178 CuEloNrnN CoNSERVATToN AND BtoLocv, Volume 4, lluntber I - 2001

'unidentified' .ltelonian were grouped into an category. The sorted Cortsen'utiort tutcl Biologt. 2001. -1( I ): 178-l 8 I O 2001 by Chelonian Research Foundation samples were allowed to air dry for 2448 hrs and air-dry mass was then recorded for each pooled sample group. Percent occurrence and percent air-dry mass was also calcu- Foraging Feeding Summertime and lated for each species present. by Immature Loggerhead Sea Turtles Microscopic investi-eation (magnification up to 1000x) (Csretta carettq) from Georgia was used to locate nematocysts within the digestive tract. Digestive tract scrapings and the remains of isopropyl alco- MTcHAEL G. FnrcKr, KnrsrrNA L. WTLLTAMSI, hol used for storage and rinsin-e of the digestive tract, were utilized for microscopic analysis (see Van Nierop and den AND LNNNP PInnRARD2 Hartog [1984] and den Hartog and Van Nierop [1984] for similar methodology). I Caretta Research Project, Twelve loggerheads were collected from P.O. Box 984I , Savannah, Georgia 3 14I2 USA Results three barrier islands in Georgia from 1995 to I 999: Wassaw I Fax: 9 ] 2-447-8656; E-mail: Caretta}S @ aol.cornl ; 2T-vbee Marine Science Center, Island (n - 9), Little Cumberland (n - 2), and Blackbeard Tybee Island, Georgia 3I328 USA Island (n - I ). Specimens were found throughout the sum- mer from May to July. All loggerheads sampled were Bjorndal ( 1 997) thoroughly reviewed the literature on immature with curved carapace lengths ranging from 59.4to the foraging ecology and nutrition of sea turtles. However, 71 .0 cm (me&n = 66.9 cm). Sex was determined, via gonadal much work is still needed to determine the feeding and examination, for 8 of the 12 turtles sampled. A female biased foraging habits of sea turtles throughout the southeastern sex rati o (7 females: 1 male) was observed. Judging from the USA north of Florida and south of Virginia. Several studies amount of body fat present, visual inspection of the visc era,, have reported diet selection in sea turtles from Georgia, and relatively little barnacle fouling, all of the turtles exam- USA. (Ruckdeschel and Shoop, 1988; Frick, I99l; Creech ined appeared to be healthy prior to death. and Allman, 1998; Frick and Mason, 1998; Pete and Winn, Air-dry masses of digestive tract contents from indi- 1998; Frick et aI.,1999; Shoop et al., 1999). One of these vidual Caretta ranged from 0.41 to 3452.81g (mean = 3 10.6 studies presented data on the quantity of prey items ingested g; total mass = 10.03 kg). Contents obtained from the by Lepidochellts kempii (Creech and Allman, 1998). The esophagus and stomach proved to be the most identifiable as other studies either reported stomach contents with no sup- most food items were only partially digested. Intestinal porting quantitative data or they only reported anecdotal contents were much more fragmented and digested. How- feeding observations. Here we report and quantify the diges- ever, a large percentage (ca. 607o) of the contents collected tive tract contents of 12 immature loggerhead turtles (Caretta from the intestines were identifiable and were included into caretta) from Georgia, USA. Additionally, we suggest the calculations with the stomach and esophageal contents. possible methodology utilized by Caretta to obtain the Anthropogenic debris was absent from all digestive tracts. observed prey items. Seven different prey groups were recognrzed from the Methods Carcasses of Caretta caretta were col- digestive tract contents of Caretta (Table 1). Crabs consti- lected from Wassaw Island (3 1"53.4'N, 80"58.4'W), Little tuted the highest percentage of all digestive tract specimens Cumberland Island (30'57.2' N, 8 1'25.5' W), and Blackbeard encountered, occurring in 1007o of the samples and repre- Island (3 1"28.4'N, 8l'1 3.1' W) along the Georgia coast from senting 18.27o of the pooled total air-dry mass. Nine differ- 1995 through 1999. Data collected from stranded turtles ent crab species were recognized overall (Table 2). Spider included: date, time, location, species identity, condition of crabs (Libinia sp.) and stone crabs (Menippe mercenaria) specimen, curved carapace length and width, and if possible, represented the bulk of the air-dry mass from crabs at23.47o sex (via gonadal examination). In order to record the loca- and 22.lTo,respectively. These two species also had a high tion within the digestive tract where a particular prey item percent occurrence relative to other crab species present was found, twist ties were placed to isolate the esophagus, (58.37o and 41.77o, taspectively). In addition to the afore- the stomach, and the intestines (see Wolke and George, mentioned species, other walking crabs were identified as 198 I ). These sections of the digestive tract were then re- major prey items (Calappa flamntea, Hepatus epheliticus, moved and placed into separate collection bags containing and Persephona mediterranea). Portunid, or swimming 70Vo isopropyl alcohol. Contents were emptied and sieved crabs (Callinectes sapidus and Portunus spinimanus), repre- through a l -mm fine mesh colander and rinsed with isopro- sented the least common crab species prey item except for pyl alcohol into the collection bug, which originally housed the flat-clawed hermit crab (Pagurus pollicaris). Interest- the sample. Sections of the digestive tract that were emptied ingly, striped hermit crabs (Clibinarius vittatus) were found were then placed back into their original collection bag until in 33.3Vo of the digestive tract samples. This represents the a microscopic analysis could be performed to determine the highest percent occurrence of hermit crabs reported from presence of digested jellyfish via nematocyst occurrence. Caretta thus far. Sieved samples were identified and grouped to the Mollusks were the second highest ranked prey item lowest taxon possible. Items that could not be identified found in digestive tract samples (Tables I and 3). Gastro- Norns AND Fmlo REponrs 179

Table 1. Percent occurrence and percent air-dry mass of prey items shrimps (Callianassa bifurmis) also ranked fairly low in identified from digestive tract contents of immature Caretta caretta their importance as a prey item consumed by loggerheads. in Georgia, US A (n l2). Total mass for percentages = 10.03 kg. - Barnacles were commonly found affixed to the carapace Prey group 7o Occurrence Vo Ak-dry mass fragments of crabs and on the shell fragments of gastropod Crabs 100.0 78.2 mollusks. Interestingly, ghost shrimp were only found from Mollusks 83.3 I1.1 digestive tract samples that also contained the anemone Horseshoe crab (Limulus polyphemus) 75.0 8.3 samples containing Barnacles 41.7 0.2 Paranthus duplicatus. However, some Anemones 25.0 0.t P. duplicatus did not contain ghost shrimp. Ghost shrimp (Callianassa biformis) 25.0 0.2 Discussion The prey items consumed by the logger- Jellyhsh 25.0 0.6 nearshore Unidentified 17.2 0.1 heads examined are common inhabitants of the environment along the southeastern USA coast (South Caro- pods represented the bulk of observed mollusks. Knobbed lina, Georgia, and north Florida) during the summer months whelks (Busycon carica) andAtlantic moon snails (Polinices of May through July (Ruppert and Fox, 1988). Walking duplicatus) represented 86.4Vo of the total air-dry mass from crabs and horseshoe crabs appear to be an important prey mollusks. item for loggerheads. Portunid or swimming crabs are also Horseshoe crabs ranked high in occurrence (l5Vo) but sought, but to a lesser extent; perhaps due to the difficulty accounted for only a minor percentage of the overall air-dry that may exist in catching these relatively fast swimmers. mass (8.3Vo).Such a low air-dry mass percentage may be The presence of hermit crabs within the digestive tract attributed to the amount of each individual horseshoe crab contents suggests that certain gastropod mollusks may be consumed. Only the legs of L. polyphemus were observed in misrepresented as important prey items. Perhaps the shells most samples. of gastropods may have actually housed hermit crabs rather Cnidarians (ellyfish and anemones) occurred in 507o of than snails, although the shells of gastropod mollusks were the digestive tract samples but only accounted for I.3Vo of found in samples not containing hermit crabs. However, the total air-dry mass (Table 1). Due to the high digestibility observations of captive loggerheads feeding upon hermit of cnidarians, likely their importance as a prey item is crabs have shown that often times a hermit crab will flee its underestimated. Anemones (Aiptasia pallida, Calliactis tri- shell as a turtle manipulates the shell in its jaws (MGF, pers. color, and Paranthus rapifurmis) were identified from indi- obs.), and in most cases the turtle continued to consume the viduals found within the esophagus and stomach. In one gastropod shell even though the crab had fled. Additionally, sample, a single jellyfish (Aurelia aurita) was obtained from most samples containing gastropod shells lacked the horny the esophagus and subsequently identified from the gonadal plates, or opercula, used by Busycon and Polinices to oc- structure. The presence of all other jellyfishes (Aurelia clude their shell opening. The presence of opercula confirms aurita and Chrysaora quinquecirrha) was determined that gastropod shells found within the digestive system through microscopic analysis of nematocysts within the contained the animals that constructed the shells and not digestive tract lining and/or from the digestive fluid/ hermit crabs, a secondary tenant of gastropod shells. On preservative mixture left over in the collection bags. A only three occasions were opercula found within diges- ctenophore, Mnemiopsis mccradyi, was found in one of tive tract samples that also contained Busycon and the samples and was identified by the presence of 'ctenes' Polinices shells. Only one of the aforementioned samples or comb plates. contained Busycon shells, opercula, and hermit crabs. Barnacles (Chthamalus fragilis) occurred in 4l.l%o of However, it should be noted that the opercula associated the digestive tract samples but contributed the lowest per- with Polinices are much thinner and more fragile than centage to the overall air-dry mass (0.2Vo; Table 1). Ghost those of Busycott and may have been too fragmented during consumption and digestion for positive identifi- cation. Nonetheless, the possibility exists that the impor- Table 2. Percent occuffence and percent air-dry mass of crabs tance of gastropod mollusks in the diet of Caretta may be identified from digestive tract contents of immature Caretta caretta in Georgia, US A (n = l2). Total mass for percentages = 7.84 kg. misrepresented within this study and possibly also in hey species Vo Occurrence Vo An-dry mass

Spider crabs (Libinia sp.)* 58.3 23.4 Table 3. Percent occurrence and percent air-dry mass of mollusks Calico box crab (Hepatus epheliticus)* 41.7 20.3 identified from digestive tract contents of immature Caretta caretta Stone crab (Mentppe mercennria)* 4t.t 22.1 in Georgia, US A (n = 10).Total mass for percentages = 1.11 kg. Strip"d hermit crab (Clibinnrius vinans) 33.3 0.9 Flamed box crab (Calappaflammea)* 25.0 I 1.0 Prey species Vo Occunence Vo Air-dry mass Mottled purse crab Knobbed whelk (Busycon carica) 41.7 83.2 (P e r s e p horn me dit e rcane a)* 2s.0 2.1 (Ilyannssa Blotched swimming crab Mud snail obsoleta) 4r.7 0.5 (Polinices (Porturun spinimnnus)* 16.7 0.9 Moon snail duplicatus) 33.3 3.2 (B Blue crab (Callinectes sapidus)* 16.7 1.0 Channeled whelk usy con cannliculatum) 25.0 0.5 (C Flat-clawed hermit crab (Pagurus pollicaris) 8.3 0.3 Slipper snail repidula fornicata) 25.0 0.3 (Sphenia Unidentified 41.1 18.0 S oft-shelled clam antilliensis) 25.0 0.1 Great heart cockle (Dinocardium robustum) 8.3 0.7 * brachyuran crab Unidentified 17.8 I 1.6 ll,--

180 CuEI-oNraN CoNSERVATToN AND BroLocy, Volume 4, Number I - 2001

past studies that report both gastropod mollusks and Texas was during the summertime. This also coincides hermit crabs from digestive tract samples. Determining with the breeding of brachyurans from the Gulf of Mexico whether or not Carettais ingesting mollusks or hermit crabs (Heard ,1982). However, more detailed studies are needed is important to clarify if one is to investigate the amount of to determine whether or not Caretta is in fact consuming nutrition a turtle receives from its overall diet. gravid crabs. Additionally, more studies are needed to Abdomens from consumed crabs were also present determine if Caretta feeds on brachyurans outside of the within the stomach contents of ca.70Vo of the loggerheads summer months in Georgia. In areas where loggerheads examined. Although we were unable to use abdomens for feed upon brachyurans year-round, it may be possible to species identification, they did allow us to determine the sex determine to what extent loggerheads benefit from the of some of the consumed crabs. Abdominal segments play crab's reproductive cycle, if at all. an important role in reproduction for brachyuran crabs (see Bivalve and gastropod mollusks were also highly repre- Table I for brachyurans consumed; Williams, 1984). The sented in the digestive tract contents of Caretta. Whelks abdomen houses the male crab's reproductive organs and (Busycon sp.) and Atlantic moon snails (Polinices duplicatus) supports brooding eggs in females. Abdomens appear elon- are common within the marshes and nearshore environ- gated in male brachyurans and rounded in females. During ments at depths up to 18 m (Ruppert and Fox, 1988). As the summer, a female crab may contain millions of spongy previously mentioned, their occuffence in the diet of Caretta eggs beneath its abdomen (i.e. , Callinectes sapidus; Will- may be misrepresented since hermit crabs are often times iams, 1984). Interestingly, ca. 657o of the abdomens found found utilizing their discarded shells (Williaffis, 1984). within Caretta digestive tracts belonged to female crabs. It However, we have no reason to believe that the animals is quite possible that these female crabs may have been themselves are not being consumed since their appearance to brooding eggs when they were consumed even though no a turtle would not be unlike that of a hermit crab. With the brachyuran eggs were found in the microscopic analysis of exception of a single great heart cockle (Dinocardium the digestive tract. The high digestibility of crab roe would robustum), it is our belief that the other ingested mol- account for its absence in our samples. Additionally, the lusks (Crepidula fornicata, Ilyanassa obsoleta, and majority of L. polyphemus legs encountered in our samples Sphenia antilliensis) as well as barnacles (Chthamalus appeared to have belonged to females. Male L. polyphemus fragilis) were consumed incidentalty through the sedi- have a modified pair of hooked walking appendages (pedi- ment surface or as epibionts on larger mollusks or even palps) that are used to attach to the female during amplexus hermit crabs. (Ruppert and Fox, 198s). Females have unhooked pedi- The presence ofjellyfish and their nematocysts suggest palps. Hooked pedipalps were absent in ca. 75vo of the that loggerheads also feed within the water column during samples containing horseshoe crab legs. It is possible that the summer months. Moon jellies (Aurelia aurita) and sea Caretta did not ingest hooked pedipalps during the con- nettles (Chrysaora quinquecirrha) are the most common sumption of male Limulus. Yet, chelicerae (feeding append- jellyfish in the nearshore waters of the southeastern USA ages located in front of the pedipalps) and walking legs from May through August (Calder, 1977). Digestive rract (located immediately behind the pedipalps) were present in samples containing jellyfish or their nematocysts also con- all samples containing horseshoe crab legs. This suggests tained brachyuran crabs. However, the brachyurans associ- that the absence of hooked pedipalps would indicate that ated with Aurelia and Chrysaora were juveniles (Libinia sp. primarily female L. polyphemus were consumed. Horseshoe and Persephona mediterranea) that have been documented crabs also brood eggs during the summer and may have been as commensals on jellyfish until their adult lifestage where gravid when consumed. A gravid crab would undoubtedly they then shifttowards abenthic existence (Williams, 1984). provide a much more substantial source of nutrition than one Some digestive tracts containing Chrysaora also contained without eggs. the warty comb jelly, Mnemiopsis mccradyi, a common prey Ruckdeschel and Shoop (1988) surveyed the gut con- item of Chrysaora (Ruppert and Fox, 1988). Only one turtle tents of 137 loggerheads stranded on Cumberland Island, was sampled that had consumedjellyfish (Aurelia) aswell as Camden County, Georgia, and found no Limulus legs from benthic gastropods. Apparently loggerheads shift between any of their samples. However, Lutcavage and Musick pelagic foraging and benthic foraging within the same sea- ( 1985) found Limulus to be a common prey item of logger- son. Shifts in their foraging methodology may occur due to heads sampled in Virginia and Layne (1952) provided evi- individual turtle preference or because of changes in prey dence of Limulus consumption by Caretta. Similarly, Shoop abundance. et al. ( 1998) indicated that fish, fish bones and (commercial) Other cnidarians consumed by Carettawere the anemo- shrimp (possibly shrimp trawler by-catch or trawl net de- nes Aiptasia pallida, Calliactis tricolor, and Paranthus rived) are occasionally found packed within the gastrointes- raptfurmis. Aiptasia and Calliactis arecommon epibionts of tinal tract of some stranded loggerheads examined on hermit crabs. Both species were found from digestive tracts Cumberland Island. None of the loggerheads examined also containing hermit crabs (Clibanarius vittatus and during our study contained commercial shrimp or fish. Pagurus pollicaris). Paranthus is a burrowing anemone Plotkin et al . (1993) reported thar rhe highesr occur- commonly referred to as the 'sea onion' due to its appear- rence of brachyuran crabs consumed by loggerheads in ance when it contracts its body into a tight ball and floats to Nores AND FIeln Rsponrs 181 a new burrowing location via wave action (Ruppert and Fox, South Carolina Sea Grant Marine Advisory Bulletin I 1:l -L2. 1988). It is easy to picture aloggerhead seizing the opportu- CneEcH, L. RNo ArrunN, P.E. 1998. Stomach and gastrointestinal (Lepidochelys kempii) sea nity to prey upon Paranthus as it floats passively within the contents of stranded Kemp's ridley In: Epperly, S.P. and Braun, J. (Eds.). Proceed- water column. However, some specimens of Paranthus turtles in Georgia. ings of the Seventeenth Annual Sea Turtle Symposium. NOAA were found in digestive tract contents also containing ghost Technical Memorandum NMFS-SEFSC- 415, pp. I 56- 1 58. shrimp (Callianassa btfurmis), and other burrowing species DpN Hnnroc, J.C. AND VnN NmRoe, M.M. 1984. A study on the gut (Busycon carica, Dinocardium robustum, Menippe contents of six leathery turtles Dermochelys coriacea (Linnaeus) mercenaria, and Polinices duplicatus). These prey items (Reptilia: Testudines: Dermochelyidae) from British waters and from suggest that loggerheads in Georgra (or other areas in the the Netherlands. Zoologische Verhandelingen Iriden 209: 1-36. southeastern U.S) might actively dig for prey items similar Fox, R.S. AND Rupprnr, E.E. 1985. Shallow-water marine benthic to loggerheads performing infaunal mining in Australia macroinvertebrates of South Carolina. University of South Caro- (Preen,1996). Callianassa seems to be too small (3 cm) to lina Press, Columbia, 330 pp. Frucr, M.G. 1991 . Lepidochelys kempi (Kemp's ridley), Caretta be actively sought by a subadult loggerhead and may have c ar e tt a (lo ggerhead), and M al ac I e my s t e r rap in c e nt r at a (C arol ina been ingested inadvertently while the turtle was consuming diamondback terrapin). Diet and predation. Herpetological Re- other Possibly burrowing organisms. such burrowing speci- view 28:149, mens were initially displaced by shrimp trawlers and subse- Fzucx, M.G. AND MasoN, P.A. 1998. Lepidochelys kempi (Kemp's quently consumed by Caretta. ridley sea turtle). Diet. Herpetological Review 29:166-168. During the summer months (May-July) in Georgia and FRrctr, M.G., QuwN, C.A., AND SLny, C.K. 1999. Dermochelys possibly in similar and adjacent marine habitats in South coriacea (leatherback sea turtle), Inpidochelys kempi (Kemp's Carolina and north Florida, immature loggerheads forage in ridley sea turtle), and Caretta caretta (loggerhead sea turtle). pelagic and benthic habitats of the nearshore environment, Pelagic feeding. Herpetological Review 30: 165. HeRRn, R.W. 1982. Guide to the common tidal marsh invertebrates including the salt marsh. Based upon the natural history of of the northeastern Gulf of Mexico. Mississippi-Alabama Sea the prey items observed Caretta may obtain its food in a Grant Consortium MASGP-79-004, 82 pp. variety of ways. By perusing and scanning the benthic LRvNE, J.N. 1952. Behavior of captive loggerhead turtles Caretta c. environment a loggerhead could easily obtain the crabs and caretta (Linnaeus). Copeia 1952:1 15. mollusks that constitute the bulk of its diet. However, the Pere, S.J. AND WINN, B. 1998. Leatherback turtle (Dermochelys majority of these animals have also been documented to coriacea) strandings in Georgia: 1982-1996.In: Epperly, S.P., and construct shallow burrows within the bottom sediment (Fox Braun, J. (Eds.). Proceedings of the Seventeenth Annual SeaTunle and Ruppert, 1985). Additionally, the occurrence of deeper Symposium. NOAA Technical Memorandum NMFS-SEFSC- burrowing shrimp and anemones within the loggerhead's 415, pp.245-246. PI-orzuN, P.T., WIcrstEtN, M.K., AND Atvtos, A.F. 1993. Feeding diet also suggest that digging may be performed to obtain ecology of the loggerhead sea turtle Caretta caretta in the north- prey items. Future studies are needed to address diet selec- western Gulf of Mexico. Marine Biology I 15: I -5. tion and foraging by loggerheads during the months outside PneEN, A.R. 1996. Infaunal mining: a novel foraging method of of the summertime in Geor gLa. Moreover, studies are needed loggerhead turtles. Journal of Herpetology 30:94-96. which address diet selection in all adult sea turtles occurring RucrnESCHEL, C.A. AND SHoot, C.R. 1988. Gut contents of logger- in Georgia. heads: findings, problems, and new questions. In: Schroeder, B.A. (compiler). Proceedings of the Eighth Annual Workshop on Sea Acknowledgments. We are grateful to Karen A. Turtle Conservation and Biology. NOAA Technical Memoran- Bjorndal and two anonymous reviewers for providing com- dum NMFS-SEFC-214, pp. 91-98. Ruppenr, E.E. AND Fox, R.S. 1993. Seashore Animals of the ments that greatly improved the original manuscript. We South- east. University of South Carolina Press, Columbta,,429 pp. would also like to thank the following individuals and SHoor, C.R., RucroESCHEL, C.A., AND KENNev, R.D. 1998. Female- institutions their support: Randy Isbister, Charles for biased sex ratio of loggerhead sea turtles in Georgia. Chelonian Warnock, David C. Veljacic, John Robinette, Deb Barnard, Conservation and Biology 3:93-96. Sam Drake, Peter Range, Robert A. Moulis, Mark Dodd, VnN NrEnor, M.M. AND DEN Hnnroc, J.C. 1984. A study on the gut Adam MacKinnon, Barb Zoodsma, Jim Richardson, Rebecca contents of five juvenile loggerhead turtles, Caretta caretta Bell, Rebecca Clark, Anne Frick, John Crawford, the (Linnaeus) (Reptilia, Cheloniidae), from the southeastern part of Courtney Knight-Gaines Foundation, the Pentecost Ecol- the north Atlantic Ocean, with emphasis on coelenterate identifi- ogy Trust Fund-Savannah Presbytery, the Shared Earth cation. Zoologtsche Mededelingen 59(): pp.35-54. Wrrrnus, A.B . 1984. Shrimps, Lobsters, and Crabs the Atlantic Foundation, the PADI Foundation, and the Turner Founda- of Coast of the Eastern United States, Maine to Florida. tion. Smithsonian Institute Press, Washington, 550 pp. WoLre, R.E. AND GEoncE, A. 1981. Sea turtle necropsy manual. LlrrnaruRE Crrpo NOAA Technical Memorandum NMFS-SEFSC -24:l-24.

BronNneI-, K.A. 1997. Foraging ecology and nutrition in sea turtles. Received: 24 January 2000 In: Lutz, P.L. and Musick, J.A. (Eds.). The Biology of Sea Turrles. Reviewed: 16 July 2001 CRC Press, Boca Raton, pp. 199-232. Revised and Accepted: 6 September 2001 CnrnEn, D.R. 1917. Guide to common jellyfishes of South Carolina.