Understanding predator–prey relationships is important when undertaking translocations because translocations concentrate animals in relatively small areas, causing an increase in local abundance, attracting predators, and occasionally resulting in intense predation (Esque et al. 2010; Cypher et al. 2018). Additionally, captive-born offspring are often naïve and more likely to fall victim to predators, resident predators may exhibit surplus killing behavior upon encountering easily captured naïve prey, and translocated individuals are disoriented when released into unfamiliar environments, making them more vulnerable to predators (Letty et al. 2007; Cypher et al. 2018). Because even a few predators can decimate populations of naïve prey (Kruuk 1972), predation has the potential to influence the outcome of wildlife translocations (Short et al. 1992; Letty et al. 2007).

The Burmese star tortoise (Geochelone platynota) is a critically endangered tortoise endemic to the semiarid Dry Zone of central Myanmar (Platt et al. 2011). Historical sources suggest that G. platynota was once abundant within this region (Blyth 1863; Theobald 1868), where it occurred in a variety of habitats (Platt et al. 2011). Chronic subsistence harvesting by rural Burmese, coupled with habitat loss to agriculture, resulted in long-term albeit gradual population declines that suddenly accelerated in the 1990s when poaching became rampant in response to demand from the burgeoning transboundary wildlife trade with China (Behler 1997; Platt et al. 2011). By the early 2000s, viable populations could no longer be found even within protected areas, leading Platt et al. (2011) to conclude that G. platynota was functionally extinct in the wild. A conservation breeding program launched in the mid-2000s staved off biological extinction (Platt et al. 2017), and translocations of captive-bred and head-started G. platynota are currently underway at 2 wildlife sanctuaries in the Dry Zone (Platt and Platt 2019). Successful restoration of G. platynota as an ecologically functional member of Dry Zone ecosystems ultimately depends on the survival of translocated tortoises (Platt et al. 2014). Here, we document predation of translocated G. platynota by Eurasian golden jackals (Canis aureus; jackal hereafter) and wild pigs (Sus scrofa) at a wildlife sanctuary in central Myanmar.

Study Area. — Our observations were made at Shwe Settaw Wildlife Sanctuary (SSWS; 20°12′N, 94°33′E), a 37,500-ha protected area in the Magway Region of central Myanmar (Beffasti and Galanti 2011) established in 1940 for the protection of the endemic Eld's deer (Recervus eldii) (Bowler et al. 2019). SSWS is characterized by a largely intact, relatively open, dry deciduous forest with a dense grass understory (Platt et al. 2001; Bowler et al. 2019). SSWS encompasses one of the largest tracts of dry deciduous forest remaining in Myanmar (Beffasti and Galanti 2011). Mean annual rainfall in this region is approximately 900 mm with most precipitation falling between late May and October, followed by a prolonged dry season extending from early November through May (Platt et al. 2001). The sanctuary is surrounded by 39 villages with 26,000 inhabitants, and the adjacent lands are a mosaic of village agricultural fields, scrub, and small forest patches (Bowler et al. 2019). The physical and cultural environment of SSWS is described in greater detail elsewhere (Platt et al. 2001; Bowler et al. 2019). Geochelone platynota occurred in SSWS until the early 2000s when populations were extirpated by intensive poaching (Platt et al. 2001, 2011). Given on-going security concerns regarding the potential theft and poaching of head-started tortoises (e.g., Platt 2016), specific locations within SSWS are not disclosed in this article.

Tortoise Translocations. — We began translocating captive-bred and head-started G. platynota to SSWS during late 2016 (Platt 2016) in accordance with a national conservation action plan (Platt et al. 2014). Our translocation protocols are described in greater detail elsewhere (Platt and Platt 2020); however, to briefly summarize, we source 3- to 6-yr-old head-started tortoises (carapace length [CL] ∼ 120–200 mm) from 3 captive-breeding centers in Myanmar (Platt et al. 2017) and use a soft-release strategy (e.g., Tuberville et al. 2005); i.e., tortoises are confined for 12 mo in 1-ha acclimation pens at the release site before being liberated. We attach a very high frequency (VHF) radio transmitter (Ri-2B; Holohil Systems Ltd., Carp, Ontario, Canada) to 20–30 tortoises in each translocation cohort and follow these individuals for the life of the transmitter batteries (∼ 24 mo). From 2016 to 2018, we successfully released 1038 head-started G. platynota into the wild. On 19 February 2019, we transferred an additional 650 captive-bred tortoises into 3 acclimation pens with the intention of releasing this cohort in late January 2020. The release cohort consisted of tortoises aged 3–5 yrs old with a mean (± 1 SD) CL of 152 ± 14 mm (range = 115–198 mm).

Jackal Predation. — While radio-tracking tortoises in February and March 2019, we found the fresh remains of 28 G. platynota killed by predators. Most tortoise remains consisted of disarticulated pieces of the carapace and plastron, with the muscle and viscera largely consumed by the predators. We found VHF transmitters at 16 kill sites, allowing us to identify individual tortoises (8 males and 8 females), all of which were released from acclimation pens on 19 January 2018. The mean (± 1 SD) CL of these 16 tortoises immediately prior to release was 168 ± 16 mm (range = 142–203 mm), although additional growth undoubtedly occurred during the 13–14 mo these tortoises were living in the wild. The remains of 12 other tortoises that we found during this period could not be assigned to particular individuals. However, the proximity (< 1 km) of these kill sites to the acclimation pens suggest these tortoises were likewise among those released in 2018.

We found canid tracks at most kill sites and the bite marks on tortoise remains were consistent with descriptions of canid predation (Causey and Cude 1978), strongly suggesting these tortoises were killed by jackals, the only canid known to occur within SSWS (Linnell et al. 2014). Moreover, an automated game camera (Reconyx PC800 Hyperfire Professional; Reconyx, Holmen, WI) deployed in the area soon after we discovered the first instances of predation captured images of a jackal carrying a tortoise on 10 March 2019 (0419 hrs) (Fig. 1). Because scavenging can be difficult to distinguish from predation, except under unusual circumstances (e.g., Platt et al. 2010), the possibility that we misinterpreted scavenging events as predation is acknowledged but considered unlikely. We are unaware of any further tortoise predation by jackals subsequent to the last documented killing on 10 March 2019.

View Figure

• Download as Powerpoint
• Download Figure

Wild Pig Predation. — During heavy rains on the night of 14 July 2019, a sounder of wild pigs (Fig. 2) penetrated the bamboo fence of an acclimation pen containing 217 tortoises awaiting release (scheduled for January 2020) and killed an undetermined number of these animals. On the following night (15 July 2019), amidst another heavy monsoonal thunderstorm, wild pigs returned, broke into a second acclimation pen containing 217 tortoises, and again killed an undetermined number of tortoises. One of us (S.H.N.A.) arrived the following day and installed metal flashing around the base of the 3 acclimation pens that effectively deterred further incursions by wild pigs and prevented the escape of surviving tortoises.

View Figure

• Download as Powerpoint
• Download Figure

Together with Forest Department Staff and Community Conservation Volunteers, we then searched the 2 compromised pens for tortoises. Our search recovered the remains of numerous tortoises, most of which had been crushed and consumed by wild pigs. Obtaining an accurate count of mortalities proved difficult because shell (carapace and plastron) fragments from multiple tortoises were mixed together in a confusing jumble at most kill sites within the pens. Based on the forensic reconstruction of these bony elements, we estimate that wild pigs killed at least 200 tortoises in the 2 acclimation pens. We recovered 26 surviving tortoises from the 2 pens which originally held 434 tortoises. We were unable to account for the remaining 208 tortoises, leaving us to conclude that either our estimates of mortality are conservative or, more likely, many surviving tortoises escaped before the pens could be searched.

Discussion. — To our knowledge, these observations constitute the first reports of predation on G. platynota (see review by Platt et al. 2011). Jackals are mesocarnivores (5–15 kg) typically considered to be opportunistic, generalist predators (Nowak 1999; but see Hayward et al. 2017). Small mammals comprise the bulk of the diet (Hayward et al. 2017), but jackals also consume carrion, plants and fruit, ground-dwelling birds, small vertebrates, and invertebrates (Corelett 2011). Although “reptiles” are reported in many dietary studies (e.g., Mukherjee et al. 2004; Ćirović et al. 2016), specific accounts of turtle predation by jackals appear rare, but include predation on nesting marine turtles (Akeinar et al. 2006; Özdilek et al. 2015) and crowned roof turtles (Hardella thurjii) moving overland to escape drying ponds (Das and Bhupathy 2009). That said, predation on G. platynota by jackals is not unexpected given that predation on Gopherus tortoises by feral dogs (Canis lupus familiaris) and the ecologically similar coyote (Canis latrans) is well-documented in North America (Causey and Cude 1978; Esque et al. 2010).

Wild pig predation of G. platynota at SSWS occurred under conditions of confinement where tortoises were concentrated with little opportunity for concealment or escape, which no doubt accounts for the extensive mortality that resulted. Although excessive killing of confined and disadvantaged prey by predators is common (Short et al. 2002), we see no reason to assume our observations are an artifact of captivity and that wild pig predation of free-living G. platynota does not occur. Wild pigs are opportunistic, generalist omnivores that at times can be aggressive predators (Mayer 2013). Wild pigs have an extremely varied diet consisting largely of plants, but also includes invertebrates, small and large vertebrates, and carrion (Ditchkoff and Mayer 2009). The bony carapace of chelonians offers little deterrence to wild pigs, which crush and consume even large bones when eating mammal carcasses (Mayer 2013). Indeed, wild pig predation of eggs, juveniles, and adult turtles (Culbertson 1907; Fordham et al. 2006; Leary et al. 2008), including other Testudinidae (Taylor and Hellgren 1997), is widely reported and considered a threat to the survival of some populations.

We speculate the extirpation of apex predators such as tiger (Panthera tigris) and leopard (Panthera pardus) from SSWS 35–50 yrs ago (Linnell et al. 2014) led to an expansion of jackal and wild pig populations (i.e., mesopredator release; Crooks and Soulé 1999), ultimately resulting in a heightened risk of predation for G. platynota. The relationship between jackals and apex predators is complex (Hayward et al. 2017). On one hand, apex predators provision jackals with carrion (Hayward et al. 2017), but on the other hand jackals are killed and eaten by larger predators (Palomares and Caro 1999). Where apex predators are absent, jackals have few competitors, exploit a wider variety of prey (including smaller species such as tortoises), and often attain high population densities (Singh et al. 2016; Hayward et al. 2017). Similarly, wild pigs are among the preferred prey of large cats (Karanth and Sunquist 1995) and often become hyper-abundant in areas where these predators are extirpated (Boomgaard 2001; Ickes 2001).

A basic assumption underlying the head-starting of chelonians is that survivorship in the wild will be increased by rearing individuals to a body size less vulnerable to predation (Daly et al. 2018). To this end, we head-start G. platynota for 3–6 yrs with a target CL of approximately 150 mm (about 50% of adult body size). However, our observations at SSWS confirm that jackals are capable of killing tortoises with a CL > 200 mm. Moreover, coyotes are capable of killing adult Gopherus (Esque et al. 2010; Walkup et al. 2019), which are somewhat larger than G. platynota, suggesting that all size classes of the latter are probably vulnerable to jackal predation. Furthermore, even the largest adult G. platynota (Platt et al. 2019) would seem unlikely to survive an encounter with wild pigs, which are known to kill and consume other similar-sized chelonians (Taylor and Hellgren 1997). Head-started tortoises could be at a further disadvantage in predatory encounters because rapid growth under captive conditions can result in less dense and softer shells (Daly et al. 2018). Nevertheless, we consider it worthwhile to continue this practice because head-starting likely confers protection from smaller predators such as snakes, rodents, and birds (Holbrook et al. 2015).

Although our observations confirm that head-started G. platynota are at risk from jackals and wild pigs, overall losses to these predators have thus far been minimal. Since late 2016, we have released 1038 head-started G. platynota into the wild at SSWS and, of these, 28 (2.7%) fell prey to jackals during a 5-wk period in 2019. The drivers underlying this brief, but intense, spate of jackal predation are unknown; however, canids frequently respond to declines in preferred prey by opportunistically shifting predation to other species (Cypher et al. 2018). The loss of a free-ranging tortoise to wild pigs has never been documented at SSWS despite the fact that wild pigs are known to actively seek out and kill concealed chelonians (Fordham et al. 2006). Although we cannot rule out the possibility that other predation events have gone undetected, our continuous post release monitoring of translocated tortoises reinforces the conclusion that predation of head-started G. platynota is uncommon and, at this point, seems to pose little threat to establishing G. platynota in SSWS.

Despite minimal losses to predators to date, measures to mitigate predation risk when translocating tortoises are warranted, particularly with regard to wild pigs. Because tortoises are most at risk from wild pigs when large numbers are concentrated in acclimation pens, predatory incursions are best countered by structural reinforcement of pens (to include electric livestock fencing) and heightened vigilance of the attending staff, augmented with electronic surveillance systems. Mitigation of jackal predation could prove more challenging, especially if particular individuals become “habit predators” that repeatedly target translocated tortoises in the release area (Quinn et al. 2018). Several nonlethal techniques used with varying degrees of success against other canids might prove effective in deterring jackals, including the use of guard dogs and biological odor repellants (Shivik 2004). With respect to the latter, we suggest that the experimental use of tiger or leopard urine, feces, and odor-saturated bedding material (obtained from captive animals) deployed at the translocation site might result in avoidance of the area by jackals and wild pigs (e.g., Apefelbach et al. 2005). Finally, ensuring that robust populations of alternate prey remain available in SSWS by reducing poaching is an obvious first step in ameliorating the loss of tortoises from predation.