Norss AND Ftplo Rsponrs 105

Acknowledgments We wish to thank Patrick T.F.H. Publications, 895 pp. The Alligator Snapping Turtle: Biology and Bergeron, Steven George, Chris Holt, and Steve Jensen for PnlrcHnnn, P.C.H. 1989. Conservation. Milwaukee: Milwaukee Public Museum, 104 pp. their help throughout this study. Jeffrey Lovich deserves SErcEL, R.A. 1980. Courtship and mating behavior of the dia- We thank Sherry Harrel, thanks for his helpful suggestions. mondback terrapin Mctlaclemys terrapin tequesta. J. Jan J. Hoover, and Anne-Marie Murphy for suggestions and Herpetol. 14:420-421 . for reviewing our manuscript. Special thanks goes to Tony Tnylon, E.H. 1933. Observations on the couftship of tuftles. Univ.

Trichel and Toby Trichel for helping capture the turtles and Kansas Sci. Bull.2l:269-271 . Mark Farr for his help in making observations. Received: 2l August 1995. Acceptecl: 3 December 1995. Literature Cited

'lrclonittrt Conserwttittrt untl Bitiltt,qt', 1996. l(l ):105-108 O 1996 by Chelonian Research Foundaticrtt ALLEN, E.R., AND NEILL, W.T. 1950. The alligator snapping turlle Mqcrochelt,s temminckii in Florida. Ross Allen's Reptile Inst., Spec. Publ. 4: I - 15. Meat on the Move: BnreR, R.E., AND GtlltxcHAM, J.C. 1983. An analysis of courlship Diet of a Predatory Turtl e, Deirochelys b g i .Herpetologi ca behavi or in B I andi n g' s tuft le, Enn do ide a landin reticularia (Testudines : Emydidae) 39 166-173. BELS, V.L., AND CRnvR, Y.J.M. 1994. Quantitative analysis of the courtship and mating behavior in the loggerhead musk turtle Dnlp R. Jl.cKSoNr Sternotherus minor (Reptilia: Kinosternidae) with comments on couftship behavior in tufiles. Copeia 1994:676-684. I Florida Natural Areas Inventorr'''^', The Notare Conservonc)', BERny, J.R., AND SutNe, R. 1980. Sexual size dimorphism and sexual I0l8 Thomasville Road, Suite 200-C, selection in turtles (Order Testudines). Oecologia (Berl.) 44:185- l9l . Tallahassee, Florida 32303 usA IFax: 904-68] -93641 BootH, J., RNp PEIERS, J.A. 1972. Behavioral studies on the green

turtle (Chelonia nn,clcls) in the sea. Anim. Behav. 20:808-812. The chicken turtle , Deirochelys reticularia, is a charac- Cox, W.A., Nownn,M.C., AND MnntoN, K.R. 1979. Observations on teristic but rarely abundant inhabitant of shallow, lentic in the musk turtle, Sternotlterus courtship and mating behavior waters in coastal lowlands of southeastern United States. ntinor. J. Herpetol. l4:200-204. Inadequate knowledge of its diet has led to widespread DRvrS, J.D.,, AND JncrcsoN, C.G., Jn. I 970. Copulatory behavior in the speculation that the species is broadly omnivorous (Carr, red-eared turtle, P s e u cl e rn), s s c r i p t u e I e g curs (Wied) . Herpetol ogi c a a few 26:238-240. 1952; Campbell, 1969; Ernst et al., 1994), although sug- DonrE, J.L. I 971 . Reproduction and growth in the alligator snapping authors (Jackson, 1978, 1988; Pritchard, l9l9) have tuftle, Mctcroclen'ty s tentntinckii (Troost). Copeia 191 l:645-658. gested a more strictly carnivorous diet. Very limited evi- EnNst, C.H., LovtcH, J.E.; AND Bnneoun, R.W. 1994. Turtles of the dence supporting the latter position is provided by Marchand United States and Canada. Washington, DC: Smithsonian Institu- (1942), Carr (1952), and Mitchell (1994), who reported tion Press, 578 pp. glass shrimp, crayfish, tadpoles, a snail, a beetle, and only a nuo TrNrLe, D.W. 1971. Biology of the Reptilia. Vol. 7. GRrus, C.', trace of aquatic plants in chicken turtle natural diets. Ecology and Behavior. New York: Academic Press. Method.s. In conjunction with a study of the reprodllc- GRtHapE, R. 1987 . Maintenance, behavior, and reproduction of the - tive biology of the species (Jackson, 1988), I examined the alligator snapping turtle, Mac'roc'lemr-s temminckii at the Tulsa feces were collected frorn Zoological Park. Bull. Oklahoma Herpetol. Soc. 12(l-4):l-6. gut contents of 24 chicken turtles; The JncrcsoN, C.G., JR., Rttp DRvts, J.D. 1972. A quantitative study of the another that was retained alive for behavioral study. courtship di spl ay of the red-eared turtle, Ch n,s e rny s s c ripta el e gan s sample included five adult males ( 103- 137 mm plastron (Wied). Herpetologica 28:5 8-64. length tPLl , 1921a8 g) ,19 adult female s (147 -190 mm PL, KnnuEn, M., AND FRttz, U. 1989. Courlship of the turtle Pseudenl\:s 645-1410 g), and one subadult female (137 mm PL, a10 g) ne lsoni. J. Herpetol. 23: 84-86. collected from 197 4 to I 971; shells of 12 were deposited in 1941. Fall mating and courtship of the musk turtle. LnGLER, K.F. the Florida Museum of Natural History herpetology collec- Copeia 194I:268. rion (UF 31555, 4421044216, 4421844220, 4423 I ). All LEclER, J.M. 1955. Observations on the sexual behavior of captive but one of the turtles were collected on land, either in tuftles. Lloydia I 8:95-99. with nesting activity or with terrestrial wander- LovrcH, J.E., Gnnsrrcn, W.R., AND CoopEn, W.E., Jn. I 990. Female association participation in courtship behavior of the turtle (Trachen4's s. ing that presumably represented migration between bodies scripta). J. Herpetol. 24:422-424. of water; several specimens were road-kills. Most of the MnHvoun, I.Y. I 961 .Courtship behavior and sexual maturity in four turtles were from Alachua County ( 12) or nearby counties species of kinosternid tuftles. Copeia 1967:314-319. (Baker, Dixie, Levy, Putnam, and Marion) in northern Munpuy, J.8., AND LauonEAUX, W.E. 1918. Mating behavior in three peninsular Florida, where the subspecies D. r. chrysea and Australian chelid turtles (Testudines: Pleurodira: Chelidae). D. r. reticularia intergrade. The sample also included one Herpetologica 34:398-405. specimen from the Florida panhandle (Wakulla County), PruvvEn' M.V. 1977 . Notes on the counship and mating behavior of from the Florida east coast (Brevard County), and two the softshell turtle, Trionrux rtrttticus (Reptilia, Testudines, two (Ware and Charlton counties). The Trionychidae). J. Herpetol. I l:90-92. from southern Georgia PnrrcHRRn, P.C.H. 1979. Encyclopedia of Tuftles. Neptune, NJ: entire alimentary tract of each animal was examined fresh 106 CuEr-oNraru CoNSERVATToN AND BtoLocv, Volume 2, Number I - 1996 within 24 hours of collection or refrigerated or frozen for Table 1. Gut contents of 25 Deirochelys reticulctriu mostly from northern peninsular Florida (see text for details). Frequency refers analysis within a few days or weeks. Most food items were to number of turtles containing the item. Table does not include first identified to family, then preserved in 40Vo isopropyl many insects too finely chopped to permit ready identification. alcohol for subsequent study. Taxa followed by an asterisk (*) are generally terrestrial though they include species found in close association with water. Lengths of the three major segments of the digestive tract (stomach, small intestine, and colon) were measured Taxon No. Individuals Frequency Deirochel;l's (seven females, for eight adult PL I4l-190 Arthropoda 94+ I 4 mm; one male, PL I l0 mm) that had been dissected for Malacostraca: Decapoda 7 4 reproductive information. For comparison, similar data were Palaemonidae: P al aenton ete s 3 I Cambaridae: Procantbarus sp. 5 4 series recorded from small of adult emydid turtles of three Insecta 53+ ll + pri mari ly herbivorou s, potenti al I y micro s ympatric fre shwa- Odonata l4 5 ter species: Pseudemys floridana peninsularis (three fe- Aeschnidae: Anuxo 5 3 Libellulidae: Libellulab I I males, PL269-332 mm; one mAle, PL 150 mm), P. nelsoni Hemiptera t2 3 (two females, PL 283-303 mm; one male, PL224 mm), and Belostomatidae : Lethoc e rus 3 2 Trachenrys s. scripta (one female, PL 183 mm; one male, PL Corixidae: cf. Hespe rocorixct 2 I Naucoridae: Pelocoris 3 2 the 154 mm). For analysis, I compared the relative length of Nepidae: Ranatra 5 2 digestive tract of each of the four species as a percentage of Orthoptera * 5 2 * plastron length. Gryllidae I I Tettigoniidae * 4 I To observe Deirochelys feeding behavior, I maintained Coleoptera 14+ 6+ one large adult female with as many as four adult males in a Curculionidae * 3 I 165 liter aquarium. Turtles were offered small amounts of a Dytiscidae 4 2 Elateridae * 2 I variety potential foods (only one at a of kind time) including Hydrophilidae 5 2 crayfish, insects, earthworrns? raw fish, and aquatic vegetation. Lepidoptera: cf. Noctuidaec* 7 2 Results Nearly half of the digestive tracts were Diptera: Stratiomyiidaeo I I Arachnida * 34+ 2 relatively empty, which may relate to the biased collection Plant matter unquantified 6 of reproductively active females during the cooler months of a the year (see Jackson, 1988). The guts from 8 of 12 gravrd larvae or nymphs b adult females as well as all 3 pre-gravid females (i.e., bearing ' all adults but one enlarged preovulatory follicles) were essentially devoid of food; the remaining 4 gravid females each contained only a species was as follows: De irochelys,,390Vo (range 320430, small number of prey items. In contrast, the guts of all4post- SD 38); Pse udenrys floridana,8507o (range 630-1030; SD reproductive females (March-June) were moderately full. 180); P. nelsoni, 920Vo (range 140-1120, SD 191); and All 5 of the males and the subadult female (collected late Trachemys scripta, J207o (range 580-860, SD 198). Only May - early October) contained several items each. the differences between Deirochelys and the other species Table I summarizes the identifiable food items found in were significant (t-tests, P < 0.05). the sample; the diet consists principally of aquatic insects In laboratory observations of feeding behavior, chicken and crustaceans (Malacostraca), with some representation turtles typically searched with extended necks until they of terrestrial arthropods. Prey generally were chopped into detected prey movement. Slow stalking or swimming then fragments, which partially prevented quantitative analysis brought them within striking range (ca. l0- 12 cm), at which by species. Most specimens appeared to represent prey in the point the head remained stationary while the neck recoiled range of 8 to 20 mm in length, although a few were as small into the still-approaching body. The ensuing rapid strike as 4 mm (Naucoridae and Hydrophilidae) and others as large usually occurred when the prey was stationary and followed as 55 mm (odonates and crayfishes [Cambaridae]). No the pattern of "pharyngeal feeding" described in detail by substantive dietary differences were detected between the Bramble (1913) for Chelydra. Small prey were captured small sample of males and the larger female sample. from all angles, but large and potentially dangerous prey Plant matter was limited to < 3 cc of green algae in the such as crayfish generally were circled until alateral strike colons of each of 5 individuals, a I cm woody twig fragment, could be attained; adult turtles of both sexes fed re adily on and parts of the roots and leaves of a water hyacinth crayfish in this manner. Dangerous prey such as large (> 35 (Eichhornia crassipes) in an individual that contained nu- mm) adult and larval dytiscid beetles and belostomatid bugs merous insect, spider, and decapod prey that typically dwell were avoided, sometimes after a single painful encounter, in the water hyacinth microhabitat. The guts of at least three though some smaller individuals of those taxa were con- individuals contained small semi-flattened stones that may sumed in the wild (Table I ). Turtles watched but spent little have represented gastroliths from crayfish prey. Four turtles to no effort in pursuing highly mobile prey such as sunfish also contained small, yellowish, gelatinous spheres (one as (Lepontis), nor did they attempt to consume aquatic vegeta- many as 20) that may have been amphibian eggs. tion. However, most individuals readily accepted pieces of Mean digestive tract length (stomach plus intestine and raw fish and even scavenged dead crayfish after a period of colon) as a percentage of plastron length for each of the four acclimation to captivity. Norps AND FIsln Rsponrs r07

Discussion.- Despite statements by some authors that herbivorous emydids, Pseudemys floridana and P. nelsoni, Deirochelys at least occasionally ingests plant matter (Carr, which occupy some of the same bodies of water as chicken 1952; Campbell, 1969; Ernst et al., 1994), my data and turtles. A higher level of overlap exists, however, with the observations indicate that this turtle is a carnivore that sympatrtc Trachemys s. scripta, anomnivore that feeds upon specializes on live, slow-moving prey. The diet consists many of the same prey (unpublished data). Dietary overlap almost entirely of arthropods, principally aquatic insects and may also occur with members of other freshwater turtle crustaceans. The occasional consumption of terrestrial families (Chelydridae, Kinosternidae, and Trionychidae), arthropods that presumably have fallen into the water under- juvenile alligators, and some fishes (e.g. , Micropterus). scores the importance of allochthonous input into aquatic Among North American turtles, the feeding habits of food webs, as was noted by Georges (1982) for other D e i r o c h e lt' s are most s imi I ar to tho se of Emy do ide a b landin g ii freshwater vertebrates. The relatively anterodorsal position (Bleakney, 1963), an allopatric and distantly related but of the eyes of Deirochelys probably allows it to scan the highly convergent emydine that appears to be the "northern water surface more effectively for such prey (Pritchard, equivalent" of the chicken turtle. Elsewhere, remarkably 1984), as well as for those that normally live at the air-water parallel diets and associated morphological adaptations oc- interface (e. g., some spiders, water striders, and gyrinid beetle s). cur in even more distantly related turtles, such as the South The degree to which individual turtles "specialized" on American chelid Hydromedusa maxintiliani (Souza and certain prey types, at least temporarily, was conspicuous. Abe, 1995). Thus, different individual turtles were responsible for nearly Interestingly, no fish were confirmed as prey of the entire consumption of glass shrimp (Palaemonidae), Deirochelys, even though several species of small body size grasshoppers, moths, and spiders, respectively. Either indi- were abundant in many of the waters from which the sampled vidual turtles formed search images for certain prey types, or turtles emerged. However, their lack of relatively large, otherwise they foraged in localized microhabitats where indigestible parts comparable to the chitinous exoskeleton those prey types were especially abundant. Dietary diversity of arthropods facilitates rapid digestion and hence may might result from turtles moving among such microhabitats. conceal the presence of small tish in the diet. Still, fish are The small amounts of plant matter found in a few guts probably difficult prey for chicken turtles to capture and can in this study, as well as the ingestion of plant parts reported be expected to be a minor dietary component, if eaten at all. elsewhere, presumably represent cases of incidental inges- In fact, in the southeastern US, Deirochelys is one of the few tion and mistaken identity. The acclimation of some captives turtles (with Kinosternon spp.) that routinely inhabit fish- to raw fish or commercial fish chow (Ernst et al. ,1994; pers. free temporary ponds. Perhaps not coincidentally, fish also obs.) does suggest the potential for scavenging in nature. The seem to be absent from the diets of other turtles that special- fact that chicken turtles rarely enter aquatic traps baited with ize on aquatic arthropods (e.g., Hydromedusa maximiliani sardines (pers. obs.) suggests, however, that the latter activ- fSouza and Abe, 1995]). ity is infrequent. Nonetheless, Cagle (1950) believed that The relatively small amount of food in most of the chicken turtles respond best to bait in advanced stages of digestive tracts is probably attributable to collection bias. decay; geographic differences in dietary preferences might My sample over-represents turtles that had emerged to nest account for this. or to emigrate from ponds. Emigrating chicken turtles not These conclusions are consistent with this turtle's spe- only are unable to feed on land but also may be moving cialized head and neck morphology, which represents a because of low prey densities. Many of the adult females in highly evolved suite of adaptations for pharyngeal feeding, my sample were deliberately taken during the species' a mode specifically adapted to the capture of mobile, aquatic winter reproductive season (October-March) (Jackson,

prey (Bramble, 1 97 3, I97 4;Jackson, 1978; Pritchard, 1984). 1988), a time when arthropod prey may be relatively sparse. Narrow jaws, a greatly hypertrophied hyoid apparatus and Furthennore, gravid female reptiles in general often reduce cervical extensor musculature, elongated cervical vertebrae their food intake (Shine, 1980), with one explanation for this to accommo- and skull, and relatively anterodorsal eyes combine to facili- - a lack of physical space in the body cavity tate rapid strikes at living prey. My observations of captive date prey items (Gregory and Stewart, I97 5) being animals revealed the technique to be equally effective at all particularly applicable to turtles. Additional sampling of adult body sizes, which might explain the lack of pro- non-reproductive turtles (males, subadults, and juveniles) nounced sex-related dietary differences despite the two- to during the winter reproductive season is necessary to deter- three-fold disparity in body mass between typical males and mine whether the reduced feeding by reproductively active females. Nevertheless, the wider gape of large females females observed in this study is a direct response to repro- allows them to take somewhat more robust prey than can be ductive state or a more generalized Seasonal response. effectively handled by males Qters. obs.). The carnivorous food habits of Deirochelys may have Additional evidence of the carnivorous diet of several ecological consequences. Relative to other sympat- Deirochelys is provided by the relatively short and simple ric freshwater emydids (Pseudemys, Trachemys), chicken gut when compared with that of sympatric herbivores turtles are relatively small; the largest species (Pseudemys) (Pseudemys) and omnivores (Trachemys). The diet of are all herbivorous, while the intermediate-sizedT. s. scripta Deirochelys shows no overlap with those of two large is omnivorous. Though perhaps more germane to mamma- ,i!t--

108 CuEloNtnx CoNSERVATIoN AND BroLoGy, Volmne 2, Nurnber I - 1996

lian than reptilian communities, population densities of H)' cl r o nt e cl u s a nmx im i I i an i (Testud i ne s : Chelidae) i n southeastern predators are often substantially smaller than those of herbi- Brazll. Chelon. Conserv. Biol. l(4):320-322. vores, and this also seems to hold within this turtle commu- Wenn, S.D., MncFnoDEN, 8.J., AND Besrcrx, J.A. 1981. Geology and paleontology nity. Even in fossil sites where Deirochelys is well repre- of the Love Bone Bed from the Late Miocene of Florida. Amer. J. Sci .281:513-544. sented (e.g., Miocene Love Bone Bed, Alachua County,

Florida [Jackson, 1918; webb et al., 198 l]), it is srill our- Rec:eivecl: 2 September I 995. Accepted: 5 November I 995. numbered roughly 4:I by Pseudenxys. And finally, though a confirmed basker, Deirochelys seems to spend less time engaged in this activity than herbivorous species, perhaps j,l 08- I 0e reflective of a difference in the relative digestibility of plant o, ee6 ., ;il:ffi ;l ?:3f1,''Jj,l;"1,j,i and animal matter. Kemp's Ridley Sea Turtle Acknowledgm,ents. - Dick Fran z and Ken Tennessen (Lepidochelys kempii) Tracked by Satellite assisted with generic identifications of crayfishes and in- Telemetry from Louisiana to Nesting Beach sects, respectively. Laboratory space was provided by the at Rancho Nuevo, Tamaulipas, Mexico Florida Museum of Natural History and the University of Florida Department of Zoology. The manuscript was en- MlunrcE L. RnNaunr, JnuES A. CnnpBNTERl, hanced substantially by the thoughtful comments of review- A. WnrnMSl, AND Arunnn LRNDRv, ers and editors. Jo M. JR.'

I Literature Cited Southeast Fislteries Center, Nqtionctl Murine Fisheries Service, Gulveston Laborcttory, 4700 Avenue (J, Gulvestor"t, Texos 7755I USA [Fcw: 409-766-3508; E-ntctil: BrEarcNEv, J.S 1963. . Notes on the distribution and life histories of 2 renauchn@ tantug3. tcuttu. eclu I ; Deportment o,f M arine Biolo g),, turtles in Nova Scotia. Canadian Field-Nat. 77:67-76. Texas A&M Universi\, ,tt Galveston, 4700 Avenue (J, BnnNaer-e, D.M. 1913. Media dependent f'eeding in tuftles. Amer. Building 303, Galveston, Texas 77551 USA ZooL 13: 1342. BnRvnLE, D.M. 1974. Ernydid shell kinesis: biomechanics and The Kemp's ridley (Lepidochelys evolution. Copeia 1974.701-727 . kempii) is considered CRGLE, F.R. 1950. The life history of the slider tuftle, Pseuclerr?.]:s the most endangered sea turtle in the world (Groombridge, st:riptet troostii (Holbrook). Ecol. Monogr . 20:31 -54. 1982: Shaver, l99l; U.S. Fish and Wildlife Service, 1992). CnHapeell, H.W . 1969, The unsung chicken turtle. Int. Turlle Tort. It is distributed throughout the Gulf of Mexico (Liner, 1954; Soc. J.3(5):22-24,36. Carr, l95l; Carr, 1980; Hildebrand, 1982; Manzella and CRRR, A.F. 1952. Handbook of Turtles. Ithaca, New York: Cornell William s, 1992), but is most abundant in coastal waters from Univ. Press, 542 pp. Texas to Florida (ogren, 1989). It also occurs along the ERNST, C.H., LclvrcH, J.E., AND Bnneoun, R.w. 1994. Turtles of the eastern shore of North America to Newfoundland and has United States and Canada. Washington, DC: Smithsonian Inst. been Press, 578 pp. reported in the European Atlantic near the British Isles, GEoncES, A. 1982. Diet of the Australian freshwater turtle Enuclurcr Netherlands, and France (Pritchard, 1989). Despite this krffiii (Chelonia: Chelidae), in an unproductive lentic environ- widespread distribution, almost all nesting occurs on about ment. Copeia 1982:331 -336. a 60 km stretch of beach near Rancho Nuevo, Tamaulipas, GnEcony, P.T., AND SrEwenr, K.W. 1915. Long-distance dispersal Mexico. and feeding strategy of the red-sided garter snake (Thcunnophis An adult female Kemp's ridley was tracked from elegcuts) in the Interlake of Manitoba. Canad . J.Zool. 53:238-245. Cameron, Louisiana to Rancho Nuevo, Tamaulipas', Mexico Jncnsox, D.R. I 978. Evolution and fossil record of the chicken turlle by satellite telemetry for 281 days from 13 Augusr Ig94 Deirochel1,s, with a re-evaluation of the genus. Tulane Stud. Zool. through 16 May 1995. The turtle measured 65.8 cm Bot. 20:35-55. straight carapace length,64.9 cm straight carapace width, and Jncrsoru, D.R. 1988. Reproductive strategies of sympatric freshwater weighed 42.6 kg. The was emydid turrtles in northern peninsular Florida. Bull. Florida State turtle captured at Cameron, Louisiana in a Mus., Biol. Sci. 33(3): I l3-158. 91.5 m turtle entanglement net (7 .3 m deep , 12.7 cm bar MnncHnNo, L.J. 1942. A contribution to a knowledge of the natural mesh) by Texas A&M University biologists. It was fitted history of ceftain fi'eshwater turtles. M.S. Thesis, Univ. Florida, with a Telonics ST- 10 satellite transmitter on l3 August Gainesville. 1994 and tracked by National Oceanographic and Atmo- MncsEll, J.C. 1994. The Reptiles of virginia. washington, DC: spheric Administration' s National Marine Fisheries Service Smithsonian Inst. Press, 352 pp. Galveston Laboratory. It moved offshore of the upper Texas PRrrcHRRo, P.C.H. 1979. Encyclopedia of Tuftles. Neptune, NJ: coast in late November 1994 and travelled south along the T.F.H. Publications, Inc., 895 pp. Texas coast into Mexican waters through early January I 995 Pntrcsnnn' P.C.H. 1984. Piscivory in turtles, and evolution of the (Fig. I It was long-necked Chelidae. Symp.zooL Soc. London 52:8i-l 10. ). offshore of the Rancho Nuevo nesting beach SutNE, R. 1980. "Costs" of reproduction in reptiles. Oecologia (Ber- by 10 March 1995. It moved an additional 100 km to rhe lin) 46:92-100. south before returning to nest on23 April 1995 and again on SouzR, F.L., AND Ane, A.S. 1995. observations on feeding habits of L9 May 1995, both nestings being recorded by Rancho