetation regions). Records from elsewhere in the species range are grouped as "Florida" and "other" (Ar- kansas, Louisiana, South Carolina, and Mexico). Feeding Behavior The description that follows incorporates pub- lished accounts (Grijs 1898;Ditmars 1907, 1912; Clark 1949) and my observations. Feeding be- havior is discussed in six groups of sequentially and functionally related motor patterns to facil- itate future comparisons with other snakes. Encountering Prey. — Methods of encounter- ing prey should be included in discussions of feeding behavior, because snakes use species- typical postures and strategies for obtaining food. VERTEBRATE ECOLOGY AND SYSTEMATICA 149 Prey might be located by some type of searching, trail following, "sitting and waiting," (Pianka 1966), or a mixed strategy (Tollestrup 1980; Chiszar and Scuddcr 1980); each of these tech- niques might be enhanced by behavioral or mor- phological specializations. For example, search- ing and trail following utilize stereotyped poking behavior (in coral snakes, see below) and highly specialized receptor systems (e.g., facial pits in boids and some viperids). "Sitting and waiting" is probably more efficient when accompanied by camouflage (Fitch 1 960) or caudal luring (Greene and Campbell 1972). When a coral snake had not fed for several days, it crawled slowly over the substrate and poked its head in and out of the leaf litter. This involved repeated forward and lateral head movements, and was accompanied by frequent tongue flick clusters. At times a snake crawled slowly beneath a large leaf or a small piece of bark and soon emerged from the opposite side, still moving its head from side to side and flicking its tongue. When a coral snake was searching, any movement of an object in the terrarium elicited pointing and, if it was not a large object, ap- proach behavior. When an acceptable prey item caused the approach, it was seized and eaten. Unsuccessful attempts to capture prey were fol- lowed by more searching behavior. F. R. Gehlbach (pers. comm.) observed similar crawling and poking movements by two free- living coral snakes on the Santa Ana Wildlife Refuge. Hidalgo County. Texas, one of which I later used for behavioral studies. Neill (1951) described what was perhaps foraging behavior by a coral snake in Clay County, Florida. The snake crawled rapidly, moved its head from side to side, and poked its head into the surface litter. Neill also stated that the snake's tail made "con- stant rapid, probing motions" in the leaves, and that at times "the hind part of the creature was thrown nearly as far forward as the head." He observed similar behavior in a captive snake, and suggested that the head and tail movements served to flush small reptiles and amphibians from cover. These observations suggest that crawling and head-poking in ground litter are motor patterns normally used by coral snakes to locate potential prey items. However, neither Gehlbach nor I observed use of the tail in for- aging, and I doubt that it is a normal behavior, at least for coral snakes in Texas. The threshold for tail waving, an important component of coral snake antipredator behavior (Gehlbach 1972; Greene 1973b), is often very low for this species (pers. obs.), and perhaps the snake observed by Neill was responding defensively to tactile or vi- brational stimuli. Several species of small snakes deposit chem- ical trails that serve as attractant pheromones (Burghardt 1970; Gehlbach et al. 1971). and there are indications that these trails release searching and trail following behavior by coral snakes. Once two small earth snakes ( Virginia striatula) were kept in ajar of wet moss for several days before the snakes and moss were put in a coral snake's cage. The coral snake was crawling on the leaves and encountered the moss. It moved its head back and forth over the moss for approximately five minutes and frequently flicked its tongue. Then it crawled across the cage, generally fol- lowing the route taken by one of the earth snakes. The coral snake soon found the prey in a corner and ate it. During staged encounters with ground skinks {Scincella lateralis), a coral snake fre- quently paused for several seconds in the exact spot where a skink had recently rested and point- ed and tongue-flicked before searching again. Experiments with coral snakes on cloth arenas provide additional evidence that they respond to prey trails. For two trails with each of two coral snakes, a small colubrid snake (adult Storeria dekayi or I 'irginia striatula) was allowed to crawl around the alley one time. In each case the coral snakes crawled away from the central release point, paused briefly and pointed at the trail, and moved off the cloth. A second block of trials used trails laid by a small snake or a skink (adult female Eumeces fasciatus) making four circuits of the octagon in five minutes. One coral snake responded to two snake trails with pointing and then escape behavior, but followed a skink trail for one complete circuit and two additional turns on the octagon. The other coral snake followed trails laid by S. dekayi (two trials) and V. stria- tula (one trial) for one complete circuit, seven lane segments, and three lane segments, respec- tively. It followed two lane segments of a skink trail before crawling off of the cloth. These ob- servations suggest that known prey species can leave trails which are perceived and followed by coral snakes. Additional experiments using more coral snakes, more prey species, and more trials are required before comparisons with the exten- 150 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY sive study by Gchlbach et al. (1971) are war- ranted. Trail following behavior was stereotyped and similar to that described by Gehlbach et al. ( 1 97 1 ) for blind snakes. Leptotyphlops dulcis. A coral snake crawled slowly from the release site, point- ed and flicked its tongue at the trail, then turned 90° and began following it. The snake's head re- mained elevated while it crawled, and there were frequent tongue-flick clusters. At each corner it overshot 2-4 cm, paused, pointed and tongue- flicked at the cloth, moved its head from side to side, turned back onto the trail, and resumed crawling. If a wire was jiggled on the cloth in front of a coral snake it pointed and approached rapidly. The available captive and field observations imply that coral snakes actively search for prey, but the frequency and extent of foraging move- ments are unknown. There is no evidence that free-living coral snakes use a "sit and wait" strat- egy to ambush prey, but the behavior of captives suggests that they might. My snakes were fre- quently seen coiled with head raised and pro- truding from beneath the edge of a piece of bark or pile of leaves. Such snakes responded to near- by movements by pointing, tongue-flicking, and approaching. Recognition and Approach. — Recognition of prey probably begins as soon as a coral snake points toward a stimulus, and incorporates vi- sual and chemical cues. Captives approached any small movement, such as a wire jiggled in the leaves or a finger moved against the glass from outside of the terrarium. Larger moving objects, such as a hand or a piece of bark, usually elicited pointing and then rapid head withdrawal and crawling. This seemed especially likely if the ob- ject was moved suddenly. Approach was accompanied by tongue-flick clusters, which evidently convey the necessary stimuli for seizing or avoiding a potential prey item. Coral snakes quickly approached to within 2 cm of large coleopteran larvae, cricket frogs (Acris crepitans), and newborn mice, but then withdrew without seizing them. Small live water snakes (Nerodia sp.) were also approached and rejected, and in most cases they had discharged the cloacal sac contents. However, rapid prey movements seemed to result in a quicker attack and to override aversive chemical cues. Dead Nerodia were usually refused when stationary or pulled slowly but were attacked when pulled more rapidly. In 10 incomplete feeding sequences, a prey item was grasped and immediately released, or maneuvered for a short time and then re- leased. This suggests that a coral snake continues to receive input from the prey after it is seized, perhaps via either oral sensory papillae (Burns 1969; Greene, unpublished) or the Jacobson's Organ (cf. Burghardt 1970). Capture and Immobilization. — Approach was usually slow if the prey snake was moving slowly, and rapid if it crawled away quickly. Prey was seized with a quick forward movement of the anterior part or entire body of the coral snake, usually from a distance of several centimeters. In some cases a coral snake crawled parallel to a moving snake, flicked its tongue several times, and then seized the prey by turning its head sharply to the side and down. Coral snakes have relatively small eyes (Marx and Rabb 1972) and apparently cannot strike very accurately. Live Scincella lateralis proved difficult for them to seize, perhaps because of the coral snakes' relatively poor vision and the skinks' small size and erratic escape behavior (Lewis 1951). Also, ground skinks seemed to perceive an approaching predator at a distance of several centimeters and often slipped away unseen. Dur- ing 1 1 attempts on these lizards by a coral snake, I observed eight misses, two tail autotomies (skink escaped unharmed), and one capture. These were during staged confrontations on a 32 x 62 cm substrate of gravel and scattered leaves, and the only capture occurred when the snake trapped a skink in a corner. Small live prey snakes pre- sented a slower and more elongate target, and were captured without difficulty; each of 23 at- tempts was successful. Ditmars (1907) and Clark (1949) stated that Micrurus fulvius immobilizes its prey with ven- om before swallowing, but Ditmars (1912) re- marked that the venom is of little value in sub- duing "cold blooded" animals. My observations indicate that this species typically holds prey at the point of seizure until paralysis and then be- gins pre-ingestion maneuvers (see below). Slight movements of the prey were sometimes seen even as the tail was swallowed, suggesting that it is immobilized but not immediately killed by the venom. Coral snakes usually dragged their prey a few centimeters backward or forward before pausing, seemingly in response to its struggles. VERTEBRATE ECOLOGY AND SYSTEMATICA 151 Table 1. Pre-ingcstion latencies (in seconds) for coral snakes. Micrurus fulvius, dealing with live and dead prey. Ranges, means, standard deviations, and sample sizes are given. Latency Snake no. 4 Snake no. 5 Time between seizure and onset of pre-ingestion maneuvers (live prey) Time between seizure and onset of pre-ingestion maneuvers (dead prey) Time between last prey body movement and onset of pre-ingestion maneuvers 290-595 (jc = 434.2 ± 132.5) N = 6 0-85 (a- =28.6 ± 32.3) N= 10 63-190 (v = 99.2 ± 52.6) N = 5 70-940 (a- = 400.9 ± 334.9) N = 8 0-290 (a= 73.7 ± 86.5) N= 10 0-152 (a =71.8 ± 73.3) N = 4 This tended to untangle a small, writhing snake, and it might also imbed the fangs more deeply. During envenomation. the temporal region of the coral snakes sometimes appeared shriveled: this was probably caused by contraction of the M. adductor mandibulae externus superficialis, which has been shown to force venom out of the main venom gland in an elapid, Bungarus cae- ruleus (Rosenberg 1967; see also Savitzky 1978). In two instances a coral snake bit and quickly released an adult female Eumeces fascial us that struggled violently. One of the skinks was im- mediately recaptured. The other lizard crawled slowly for several centimeters and went under a piece of bark. It was soon followed by the coral snake and regrasped. Both skinks subsequently- made only feeble movements and were eventu- ally eaten. Pre-ingestion Maneuvers. — Coral snakes nor- mally do not release prey prior to swallowing it. Pre-ingestion maneuvers are probably evoked by tactile and/or chemical cues (cf. Nalleau 1966) and inhibited by prey movements. If prey move- ments inhibit the coral snake, the time between seizure and the onset of pre-ingestion maneuvers should be longer with live prey than with dead prey. The mean pre-ingestion handling times with live and dead prey (Table 1) differed significantly for each of two coral snakes (P < .01, Mann- Whitney c'test). If prey movements inhibit the snake, the time between the last prey movement and the onset of preingestion maneuvers should be similar for live and dead prey. These times were significantly different for one snake (P < .01, Mann-Whitney c'test) but not for the other snake (P > .90). I interpret the large variances and the equivocal results of the last comparison as resulting from individual differences and from the use of different sizes and species of prey in the feeding trials. Captive and free living coral snakes almost always swallowed prey head first, and scale over- lap on the prey item was used as a cue in locating its anterior end (Greene 1976). Alternating jaw movements, typical of snakes (Gans 1961). were used to shift along the prey's body prior to swal- lowing. In one instance a small stick in the mouth of a coral snake prevented it from shifting over a snake's snout to begin swallowing. The coral snake released the prey, removed the stick by jaw movements and rubbing its head on the sub- stance, regrasped the prey by the snout, and swal- lowed it. In all other feeding sequences, prey snakes were not released before they were swal- lowed. Swallowing.— After the prey's head had been shifted down the throat, it was swallowed by re- peated series of alternating jaw movements. These were separated by brief pauses and accompanied by lateral movements of the entire head. Ac- cording to McDowell (1970), Micrurus belongs to a group of elapids in which "the palatine is erected along with the maxilla during maximum protraction of the palate." This presumably oc- curs when a coral snake's head is rotated back and forth across a prey snake's long axis during swallowing movements. I could not observe the action of the palatine bones in live coral snakes, but frequently saw the maxillary fangs depress (and penetrate?) a prey snake's skin during swal- lowing. During swallowing a coral snake sometimes rolled about its long axis, perhaps using the prey's inertia to achieve better contact between its teeth 52 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 2. Frequency of prey items by taxon in eastern coral snakes, Micrurus fulvius. Abbreviations refer to east Texas (E), north central Texas (N), central Texas (C), south Texas (S), unknown localities in Texas (U), Florida (F), other parts of the species range (O), and total for all localities (T). Number of coral snakes containing prey for each sample is in parentheses. Prey species E N C s u F O T (46) (1?) (16) (14) (10) (57) (14) (177) Amphibia Anura Unidentified Reptilia Unidentified Amphisbaenia Rhineura floridana Sauria Unidentified Anguidae Ophisaurus sp. Teiidae Cnemidophorus gularis Iguanidae Sceloporous sp. S. undulatus Scincidae Unidentified Eumeces sp. E. fasciatus E. inexpectatus E. tetragrammus Neoseps reynoldsi Scincella lateralis Serpentes Leptotyphlopidae Leptotyphlops dulcis Colubridae Unidentified Arizona elegans Coluber constrictor Diadophis punctatus Elaphe guttata E. obsoleta Farancia abacura Ficimia olivacea Lampropeltis calligaster Opheodrys aestivus Salvadora grahamiae Seminatrix pygaea Sonora semiannulata Stilosoma extenuatum Storeria dekayi S. occipitomaculata Tantilla sp. T. coronata T. gracilis T. planiceps T. relicta T. rubra Thamnophis sp. T. marcianus T. proximus Tropidoclonion lineatum Tropidodipsas sartoni 12 15 1 3 2 1 1 5 1 13 1 1 1 2 1 9 3 1 2 21 2 3 4 4 1 19 1 5 1 5 1 8 2 11 1 1 1 3 4 1 2 3 3 1 1 9 1 3 3 1 2 1 1 6 2 1 2 3 11 1 1 2 3 2 3 2 1 1 4 1 3 1 7 1 1 1 7 1 1 3 1 4 1 VERTEBRATE ECOLOGY AND SYSTEMATICS 153 Table 2. Continued. E N C s u F ( > T Prey species (46) (15) (16) (19) (10) (57) (14) (177) I irginia striatula 8 9 1 1 19 V. valeriae 2 2 S. semiannulata or Tantilla sp. 1 1 4 6 S. dekavi or V. striatula 5 2 7 Elapidae Micrurus fulvius 2 3 5 Viperidae Agkistrodon contortnx 2 2 Mammalia Rodentia Unidentified 1 1 2 Totals 60 17 17 25 13 67 22 221 and the prey's skin. As swallowing neared com- pletion, lateral bends of the anterior part of the body moved the prey posteriorly in the gut. Sometimes the snake attempted to maneuver the prey's tail into its mouth by head rubbing or snout pushing against the substrate. As the prey's tail was swallowed, a coral snake usually raised its head almost vertically and two to ten centi- meters from the substrate. Post-ingestion Behavior. — Swallowing was al- ways followed by tongue-flick clusters, and sometimes by yawns. These usually occurred prior to lowering the head, and were followed by searching behavior. Occasionally a snake rubbed its head on the substrate after swallowing was completed. After feeding, coral snakes always re- sponded to small movements (e.g., a wire jiggled in leaves) by pointing, tongue flicking, and ap- proaching. If another prey item was offered it was seized and eaten. Discussion. — Snakes exhibit several behavior- al grades of prey immobilization. Most nonven- omous colubrids probably rely on a weight ad- vantage and simply grasp and swallow living prey. Some other colubrids and probably all boids use constriction to bring about immobilization by suffocation (Greene and Burghardt 1978). Many viperids and at least one sea snake release and then relocate relatively large prey before swal- lowing (Loop and Bailey 1972; Chiszar and Scudder 1 980; Radcliffe and Chiszar 1 980; Jacob and Greene, unpublished). Micrurus fulvius typ- ically uses a variation on the simple colubrid pattern, in that prey is seized, held, and immo- bilized by venom injection and then swallowed. Similar behavior occurs in other species of Mi- crurus (Greene 1973a) and in sea snakes of the genera Enhydrina, Hydrophis, and Pelamis (Pickwell 1972; Voris et al. 1978). However, M. fulvius (this paper), M. frontalis (Lankes 1928; Mertens 1956), and perhaps M. lemniscatus (Mole 1898; Lankes 1938) occasionally release prey if it struggles violently or if some other dif- ficulty is encountered in the preingestion phase; they then resume maneuvers at the same stage in the feeding sequence. It is interesting to note that the variable feeding behavior of these Mi- crurus combines a simple pattern seen in colu- brids and some proteroglyphs (seize, hold, swal- low) with a more complex sequence obtaining in some proteroglyphs (Armitage, 1965; Lambiris, 1967) and in many solenoglyphs (strike, release, relocate, seize, swallow). Unlike Ditmars (1912), I do not imply that this pattern mirrors the phy- logeny of colubrids, elapids, and viperids. Ad- ditional information on feeding in other species and on intergeneric and interfamilial relation- ships is required before these observations can be meaningfully applied to scenarios of snake evolution. Diet Ecology Taxonomic Composition and Seasonal I 'ari- ation. — Eastern coral snakes of all sizes are spe- cialized tertiary consumers; snakes, amphisbae- nians, and elongate lizards comprised 97% of 22 1 prey items from throughout the species range (Table 2). Seventy of 132 items (53%) in Texas snakes were colubrid or leptotyphlopid snakes of 154 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY Table 3. Seasonal incidence of three prey types in coral snakes, Micrurus fulvius, from Texas. Number of coral snakes containing prey per season are in parentheses following months. Decimal fractions indicate contribution of each prey type to the total prey sample for each season. Seasons Skinks Juvenile large snakes Other prey Total Spring. March-May (33) Summer. June-August (16) Fall, September-December (33) Total (82) 12 (.29) 2 (.05) 27 (.66) 41 2 (.11) 3 (.17) 13 (.72) 18 9 (.22) 10 (.24) 22 (.54) 41 23 15 62 100 the genera Diadophis, Leptotyphlops, Sonora, Storeria, Tantilla, Tropidoclonion, and Virginia; undoubtedly, some of the unidentified snakes also belonged to these genera. These secretive snakes are normally found in litter or beneath logs or rocks (Wright and Wright 1957; Kassing 1961; Clark 1964). Diadophis. Sonora, Tropidoclo- nion, and Virginia usually have TLs of 20-40 cm and weigh 3-10 g. Leptotyphlops and Tantilla are shorter (TL < 30 cm), more slender, and weigh less (1-3 g). Five coral snakes had also eaten green snakes (Opheodrys aestivus); these are relatively long arboreal snakes T (TL < 80 cm), but no thicker than an adult Tropidoclonion. Nine Eumeces made up 7% of the total Texas sample. The E. fasciatus were females or sub- adult males (TL = 1 3 cm), and probably weighed 5-7 g (Fitch 1954). Adult E. tetragrammus (TL = 12 cm) are more slender than E. fasciatus and probably weigh slightly less. Eighteen Scin- cella lateralis comprised 14% of the prey items, but this species probably makes only a small contribution to total prey biomass of Micrurus fulvius. Adult S. lateralis (TL = 8 cm) weigh ca. 1 g, and only four records represent confirmed ingestion of an entire skink. In 12 cases only the tail was found; six of these had clearly been au- totomized, and five others looked as if they had been also. It seems likely that in most cases the skinks escaped, and observations on captive en- counters (see above) support this conclusion. The monthly distribution of 23 skinks {Eumeces and Scincella) in stomachs suggests that they are more frequently eaten in the spring and fall (Table 3). Most of the other prey were young or subadults of large terrestrial snakes (E/aphe, Lampropeltis, Salvadora, Thamnophis, Agkistrodon). Aquatic snakes (Nerodia, Thamnophis). amphibians, and mammals were very rarely eaten, and the three items in the latter two classes might have been secondarily ingested. Five Micrurus fulvius made up 2% of the total 221 items from throughout the species range. Curtis ( 1952) suggested that two males from An- gelina County, Texas attempted to swallow a Storeria dekayi from opposite ends, and that the smaller coral snake was then eaten by the larger one. This seems unlikely, because it implies that two free-living coral snakes found a single item almost simultaneously and, more importantly, that one of them attempted to swallow the prey tail-first (cf. Greene 1976). Ardrey (1970) mis- represented Loveridge's (1944) account of can- nibalism in M. fulvius and suggested it as an example of inability to control aggressive social behavior, an unsupported speculation. There is no evidence for size of sex as an explanation, since the following combinations of predator and prey were involved: two adult males had eaten other adult males, one adult male had eaten a juvenile male, one adult male had eaten a gravid female, and one adult female had eaten another adult female. Geographic Variation. — Regional variation in the diet of Micrurus fulvius largely reflects the distributions of particular prey species (compare Table 2 with maps in Conant 1975), rather than shifts in the general size and type of prey taken. Each Texas subsample included small secretive snakes, skinks, and the young of large snakes. Coral snakes frequently ate small colubrids in each region: I 'irginia and Storeria in east Texas; these genera and Tropidoclonion in north-central Texas; and Sonora, Storeria, and Tantilla in the more xeric southwestern and southern parts of the state. An obvious exception is that predation on scincid lizards occurred more often in the mesic forests of the eastern parts of the state. Eighty-nine items in 7 1 Micrurus fulvius from elsewhere in the species range (Table 2) are con- sistent with the Texas data. Forty small, secretive colubrids of the genera Diadophis, Ficimia, VERTEBRATE ECOLOGY AND SYSTEMATICA 155 500- • ^_^ E E • ^-» 400- X h- O z LU 300- _l • • • _l •• • < • «• h~ • • O 200- • • • • \- • • • • • >■ • • o LU • cc nn • o m Q_ 100- o o . "H 200 300 400 500 600 700 800 PREDATOR SNOUT-VENT LENGTH (mm) Fig. 1. Relationship between snout-vent length of coral snakes, Micrurus fulvius, and total length of prey. • indicates snake prey and o indicates lizard prey. Seminatrix, Stilosoma, Storeria, Tantilla, Tro- pidodipsas, and Virginia accounted for 45% of the sample. The remaining items included 12 limbless anguid lizards (14%), 12 scincid lizards (14%), five limbless amphisbaenians (6%), and 21 other snakes (22%). The only mammal was represented by a few hairs in a specimen from Tamaulipas, Mexico, and was perhaps second- arily ingested. Predator Prey Size Relationships. — Even very small coral snakes eat skinks and small colubrid snakes. Seven specimens from Texas (SV 25-29 cm) contained three Scincel/a lateralis (tails only), scales of an unidentified skink, one Opheodrys aestivus, one Storeria dekayi, and one Virginia striatula. A snake hatched in captivity began feeding on S. lateralis at an age of two months (Campbell 1 973), and another very small captive also ate these lizards (Sochurek 1955). Coral snake SV was positively correlated with prey TL (Fig. 1). although large coral snakes sometimes ate small prey. On some occasions Micrurus fulvius eats colubrids that are quite large (Matthes 1860; Mitchell 1903), but such meals are infrequent and can even be fatal (Neill 1 968). If a 45 g adult M. fulvius ate a 9 g prey snake, the meal would equal 20% of the coral snake's weight. In six cases where adult coral snakes (23.5-64. 1 g, x = 4 1 .5 g) and their prey could be weighed, the items comprised 2-131% (x = 41.7%) of the predators' weights. Discussion. — My results demonstrate that throughout its range Micrurus fulvius feeds al- most entirely on small snakes, elongate lizards, and amphisbaenians. Within these broad taxo- nomic. size, and shape limits, an interesting va- riety of prey are eaten. Factors that can affect diet composition include demographic, behavioral, and morphological characteristics of the preda- tor and all of its potential prey species (Holling 1959; MacArthur and Pianka 1966; MacArthur 1972; Schoener 1971). These components are in- cluded in the concepts of availability and vul- nerability, where "availability means that [par- ticular] prey organisms are present, and vulnerability encompasses all physical and bio- logical conditions that cause one species to be preyed on more heavily than another" (Hor- nocker 1970:29; availability is relative to the densities of particular prey populations as well as absolute in terms of their presence and ab- sence). Demographic data on M. fulvius and its prey are lacking, but available information on the natural history of these animals suggests ways in which availability and vulnerability interact to determine the diet. Small fossorial and terrestrial snakes are es- pecially vulverable to capture by coral snakes 156 SPECIAL PUBLICATION-MUSEUM OF NATURAL HISTORY PREY WEIGHT Fig. 2. Costs and benefits for coral snakes feeding on skinks (SK) versus snakes (SN). The straight line (F) indicates food value. See text for details. because of their size, absence of effective anti- predator behavior, and preferred microhabitat. Predation on juveniles of larger species of snakes is presumably more restricted because of their seasonal availability in temperate climates (Ta- ble 3; Fitch 1970). Size, defensive capabilities, and microhabitat probably also influence coral snake predation on lizards. It appears that skinks are not as vulnerable as snakes because of their smaller total length, agility, and capacity for tail autotomy. Whiptails {Cnemidophorus sp.) are sympatnc with Micrurus fulvius throughout its range, but these highly mobile lizards prefer hot, open areas (Fitch 1958) and are probably rarely encountered by coral snakes. Small iguanids {Anolis sp., Sceloporus sp.) are abundant in some places and anoles are sometimes accepted as food by captives (pers. obs.); however, these lizards are probably not important in the diet of coral snakes because they are largely arboreal and would not often be found by a foraging Micrurus. What follows is a post hoc consideration of "ideal" prey size, "ideaF prey type, and two as- pects of variation in the diet of Micrurus fulvius (see Appendix, Note 1). For this purpose, loca- tion costs include the energy expenditure and risk required to bring a snake within attack distance of its prey, and handling costs include the energy expenditure and risk involved in capturing and ingesting an item (these terms include search time and pursuit time, respectively, of MacArthur and Pianka 1966). Food value includes the energy and other nutritional factors present in a prey UJ 4 O HI a o 3 LL O 2 cc LU CO ^ 1 •:•:•:•:•:•: 15- 29 30- 44 45- 59 60- 74 75- 89 90- 104 105- 119 120- 134 135- 144 AVERAGE TOTAL LENGTH (cm) Fig. 3. Frequency distribution of average total length for snake species that are sympatric with Mi- crurus fulvius in eastern Texas. Cross hatching indi- cates species in genera frequently eaten by coral snakes. Stipple indicates species that are relatively unavailable (rare) or invulnerable (too stout). See text and Appen- dix Note 3 for details. item; I assume that food value is a positive func- tion of weight and equal on a per gram basis for all items (e.g., skinks versus snakes). The overall value of a prey item is thus some function of location costs, handling costs, and food value. Following MacArthur and Pianka (1966), I as- sume that prey items differ in terms of handling costs (due to differential characteristics of pred- ator and prey), and that the addition of different prey types to the diet results in lower location costs (due to increased prey density). The overall value of a prey item increases as a function of weight (as does food value. Fig. 2) up to some point for a coral snake and then drops rapidly to zero. This is because as the diameter of the prey approaches and then exceeds the gape of the predator, ingestion will rapidly become difficult and dangerous (Appendix, Note 2). Ideally, coral snakes should take items that are to the left of but as close as possible to the food value-han- dling cost intercept in Fig. 2. It is true that large coral snakes sometimes eat larger prey than do small individuals (Fig. 1 ), but large individuals also frequently eat small prey. Similar predator/prey size relationships have been demonstrated in a sea snake (Voris and Moffett 1981), males of a small, burrowing colubrid (Seib 1981), primitive snakes of the family Aniliidae (Greene 1983), and among small and medium size classes of a viperid (Beavers 1976). At least three factors might account for the fact that large VERTEBRATE ECOLOGY AND SYSTEMATICS 157 coral snakes sometimes eat small prey: (i) Be- cause of the negative allometry of metabolic rate that obtains in most snakes (Bennett and Dawson 1976), an item of a particular relative weight might contribute proportionately more to the to- tal energy budget of a large snake than it would to that of a smaller individual. Whether this is actually true for large and small coral snakes is not known, (ii) Occurrence of young individuals of large prey snakes is seasonally restricted, and they are thus not a predictable resource for coral snakes at all times of the year, (iii) The size con- figurations of terrestrial snake communities in temperate forests can be discontinuous; in east Texas they contain several species of very small, moderate, and large snakes, but very few me- dium-small species relative to adult M. fulvius (Fig. 3: Appendix, Note 3). In other words, be- cause of (ii) and (iii), large coral snakes in the southeastern United States probably rarely en- counter prey snakes proportionately as large as those eaten by small individuals. Skinks are more heavy bodied than small snakes and as a result their food value-handling cost intercept occurs at a lower weight (Fig. 2). The disparity is increased by the lower vulner- ability of skinks (see above and Vitt et al. 1977) and perhaps by their capacity for inflicting a pow- erful bite on the predator. In other words, skinks are probably more costly to handle than small snakes of equivalent weight and provide less food value than small snakes of equivalent handling cost. Ideally, coral snakes should add skinks to their diets only when location costs are reduced proportionate to the increased handling costs these lizards impose. This suggests an explana- tion for the increased predation on skinks in east Texas: quantitative data are lacking, but my field experience is that skinks are much more com- monly encountered there than in other parts of the state where coral snakes occur. In any case, the stomach contents and behavioral observa- tions certainly imply that Micrurus fulvius often attacks skinks and that these encounters fre- quently result in little or no net energy gain for adult coral snakes (three of 12 records of skink tails were for very small M. fulvius, for which they might have been proportionately large items). Either skinks (or skink tails) are propor- tionately more valuable than small snakes in per gram food value (cf. Clark 1 97 1 ; B. E. Dial, pers. comm.) or the overall expectation of finding "better" items is low enough to make them worth chasing in spite of the very low average payoff per attempt. These considerations suggest that eastern coral snakes attack and sometimes eat substantial numbers of intuitively non-ideal prey (skinks, relatively small snakes). That they do so is per- haps surprising, because many snakes apparently feed infrequently on relatively heavy items (Greene 1983, MS) and such predators might be especially able to defer feeding until a highly prof- itable prey could be located. There are at least two plausible, non-exclusive reasons why Mi- crurus fulvius does not meet this prediction: (i) Coral snakes might forage so as to minimize the time required to find and consume a given amount of food, rather than to maximize the intake of energy in a given time period or prey- encounter (Schoener 1969; Morse 1980). In doing so they would reduce the time of exposure to predators and gain time for other activities, but the importance of either factor in co