Anatomy of a Turtle Confiscation: What to Expect When You’re Not Expecting and Recommendations for Improving Outcomes

Approximately one-quarter of all terrestrial vertebrate species are traded in the global wildlife market, with millions of animals sold each year (Scheffers et al. 2019). Much of that trade is unregulated or underregulated, with only a small proportion of species protected by the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES; Auliya et al. 2016; Marshall et al. 2020). Even trade of CITES-listed species has increased, quadrupling from 25 million individuals in 1975 to 100 million individuals in 2014 (Harfoot et al. 2018). The wildlife trade is recognized as one of the biggest contributors to biodiversity loss and species extinction risk worldwide (Maxwell et al. 2016), largely because many individual animals are wild caught rather than captive bred. For example, it is estimated that approximately 90% of the reptile species and half of all individuals traded originated in the wild (Marshall et al. 2020).

Turtles have been hit especially hard by the live wildlife trade. Over a 20-yr period, 2 million wild-caught turtles were legally traded (Luiselli et al. 2016). During 2000–2019, 320 of the world’s ∼350–360 turtle species were detected in the global wildlife market (Marshall et al. 2020; Stanford et al. 2020). In addition, turtles are more likely than other taxa to be sourced from wild populations, because wild-caught turtles often sell for more than captive-bred conspecifics and long generation times require long-term investments into captive stock (Sung and Fong 2018; Marshall et al. 2020). Turtles are already among the most threatened vertebrate taxa, with > 60% of the 274 species that have been evaluated by the International Union for Conservation of Nature (IUCN) being considered threatened with extinction (IUCN 2024), primarily by over-exploitation (Stanford et al. 2020). Furthermore, their “slow” life history—i.e., delayed reproductive maturity, low egg and juvenile survivorship, and high adult survivorship—makes their populations vulnerable to chronic losses of adults (Congdon et al. 1993, 1994; Spencer et al. 2017). It has yet to be demonstrated that turtle populations can be harvested sustainably (Sigouin et al. 2017; CEC 2019), making each individual (especially adults) valuable for population persistence.

Growing awareness of the plight of turtles has stimulated stronger legislation and greater enforcement of wildlife laws by wildlife agencies. The increased enforcement has resulted in an escalation of confiscations, necessitating a coordinated and strategic approach to managing these live, but often moribund, animals (D’Cruze and Macdonald 2016; Gray et al. 2017; Rivera et al. 2021). Confiscated animals are typically either euthanized, maintained in long-term captivity, or—less commonly—released back into the wild. The IUCN has historically viewed euthanasia as preferable (because it is the cheapest and least risky option) and a return to the wild as the least desirable option, primarily because of disease risks to wild populations or animals being of unknown genetic origin (IUCN 2002). More recently, the IUCN (2019) has acknowledged that the growing number of animals being removed from the wild is placing populations at risk of extinction. As such, managing live animals confiscated from the trade will be critical for conserving global biodiversity, and best practices are needed to determine how confiscated live animals can best contribute to conservation (IUCN 2019). Given the finite capacity for permanent captivity of confiscated animals, understanding the suitability of confiscated animals for release into the wild could alleviate strain on captive facilities (Gray et al. 2017; Rivera et al. 2021). A better understanding of the fate of confiscated animals, particularly those released into the wild, would also facilitate informed decision making and potentially improve the outcome of future confiscations, but few such data exist (D’Cruze and Macdonald 2016; Gray et al. 2017; Rivera et al. 2021).

We present as a case study of a confiscation of > 200 illegally traded eastern box turtles (Terrapene carolina carolina) and our efforts to repatriate those turtles back to the wild. The eastern box turtle is currently thought to be the US turtle species most heavily traded in the global illegal wildlife market (Easter et al. 2023) and is frequently seized during confiscations. For each phase of the project from the time turtles were seized to release into the wild (Fig. 1), we describe the challenges we experienced and the lessons learned. Our objectives were to 1) document the outcomes of the confiscation to serve as a baseline for subsequent postrelease monitoring of these turtles, 2) communicate our experiences so those likely to be involved in handling confiscated turtles have a better idea of what to expect, and 3) identify ways to improve the health and survival outcomes of turtles confiscated in the future.

On 14 August 2019, South Carolina Department of Natural Resources Law Enforcement (SCDNR LE) assisted federal investigators from the United States Fish and Wildlife Service to execute a search warrant at a private residence in Chester County, South Carolina, related to illegal reptile exports from the state (DOJ 2022). A herpetologist was engaged to identify the species and assist with any confiscated animals. Though many species of native (presumably wild-caught) and exotic herpetofauna were discovered, eastern box turtles made up the bulk of the collection.

Box turtles were being held outdoors in 15 round, black plastic 1136-L tubs in groups of 2–25 + per tub (Fig. 2a). Although most tubs were partially shaded by trees, local summer temperatures are routinely > 35°C. Turtles were not provided any substrate, and there was no evidence that they were being fed or watered (Fig. 2b). Initial inspection revealed a total of 4 to 6 dead box turtles and many other underweight turtles; some adults had closed or puffy eyes and aural swellings. All live box turtles were seized, consolidated into 3 large tubs, and then moved to an offsite location for temporary holding by SCDNR LE. Mulch substrate was added to tubs and watered to rehydrate and cool the turtles, many of which drank immediately and copiously. On 15 August 2019, turtles were rehydrated again and given raw strawberries and mushrooms. Box turtles were transported on 16 August 2019 to the University of Georgia’s Savannah River Ecology Laboratory (SREL) in Aiken County, South Carolina, approximately 175 km away. No turtles died between 14 and 16 August 2019.

Our primary objectives in Phase 1 were to 1) coordinate with and directly assist law enforcement with species identification and 2) facilitate rapid transfer to a temporary holding facility. This confiscation represents the largest seizure of reptiles led by SCDNR LE, posing the first challenge to mounting a rapid response to this unexpected event. The close working relationship of agency LE with agency biologists helped expedite the response. Once on scene, agency biologists were able to assist with species identification, thereby supporting LE in determining which statutes may have been violated. Biologists and LE then developed a plan for the agency to temporarily hold animals until a longer-term holding facility could be identified. The ability of LE to also seize the tubs turtles were housed in simplified care while animals were held by LE and aided in transport to longer-term holding facilities.

SCDNR does not have facilities or holding areas that can house large numbers of confiscated or surrendered wildlife. SCDNR explored several options while permissions were being secured but encountered the following challenges: 1) few potential organizations with appropriate indoor and/or outdoor animal holding facilities with enough space to absorb additional animals, 2) recipient organizations and facilities not being able to fundraise because of concerns about releasing sensitive information regarding the legal case, 3) chain of custody requirements and restrictions, 4) uncertainty about timeline for release, 5) uncertainty about suitability of animals for release, 6) the potential need to provide permanent captive care should animals be deemed not releasable, and 7) ambiguity as to whether agency funds could be used to cover costs.

Having an established network of holding facilities available and willing to receive animals would facilitate a rapid response to future turtle confiscations. A network of facilities would additionally provide a framework for anticipating transportation logistics—also key to a rapid response. Agency LE personnel are more likely to initiate confiscations when they can be assured that animals will be quickly transferred to appropriate holding facilities. In addition, the quicker animals can be transferred to longer-term holding facilities, the more quickly turtles can be assessed in terms of health and receive specialized husbandry and veterinary care. Other ways to increase capacity include turtle identification guides or training workshops for law enforcement, particularly when agency biologists might not be able to respond immediately.

Immediately upon the turtles’ arrival at SREL, biologists from SCDNR and SREL started inventorying turtles and sorting obviously sick turtles from putatively healthy turtles. Each turtle was sexed, weighed to the nearest 2 g, measured (midline carapace length, MCL) to the nearest 1 mm, and assigned a permanent and unique notch ID. To facilitate monitoring and minimize handling, notch IDs were also painted on carapaces. Any abnormalities or injuries were noted, including missing limbs, shell damage, aural abscesses, swollen or puffy eyes, nasal discharge, or lethargy.

Of the 208 live box turtles received, 61 were adult females, 90 were adult males, and 57 were immatures. Turtles ranged in size from 72 to 162 mm MCL. The absence of smaller juvenile classes, the prevalence of shell wear and damage, and the frequency of individuals that were missing limbs (n = 11) all provided evidence that the turtles were likely collected from the wild. Body condition index (BCI) at intake, based on a published formula for box turtles (DePersio et al. 2019), ranged from 2.75 to 15.06 (mean = 9.30; n = 175). In comparison, on average, wild adult box turtles have an average BCI of 11.67 ± 2.34 (mean ± 1 SE), based on records of > 3500 individuals (M. Allender, pers. comm. 2023).

Our primary objective during Phase 2 was to inventory and triage confiscated turtles while minimizing handling time. One of our biggest challenges was balancing the desire to collect data with the need to quickly improve animal holding conditions. We prioritized assigning permanent, individual IDs, which were critical for subsequent monitoring of captive individuals. Given the above-mentioned signs of emaciation and dehydration, we expected that weight would be an informative health metric. Recording sex and size (MCL) were useful in confirming IDs at subsequent handlings. We elected to not collect any biological samples at intake, which would have greatly increased processing time. We also felt that sample collection and more invasive marking procedures would cause additional stress to animals that had already experienced chronic stress (Norton and Allender 2020).

The other major challenge was the lack of established criteria for triaging animals. Such criteria would have helped us 1) quickly identify the most suitable candidates for immediate release, 2) segregate ill animals needing specialized care (and a more thorough health assessment), and 3) set benchmarks for when euthanasia was the most humane disposition. The personnel involved in the initial triage of turtles had ample experience to identify abnormal behavior and some common clinical signs of poor health but not the expertise to evaluate their significance or potential causes. Initial triage efforts would have been greatly facilitated by easy access to a training manual or guide with a description of clinical symptoms for which to screen, supporting photos or other illustrations, and a decision tree for how to triage animals based on presence or severity of specific clinical signs. Ideally these materials would be aimed at turtle biologists without veterinary training and would focus on symptoms that could be visually detected during a brief, external physical examination. Valuable health assessment manuals (e.g., Wendland et al. 2009; USFWS 2019) have been developed for some species of turtles, and several references (e.g., Jacobson et al. 1999; Norton and Allender 2020) present criteria to identify nonreleasable turtles in the context of translocation and wildlife rehabilitation. To our knowledge, however, no general, readily accessible guide is available that could easily be implemented by non-veterinarian staff in the context of turtle confiscations, which are likely to include a wide variety of species as well as individuals that differ markedly in their initial health status.

After we inventoried and triaged turtles, we placed most turtles in a 10-m × 15-m outdoor predator-proof aviary (Fig. 3a) with an artificial pond and stream system (Fig. 3b, c) and abundant terrestrial vegetation along the margins for shade and shelter sites, which we supplemented with hay. We provided supplemental water via sprinklers as needed and supplemental food (strawberries, blueberries, Mazuri 5M21 tortoise diet, and 5M87 aquatic turtle diet [Mazuri Exotic Animal Nutrition, St. Louis, MO]) 3 or 4 times per week. Turtles could also forage for invertebrates in soil substrate. Many turtles immediately drank and began eating; some were even soaking and swimming across the pools of water by the time we finished placing turtles in the aviary (Fig. 3d).

We held extremely lethargic turtles (n = 20) separately in an indoor animal care facility where we could maintain temperatures at approximately 27°C. We held lethargic turtles in 190-L tubs (132 × 79 × 30.5 cm; Newell Rubbermaid, Atlanta, Georgia) in small groups of 5 or fewer per tub (Fig. 4), allowing us to monitor them more closely and to quarantine them from turtles we deemed healthy. Over each tub, we suspended a 100-W Zoo Med PowerSun mercury vapor UVB-UVA bulb and a 50-W Exo Terra Infrared Basking Spot (Rolf C. Hagen Group, Montreal, Canada) controlled by automatic timers. We soaked animals 1–2 times per week and provided supplemental food as described above.

We monitored turtles daily. Any animals in the aviary later discovered to be lethargic were transferred indoors for more intensive monitoring. Any animals that died in captivity were collected and recorded. The first 15 carcasses were refrigerated (but not frozen) and promptly submitted to the Southeastern Cooperative Wildlife Disease Study (SCWDS) at the University of Georgia for necropsy; the rest were frozen for future potential study. Of 208 box turtles transferred to SREL, 188 (90.4%) were placed in the outdoor aviary and 20 (9.6%) in the indoor facility, although more individuals would likely have been placed indoors had there been space and personnel capacity. Turtles were maintained in captivity until early November while a release site was being identified and secured. A total of 53 turtles (25.5% of all turtles received) died during this phase, including all 20 that were held indoors. Immature turtles experienced higher mortality (n = 23; 40.4%) than adults (males: n = 15, 16.7%; females: n = 15, 24.6%). Surviving turtles gained an average of 10.7 g (range: –60 g to +114 g) during the recovery period, corresponding to an average gain of 4.8% (range: –15.6% to +40.7%) over their intake weight. The BCI of surviving animals ranged from 3.55 to 15.81 (mean = 9.50).

Although our staff had basic husbandry expertise, none had veterinary training to identify and manage critical cases. Training or consultations in basic supportive care such as administering fluids, nutritional support, or antibiotics (when warranted) might have improved conditions experienced by turtles, potentially increasing their survival during the recovery phase. Supportive care might also have resulted in turtles being released in better condition, thereby improving outcomes during subsequent phases (Raphael 2019; Innis et al. 2022). Although inability to access veterinary care was one of the greatest challenges we experienced, we were able to use the necropsy and diagnostic services of SCWDS for a subset of turtles that died early during the captive period to elucidate potential causes of death.

Our access to suitable indoor and outdoor holding facilities helped us quickly improve husbandry conditions. However, such facilities are not reliably available, which can be a major constraint in responding to confiscations that are unpredictable in timing and scope. In our case, having access to the outdoor holding facilities only until the end of October required us to rapidly identify a release site and develop a release plan, which was further complicated by the approaching winter dormancy season for turtles in our region. The need to expedite the transfer of turtles out of short-term holding facilities precluded our giving turtles more time to recover. In this case, the turtles had been surrendered to SCDNR LE and did not require long-term holding during adjudication of the case. This allowed us to move directly to the release phase as time constraints for the holding facilities approached. Chain-of-custody issues and requirements for maintaining confiscated turtles during the legal process may pose additional challenges in some cases.

SCDNR explored several options for potential release sites but encountered similar problems as when trying to identify captive holding facilities, including concerns about the potential for disease transmission to resident wildlife. Furthermore, although the confiscated box turtles were presumed to have been collected from the wild in South Carolina, no range-wide genetic database was available against which to compare the confiscated turtles to confirm their origin. In addition, few large, relatively unfragmented protected sites are found within the state where the agency could be assured that released turtles would not be poached again, particularly if held in soft-release enclosures. Ultimately, the Savannah River Site (SRS) was chosen due to its large size, restricted access, designation as a National Environmental Research Park (Shearer and Frazer 1997), and affiliation with willing partners for developing a release plan and postrelease monitoring.

The SRS is an 800 km² United States Department of Energy reserve in Aiken, Barnwell, and Allendale counties in the Upper Coastal Plain of west‐central South Carolina, constituting 1% of the state. Only 10% of the SRS is developed for site operations, with the rest supporting a wide range of terrestrial and freshwater habitats (White and Gaines 2000). Since its establishment in the 1950s, the SRS has been closed to public access. The natural resources of SRS are managed by the United States Department of Agriculture Forest Service–Savannah River (USFS‐SR), and the SCDNR has onsite wildlife managers who work with USFS‐SR to manage the wildlife. Because of its long history, the ecology of the SRS is also well characterized (Kilgo and Blake 2005).

Using historical information on the distribution of reptiles and amphibians on the SRS, combined with information on forest type and planned future timber harvests, we identified several candidate release sites within the SRS. We selected 1 in the northwestern quadrant of the SRS along the Upper Three Runs and Tinker Creek drainages that occurs in a special protection zone and supports some of the highest quality hardwood forests on the SRS (Davis and Janacek 1997). It includes seepages, flatwoods-like pine forest, stream heads, stream swamps, stream levees, and moist to wet pine-hardwood upland forest transitions. Although the SRS has been the subject of extensive herpetological studies (Gibbons et al. 1997; Buhlmann et al. 2005), we had no baseline data on abundance, demography or health of resident eastern box turtles at the release site. Ranavirus, a pathogen of concern in turtles, had been previously detected in amphibians and an eastern mud turtle (Kinosternon subrubrum subrubrum) on the SRS (Winzeler et al. 2015; Love et al. 2016).

From 28 October to 3 November 2019, we collected turtles from the outdoor aviary, performed health assessments, and collected biological samples. We reweighed them and recorded MCL and carapace width. We collected blood (< 1% body weight; 1–1.5 ml) from the subcarapacial sinus (Hernandez-Divers et al. 2002). After making 3 blood smears for hemoparasite and leukocyte counts, quantifying lactate with a lactate meter (Nova Biomedical, Waltham, MA), and adding 3–5 drops to lysis buffer (100 mM Tris pH 8.0, 100 mM EDTA, 150 mM NaCl, 1% SDS) for genetic banking, we transferred approximately 75 µl of whole blood to a microhematocrit tube, which we centrifuged to quantify packed cell volume. We centrifuged any remaining whole blood, pipetted off plasma supernatant and distributed among 125-µl aliquots, and retained the red blood cell pellet as back-up genetic samples. Archived plasma samples are useful for various health assays (e.g., nutrition, immune status, hydration levels, see Table 1; Hamilton et al. 2016; Haskins et al. 2021) and for quantifying reproductive and stress hormones (Currylow et al. 2013; Candal 2021; Loope et al. 2022). Detailed health data will be reported elsewhere, but the methods are described here to indicate how we prioritized sample collection (Table 1).

We focused on collecting samples that we 1) had the expertise and experience to collect, 2) could analyze in house, or 3) knew were valuable and should be banked for future analysis as funds became available. In retrospect, we would have also used blood samples to quantify glucose in a point-of-care device (Perrault et al. 2018). Several personnel had experience collecting blood samples and performing some assays, and we had the necessary supplies and sample storage space. However, many assays or health assessment components required specialized expertise (e.g., hemogram analysis of blood films) or were prohibitively expensive to have analyzed by an external fee-for-service diagnostic lab, limiting the types of health-related data we were able to obtain in “real time.” Archived samples, although available for future analyses, can provide only a retrospective health assessment of the confiscated turtles.

The primary challenge we encountered during this phase was the ability to obtain pertinent health information in time to triage, treat, and make prudent and timely decisions regarding the disposition of turtles. Recommendations from veterinarians might have helped refine the types of data or samples we collected, providing a more comprehensive picture of individual turtle health. Consensus on the key health components would aid prioritization and facilitate comparison among confiscations. The greatest impediment to obtaining “real-time” data was access to funds to support sample analysis.

Because soft release has been shown to increase site fidelity in other terrestrial turtles (Tuberville et al. 2005; Tetzlaff et al. 2019), we installed a 1-ha soft-release pen in the upland hardwood forest within the Research Set-Aside Area (Fig. 5) on 31 October 2019. It was constructed of 91.4-cm-tall Department of Transportation–grade silt fencing, buried approximately 25 cm, and reinforced with hardwood stakes and UV-resistant cable ties.

From 1 to 3 November 2019, we placed 156 turtles inside the soft-release pen to allow them to acclimate to the release site and facilitate monitoring until turtles emerged from winter dormancy. But, first, we attached passive integrated transponders (PIT tags; AVID, Norco, CA) to the anterior carapace of all turtles using waterproof epoxy. We monitored the pen weekly to ensure the fencing remained intact and we conducted monthly transects of the pen interior searching for turtles on the surface. We recorded their notch ID, PIT tag ID, and GPS coordinates (± 3 m) and tucked turtles under leaf litter or other cover. We collected any turtle carcasses and froze them.

During 1 April–14 May 2020, after turtles had emerged from winter dormancy, we searched the pen multiple times to inventory all survivors. We recorded their notch and PIT IDs, weight to the nearest 1 g, GPS location, and behavior (alert or lethargic). We attached radio transmitters to 20 male and 20 female turtles to monitor their postrelease movements (reported elsewhere; Browning 2022). On 15 May 2020, we released turtles from the pen by opening eight 1-m gaps in the pen walls.

We confirmed 73 turtles (24 females, 33 males, 16 immatures) survived until spring emergence, representing 47.1% of animals placed into the pen and 35.1% of the total confiscated (Fig. 6a, b). We could not confirm the fate of 14 turtles that were not seen after the onset of winter dormancy. During the ∼ 7-mo penning, we documented 68 mortalities (20 females, 36 males, 12 immatures; Fig. 6a, b), corresponding to 43.9% of those placed in the pen and 32.7% of all turtles confiscated. We attributed 3 mortalities to predation; we found another 5 upside down but could not determine if they had been depredated or scavenged. We could not identify the cause of death for the other 60 individuals but did observe surface activity of some turtles throughout the winter. Translocated turtles often exhibit increased surface activity or movement immediately following release (Pille et al. 2018; Quinn et al. 2018; Richter et al. 2024), which could expose them to potentially lethal temperatures and/or signal an underlying infection (McGuire et al. 2014; Agha et al. 2017; Cozad et al. 2020). On average, turtles lost 8.7 g during winter dormancy (n = 58), corresponding to 2.8% of their predormancy body weight. Turtles exhibited high variation among individuals, with weight change during dormancy ranging from –66 g to +42g, or –17.2 to +9.7% of their predormancy body weight.

Our main challenge during soft release was that we secured release-site permission only in late October—very shortly before they would normally enter winter dormancy (i.e., as early as 1 November for populations within 25 km of our site; DeGregorio et al. 2017), giving them little time to find an appropriate dormancy location. The proximity of our research facility to the release site facilitated regular monitoring of the soft-release pen, but in cases where the distance is greater, frequent monitoring may not be practical.

We have reported on the fate of individuals in the largest confiscation of native turtles with which any of the authors had ever dealt. Thus, we neither knew what challenges to anticipate (Fig. 1) nor had a response network or established protocols to guide decision making. Consequently, our decisions were based on expedience, existing capacity, and the desire for a positive conservation outcome. Below we offer recommendations based on hindsight.

The need for expedience prompted actions that improved the immediate welfare of confiscated turtles and laid the groundwork for responding to future confiscations. Although it would have been preferable to have been able to analyze many of the biological samples we collected for triage or treatment purposes before releasing turtles, the choice to bank as many samples as we could feasibly obtain will allow us to retrospectively evaluate the initial health of confiscated turtles (Table 1; McKee et al. 2022). The use of soft-release pens—in addition to any potential positive outcomes on postrelease site fidelity—facilitated monitoring of turtles at the release site and could also be used as an additional in situ quarantine opportunity (McKee et al. 2022).

The need for expedience, however, also elicited decisions that, in retrospect, we might not make in future confiscations. For example, if holding facilities had been available for a longer period, we would likely have elected to delay transferring turtles to the release site until the following spring. The delay would have provided turtles with a longer recovery period and provided the opportunity to more critically evaluate health status of individual turtles before release. Although such a delay would have required resources to which we did not have access, longer recovery and rehabilitation would undoubtedly have improved the outcomes of this confiscation.

The box turtles in this case study had been exposed to an extended period of stressful and unhealthy conditions, including overcrowding, extreme temperatures, and limited or no access to food or water before being confiscated by SCDNR LE. Individual animal health and mortality rates while in captivity can vary widely among confiscations depending on conditions to which turtles have been exposed (Raphael 2019). During initial intake (Phase 2), our objective was to quickly triage turtles into a few rudimentary health categories (dead, obviously sick, apparently healthy) based on external symptoms while minimizing handling time. Our approach parallels that of veterinarians that have been involved in turtle confiscations or emergency care, who generally recommend delaying sample collection and specialized treatment until animals are stabilized (Norton 2005; Raphael 2019; Norton and Allender 2020). After triage, trained personnel can start to administer basic supportive care, such as fluid therapy, antibiotics, and supplemental nutrition (Norton 2005; Raphael 2019; Norton and Allender 2020).

Veterinarians trained in chelonian medicine can serve key roles in the medical management of confiscations, including performing comprehensive physical exams, running and interpreting diagnostic tests, developing individualized treatment plans, conducting disease surveillance, and making recommendations regarding the final disposition of individuals or groups of confiscated turtles (Raphael 2019; Norton and Allender 2020; Innis et al. 2022). Necropsies of animals that die early in a confiscation event can identify pathological processes that should be considered when treating surviving individuals (Raphael 2019; Innis et al. 2022). In the United States, state or federal partners may be eligible to submit carcasses to contracted diagnostic facilities, such as the Southeastern Cooperative Wildlife Disease Study at the University of Georgia (state partners) or the United States Geological Survey’s National Wildlife Health Center (federal partners). Veterinarian involvement can improve the welfare of individual turtles and their prognosis for release, as well as help assess and manage risks inherent with releasing confiscated turtles back to the wild.

Ultimately, the decision to release confiscated turtles will depend on a variety of factors that will vary across contexts, including the risk tolerance of involved partners, the origins and history of the confiscated turtles (if known), and what is known about the population size, genetic uniqueness, and health of potential recipient populations (Gray et al. 2017; Norton and Allender 2020). Health data obtained from confiscated turtles can be used to identify candidates suitable for release. However, for most turtle species, little is known regarding what constitutes “normal” ranges for healthy wild turtles for many health metrics. When known, these reference values can be used to distinguish between confiscated turtles that are healthy enough for release and debilitated individuals that either are not candidates for release or will require additional rehabilitation (Innis et al. 2022). In practice, there may be multiple time points at which some turtles could potentially be suitable for release, and release criteria may differ at different phases of the confiscation.

The biggest capacity bottlenecks we experienced were expertise in providing specialized care and the inability to serve as a longer-term holding facility. We were well versed in turtle ecology, rudimentary health assessments, and sample collection. These skills, if combined with a few key resources that were unavailable at the time, would have increased our capacity to manage the confiscation. First, we would have benefited from a manual aimed at providing biologists with biosecurity protocols, triage criteria, clinical signs to screen for, recommendations for priority samples and associated diagnostics, and instructions for providing basic supportive care. A manual (or training workshops)—broadly applicable across turtle species and placed in the context of confiscations—would be an extremely valuable resource. Likewise, the ability to consult remotely with a chelonian veterinarian—along the lines of the human telemedicine model—would have provided critical insights into the health of confiscated turtles and guidance for how to manage them. These options do not replace the value of on-site veterinarian expertise, which we lacked, but they would increase the capacity of individual facilities and allow veterinarians to focus on the most critical cases.

In addition to increasing capacity of individual organizations, the other key need is to establish a network of designated partners poised to respond quickly to confiscations (Rivera et al. 2021; Sevin et al. 2022). During a 2022 workshop on the illegal turtle trade in the United States, 80% of 78 poll respondents indicated that they did not have a plan in place to deal with confiscations of ≥ 50 turtles, and 40% shared that a confiscation of > 100 turtles would present a serious challenge (Sevin et al. 2022; Wixted and Christman 2022). Several initiatives, however, are working to fill gaps in the response network and to increase capacity. The Collaborative to Combat the Illegal Trade in Turtles (CCITT), formed in 2018, is an “interdisciplinary collaboration among agency biologists, law enforcement, legal professionals, academics and nongovernmental organizations” that share a common goal of addressing the illegal trade of North American turtle species (Sevin et al. 2022). In addition to focusing efforts on prevention and enforcement, CCITT serves as a clearinghouse of resources for partners responding to confiscations. Since the 2019 confiscation of > 200 box turtles that we experienced, the CCITT and its partner members have developed a draft confiscation response plan, secured funding for range-wide genetic surveys of commonly traded species to facilitate population assignment of confiscated turtles, pursued dedicated funding for pathogen screening and genetic testing of confiscated turtles, and worked to develop a coordinated response network (Sevin et al. 2022). These efforts should address many of the challenges we encountered (Fig. 1) and that are common to confiscation events.

To continue to improve confiscation management and decision-making, the details and outcomes (both positive and negative) of individual confiscations need to be reported. Such data are generally lacking (D’Cruze and Macdonald 2016; Rivera et al. 2021; Innis et al. 2022). In addition, some standardization of minimum data reported would facilitate comparison across confiscations. One of the major objectives of the Association of Zoos and Aquariums’ American Turtle SAFE (Saving Animals from Extinction) Program is to help identify the best pathways for confiscated turtles to contribute to conservation (Sevin et al. 2022). Toward that goal, the American Turtle SAFE Program helps track confiscation case studies and works with CCITT to communicate outcomes among their members and partners. As more data become available, they can be used to refine decision trees and adapt them to specific situations (Gray et al. 2017).

Many partners may be reluctant to release confiscated turtles back to the wild. In addition to justifiable concerns regarding the risk of pathogen transfer to wild populations, their reluctance stems from uncertainty regarding how well confiscated turtles will perform and whether they will ultimately contribute to recipient populations. Virtually no data are available on postrelease health, survivorship, site fidelity, or reproduction of confiscated turtles (Gray et al. 2017; Norton and Allender 2020). Waif gopher tortoises—those displaced from the wild and often of unknown origin and having spent at least some time in captivity—share many characteristics with confiscated turtles. A recent study by McKee et al. (2021) showed that waif tortoises exhibit long-term postrelease survivorship on par with wild tortoises. Similarly, Paterson et al. (2021) demonstrated that release of rehabilitated injured turtles can have positive population-level effects. These studies suggest that—provided the risk of pathogen transmission can be managed—healthy confiscated turtles can potentially contribute to the recovery of wild turtle populations. However, the lack of data on postrelease fate of confiscated turtles remains a major impediment to identifying the best conservation pathway for these turtles.

It is important to manage expectations regarding the outcomes of individual confiscations. Typically, the benchmark for “success” for translocations is that translocated animals behave and survive similarly to their wild counterparts, barring any initial “release costs” (e.g., Tuberville et al. 2008; Bertolero et al. 2018). We argue that this expectation is both unrealistic and inappropriate for confiscated turtles, which—in addition to experiencing the normal sources of stress associated with translocation (Teixeira et al. 2007)—have in most cases also been subjected to long periods of extremely stressful conditions before being seized. In contrast, we contend that a better measure of conservation outcome is comparison with the cost of not intervening, which is the complete demographic and genetic loss of those individuals from wild populations. Thus, success should be measured by how many confiscated turtles (or their offspring) are responsibly and successfully returned to the wild. Given the sheer number of turtles being removed from wild populations as part of the illegal wildlife trade, every healthy turtle recovered via confiscation needs to contribute to conservation in some way. Therefore, the need is growing to evaluate the suitability of confiscated animals being returned to the wild and to develop protocols for reasonably managing risk while also making such repatriations practical and affordable to implement.

We are indebted to the law enforcement officers with South Carolina Department of Natural Resources (SCDNR) who initiated the confiscation and assisted with the initial holding of turtles and their transport to the Savannah River Ecology Laboratory. We thank L. Lee for helping secure site use permits, L. McCallie for assistance with animal handling and care, H. Forehand for providing food for turtles while in captivity, and a South Carolina Institute of Archaeology and Anthropology archaeologist who expedited their surveys to facilitate the release. The USDA Forest Service (USFS-SR) was instrumental in the selection of release sites, assisting with site use and coordination of forest management activities nearby, and installation of the soft-release pen. SCDNR provided materials for the soft-release pen. The Southeastern Cooperative Wildlife Disease Study provided valuable necropsy services. We thank Chris Shepherd for leading this special volume. Sarah Stoner and 1 anonymous reviewer provided helpful comments on previous versions of this manuscript. This manuscript is based on work supported by the Department of Energy Office of Environmental Management under Award Number DE-EM0005228. All methods followed protocols approved by the University of Georgia Institutional Animal Care and Use Committee (A2019-12-007) and SCDNR (SC-08-2019, SC-08-2020).