Disarticulation of turtle shells in north-central Florida: how long does a shell remain in the woods?
When a vertebrate animal dies, the most obvious record of its existence is to be found in the bones that it leaves behind. Skeletal remains allow researchers to determine phylogenetic history by providing insights into morphological evolution (Carroll, 1988). Characteristics of fossils, such as the number of blood vessels supplying the bone, the type and number of offensive and defensive structures, and the size and positioning of eyes and olfactory chambers, give important clues to the physiology, ecology and behavior of extinct species. By studying suites of fossils, community profiles can be developed allowing speculation into relative ratios of predators and prey, biomass and social structure.
Although the importance of fossils is recognized and the process of fossilization is well-understood, the idea that the position of the fossil can tell important facts about the animal or its environment has been recognized only fairly recently (Olson, 1980). The science of taphonomy (literally, 'the laws of burial,' Olson, 1980) examines the positioning and decomposition of carcasses immediately after death. Taphonomic studies provide insights into paleoecology (Behrensmeyer, 1978; Pratt, 1986), the cause of death, and how long bones are exposed before burial. These data are useful both in studies of fossil and subrecent remains as well as in anthropological and ethnoarcheological studies (Behrensmeyer and Hill, 1980).
Turtles (Reptilia: Testudines) would seem to be good subjects for taphonomic studies. Their shells are durable and easily observed, both in recent and fossil deposits, and turtles are relatively abundant in certain fossil beds. However, there is little information in the literature on factors affecting burial and fossilization of turtles with the exception of the classic studies of Weigelt (1927). Weigelt (1927) studied the positioning of animal carcasses from both recent catastrophes and fossil deposits, and included observations on turtles killed during a severe freeze in Texas. He did not examine disarticulation rates and processes, however. Meyer (1991) studied the disarticulation patterns of two hawksbill sea turtles (Eretmochelys imbricata) buried in a subtidal lagoon, and noted that disarticulation can occur in from a few days to a few weeks.
The only published study of terrestrial turtle shell disarticulation is that of Bourn and Coe (1979) who studied disarticulation in 31 Dipsochelys elephantina (= Geochelone gigantea) shells on Aldabra Atoll over an 18-mo period. They divided shell carcasses into eight stages, and conservatively estimated that larger shells persisted ca. 2 yr until complete disintegration. K. H. Berry (pers. comm.) used five stages of scute condition and seven stages of nonlimb bone condition to estimate carcass age in an unpublished study of 42 desert tortoise (Gopherus agassizii) shells. Depending on size, some carcasses were estimated to remain up to 8 yr before disarticulation. Both Bourn and Coe (1979) and Berry found substantial individual variation in disarticulation rates.
In the present study, turtle shells often were observed far from the nearest lake or wetland during several years of extensive herpetological inventory and field research on the Katharine Ordway Preserve, Putnam County, Florida. Shells of three species [Gopherus polyphemus, Pseudemys nelsoni, P. floridana (= P. peninsularis of Seidel, 1994)] were found most often. Gopherus polyphemus is a terrestrial species, whereas the other species (including Apalone ferox, Deirochelys reticularia and Trachemys scripta) are inhabitants of lakes and ponds. Shells were found singly and did not appear to result from any mass mortality event.
Finding turtle shells scattered through the N Florida sandhills led to several questions. Had the individual turtles died at the same time, or had mortality been extended over a period of time? Did the different species die at the same or different times? In any case, could the stage of disarticulation at which a turtle shell was discovered be used to determine its approximate time of death and, if so, could the cause of death be identified? In this study, I describe and quantify the patterns and rates of turtle shell disarticulation.
Shells of six turtle species were collected during herpetofaunal fieldwork on the 3750-ha Katharine Ordway Preserve/Swisher Memorial Sanctuary, Putnam County, Florida (29 [degrees] 41 [minutes] N, 82 [degrees] 00 [minute] w). A few additional shells from nearby areas were provided by colleagues. All turtles used in these observations were found dead; no live specimens were sacrificed. Initial disarticulation stages (Table 1) were as follows: five in stage 1, 23 in stage 2, 17 in stage 3, 17 in stage 4, nine in stage 5, six in stage 6, two in stage 7. Species examined included Apalone ferox (n = 2), Deirochelys reticularia (n = 3), Gopherus polyphemus (n = 34), Pseudemys nelsoni (n = 12), P. floridana (n = 27) and Trachemys scripta (n = 2). Before the start of observations, carapace length (CL) was measured with calipers to the nearest mm.
Shells were arranged approximately 1 m apart in linear rows in a sandhill habitat adjacent to Breezeway Pond (see Dodd, 1992, for a description of this site). The rows were parallel, approximately 2 m apart. All shells were complete, i.e., they consisted of a carapace, plastron and associated bridge. Shells were placed carapace up so that they rested flat on the plastron. The substrate was bare white sand interspersed by small clumps of maidencane (Panicum sp.). Shells were exposed to direct sunlight for at least the first two-thirds of the day. A flag with the turtle's identification number etched on an aluminum tag was placed next to each shell.
Shells were photographed at the beginning of each month from the time that they were placed in the row until the shell completely disarticulated and the bones became widely scattered. Photographs were taken perpendicularly to the shell at a distance of approximately 75 cm. When a shell completely disarticulated, the remaining bones were removed from the area to minimize disturbance to adjacent shells. Observations began in December 1985 and terminated in October 1991. The longest time a single shell was monitored was 54 mo.
I initially examined the photographs of 39 shells and wrote detailed descriptions of the sequences in which the bones of the shell, both carapace and plastron, fell apart. Based on these descriptions and later refinements during classification, I recognized nine stages of [TABULAR DATA FOR TABLE 1 OMITTED] disarticulation (Table 1, [ILLUSTRATION FOR FIGURE 1 OMITTED]). Each of the 80 shells was then classified monthly to disarticulation stage from its photographic series. A computer spread sheet was used to track the monthly disarticulation stage and, later, these data from the sheet were used to diagram individual and species-specific patterns of disarticulation.
Inasmuch as the shells varied in disarticulation stage at the beginning of the observations (termed stage), I only used observations recorded after the shell went from [stage.sub.i+1] to [stage.sub.i+2] to compute temporal aspects of disarticulation. Hence, observations on the amount of time a shell took to go from one stage to the next are not biased by the initial stage of disarticulation. Scute and bone nomenclature follows Ernst et al. (1994). The analysis of descriptive statistics was performed using the SAS program for microcomputers (SAS Institute, 1988).
RESULTS AND DISCUSSION
Both juvenile and adult shells were examined and encompassed a wide range of sizes (Table 2). Carapaces proved more useful than plastrons to examine temporal aspects of disarticulation. The relatively flat plastrons of these turtles remained articulated far longer than the domed carapaces and they were often intact well after the carapaces completely fell apart. Otherwise, disarticulation patterns of the rest of the shell varied among the three families of turtles (Emydidae, Testudinidae, Trionychidae), although the patterns within each family were similar. Therefore, disarticulation patterns are described by family.
TABLE 2. - Descriptive statistics for disarticulating turtle shells. CL = Carapace length (mm); MO = total time shells were monitored (in months)
Species Variable n Mean (Range)
Deirochelys reticularia CL 3 176.0 (134-198) MO 3 13.3 (5-18) Pseudemys floridana CL 25 256.9 (133-337) MO 27 29.0 (11-50) P. nelsoni CL 11 298.6 (148-379) MO 12 31.8 (11-45) Trachemys scripta CL 2 166.5 (146-187) MO 2 17.5 (12-23)
Gopherus polyphemus CL 30 222.8 (79-284) MO 34 25.4 (8-54)
Apalone ferox CL 2 265.0 (257-273) MO 2 16.5 (13-20)
Emydidae (Deirochelys reticularia, Pseudemys floridana, P. nelsoni, Trachemys scripta). - The scutes of emydids are initially shiny but quickly lose their luster. They detach completely by curling up at their margins giving the appearance of thin rolled-up scrolls. The scutes peel off individually, often for a long time. Underlying pigmented dermis often remains attached to the carapace giving it a blotchy pattern that appears dark or dirty [ILLUSTRATION FOR FIGURE 1 OMITTED]. Vertebrals and pleurals flake simultaneously, but the marginals are often the first group to detach completely. Shedding may be localized (for example, only on the left front side) or may occur simultaneously over the entire carapace.
After several weeks or months, small separations develop at the sutures between the carapacial bones, and the bones begin to take on a weathered appearance (Table 1, [ILLUSTRATION FOR FIGURE 1 OMITTED]). Sutures widen until eventually the neurals fall in and/or the costals separate and disarticulate from one another. Sometimes, a single large crack expands to the point where the carapace breaks into two large sections (anterior/posterior). These large sections break down independently of one another and often persist far longer than they would have if they had remained part of a single carapace.
There is no consistent pattern of shell disarticulation; i.e., the neural bones do not always disarticulate first, followed by the costals, or vice versa. Neither does disarticulation occur earlier on the right or left side of the shell. Neurals 2-6 appear most prone to disarticulation. Disarticulation also is prevalent at the sutures between costals 3/4, 4/5, 5/6 and 6/7. Posteriorly, the pygal, suprapygal, and adjacent peripheral bones commonly disarticulate quickly, although patterns vary greatly from one shell to the next. Peripheral bones appear to detach randomly. After stage 6 begins, bones may remain articulated for a long time (Table 3), and the duration of stages 6 and 7 is usually much longer than the earlier stages of disarticulation. Once stage 8 is reached, bones of the shell collapse rapidly. Bones were never observed to separate (i.e., break) other than at sutures.
TABLE 3. - Mean duration, in months, in the amount of time that it takes to progress from one shell disarticulation stage to the next disarticulation stage in gopher tortoises, Gopherus polyphemus, and emydid turtles, Pseudemys floridana and P. nelsoni, in N-central Florida
Disarticulation stages 2-3 3-4 4-5 5-6 6-7 7-8 8-9
n 8 13 17 21 18 16 18 Mean 2.9 3.5 3.8 5.1 7.4 11.8 5.8 SD 1.6 2.4 2.0 4.5 5.7 17.5 5.1
n 8 13 16 19 16 13 13 Mean 1.6 1.9 3.4 6.3 4.5 5.0 5.5 SD 1.1 0.8 3.3 5.4 5.6 5.6 5.9
n 3 6 7 7 6 6 6 Mean 1.3 6.8 4.6 2.3 5.8 7.2 6.8 SD 0.6 5.6 5.3 2.6 8.5 4.2 4.4
Testudinidae (Gopherus polyphemus). - In gopher tortoises, scutes are initially shiny but begin to dull after 1 or 2 mo. Scutes warp at their margins where they contact other scutes, and turn upwards. Eventually, the amount of shell surface area contacted shrinks and the scutes loosen and bow upwards in the center. The scutes are rigid and detach singly in their entirety; detached scutes may remain for a long time close to the carcass. Vertebral scutes usually detach first, followed by the pleurals. There appears to be no set pattern in which the vertebrals or pleurals detach, i.e., anteriorly or posteriorly. Marginal scutes generally remain attached for longer periods, although in some tortoises the scutes seem to detach quickly and in unison. Once scutes detach, there is no underlying connective tissue or pigmentation and the bone is uniformly white.
The pattern of gopher tortoise shell disarticulation is similar to that of emydids, except that the pygal and suprapygal bones usually disarticulate even faster than the neurals. The posterior part of the shell tends to collapse sooner than the anterior shell. There is less tendency for the shell to split into large sections which then remain intact. If any part of the shell remains intact, it is the anterior section. Peripherals disarticulate randomly.
Trionychidae (Apalone ferox). - Softshell turtles have a leathery skin covering their shell rather than the whole horny scutes of other turtles. When a softshell turtle dies, its skin dries, shrinks, and becomes extremely tough. As the skin dehydrates, tension is placed on the margin of the shell, which lacks peripheral and pygal bones and is where the skin is thickest, causing the carapace to bow upward and rupture along costal and neural sutures. This process occurs quickly, hence, nearly all softshell turtle shells are past stage 6 of disarticulation when encountered. Bones of the carapace quickly separate and are easily scattered unless held close to one another by the mummified skin. Softshell turtles do not pass through a linear sequence of disarticulation stages as outlined in Table 1. Based on present results and unquantified observations of softshell carcasses on the Ordway Preserve, most carcasses likely reach disarticulation stage 9 in under 12 mo.
Temporal aspects of turtle shell disarticulation. - Shells can be categorized fairly easily into nine stages described in Table 1. However, there can be a great deal of individual variation in the rate of disarticulation [ILLUSTRATION FOR FIGURE 2 OMITTED]. Thus, categorizing a shell to a disarticulation stage does not provide a precise estimate for the total amount of time since the turtle's death, especially in the latter stages of disarticulation. Although it is relatively simple to compute a line graph depicting mean temporal aspects of shell disarticulation [ILLUSTRATION FOR FIGURE 3 OMITTED], the results must be used cautiously.
Perhaps the most important shell disarticulation stages are stages 1, 5-6, and 9. A turtle in stage 1 has died very recently, although the small sample size precluded determination of specific differences in exactly how long a carcass remained in that stage. For the five very fresh carcasses (mixed species), each was quickly disturbed by scavengers and decomposers and all nonbone tissue (excluding dried and hardened skin and connective tissue) disappeared in less than a week. Small fresh carcasses quickly disarticulated and the bones were scattered usually in less than a few weeks.
The transition between stages 5 and 6 is important because until this time, the shell is entire and usually easily observed under field conditions. Before this stage, there is relatively little variation in the amount of time between disarticulation stages (Table 3). Examination of individual disarticulation patterns suggests that the shift from stages 5 to 6 is reached at 22 mo in adult Gopherus polyphemus, under 30 mo in adult Pseudemys nelsoni, [ILLUSTRATION FOR FIGURE 2 OMITTED], and from 12-15 mo in adult P. floridana, despite the impression provided by the cumulative line graph [ILLUSTRATION FOR FIGURE 3 OMITTED]. Entire shells of these species remained intact a minimum of 12-30 too, on average, depending on species. Except in P. nelsoni, standard deviations increased after stage 5-6 (Table 3), which suggests greater temporal variation within disarticulation stages after the shells began to collapse. The thick robust shells of P. nelsoni probably made them less prone to initial weathering than the thinner shells of the other species.
Stage 9 is important because at this stage the bones are completely disarticulated and are rapidly scattered throughout the surrounding litter. Stage 9 is usually reached by ca. 30 mo in adult emydids, and by 40 mo in adult tortoises [ILLUSTRATION FOR FIGURE 3 OMITTED], although there are notable exceptions ([ILLUSTRATION FOR FIGURE 2 OMITTED] for examples among Pseudemys nelsoni). Examination of individual disarticulation patterns, however, shows that stage 9 may be reached as quickly as 10 mo. Because they are thick, P. nelsoni shells persist longer than the other emydids and disarticulate completely after approximately 35 mo.
Factors affecting disarticulation. - Many factors influence the disarticulation and scattering of bones (Weigelt, 1927; Behrensmeyer, 1978). Scavengers (rodents, small carnivores, birds) gnaw or break bones apart or bore into shells, bones may be trampled by ungulates, and weather (exposure, humidity, rainfall, drought) likely influences rates of disarticulation (K. Berry, unpubl. data). In the present study, none of the shells showed signs of rodent gnawing. Inasmuch as all shells were placed close to one another, it is unlikely that weather or substrate conditions affected disarticulation differentially. The similarity under which these observations were conducted underscores the importance of individual patterns of variation in disarticulation rates.
In addition to external factors, species-specific shell morphology probably exerts constraints on shell cohesion. One might expect similar disarticulation rates between the tortoises Gopherus polyphemus and G. agassizii because of similar shell shape (Germano, 1993). However, K. Berry (pers. comm.) found that large desert tortoise (G. agassizii) shells may last well beyond 4 yr under natural conditions. The longer duration to disarticulation in desert tortoises is undoubtedly related to dry desert conditions.
In Florida, gopher tortoise shells likely remain intact longer than the emydids because of undetermined differences in morphology which make them more mechanically resistant to weathering factors. Florida red-bellied turtles (Pseudemys nelsoni) are more resistant to disarticulation than peninsula cooters (Pseudemys floridana) probably because their high [equivalent to] domed shells are much thicker, possibly an adaptation against alligator attacks (Jackson, 1989). Small shells of all species break down rapidly, presumably due to their delicate and incompletely ossified shells.
The use of disarticulation stages to estimate time since death may be valuable in cases where large numbers of turtles die over a short period of time. Mass mortality may occur as wetlands dry (D. Jackson, pers. comm.; C. Dodd and R. Franz, pers. observ.) and are exposed to sudden freezes (Weigelt, 1927), as nesting turtles are caught away from wetlands during sudden cold spells (e.g., some freshwater turtles in Florida nest during the winter, including Deirochelys reticularia and Pseudemys floridana; Jackson, 1988), as the result of a disease epidemic (Jacobson et al., 1991), or for unknown reasons. In such cases, comparative skeletal material used with general estimates of rates of disarticulation may allow a researcher to identify a possible cause and/or its timing. For the majority of shells encountered, the use of disarticulation stages will provide only general estimates of time since death. Estimates of elapsed time to a particular disarticulation stage also must be determined for a variety of species at different localities to adjust for regional environmental differences. In short, there is no easy answer to how long turtle shells remain in the woods.
Acknowledgments. - I thank Kristin Berry for allowing me access to her unpublished data on desert tortoises, and the Katharine Ordway Preserve-Swisher Memorial Sanctuary for allowing me to conduct these observations on the Preserve. Richard Franz, Dale Jackson and George Zug provided thoughtful comments on the manuscript.
BEHRENSMEYER, A. K. 1978. Taphonomic and ecologic information from bone weathering. Paleobiology, 4:150-162.
----- AND A. P. HILL. 1980. Fossils in the making. Vertebrate taphonomy and paleoecology. Univ. Chicago Press, Chicago. 338 p.
Bourn, D. And M.J. Coe. 1979. Features of tortoise mortality and decomposition on Aldabra. Philos. Trans. R. Soc. Lond. Ser. B. Biol. Sci., 286:189-193.
Carroll, R. L. 1988. Vertebrate paleontology and evolution. W. H. Freeman, New York. 698 p.
DODD, C. K., JR. 1992. Biological diversity of a temporary pond herpetofauna in north Florida sandhills. Biodiversity Conserv., 1:125-142.
ERNST, C. H., R. W. BARBOUR and J. E. Lovich. 1994. Turtles of the United States and Canada. Smithsonian Inst. Press, Washington. 578 p.
Germano, D.J. 1993. Shell morphology of North American tortoises. Am. Midl. Nat., 129:319-335.
JACKSON, D. R. 1988. Reproductive strategies of sympatric freshwater emydid turtles in northern peninsular Florida. Bull. Fla. State Mus. Biol. Sci., 33:113-158.
-----. 1989. Turtle's use of alligator nests, p. 145. In: C. A. Ross (ed.). Crocodiles and alligators. Facts on File, New York.
JACOBSON, E. R., J. M. GASKIN, M. B. BROWN, R. K. HARRIS, C. H. GARDINER, J. L. LAPOINTE, H. P. ADAMS and C. REGGIARDO. 1991. Chronic upper respiratory tract disease of free-ranging desert tortoises (Xerobates agassizii). J. Wildl. Dis., 27:296-316.
MEYER, C. A. 1991. Burial experiments with marine turtle carcasses and their paleoecological significance. Palaios, 6:89-96.
OLSON, E. C. 1980. Taphonomy: its history and role in community evolution, p. 5-19. In: A. K. Behrensmeyer and A. P. Hill (eds.). Fossils in the making. Vertebrate taphonomy and paleoecology. Univ. Chicago Press, Chicago.
PRATT, A. E. 1986. The taphonomy and paleoecology. of the Thomas Farm local fauna (Miocene, Hemingfordian), Gilchrist County, Florida. Unpubl. Ph.D. Dissertation, Univ. Florida, Gainesville. 487 p.
SAS INSTITUTE, INC. 1988. SAS/STAT user's guide. Release 6.03 ed. SAS Institute, Inc., Cary, N.C. 1028 p.
SEIDEL, M. E. 1994. Morphometric analysis and taxonomy of cooter and red-bellied turtles in the North American genus Pseudemys (Emydidae). Chelonian Conserv. Biol., 1:117-130.
WEIGELT, J. 1927. Rezente Wirbelterleichen und ihre palaobiologische Bedeutung. Verlag von Max Weg, Leipzig. (1989 English translation published by Univ. Chicago Press, Chicago. 188 p.)
|Printer friendly Cite/link Email Feedback|
|Author:||Dodd, C. Kenneth, Jr.|
|Publication:||The American Midland Naturalist|
|Date:||Oct 1, 1995|
|Previous Article:||Reproductive cycles and temporal variation in fecundity in the black swamp snake, Seminatrix pygaea.|
|Next Article:||Distribution of Van Dyke's salamander (Plethodon vandykei van Denburgh).|