BULLETIN OF THE FLORIDA STATE MUSEUM BIOLOGICAL SCIENCES Volume 14 Number 3 THE COTTONMOUTH MOCCASIN ON SEA HORSE KEY, FLORIDA Charles H. Wharton UNIVERSITY OF FLORIDA Gainesville 1969 Numbers of the BULLETIN OF THE FLORIDA STATE MUSEUM are pub- lished at irregular intervals. Volumes contain about 300 pages and are not neces- sarily completed in any one calendar year. WALTER AUFFENBERG, Managing Editor OLIVER L. AUSTIN, JR., Editor Consultant for this issue: J. c. DICKINSON, JR. Communications concerning purchase or exchange of the publication and all manuscripts should be addressed to tlie Managing Editor of the Bulletin, Florida State Museum, Seagle Building, Gainesville, Florida 32601. Published 31 December 1969 Price for this issue $.80 THE COTTONMOUTH MOCCASIN ON SEA HORSE KEY, FLORIDA CHARLES H. WHARTON1 SYNopsis: From 1954 to 1957 the author studied the cottonmouth mOccasin (Agkistrodon pisciuorus) population of the islands near Cedar Key, Levy County; Florida. Sea Horse Key, one of the outer islands, approximately 1 mile long and 7 miles from the mainland, then supported a total population of roughly 600 cottonmouths. The cottonmouths aggregate under the breeding colonies of cormorants and herons on the outer Cedar Keys and scavenge the fish the birds drop accidentally from their nests. After post-nidal feeding stops in August, they eat birds, rats, and squirrels. An abundant skink is the principal food of young cottonmouths. The bird rookeries control distribution and food habits of three-fourths of the island's snakes, serving snakes up to 150 meters distant. Activity ranges are remarkably small ( males average 0.43 acre, females 0.35 acre). Although the main ridge lacked rookeries, ranges were similar. No evidence of territoriality was found. Snake Key rookeries also support a large population of cottonmouths. Atsena Otie and North Keys have no rookeries and few cottonmouths. Island snakes den in shallow stump holes and under debris. Frequent warm winter periods make emergency demands on stored fat, and 77 per cent of Sea Horse Key snakes are in danger of starvation during winters with a mean tem- pdrature of 16.3' C, compared to 36 per cent of the fatter snakes on Snake Key. The fat bodies of snakes apparently function as reserve food during periods of high temperature. Fat-body weights, length/weight ratios, and feeding behavior suggest that Sea Horse snakes are at a critical survival level. When removed from their winter dens, cottonmouths apparently shun all holes for periods of up to 2 years; a persistant site memory is postulated. At lowest cloacal temperatures (4.0'C) cottonmouths are passive; they are able to strike at 4.5'C, crawl at 12.5'C, and feed at 14.5'C. Variation in ob- served rates of heating and cooling suggests some ph-ysiological control. Up to 36 per cent of cottonmouths showed aggressive behavior at low temperatures con- trasted to 4.1 per cent at high readings (21'C and above). The major cause of mortality is starvati6n. Adults have no enemies other than linguatulid parasites and man. Two healthy and apparently genetically eyeless snakes indicate the relative importance of olfaction in the island environ- ment. The ambithermal cottonmouth, with its endogenous biennial sexual cycle, its vagile nature, and its keen olfaction seems well adapted to preempt the island niche of a terrestial carnivore-scavenger. 1The author is Professor of Biology at Georgia State College, Atlanta, Georgia. He submitted an earlier version of the material presented here to the University of Florida in partial fulfillment of the Ph.D. degree ( 1958). Wharton, Charles H. 1969. The cottonmouth moccasin on Sea Horse Key, Flor- ida. Bull. Florida State Mus., vol. 14, no. 3, pA 227-272 228 BULLETIN FLORIDA STATE MUSEUM Vol. 14 TABLE OF CONTENTS INTRODUCTION __.______.____---*...-_._._.-_228 FOOD AND FEEDING ._..247 ACKNOWLEDGMENT ._......_...._.__.-_...231 HIBERNATION -_._-_-___....__-_--__-_ 253 MATERIALS AND METHODS _____..__.231 TEMPERATURE AND BEHAVIOR ___. 259 AGGREGATION __._..... ...232 BODY FAT AND SURVIVAL ____ 263 EFFECT OF ROOKERIES ON MOVE- MORTALITY AND PREDATION . 268 MENT, ABUNDANCE, AND DISCUSSION 269 DISTRIBUTION .-_.--_.-_.--_____....236 LrrERATURE CITED 270 AcTIvrry RANGE AND MOVEMENT _..,242 INTRODUCTION The cottonmouth moccasin, Agkistrodon pisciuorus, is the domi- nant terrestrial vertebrate on two islands 6f the Cedar Keys, Levy County, Florida ( Fig. 1). The present study, made from 1954 to 1957, sought to discover factors resp6nsible for the high population density on Sea Horse Key, which offers an environment markedly different from the species' usual mainland habitats ( Wharton, 1958). This island, lying on an immense sand flat 5 to 7 miles offshore, is an old, high, sand dune covered with a modified climax forest. Its lee shores support mostly black mangrove ( Auicennia niti(la) ( Fig. 2) with some white mangrove ( Laguncularia'racemosa) and red man- grove ( Rhizophora mangle ). I have reported elsewhere ( 1966 ) on reproduction and growth of the cottonmouths of the island. The present study deals principally with the behavior of the snakes in their adaption to this unusual and seemingly hostile environment, almost entirely without fresh water. Arriving by drift or storm, the cottonmouths have adapted sur- prisingly well to these offshore islands. Some protection is afforded the snakes by the fact that the three outermost islands, North, Snake, and Sea Horse, lie within the Cedar Keys National Wildlife Refuge, established to protect breeding colonies of birds, primarily white ibis ( Guara alba), American egret ( Casnierodius albus), snowy egret C Leucophoyx thula) and several heron species. The islands also afford protected nesting sites for cormorants ( Phalacrocorax auritus), osprey ( Pandion haliaetus) and bald eagle ( Haliaetus leucocephalus). Cormorants and larger herons are especially importaht provisioners of the cottonmouth population. It is generally conceded that on the mainland cottonmouths are largely aquatic or semi-aquatic. On the Cedar Keys islands they lead -- a- terrestrial life,-seldom, if ever, venturing into the surrounding salt water. I questioned 14 long' term residents and many fishermen at Cedar Keys; only one, James Richburg, saw a snake of this species 1969 WHARTON: COTTONMOUTH MOCCASIN 229 Z A- W 0 ) W W ZX X W Ul 62 *1 r: nw ~ , 2-inrh. - ¢4-:..: Y . . , ./------------ i *31 3SH·OH V35 . 0 -% ---1. ./ , i -- 1 /0 \ I , , I . I ·~ i 1 -1- >; i ' 44= STV55 - 3 1 I K - - FIGURE 1. The major islands of the Cedar Keys. Snake, Sea Horse, and North Keys comprise the Cedar Keys National Wildlife Refuge. Dashed lines indicate main ship channels. GARDNER'S SCALE LOW i POINT o IOO METERS PENINSULA 0 328 FEET /,' CHANNEL /// 4 ENTRANCE M- LAGOON ); ''S, 9 ROOKERY 03 MANGROVE (M) 050 , ~u 51 PALMETTO HIGH MARINE LAB OSI SANDFILL RIDGE BEACH LIGHTHOUSE FIGURE 2. Maps of Sea Horse Key showing the extent of rookeries, mangroves, and saw palmetto ( Serenoa repens). 1969 WHARTON: COTTONMOUTH MOCCASIN 231 swimming ( at Snake Key). Kirk Strawn, who studied marine life on the Rats around Sea Horse Key, never saw a cottonmouth in salt water. I worked the shore of Sea Horse Key many times on foot and by boat, both day and night, and I never saw a cottonmouth swim- ming in the salt water other than when they were provoked by handling. Nor have I seen any signs of snakes moving to or from the beach. Sometimes cottonmouths coil at the strandline; presence of cover, beach-scavenging rats, or an activity range adjoining the beach may explain these instances. I found no instance of their feeding on fish that could not be attributed to the nesting, feeding, or roosting of birds. That these snakes do not directly exploit the plentiful fish life in the shallow, calm waters of Sea Horse Key is difficult to reconcile with the habits of the cottonmouth on the mainland. Only a few notes have been published on marine island cotton- mouths. Sass ( 1926) thought that cottonmouths on the Isle of Palms, South Carolina, rarely, if ever, fed on marine fishes. Wood ( 1954) noted mass migration from Virginia's barrier islands to the mainland. Carr ( 1936) and Wharton ( 1958, 1960, 1966) have made the only previ6us reports dealing directly with the Cedar Keys cottonmouths. ACKNOWLEDGMENTS This study was made possible by the cooperation of Kent Myers and other personnel of the Refuge Division of the U.S. Bureau of Sports Fisheries and Wildlife. For guidance during the course of this study I am especially in- debted to the faculty of the Zoology Department, University of Florida, particu- larly J. C Dickinson, Jr., and Archie Carr. E. Lowe Pierce, Director of the Sea Horse Key Marine Biological Laboratory, extended me many courtesies ahd transportation through his caretaker, Doyle Folks of Cedar Keys. John N. Ham- let kindly allowed me the use of his boat for the first half of my study. I ac- knowledge the secretarial and other assistance of Carol Ruckdeschel and Patricia Parker provided by the Biology Department of Georgia State College, and of Taylor Murray of the Georgia State computer center. MATERIALS AND METHODS Most of the cottonmouths occupied the island's western end. The thick vegetation made it necessary to cut trails through this area; and 206 trail stations were designated with stamped aluminum tags nailed on trees at 25 meter intervals ( Fig. 3). Snakes were approached carefully and restrained gently by a leather noose around the neck. By the use of subcaudal scale- clipping, (Blanchard and Finster, 1933), 402 cottonmouths were marked individually. These afforded 510 recaptures, counting dissec- tions, and a total of 545 individuals were handled. Data recorded on 232 BULLETIN FLORIDA STATE MUSEUM Vol. 14 a keysort card for each capture included length, weight, cloacal tem- perature, and heart rate (if the specimen was quiet). Nearby soil, air, and leaf mold temperatures, humidity, body position, behavior GARDNER'S POINT SCALE 16#--4~~1~5162 0 50 100 '/ -0-9 . 160 lili lilli tf . 8 METERS '|,~ ~29 ~ I 7125.. ~~ AL "-*' > Yf*-il¥ 5241 \,iso r:, ..---0-0 , N O R TH W E S T C O VE '~5- 127 \, \1 MANGROVE*d'~.e,4 , INLE1~0-00 - I .- -- I \ ,75,52.'6 200 'r ' i173-*.$- 79 \. ' 136 i; f K, 93 2RS I i' LAGOON ~» 3 1791'JO:21 ~05\ 10 +0./ 0-*t \485 d_e 149 20 N BEACH FIGURE 3. Sea Horse Key trail system. Points are stations 25 meters apart. Stations 1-30, high ridge of island. Dashed line, edge of mangroves. ( activity in which the snake was engaged), and demeanor before noosing were also noted, For estimating the size of cottonmouths found dead in the field the largest vertebrae and ribs were compared with a skeletal series of 33 specim6ns of known size and weight. Cloacal temperatures were taken with Schultheis rectal thermom- eter; soil and air temperatures with either the Schultheis instrument or a standard mercury thermometer. A maximum-minimum ther- mometer was kept at ground level in the forest of the central ridge throughout the study. Temperature and humidity in den cavities and in adjacent microclimates were taken with a battery-operated Aminco instrument with a temperature-humidity probe. AGGREGATION While records of snake aggregations in the literature apply chiefly 1969 WHARTON: COTTONMOUTH MOCCASIN 233 to denning such as the record by Noble and Clausen (1986),or to anusual concentrations of snakes caused by factors such as rising water ( Slevin, 1950), one of the most spectacular and seldom docu- mented aggregations of vertebrate life is that brought about by the odorous and noisy bird rookery. With the fragility of the nests of many wading birds and the frequency of dropped eggs, fish, and fallen nestlings, a variety of scavengers and predators find bird colonies a happy hunting ground. Host ( 1955) reports a gathering 6f alligators near a large rookery of herons and ibises in the Everglades. Nor is this food, source ignored by mammals, including man. I have a report that over 20 jaguars aggregated under a great jabiru stork rookery in central South America in 1955. Austin and Kuroda ( 1953) describe the remarkable instance of humans subsisting on salmon dropped by cormorants and herons nesting around the Suruga Shrine in Aomori Prefecture, Japan. The cottonmouth, by virtue of its adaptability to almoSt· any ayail- abie food source, dead or alive, becomes a resourceful opportiinist when bird rookeries are nearby. Here it scavenges beneath the nest trees, apparently attracted by the odor of the excreta and the fish re- gurgitated by annoyed parent birds or dropped by clumsy nestlings. Chapman ( 1908) was one of the first to note what may have been aggregating cottonmouths under a nesting colony of anhingas and herons near St. Lucia, Florida. In the early 1920's, Sass ( 1926), on the Isle of Palms near Charleston, South Carolina, saw scores of cot- tonmouths gathered under a great blue heron "village" apparently attracted by dropped fish. Host ( 1955) noted many cottonmouths beneath a large rookery on the edge of Lake Okeechobee, Florida, and assumed that they were after the eggs and young. Carr ( 1936) was the first to report aggregations of cottonmouths on the islands of the Cedar Keys group, where he found 11 of 13 captured snakes coiled beneath nest trees. In May 1942 Mr: and Mrs. Allen D. Cruickshank (pers. comm.), encountered a concentration of cotton- mouths and banded water snakes, Natrix fasciata, under a heron and glossy ibis rookery at King's Bar Reef, Lake Okeechobee, Florida, where 3 to 10 cottonmouths swam in front of the blind constantly during the week they were there. This couple also report seeing frOm 8 to 12 of these snakes on 20 April 1942 in the vicinity of anhinga and egret nests near the present Anhinga Trail in Everglades Na- tional Park, Dade County, Florida. W. T. Neill ( pers. comm.) found cottonmouths very common under nests of about 95 little blue herons at McKinney's Mill, a small bonnet-choked cypress pond in Emanuel County, Georgia. Neill had noted that cottonmouths were not nor- 234 BULLETIN FLORIDA STATE MUSEUM Vol. 14 mally abundant there, and believes they came·from other nearby ponds, drawn by the odor of the fish and excreta. John N. Hamlet ( pers. comm.) saw many large cottonmouths in an active anhinga and heron rookery near Pritchardville, South Carolina, from 1951 to 1953. On the Florida mainland eottonmouths occur most commonly in shallow ponds, marshes, and swamps. When water levels drop the life concentrated in them provides a considerable food supply for the cottonmouths. Allen and Swindell ( 1948) report cottonmouth ag- gregations at pools left in dryihg marshes. Henry P. Bennett ( pers. comm.) encountered cottonmouth concentrations in 1955 in drying sloughs and ponds in the Corkscrew Swamp sanctuary, Collier County, Florida. When such ponds dry completely the cottonmouths apparently seek out, deeper ponds or subsist on dry land until the water table rises again. Walter Auffenberg (pers. comm.) collected reptiles from 1949 to 1951 between DeLand and New Smyrna, Volusia County, Florida. He noted that when water levels fell, aggregations of cottonmouths fed ravenously on the abundant life concentrated in the remaining pools of roadside ditches, consuming Bsh, frogs, rodents, and reptiles. Sometimes they ingested inanimate objects as well, for the stomachs of two specimens dissected after death were packed with mud and some with sticks up to 12 mm in diameter and 75 mm long. Burkett ( 1966) remarks on the ingestion of plant material by midwestern cottonmouths. Sometimes cottonmouths move considerable distances. Hamilton and Pollack ( 1955) record a cottonmouth killed 1 mile from the nearest water at Fort Benning, Georgia. Probably such overland journeys are aided by a keen olfactory sense. Noble and Clausen ( 1936) have shown that the sense of smell is very important in some snakes. Fish odor apparently attracts cottonmouths, for Neill ( pers. comm.) noted cottonmouths drawn to a fish-cleaning platform 50 m from a lakeshore. I conducted several tests to verify the cottonmouths' scenting ability on Sea Horse Key. A crushed sardine and oil trail was ineffec- tive in February. In early March 1967 near Gardner's Point I placed a known number of pieces of cut mullet at points where no snakes were present within a radius of 30 m. The next morning three pieces were missing and I found two snakes in dense cover about 20 m away; one, a male, contained one piece of mullet; the other, a female, two. At another locality I caught two snakes, including a 356-mm juvenile, in a small funnel trap baited with mullet. 1969 WHARTON: COTTONMOUTH MOCCASIN 235 ---- 0 / i /0 aa O 0 ie O aaa IA 01955 01956 0 t •1957 01955- 56 I 1956 - 57 61955-56 -57 A 1955 AND 1957 FIGURE 4. Location of nest tree5 on Sea Horse Key used by birds ( principally cormorants). Symbols show years occupied. 236 BULLETIN FLORIDA STATE MUSEUM Vol. 14 In early April I placed five piles of 10 chunks of muliet on the ground at 2100 between Stations 95 and 115 and visited them re- peatedly until 2300. A light breeze blew from the east. Shortly after 2100 near Station 115, a snake was seen to eat one piece and crawl away; another snake was di5covered swallowing one piece 30 minutes later, and a third snake ate two pieces of fish 1 hour and 20 minutes later. As the last two individuals were spattered with avian excreta, they were presumably drawn a minimum distance of 35 m from the nearest nest tree. It is puzzling that the snakes ate only one or two pieces' of cut fish each instead of gorging themselves as they might have. EFFECT OF ROOKERIES ON MOVEMENT, ABUNDANCE, AND DISTRIBUTION Spring aggregations of cottonmouths beneath nest trees are com- prised of individuals arriving from varying distances ( Fig. 4,). Thirty snakes captured during the previous non-nesting period; 1 September- 28 February, and found beneath active nest trees had moved an aver- age of 62 m. Only two had presumably arrived from distances ex- ceeding 150 m., only five from over 100 m. During the rookery season 16 snakes did not move to a nest tree; the closest point of their ac- tivity ranges to an active nest tree averaged 139 m. These data sug- gest that Sea Horse Key cottonmouths are seldom lured by scent to a nest tree further than 150 m. Table 1 summarizes the distances between points of capture for 483 snakes and large annually active nest trees ( six or more nests per tree). Except for two osprey nests, all trees were cormorant rook- eries. Two-thirds of the cottonmouths taken in the nesting season were found within 32 m of an active nest tree, and evidently more than half of the snakes remain with in 32 m of these trees during fall and winter months. Table 1. NUMBERS OF COTTONMOUTHS RECAPTURED AT VARYING DISTANCES FROM ACTIVE NEST TREES DURING NESTING AND NON-NESTING.SEASONS Distance from nest tree Nesting Season Non-Nesting season (,in meters) April 1-August 31 Sept. 1-March 31 0- 16 42 56 17- 32 54 147 32-150 42 142 TOTAL 138 345 Of 402 original captures, 234 snakes were taken on the low penin- sula during the nesting season ( 1 March-31 August). Of these 121 1969 WHARTON: COTTONMOUTH MOCCASIN 237 ,---. \ 1 U 0 402 0----0 0 OA oo 0 0 0 i 00089 0 / 0 /1 1 0 O o ao V 00 0 0 000 0 0 0 0 O 00 0 0 0 0 - FIGURE 5. Midpoints of cottonmouth activity ranges on the low peninsula of Sea Horse Key. 238 BULLETIN FLORIDA STATE MUSEUM Vol. 14 - were taken under, or had been under ( bore avian excreta) active nests; 172 ( 73.5%) captures were within 32 m of the nest trees. The midpoints of snake activity ranges ( Fig. 5) plotted on a map of the low peninsula show that the bulk of the population is concen- trated between Stations 68 and 77 on the west side of the peninsula, and from Station 92 west around the entire Gardner's Point area. A 60-m circle ( mean distance moved) drawn around each active nest tree excludes only 15 per cent of range midpoints; 8 per cent of the range midpoints are more than 100 m from active nest trees. Evi- dently most individuals live within the mean distance at which the majority are attracted to nest trees. Thus the location of rookery trees appears to control the distribu- tion and food habits of almost three-fourths of the Sea H6rse Key cottonmouths. With the snakes so dependent on the birds, obviously their nesting habits must effect the welfare of the cottonmouth popu- lation. Well developed olfaction is one facet of a generalized behavior pattern enabling cottonmouths to adapt to the environment of Sea Horse Key. Were it not for their ability to locate the principle food sources of the island from a considerable distance, these snakes would doubtless be far less abundant, and the density ( 22.27 per acre) would more nearly approach that of the main ridge ( 1.85 per acre ). On the ridge the reptiles apparently feed chiefly on rodents. During hot summer days cottonmouths are seldom found in the open rookery areas. At these times they remain tightly coiled in the nearest heavy cover, crawling to the nest trees at dusk. Of 232 night captures, 141 were taken under active nest trees, 47 in 1955, 42 in 1956, 52 in 1957, distributed monthly as follows: March 2, April 26, May 43, June 55, July 2. This coincides roughly with the availability of dropped fish. I have watched 17 different snakes in the act of feeding on Bsh or fish remains. During April 1955 and May 1956 snakes aggregated under the nest trees about 15 minutes after dark, gliding about, searching and feeding. I have seen their reactions to dropped fish several times. A falling fish instantly alerts them. With the head raised about 125 mm they glide quickly near to where the fish landed, then lower the head and try to locate it, apparently by use of the tongue. The snakes generally find and swallow a fish within 5 to 10 minutes after it falls. While feeding beneath nest trees cottonmouths are quite active and easily disturbed by the flashlight beam, moving quickly for the nearest cover. By working carefully it is possible to move among a 1969 WHARTON: COTTONMOUTH MOCCASIN 239 feeding aggregation, and I have captured as many as 9 individuals beneath a single tree. When beneath active nest trees the snakes are quick to investigate any commotion in the leaves. Sometimes they may be lured by throwing sticks or other objects to simulate fall- ing Bsh. I once tapped my catch stick on the ground and a snake came forward with upraised head and struck at it. Young cormorants occasionally fall to the ground. Fallen fledg-_ lings I placed near searching cottonmouths were never eaten, though I once found a snake that had swallowed a cormorant leg and another that had ingested a wing, both probably scavenged from a decaying carcass. The remains of many well feathered nestlings are found on the ground where it is apparently difficult or impossible for parent cormorants to feed their young. Carr ( 1936) reports a portion of an eggshell in a cottonmouth on Snake Key. It is extremely rare to find an unbroken egg on the ground. Fallen nestlings and eggs are un- doubtedly insignificant food sources. The small size of some dropped fish does not deter feeding cotton- mouths. Carr (1936) found several tiny flsh in a sizable cottonmouth from Snake Key. I once saw a 1422-mm male swallow two fish 50 mm long. Small cottonmouths may attempt to swallow fish heavier than themselves. I found a 711-mm snake weighing 269 grams attempting to swallow a 305-mm, 354-gram serranid dropped by an osprey. Some- times parts of fish skeletons are eaten. One small cottonmouth had about 25 mm of the skull of a catfish, Bagre marinus, protruding from its body about 100 mm behind the head. A 559-mm male was seen eating the skull of another Esh. Large cottonmouths sometimes eat small and unusual prey. Carr ( 1936) reported that a 1473-mm specimen from Snake Key had eaten a skink. A 1016-mm specimen from Snake Key that I took had eaten a caterpillar and a 1295-mm snake from North Key had eaten a ter- restrial snail about 3 inches long. Burkett ( 1966) notes that cotton- mouths feed on both insects and gastropods. Cottonmouths are oc- casionally ophiophagous. In the fall of 1955 I saw two water snakes ( Natrix sipedon ssp.) at Station 160, at Gardner's Point and caught one; a few days later I took a cottonmouth containing a Natrix from the sanne area. Cottonmouths may compete for food under nest trees. Two snakes were once found engaged in a violent tug-of-war over a 125-mm dog- Rsh, ( Opsanus sp.). On another occasion I found two large male snakes, both over 1270 mm, struggling over a fledgling crow. Small cottonmouths are uncommon beneath nest trees and no 240 BULLETIN FLORIDA STATE MUSEUM Vol. 14 newborn snakes were found there. Of 101 original captures taken beneath these trees, only 4 snakes were under 610 mm, and all con- tained fish or were feeding when captured; 22 others ranged between 610 and 914 mm, 64 between 914 and 1219 mm, and 11 above 1219 mm. I think that undisturbed snakes must return to the same nest trees repeatedly, but my experience suggests that snakes may leave tempo- rarily the vicinity of capture, very likely because of the shock of the experience. Both males and females sometimes make a "shock" or escape run following release. In 9 instances this straight line escape behavior of large male snakes carried them out into salt water for about 30 m before each turned back to shore; I never saw a female enter the Sea. On 6 June I captured 4 snakes at Station 92; on the following night I found 5 different individuals there. I believe that my handling of the previous night caused the first 4 snakes to leave the immediate area of the tree Normally snakes wait for darkness to move into exposed places under isolated nest trees, although on hot July days they begin to gather beneath egret and night heron nests in dense canopy by 1700. Occasionally they feed under nests in the daytime. On 5 April I watched a 1041-mm female crawl several meters to a dropped fish at 1400. During daylight hours I once slid a white 6-foot carpenter's rule through the leaves to within 250 mm of a large male snake, which struck it viciously, perhaps a feeding reaction. After young cormorants leave the nest tree, they frequently return to it to be fed by the parents. In late July I took 8 snakes scavenging at night at Station 120 where young cormorants still perched in their nest trees. Later ( mid-July tb early September), these cormorants leave their nest tree and congregate in certain roost trees. The snakes seem to leave the nest tree as soon as the cormorants depart. On 27 June I noted snakes beneath the StatiOn 91 cormorant nest tree ( still occupied by young birds), whereas the Station 86 nest tree had neither birds or snakes. By 19 July only a few snakes remained in the area of the nest trees, but I caught two under new roosting trees. While ospreys are not plentiful enough on Sea Horse Key to be principal provisioners for snakes living there, they are important as they start carrying fish before other birds begin nesting. On Snake Key 24 February, I collected a 350-mm, 681-g mullet one of a pair of ospreys dropped. Courtship maneuvers of ospreys often involve carrying fsh, which are sometimes dropped from the feeding perch. One night a perched osprey dropped a, 681-g serranid at Station 110, indicating they may go to roost holding a fish. Most osprey- 1969 WHARTON: COTTONMOUTH MOCCASIN 241 caught Bsh were either trout, C~noscion sp; serranids, or mullet, Mugil sp. and large enough for a substantial meal. On Snake Key, as on Sea Horse Key, cormorants provide most of the food. Herons and egrets are minor contributors. Ibises, because they feed on small crabs and crayfish, probably contribute little. In 1966 and 1968 a large rookery of white ibis occupied most of the main ridge of Sea Horse, but no increase of cottonmouths was noted beneath the nests while at the same time snakes were markedly numerous beneath small rookeries of the brown pelican, Pelecanus occidentalis, a species that.had not nested on the island during the initial study. To ascertain how long individual snakes attended nest trees in consecutive years, 218 captures from Sea Horse Key were plotted by year, month, and capture point. Of 3 individuals captured during 3 consecutive feeding seasons, 1 was taken at the same tree each yeak and 2 had moved to other trees. Of 15 snakes captured in 2 successive feeding seasons, 7 returned to the same nest trees. Of the other 8,3 left probably because nesting stopped in the old tree, 4 changed to other nest trees ( about 50 m away) even though the original tree was still active, and 1 moved 190 m for no apparent reason. Station 76 functioned as an active nest tree for 3 consecutive years. Four snakes taken there in the winter of 1954-55 were found there again in the summer of 1955, 1 remained until the winter of 1955-56, and 2 until the summer of 1956. Of 3 more taken initially in the summer of 1955, 3 remained through the following winter, 2 persisted until the summer of 1956, and 1 stayed for the winter of 1956-57. During the same 3 years 16 snakes arrived from other nest trees. The small size of activity ranges seems to preclude a con- tinuous nomadic wandering from tree to tree, although the cotton- mouths usually move widely enough t0 include several active nest trees within their activity ranges. I conclude that while some snakes may remain at or near a nest tree for 3 consecutive years, the bulk of the population seems divided between those individuals that remain about 2 years at a nest tree and those that move to another site. If disturbance by crows and ospreys should make cormorants regurgitate heavily in a certain area, or if younger birds drop fish more often in certain trees, snakes might be drawn by scent from one nest tree to another. Also a snake on its escape run after the traumatic experience of being handled may approach within smelling range of a different rookery tree. Fish dropped by birds provide a continuous source of food through the warmer months when reptile metabolism is highest.. I believe 242 BULLETIN FLORIDA STATE MUSEUM Vol. 14 the very small number of cottonmouths I encountered on several trips to both Atsena Otie and North Keys can be explained by the absence of large rookeries. Rats and the few fish dropped by the 6 or 8 pairs of ospreys seem to constitute the only food for the cotton- mouths on Atsena Otie. As the welfare of most cottonmouths on Sea Horse and Snake Keys is dependent on the rookeries, the snake population has undoubtedly fluctuated over the years, along with the nesting bird populations. The interrelationship is not entirely one-sided. The legend that the islands swarm with venomous snakes normally protects the rookeries from invasion by man. The presence of snakes near nest trees may also discourage the roof rats that might otherwise prey upon the eggs and nestlings. In 1958, unfort-unately, an estimated 400 snakes were removed for venom research at the University of Florida Medical School. ACTIVITY RANGE AND MOVEMENTS As cottonmouths on the Florida mainland are chiefly found in aquatic habitats where their normal sphere of activity is influenced by a body of fresh water, it is interesting t6 examine the movements of cottonmouths in the terrestrial, insular surroundings of Sea Horse Key. Whether they wander at random or favor a circumscribed area has an important bearing on the ability of this limited environment to support a large number of reptiles. Activity range and movement was studied by plotting points where each snake was captured. Pro- vided they were captured more than twice, the area enclosed by lines connecting these points is regarded as an activity range as de- fined by Carpenter ( 1952). The activity ranges established in this study are based on marked and recaptured snakes. Individuals were marked and released at the point of capture. Of these 107 were captured 3 or more times, pro- viding usable ( three-dimensional) data. Six snakes were taken 5 to 6 times each, six 7 times, and two 8 times. The mean number of cap- tures, including the initial Capture, was 3.8 per snake, and the mean elapsed time between initial capture and the last recapture for each of 107 snakes was 18.6 months. Thirty-five snakes were taken 2 years or more after the initial capture, and one was retaken after 39 months. Capture-points were plotted on scale maps of the island using compass bearings from the nearest station and station to snake dis- tances. Straight lines connecting the outermost points of capture 1969 WHARTON: COTTONMOUTH MOCCASIN 243 were then drawn on the map. The areas of these triangles and poly- gons were measured in acres with a planimeter. Activity fanges for Sea Horse Key cettonmouths vary from less than 0.11 acre to more than 3 acres. Males seem to have slightly larger ranges than females (M 31 3 S -0.43 acre, M 589 9 -0.35 acre). Over 18 per cent (1599,533) have activity range of less than 0.11 acre. With two exceptions these captures spanned a 7 to 33- month period. Four snakes remained in almost the identical spot for well over 2 years. Small- activity ranges were often associated with the presence of nearby winter dens and food sources; half of 20 snakes with limited range lived very near both a den and a food source ( usually a rookery tree). Others were reasonably near a den and about 50 meters from a possible food source. Only one snake was not near a den or a food source. In 16.8% (n=18) the activity range is greater than one acre; 16 of these were excluded from the calculation of mean range, for most of these longer movements could be accounted for by travels, pre- sumably to wintering areas. Six of the ranges were greatly extended by travel to large palmetto patches, while five of these movements were by snakes that wintered on the wind-Shielded east side of the low peninsula with its thicker vegetation. One large male was cap- tured 2 successive summers at the nest tree of Station 76; each year it traveled 295 meters to a wintering area near Station 34. Three other large activity ranges can be explained by the abandonment of Station 92 as a nesting site. Any cessation of nesting at a given site may also cause permanent changes of range. An example of the complete desertion of a rookery and its effect on the cottonmouths is afforded by the area near Station 36. John D. Kilby (pers. comm. ) recalls that around 1934 a large rookery of white wading birds occupied this vicinity. In 1954 the cherry-laurel and bay trees contained old nests probably used for the last time in 1953. During the 3 years following 1954 snakes became progressively less numerous in the vicinity; seven emigrations were noted. It seems reasonable to suppose that the abandonment of the rookery caused these snakes to leave. Most evidently migrated to active nest trees elsewhere; several skeletons attested the fact that some died. The number of individuals remaining approximated that of the normal ridge population where no birds nest. As the main ridge supports far fewer snakes than the rookery- bearing peninsula, the size of the activity ranges in both habitats might be compared. Only two snakes from the main ridge were 244 BULLETIN FLORIDA STATE MUSEUM Vol. 14 captured three times; one had a range of 0.40 acre, the other 0.11 acre. Ten additional snakes from the main ridge were captured only twice. By measuring the distance between the capture points of , METERS1\ 0.50 E VoII . 0.-4--/ lilllil 11-3 -I D-4-02=-r /P I, L#/1 f4~~-*..IC0 Go .' O 0 00 ®Cb . O. 'GO 00 FIGURE 8. Location of 163 cottonmouths captured in fall and winter. Solid symbols, captures 1 Sept.-30 Nov.; open symbols, captures 1 Dec.- 28 Feb. Compare with Fig. 7. 1969 WHARTON: COTTONMOUTH MOCCASIN 257 was 7 April, with a cloacal temperature of 15.8°C. The highest air temperature recorded during denning was 18.3°C. On 18 and 19 December 1955 (air temp. 15.6°C. ) I took nine snakes from seven sub-surface dens and one from a palmetto patch. Their cloacal temperatures varied from 13.5 to 16.5°C (M- 14.9°C). Three snakes taken near Station 92 on 12 February 1955, had a mean cloacal temperature of 4.2°C (air, 5.6°C), the lowest cloacal temper- atures I recorded. These snakes were in a shallow den, and their temperature was lagging well behind that of the air. The cloacal temperatures of nine snakes taken in a den near Station 153 on 10 January 1956 were within 1.2°C (M = 10.7°C) of each other at a soil temperature of 12.8°C and an outside air temperature of 7.8°C., The low of the previous night was 5.0°C, and the highest daytime temper- ature was 8.9°C. Air temperature within the den was 9.1°C at the time of capture. This suggests that, in this instance, soil temperature kept the snakes at a point about intermediate between soil and air temperatures. Table 5. DIURNAL TEMPERATURE ('C) AND PER CENT RELATIVE .HUMIDITY oF DEN SITES COMPARED WITH GROUND SURFACES Temperature Humidity Air Soil Date and den number Den Surface Surface Den Surface ( 4" depth) 18 Dec. 1955, Den 3 13.3 16.7 - 85 .70 19 Dec. 1955, Den 3 15.6 17.8 15.0 87 75 10 Jan. 1956, Den 13 9.1 7.8 12.8 - - 11 Jan. 1956, Den 13 8.9 6.9 12.8 73 68# 18 Feb. 1956, Den 3 18.9 21.1 19.5 96 92 4 Mar., 1956, Den 3 17.8 20.0 18.3 90 84 19 Jan. 1957, Den 7 11.3 13.3 12.8 Table 5 comparing temperature and hQmidity at ground lavel and within different dens shows in five of the seven recordings the den was from 1.3° to 3.4°C colder than the outside air. The den air temperature is usually closer to that of the soil. When air tempera- tures drop the den remains warmer because of the lag of the soil temperatures ( 10, 11 January 1956) . As table 5 shows, den tempera- 258 BULLETIN FLORIDA STATE MUSEUM Vol, 14 tures usually remain colder when air temperatures rise. Dens also have a consistently higher humidity than the surface air. Thus while dens are only shallow, open stump holes, they are deep enough to ameliorate temperatures and maintain a higher humidity. As B6ne- diet ( 1932) indicates, reptiles may lose heat readily by vaporization of water; higher humidities in dens would tend to favor the conserva- tion of heat. On 11 December 1955 I found five snakes in a cedar stump-hole near Station 128, which was open and scarcely below ground level. Three snakes were coiled on one another deep in the den, and two towards the entrance. The two outer snakes had cloacal tempera- tures of 10.0°C, the same as air temperature. The inner three were distinctly warmer, with cloacal temperatures from top to bottom of 11.0°C, 11.3°C, and with the snake next to the soil 12.2°C. One of the most perplexing wintering problems is den desertion. Of 50 snakes taken in dens, only three were ever retaken again in a den and none of these in the den in which it was originally found, although each snake was released quickly into the opening of its den after weighing. Near Station 92, an area that always yielded a large number of snakes, there were several active nest trees in 1955 and at least one functional nest tree in 1956. In the winter of 1955 I took 10 snakes from three dens near Station 92 and released them near the same dens. They did not return to these dens that winter, and the dens remained deserted the next 2 years. I found two of the snakes in 1956 in a nearby den that had escaped my notice in 1955. This den contained nine snakes in all, which I released at the den mouth; none returned either that winter or the next, although two new snakes appeared there in 1957. Not only were the Station 92 dens deserted, but in 1955 I took snakes from 10 other dens, 8 of which remained deserted by all snakes for the second ( 1956), and third ( 1957) winters. On 10 January 1956 I marked nine snakes from a den near Station 153, and released each at the entrance hole. Though the next day was several degrees colder, not a single snake was present in this den. Even though they make an escape run on release, I find it diffcult to believe that these serpents could not find their way back to their den. I conclude that snakes move away from the site of a den disturb- ance, and do not re-enter the den for at least 2 years. Perhaps they move off in response to the normal shock pattern of behavior and are too cold to return, but this does not explain their absence in subse- quent years. It has been suggested ( I.L. Brisbin, Jr.-pers. comm.) 1969 WHARTON: COTTONMOUTH MOCCASIN 259 that musk sprayed during capture is a warning device. Musk would have to persist for several years to explain den desertion, and this theory does not explain why some dens are re-used by other snakes. Another theory supposes mass movements to winter in other parts of the peninsula, which the general location of winter captures ( Fig. 13) does not support. As I found the denning sites of only a small percentage of the total population, I could possibly have overlooked many of them but I felt that I had discovered most of the den sites in the Gardner's Poiht afea where I made these observations. Published accounts of den re-use are limited and conflicting. Hirth ( 1966) found that 9 out of 10 Crotalus uiridis returned to dens after displacement of 50 to 774 m. Den abandonment, or specific crevice abandonment, might be inferred from Fitch's ( 1960) data an Agkistrodon mokeson. Of 492 copperheads Fitch captured during a 10-year period, apparently only 11 were recaptured on the ledge where the original captures were made, and only 3 of these at the exact crevice. It is possible that all the snakes might have returned to the identical crevice had they not been disturbed. Fitch ( 1949) and Woodbury ( 1951) note that there may be a difference in whether a snake is allowed to emerge normally to be trapped or caught or whether it is disturbed in situ and physically removed when its instincts direct it to remain underground. I suggest that some crotalid snakes may have a persistent site memory. Rattlesnakes, successfully adapted to cold climates and . large homiotherm prey, have developed the ability to migrate long distances to and from specific dens. Cottonmouths have been known to migrate up to 1/2 mile ( Dundee and Burger, 1948) and presumably have the same instincts. If these snakes have a highly developed directional sense, perhaps the site memory is also strong and enables them to avoid disturbed dens. Evidence to support such a theory of den avoidance is weak, but if cottonmouths do have such associative powers, my data suggest that this ability may persist for at least 2 years. TEMPERATURE AND BEHAVIOR Cottonmouths appeared disindined to move at 8.0°C and below, but from 10.0 to 12.0°C followed the usual escape pattern of remain- ing until One leaves, then moving away. Generally snakes handled with a CT ( cloacal temperature) of 8.0-10.0°C were passive and only gaped, though one 1524-mm male with a CT of 9.0°C squirmed continuously on being handled and finally bit itself twice. At the 260 BULLETIN FLORIDA STATE MUSEUM Vol. 14 lowest CTs recorded ( 4.0, 4.1, and 4.5°C from three snakes 12 February 1955) all were passive and only two sprayed musk-the one with a CT of 4.0°C sprayed musk cepiously. A 1118-mm male at CT 4.5°C swelled up and struck several times after being annoyed. Warning by opening the mouth or gaping first occurred at a CT of 6.5°C, with musking and inflation. In one den three individuals with CTs of 10.0, 11.1 and 11.3°C seemed aggressive towards the snake stick. All three struck it, and two bit the leather noose. Two inflated and one attempted to vi- brate the tail, which at this temperature is a feeble and slow move- ment. These snakes were passive until touched. Two bit the stick while being withdrawn from the den. At 9.0-10°C some snakes partially coiled When touched. One snake with a CT of 11.0°C crawled out of a hole when molested. A 406-mm young, disturbed under a palmetto frond, was violently aggressive at CT 11.4°C. About half the cottonmouths threatened at this, temperature. Of Il snakes whose demeanor was recorded between 4.5 and 10.0°C, 4 were aggressive and 7 were passive. This high incidence of aggression may be due to rough handling, although aggression would have survival value when snakes are immobilized by cold. I noted a cottonmouth vibrating its tail at a CT of 12.2°C. At 15.0- 16.0°C cottonmouths were able to snap into a· coil fairly rapidly, and I noted unprovoked crawling at this temperature. Figure 9, showing numbers of snakes captured at each degree of air temperature, shows four snakes captured under nest trees at an air temperature of 14.4°C. These snakes, two of which were actually swallowing fish at the moment of capture, had CT's of 14.5, 15.8, 15.2, and 15.4°C. The leaf mold temperature was 16.8°C and the soil 19.8°C, so the snakes were feeding on a warm substrate. Island cottonmouths may thus feed at 3.8°C below the 18.3°C recorded by Allen and Swindell (1948). From these observations I conclude that cottonmouths in this area become attracted to food and are able to scavenge at a CT of 14.0°C and are quite active at a CT of 16.0°C. Figure 10 suggests that the optimum temperature for these snakes lies between about 18°C and 29. 5° C, somewhat lower than the preferred substrate temperature of 35.34°C mentioned by Bogart ( 1949). I have noted sunning activity in November between 10.0 and 20.0°C ( Fig. 10). The coldest day on which I captured a cotton- mouth sunning was 6.1°C; this snake ( CT 10°C) had crawled out of a gopher hole den at Station 40. 1969 WHARTON: COTTONMOUTH MOCCASIN 261 60 - 55 - I IN STUMP HOLES M UNDER VINES, PALM LEAVES, ETC. 50 - tl IN PALMETTO (OR OTHER COVER) 7 SUNNING NEAR DEN 45 - O LYING OUT ON FOREST FLOOR Q LYING OUT, SUNNING 0 IN ROOKERY AREA N U M BE R O F C A P TU R E S 40- 1 UNDER NEST TREE (DAY) 1 UNDER NEST TREE (NIGHT) 35- 91 AESTIVATION . . . . . . - iii f zI :::: - - 25- - - ~~~~ --- Z =- lili 2 - 20 - 2 - ..... 0 Z= I - -- - 0.... 0 - •000.0 0 15 - I Z -- •0•000 0 E - -:g:„g: FEE0 . .00..0 . 2 -: :ggs:o.: 0000•000 ih : : :tgoo:8::5 0 \ •• \\00•000 0 \\ 0 0 \\00.000 X \\000OIDO 00 0 A \\\\ \\\00\000 0000 -0 1 & '2„\,\\Ob'\09000'0000 1 ®F - 36 40 44 48 52 56 60 64 68 72 76 80 84 88 2- 22 44 6.7 8.9 11.1 13.3 15.6 17.8 20.0 22.2 244 26.7 28.9 31.1 AIR TEMPERATURE AT GROUND LEVEL FIGURE 9. Cottonmouth captures at specific air temperatures, indicating behavior or body position. The warm-up of a dark, basking cottonmouth is relatively rapid. On 10 March 1957 I placed a large, completely docile female on dried grass background of a neutral color on the open beach. The sun was at an approximate angle of.40°. The sand under the grass at a depth of 25 mm had a temperature of 15.4°C, the humidity was 78 per cent, and the snake was uniformly dark. The thermometer was gently in- serted about 36 mm into the cloaca and held in a central position. Seven readings taken from 09: 13 to 09:40 showed the snake's tempera- 262 BULLETIN FLORIDA STATE MUSEUM Vol. 14 ture rose from an initial 8.8°C (air 11.3°C) to 21.5°C, a rise of 12.7 degrees. During the first 6 minutes the temperature rose 0.90C per minute, and thence roughly 0.4°C per minute for the remaining 21 minutes. This snake was so docile, neither swelling nor moying during the entire test, that I gently moved it to a shady area beneath the man. groves and recorded the rapidity of heat loss. The air temperature remained at 12.0°C and the shaded sand covered by dead leaves also registered 12.0°C. I took 11 cloacal readings at 5-minute intervals while the snake's temperature dropped to 14.8°C, a loss of 6.7°C in 60 minutes. For the first 15 minutes the heat loss was approximately 0.2°C per minute, thereafter for 7 readings, 0.1°C per minute. Some physiological control over rate of change, as Bartholomew and Tucker ( 1963) note for a lizard, is possible. During this test the importance of central cloacal temperature was demonstrated. Turned so that it rested against the body wall on the shaded side of the snake the thermometer showed a drop of 0.3°C; but held against the side of the snake in the sun it registered 4.6°C higher than central cloacal temperature. During the winter Sea Horse Key snakes often remain in one posi- tion for considerable periods of time. I saw five individuals that had remained coiled in One spot on the ground for at least 24 hours. Two 40- KEY: 36- 0 LYING OUT ON FOREST FLOOR1 BASKING NEAR DEN 21 UNDER HIBERNATION COVER Bl UNDER NEST TREE (DAY) 9 LYING IN SUN Ill UNDER NEST TREE (NIGHT) 30- 0 IN ROOKERY AREA 0 UNDER VINES, PALM LEAVES, ETC. N U M B E R O F C A P T U R E S m IN STUMP (DEN} HOLES 9 ELEVATED OFF GROUND 26- 20- I 0 6 0 i . %.1.. .11 5, 8 810 ..5 - 0 .1 •A ./ f A I IL M Al MITFl 2 4 6 8 10 12 14 16 18 20 22 24 26 '28 30 32 TEMPERATURE (~ C 1 FIGURE 10. Observations on 387 cottonmouths captured at specific cloacal tem- peratures, indicating behavior or body position. 1969 WHARTON: COTTONMOUTH MOCCASIN 263 remained in one spot 72 hours without moving. Others seemingly moved only 5 m or less in 72 hours. Demeanor on approach was noted on 50 per cent of the cotton- mouths. Such information renders at best only a crude picture of the behavior of the snakes, as one does not always approach at the same speed. Some snakes are approached more cautiously because of ready refuge near. Others appear to be asleep. Reactions were classed in five categories: violently aggressive, aggressive, warning, passive; and escape. The first two involve striking; warning signifies opening the mouth ( gaping); escape means any attempt to crawl away. Between 4.4°C and 15.6°C 36 per cent of the snakes captured showed aggressive behavior. This contrasts with approximately 4.1 per cent between 15.6 and 32.2°C. No escape reactions were noted between 4.4 and 10.0°C. Between 10.0 and 5.6°C 6.8 per cent tried to escape, and 4.6 per cent between 15.6 and 21.1°C. However, at higher temperatures (to 32.2°C) 17.5 per cant of snakes tried to escape. Lueth (1941.), working with young water snakes of the genus Natrix, noted a tendency to escape at temperatures above 27°C. BODY FAT AND SURIVAL Volsoe ( 1944), St. Girons ( 1957), Tinkle ( 1957, 1962) and Wharton ( 1966) have discussed the relation of body fat to reproduc- tion in snakes. Body fat also forms an important energy resenve for snakes. Benedict ( 1932) shows that during - fasting, the respiratory quotient quickly becomes that of pure fat ( 0.72). Cottonmouths carry body fat in two long bundles ventral to the re- productive and excretory organs. Though it clings tightly in the vi- cinity of the spleen and gail bladder, it may be stripped out almost as a discrete unit. Figures 11 and 12 show weights of fat bodies dis- sected from snakes collected on Snake and Sea Horse Keys. Large fat bodies are white or yellowish. When depleted they appear as two coiled orange bands; these were classed as negative fat, and symbols representing them appear in the graphs along the abscissa. Far more Sea Horse Key snakes had negative fat bodies than did those from Snake Key. This supports my earlier statement ( 1966) that 50 per cent of the Sea Horse Key cottonmouths gained no weight during the study. Benedict ( 1932) dem6nstrates a direct relationship between meta- bolic rate and higher temperature in reptiles, and notes that the heat production of all cold-blooded animals is extraordinarily unif6rm. A snake's surface area (S) may be computed by the formula S= 264 BULLETIN FLORIDA STATE MUSEUM Vol. 14 Table 6, FAT WEIGHT ISLAND COTTONMOUTHS NEED FROM MARCH THROUGH SEPTnIBER AND ITS CALORIC EQUIVALENT IN AVAILABLE FISH Snake length Snake weight Calculated No. of 46-g fish yielding (mm) (g) fat weight (g) caloric equivalent 457 53 13.6 2.1 800 475 59.5 9.1 940 907 91.4 13.9 1168 1404 122.2 18.6 1422 2608 185.8 28.3 K X %W, where K is a constant ( 12.5 ) times % weight in grams. Using values from Benedict's rattlesnake curve, I computed the grams of fat needed by five cottonmouths of varying sizes ( Table 6). As the weight-length data show that Benedict's rattlers were well-fed, heavy specimens, I selected five fat cottonmouths. While the body surface 108-5 0 e96,3 769 0 64- W E G HT (G R A M S ) 0 .. 56- - 0 48- 0 - 040 - .. 32- 00: O%. 24 - B 16 - 8 -0 8- 8 --0398-t' 0 0- • A 1,0,1.111,1 101 P.QIII,1,111.1 3 4 5 6 7 8 9 10 1112 13 14 15 16 LENGTH (MM. x 100) FIGURE 11. Fat-body weight in grams versus length of Snake Key cottonmouths. Curves indicate weights of fat bodies f6r winter survival. Captures below zero grams indicate negligible fat bodies. A, cold winter ( 1954-55); B, warm winter ( 1956-57); solid symbols, nonfeeding period ( 1 Oct.-28 Feb.); open symbols, feeding period ( 1 Mar.-30 Sept.) 1969 WHARTON: COTTONMOUTH MOCCASIN 265 80- 72- . 64- W E IG H T (G R A M S ) 56- 0 0 48-- 0.0 0 40- • _ 032- .0 24- 0 0. - 90 0 8 16- 8 0002 0 /'/r/- - O¥--5az239~~----* - 0- 00 sri-- 1 40 'A' 99• 1 9 1 0, p f'*~ pOP I I p , , 1 61 1,1 345678 9 10 11 12 13 14 15 16 LENGTH (MM x 100 ) FIGURE 12. Fat-body weights in grams versus lehgth of Sea Horse Key cotton- mouths. Curves indicate weights of fat bodies for winter survival. Captures below zero grams indicate negligible fat bodies. A, cold winter ( 1954-55); B, warm winter ( 1956-57); solid symbols, non- feeding period ( 1 Oct.-28 Feb.); open symbols, feeding period ( 1 Mar.-30 Sept. ). of an emaciated snake almost equals that of a fat one, Benedict's con- stant requires snakes of reasonable similar condition. The f6rmula gives a figure which, divided by 10,000, yields the approximate calories needed by a cottonmouth of these surface areas per 24 hours. This figure was multiplied by the number of days in the seasons considered. Curves were drawn to connect points plotted for each of the five snakes mentioned above for both warm and cold winters at Sea Horse and Snake Keys ( Fig. 11, 12). Curve A, Bgure 11, shows that 6 of 30 winter captures ( 20%) on Snake Key had less fat than needed to survive a winter as cold as the 1954-55 season (monthly mean, 16.3°C), curve B that 11 of 30 winter captures ( 36%) had inadequate fat reserves for a winter as warm as 1956-57, ( monthly - mean 18.7°C). In addition, a winter as warm as the latter would en- danger 13 of 38 summer captures ( 34%) unless their fat reserves were increased. These figures may be contrasted with those of the Sea Horse Key population ( Fig. 12) where, in the colder of the two Winters 17 of 31 266 BULLETIN FLORIDA STATE MUSEUM Vol. 14 ( 55%) had inadequate fat reserve, while 24 of 31 ( 77%) had insum- cient fat reserves to survive a 5-month winter period similar to that of the 1956-57 season. On Sea Horse Key 52 of 79 summer captures ( 66%) had inadequate reserves to meet a warm winter. Lueth ( 1941) found that the rate at which snakes utilize their energy reserves during starvation varies directly with air temperature. Volsoe ( 1944) concluded that the European viper evinces no observ- able reduction of fat bodies during hibernation, but that these bodies become reduced after long starvation or during pregnancy. The cottonmouths showed no marked difference in their fat-body weights in summer and winter (Fig. 11, 12). While undoubtedly useful for both males and females in reproduc- tion, evidently the major function of fat bodies is to provide energy during warm weather when food is scarce and, although they are not markedly reduced in normal winters, they serve to offset the effects of unseasonal warmth during the cooler parts of the year when little food is available or digestion impossible. Figures 11 and 12 show that the fat body weights needed for average winter survival are not high, but a prolonged winter warm spell subjects cottonmouths to em-ergency demands on stored fat, particularly as they do not go into deep dens where the temperature remains uniform throughout the winter. Winter warm spells occur frequently in Florida-the extreme warm period in December of 1956 and the mild period that persisted into February over much of northern Florida in 1957 are examples. Volsoe ( 1944) suggests thut cold winters are more favorable to hibernating vipers than mild ones. To the island cottonmouths the winter warm spells may be critical; following the warm spell in January and February of 1957, I picked up eight dead snakes. Evidently fat reserves are more important for snakes in the variable climate of the southeastern United States than for either northern or tropical forms that live in uniformly cold or warm temperatures. The statistical relationship of fat weight to snake length was in- vestigated in snakes from both Sea Horse and Snake Keys. The sexes were treated separately and subjected to four F-tests. Only with the males of Sea Horse Key did the fat weight increase exponentially with increasing length. For females from Sea Horse Key and males from Snake Key, a linear equation best expresses the relationship between fat and total length. Neither curve appeared to Bt the data for females from Snake Key. No relationship between fat weight and total length was apparent. 1969 WHARTON: COTTONMOUTH MOCCASIN 267 Energy demands are larger in small snakes because of their pro- portionately larger surface area. During warm seasons food needs rise to high levels, and unless it eats promptly, a snake with negligible body fat must begin converting its own protein into energy. Negative fat bodies in Sea Horse Key snakes occur most consistently in spring and summer. Figures 11 and 12 have no curves for the fat require- ments of the five snakes during the warmer months. Their energy requirements ( in fat) from March through September, were com- puted by using the mean ( 25.3°C) of the three 7-month feeding seasons (Table 6). The number of fish needed to supply these amounts of energy for the warm months was calculated. Mattice ( 1950) gives the average cdorie yield of protein and fat from 6 com- mon marine fish as about 300 calories per 277 g of fish. The average length of pinfish eaten by cormorants is about 115 mm; toadfish, about 127 mm. Fish of these measurements weigh about 46 g. Table 6 lists the approximate number of fish of this size necessary 'to maintain cottonmouths at 25.3°C for the warm period from March through September, providing growth demands are ignored. Sea Horse Key snakes have various ways of offseting rapid heat loss. Undoubtedly tight coiling reduces the exposed surface area. The snakes also take refuge in holes or under thick cover at high environmental temperatures. On very hot days it is often difficult to find snakes at all. On one occasion, when snakes were avoiding the open rookery area (33.3°C, rel. hum. 54% ), I found them 50 m away beneath dense canopy ( 31.7°C, rel. hum. 62% ). I noted 3 snakes seeking den holes in excessively hot weather. The frequency of negative fat bodies and the randomness of the fat body weights in reference to the length of the snakes ( Figure 12) suggest that many Sea Horse Key snakes lead a precarious borderline existence, with competition and chance operating to pro- duce a haphazard food intake. The Snake Key population is better fed. A factor difrcult to assess is water supply. Klauber ( 1956) records rattlesnakes drinking water from their own skin and from the surface of rocks and outlines the reasons why snakes require little water. While I have seen young cottonmouths suck up water from leaves or the backs of other snakes after being sprinkled, I have not observed this behavior in adults. Adult cottonmouths should be able to drink limited amounts of rain water caught in cuYled leaves. Possibly dehydration, as well as food shortage, is an important factor in snake survival on these islands. 268 BULLETIN FLORIDA STATE MUSEUM Vol. 14 MORTALITY AND PREDATION The remains of 20 snakes found dead were measured in the field and compared with the closest matching sizes of prepared skeletal material; only one of the dead snakes would, have had a total length less than 1000 mm. The mean length of the 20 snakes was calculated to be 1291 mm, their mean weight 1634 grams. The chief cause of death is probably starvation, though dehydra- tion may also be important. Most dead snakes appeared thin and emaciated. I have found a number of individuals in a rnoribund condition with scarcely enough vitality to crawl. Death may sometimes be due to other causes. I once found a large male that appeared to have been in good condition before death. The posterior end of a linguatulid worm completely blocked the glottis. Adult cottonmouths apparently have few enemies on Sea Horse Key. Rats do not seem to disturb live cottonmouths, nor do they molest dead snakes. Possible predators of young cottonmouths are the red-shouldered hawk and various species of herons. Walter Auffenberg ( pers. comm.) once saw American egrets feeding on young cottonmouths in roadside ditches between Daytona Beach and Deland, Volusia County, Florida. Allen and Swindell ( 1948) state that herons and cranes eat young moccasins, and mention a Great Blue Heron choked to death on a cottonmouth it had stabbed through the head. Barbour ( 1956) found a cottonmouth half devoured by a raccoon. A few snakes are possibly eaten when raccoons are present on Sea Horse Key. Eyes are apparently not important to the survival of island cotton- mouths. Two healthy male snakes were captured that were uniformly and bilaterally blind, suggesting a genetic cause. There was no macroscopic vestige of pupil-the supraocular scales nearly touched the upper labials. Number 131 ( total length 1200 mm), in good con- dition, was captured 4 times between January 1955 and April 1956. Once, it was struggling for possession of a fish with another snake. Number 153 was actually fat ( total length 1245 mm, weight 1517 g) and was captured 3 times between June 1955 and November 1956. Both snakes were in as good or better condition than the aver- age island snake. If this blindness is genetic it suggests the intriguing possibility that eyelessness might compete successfully with the eyed condition on this island, as food is apparently detected best by olfac- tion. 1969 WHARTON: COTTONMOUTH MOCCASIN 269 DISCUSSION The cottonmouth is believed to have descended from an Asiatic viperid that crossed the Bering Iand bridge in late Miocene ( Neill, 1964) to become an element of the Arcto-Tertiary forest fauna. It has been present in Florida since the Pleistocene ( Brattstrom, 1953; Auffenberg, 1963) where it has evidently preserved an endogenous biennial sexual cycle inherited from Viviparous ancestral forms ( Whar- ton, 1966). This adaptation to cold climates serves the snake well in the food-scarce environment of the Cedar Keys. Various authors ( Hamilton and Pollack, 1955; Dundee and Bur- ger, 1948; Allen and Swindell, 1948) have found mainland cotton- mouths in dry habitats, often far from water. The migration of main- land progenitors from one drying pond to another has probably equipped this snake to cope with different habitats and to seek food actively on dry uplands. Such vagility, coupled with the develop- ment of excellent olfaction, enables the island snakes td find and ex- ploit both the rookeries and the indigenous vertebrates. The swamp- dwelling specialization has resulted in an ability to feed on both warm and cold-blooded prey, including dead fish, which fits it per- fectly as a rookery scavenger. Its wide temperature tolerance and the absence of notable behavioral thermoregulation is probably an adaption to a cool, amphibious swamp life, allowing it to survive in a canopied forest where feeding at low temperatures is often neces- sary. Cottonmouths have been reported not to feed in zoos in winter ( Conant, 1929). This trait may vary between populations. The Cedar Keys cottonmouths feed in winter and at that season augment their warm-season scavenging diet with litter-feeding mammals and birds. The ability to forage both by active pursuit and by ambush serves them well. on a year-round basis on the island. At the same time its odor and venom exempt it, from predation by most mammals. The young retain a primitive color pattern, shared with the cop- perhead, that affords protective coloration in upland deciduous forest leaf litter. Young cottonmouths are cryptozoic in habits, feeding largely on lizards and frogs lured by worm-imitating tail tips ( Whar- ton, 1960 ), a characteristic shared with other members of the genus. The adult cottonmouth usually loses its primitive, juvenile pattern by darkening and thus blends well with the normal swamp substrate. Increased size, more venom, secretiveness, and nocturnality enable the adult to offset its lack of protective coloration in upland habitats. The cottonmouth is able to move into deciduous forests in the 270 BULLETIN FLORIDA STATE MUSEUM Vol. 14 winter and seek underground safety in a wide variety of refuges, in habitats entirely different from its summer haunts. This ability has not only led to ease of ecesis on the Florida coastal islands, but has encouraged the development of a homing ability and, perhaps, a persistant site-memory. The Snake Key population survives well entirely by scavenging, as competing and potentially-edible rodents are absent. Both Snake and Sea Horse populations exist in high numbers partly because of the legal protection afforded the rookeries; reciprocally the presence of so many venomous snakes must deter human molestation at crucial periods in the birds' life cycle. The cottonmouth thus appears almost ideally preadapted to ex- ploit the unique habitat offered by the outer islands of the Cedar Keys. The niche might be termed that of a terrestrial carnivore- scavenger. LITERATURE CITED Adams, W. H., Jr. 1955. Water moccasin as a predator on birds. Wilson Bull. 68(20: 158. Allen, E. R. and D. Swindell. 1948. Cottonmouth moccasin of Florida. Herpetologica. 4 C first supp. ) : 1-15, Allen, E. R. and W. T. Neill, 1950. The cottonmouth moccasin. Florida Wildlife. 4: 8-9, 16. Arny, S. A. 1948. A survey of the reptiles and amphibians of the Delta Na- tional Wildlife Refuge. Unpublished M.S. thesis, NeW Orleans, Tulane Univ. Auffenberg, W. 1963. The fossil snakes of Florida. Tulane Studies Zool. 10: 131-216. Austin, 0. L., Jr. and N. Kuroda. 1953. The birds of Japan, their status and distribution. Bull. Mus. Comp. Zool. 109: 280-637. Barbour, R. W. 1956. A study of the cottonmouth Ancistrodon visciuorus leu- costoma Troost, in Kentucky. Trans. Kentucky Acad. Sci. 17: 33-41. Bartholomew, G. A. and V. A. Tucker. 1963. Control of changes in body tem- perat[tre, metabolism and circulation by the agamid lizard, Amphibolurus harbatus. Physiol. Zool. 36: 199-218. Benedict, F. G. 1932. The physiology of large reptiles with special reference to the heat production of snakes, tortoises, lizards, and alligators. Carnegie Inst. of Washington. Publ. no. 425: 1-539. Blanchard, F. N. and. E. B. Finster. 1933. A method of marking living snakes for future recognition, with a discussion of some problems and results. Ecology. 14: 334-347. Bogert, C. M. 1949. Thermoregulation in reptiles, a factor in evolution. Evo- lution. 3: 195-211. Brattstrom, B. H. 1953. Records of Pleistocene reptiles and amphibians from Florida. Ouart. J. Florida Aead. Sci. 16: 243-248. 1969 WHARTON: COTTONMOUTH MOCCASIN 271 Burkett, R. D. 1966. Natural history of cottonmouth moccasin, Agkistrodon pisciuorus ( Reptilia ). Univ. Kansas Publ. Mus. Nat. Hist. 17: 435-491. Cagle, F. R. 1942. Herpetological fauna of Jackson and Union counties, Illinois. Amer. Midl. Naturalist. 28: 164-200. Carpenter, C. C. 1952. Cornparative ecology of the common garter snake C Thamnophis s. sirtalis ), the ribbon snake ( Thamnophis s. sauritis), and Butler's garter snake ( Thamnophis butleri) in mixed populations. Ecol. Monogr. 22: 235-258. Carr, A. E., Jr. 1936. The Gulf Island cottonmouths. Proc. Florida Acad, Sci. 1: 86-90. Chapman, F. M. 1908. Camps and cruises of an ornithologist. New York: D. Appleton & Co. Pp. 1-432. Conant, R. 1929. Notes on a water moccasin in captivity C Agkistrodon pisci- vorus)( Female ). Bull. Ahtivenom Inst. Amer. III( 3): 61-64. - 1934. Two rattlesnakes killed by a cottonmouth. Science. 80 (2078): 382. Dundee, H. A. and W. L. Burger. 1948. A denning aggregation of the westefn cklttonmouth, Chicago Mus. Nat. Hist., Bot. Zool. Ser. 21: 1-2. Fitch, H. S. 1949. A study of snake populations in central California. Amer Midi. Naturalist. 41: 513-579. - 1960. Autecology of the copperhead. Univ. Kansas Publ. Mus. Nat Hist. 13: 85-288. Fitch, H. S. and B. Glading. 1947. A field study of a rattlesnake population. California Fish and Game. 33: 103-123. Gloyd, H. K. 1938. The snakes of Goose Hill Pond. The Chicago Nat. 1: 121-122. Hamilton, W. J·, Jr. and J. A. Pollack. 1955. The food of some crotalid snakes from Fort Benning, Georgia. Nat. Hist. Misc. Chicago Acad. Sci. 140: 1-4. Hirth, H. F. 1966. The ability of two species of snakes to return to a hiber- naculum after displacement. Southwestern Naturalist. 11(1): 49-53. Host, P. 1955. What the world sh6wed me. New York: Rand MeNally & Co. Pp. 1-302. Klauber, L. M. 1956. Rattlesnakes, vol. 1 and 2. Berkeley, Univ. California Press. Pp. 1-708, 709-1476. Leavitt, B. J. 1956. Water moccasin preys on pied-billed grebe. Wilson Bull. 69: 112-113. Lowe, C. H., Jr. 1948. Territorial behavior in snakes and the so-called court- ship dance. Herpetologica. 4: 129-135. Lueth, F. X. 1941. Effects of temperature on snakes. Copeia ( 1941 ), (3): 125-132, Mattice, M. R. 1950. Bridge's food and beverage analyses. Philadelphia: Lea & Febiger. Pp. 1-412. Neill, W. T. 1947. Size and habits of the cottonmouth moccasin. Herpetologica 3: 203-205. - 1948. Hibernation of amphibian5 and reptiles in Richmond County, Georgia. Herpetologica. 4: 107-114. - 1964. Viviparity in snakes: some ecological and zoogeographical con- siderations. Amer. Naturalist. 98: 35-56. 272 BULLETIN FLORIDA STATE MUSEUM Vol. 14 Noble, G. K. and H. J. Clausen. 1936. The aggregation behavior of. Storetia dekayi and other snakes with especial reference to the sense organs involved. Ecol. Monogr. 6: 269-316. O'Neil, T. 1949. The muskrat in the Louisiana coastal marshes. Louisiana Dept. Wildl. and Fisheries. Pp. 1-152. Parker, M. V. 1937. Some amphibians and reptiles from Reelfoot Lake. J. Tennessee Acad. Sci. 12: 60-86. Penn, G. H., Jr. 1943. Herpetological notes from Cameron Parish, Louisiana. Copeia ( 1943 ), ( 1): 50-59. Peterson, H. W., R. Garrett, and J. P. Lantz. 1952. The mating period of the giant tree frog Hula domicensis. Herpetologica, 8: 63. Ramsey, L . W. 1948 . Combat dance and range extension of Aglitstrodon pisci©orus leucostoma. Herpetologica. 4: 228. Sass, H. R. 1926. Adventures in green places. New York: G. P. Putnam's Sons. Pp. 1-297. Schwartz, A. 1952. The land mammals of southern Florida and the upper Florida keys. Unpublished Ph.D. dissertation, Ann Arbor, Univ. Michigan. Shaw, C. E. 1948. The male combat "dance- of some crotalid snakes. Herpe- tologica. 4. 137-145. Slevin, J. R. 1950. A remarkable concentration of dfsert snakes. Herpetologica. 6: 12-13. Smith, P. W. and J. C. List. 1955. N6tes on Mississippi amphibians and rep- tiles. Amer. Midl. Naturalist. 53: 115-125. St. Girons, H. 1957. Le cycle sexuel chez Vipera aspis ( L.) dans l'ouest de la France. Bull. Biol. France Belgique. 91: 284-350. Strecker, J. K. 1926. On the habitats of some southern snakes. Contrib. Baylor Univ. Mus. (4): 10-11. Tinkle, D. W. 1957. Ecology, maturation and reproduction of Thamnophis sauritus proximus. Ecology. 38: 69-77. ---. 1962. Reproductive potential and cycles in ·female Crotalus atrox from northwestern Texas. Copeia ( 1962), (2): 306-313. Volsoe, H. 1944. Structure and seasonal variation of the male reproductive organs of Vipera berus ( L.). Spolia Zool. Mus. Hauniensis. 5: 1-157. Wharton, C. H. 1958. The ecology of the cottonmouths, Agkistrodon viscivorus pisciuorus Lacepede, of Sea Horse Key, Florida. Unpublished Ph.D. disserta- tion, Gainesville, Univ. Florida. 1960. Birth and behavior of a brood of cottonmouths, Agkistrodon pia- civorus piscioorus with notes on tail-luring. Herpetologica. 16: 125-129. - 1966. Reproduction and growth in the cottonmouths, Agkistrodon pis- ciuOmB Lacepede, of Cedar Keys, Florida. Copeia ( 1966 ), (2): 149-161. Wood, J. T. 1954. The distribution of poisonous snakes in Virginia. Virginia J. Sci. 5: 152-167. Woodbury, A. M. 1951. Symposium: A snake den in Tooele County, Utah. Herpetologica. 7: 1-14. Contributions to the BULLETIN OF THE FLORIDA STATE MUSEUM may be in any field of biology. Manuscripts dealing with natural history or systematic problems involving the southeastern United States or the Caribbean area are solicited especially. Manuscripts should be of medium length-50 to 200 pages. Examination for suitability is made by an Editorial Board. The BULLETIN is distributed worldwide through institutional subscriptions and exchanges only. It is considered the responsibility of the author to distribute his paper to all interested individuals. To aid in this, fifty copies are furnished the author without cost. PREPARATION OF MANUSCRIPT Highly recommended as a guide is the volume: Conference of Biological Editors, Committee on Form and Style. 1964. Style manual for biological journals. ( Second Edition) Amer. Inst. Biol. Sci., Washington. 117 pp. Manuscripts should be typewritten with double spacing throughout, with ample margins, and on only one side of the paper. The author should keep a copy; the original and a carbon must be submitted. Tables and legends of figures should be typed on sheets separate from the text. Several legends or tables may be placed on a single sheet. Illustrations, including maps and photographs, should be referred to as "figures." All illustrations are reduced to a maximum of 4-1/4 by 7-1/8 inches. Size scales, wherever they are necessary, should be incorporated into the figure. References to literature should conform with the bibliographic style used in recent numbers of the BULLETIN. Spell out in full the titles of non-English serials and places of publication. Footnote material should be kept to a minimum. However, provide copy for a footnote detailing the title, a~liations, and address of the author ( see recent numbers of the BULLETIN). Manuscripts must be accompanied by a synopsis-a brief and factual summary ( not a mere description ) of the contents and conclusions, which points out the presence of any new information and indicates its relevance. In it list all new organisms described and give their ranges; indicate all taxonomic changes pro- posed. The synopsis, written in full sentences, should be concise, but completely intelligible in itself without references to the paper, thereby enabling the busy reader to decide more surely than he can from the title alone whether the paper merits his reading. The synopsis will be published with the paper. It does not replace the usual conclusions or summary sections. It may also serve as copy for the abstracting services. Manuscripts and all editorial matters should be addressed to: Managing Editor of the BULLETIN Florida State Museum Seagle Building Gainesville, Florida