Prerelease

Tortoises deemed suitable for release were those that had been at CROW for more than 2 months, to allow for any latent diseases to appear and be treated, and to wean them off a captive diet. Tortoises also had to be larger than 100 mm in carapace length to increase their chance of being able to withstand attempted predation. Because only the subspecies S. p. babcocki can be released in KZN, various morphological indicators were used to separate it from S. p. pardalis, which is found only in the Western Cape and Namibia (Loveridge 1935), and from putative hybrids of the 2 subspecies. Even though many investigators do not recognize the 2 subspecies (e.g., Branch 1998; Boycott and Bourquin 2000), there is genetic and epidemiological evidence to suggest that there is a difference (Lambiris 1998; Varhol 1998, Le et al. 2006). Shortly before release, those tortoises selected for release were deemed fit for release by a herpetologist, being certified free of injuries, transmittable diseases, abnormal loads of parasites and foreign parasites (see Armstrong 2005). Each S. p. babcocki was then fitted with a locally manufactured radiotelemeter onto the rear of its carapace by using dental acrylic. The position of the radiotelemeter on the carapace was to prevent the possibility of the tortoises catching on vegetation or being unable to get under cover. This method was used successfully in a previous study that monitored the movements of leopard tortoises in the Nama-Karoo, South Africa (McMaster and Downs 2009). Each radiotelemeter had a unique frequency (150 VHF) with a 1/4 wavelength stranded stainless steel tracer wire antenna (plastic coated) powered by a lithium 3.6 V AA battery that was sealed in a slightly flexible rubber coating (ColourGuard; Loctite). The radiotelemeters transmitted for 12 hours each day. Battery life was unknown but was estimated by the person who assembled the radiotelemeters to be between 12 and 18 months.

Study Animals

For the first release, in January 2005, 22 S. p. babcocki of 44 S. pardalis (11 males, 11 females) at the rehabilitation center were deemed suitable for release. The rest were either putative hybrids (7), females that had been placed in the same enclosure as putative hybrid males that may have subsequently mated with them (12), or were not certified as fit for release (3). One of the latter 3 had an upper respiratory tract infection; whereas, the other 2 had ticks. Because only 10 (5 males, 5 females: L1–L10) of the 22 were regularly monitored after release (because they had radiotelemeters attached), further detail for the other 12 individually marked and identifiable tortoises is not provided except where considered important. Resightings of the latter 12 were nonexistent or very irregular. The 10 postrelease monitored tortoises were mostly medium-sized and measured between 263 and 313 mm straight carapace length and weighed between 2.5 and 4.5 kg; whereas, one was larger, at 424 mm and 6.0 kg (Table 1). The movements of a large wild female S. p. babcocki (LW, Table 1) were also monitored, after finding it opportunistically in the reserve and attaching a radiotelemeter onto its carapace in a similar manner as described.

Table 1 Prerelease measurements of each Stigmochelys pardalis babcocki released in January 2005 at Leopard Mountain Game Reserve and their status at the end of the study or when they became undetectable or until their demise.^a,b

View Fullscreen

For the second release, in December 2006, 5 S. p. babcocki of 16 S. pardalis at the rehabilitation center, namely 3 females (T1–T3) and 2 males (T4, T5), were deemed suitable for release. The rest were putative hybrids (8) or too small to be identified to subspecies (3). In February 2007, only 2 female (T6, T7) S. p. babcocki of 18 S. pardalis at the rehabilitation center were deemed suitable for release. The rest were putative hybrids (9) or too young to be identified to subspecies (6); whereas, 1 female had been in a putative hybrid male enclosure. Most of the tortoises were of medium size, with straight carapace length between 255 and 328 mm and weighing between 2.8 and 5.0 kg (Table 2). Tortoise T3 was the largest (463 mm, 15.3 kg), and T6 was the smallest (181 mm, 1.2 kg) (Table 2).

Table 2 Prerelease measurements of each Stigmochelys pardalis babcocki released in December 2006 (T1–T5) and February 2007 (T6, T7) at Usuthu Gorge Community Conservation Area, as well as the percentage mass body change between initial and final measurements and their status at the end of the study.^a

View Fullscreen

All S. p. babcocki, except T6, were termed adults because they had plastron lengths more than 200-mm long (Douglas and Rall 2006). An estimate of age was determined by using the curves produced by Hailey and Coulson (1999). Thus, all except T3 and L7 were between 6 and 15 years old. T3 was estimated to be 75 years old because this was the age of a similar sized (483 mm carapace length, 13.2 kg) captive leopard tortoise (Boycott and Bourquin 2000). L7 was of similar size to T3 but half the mass and, therefore, may have been younger. Besides T1 (confiscated from the traditional medicine trade), most of the S. p. babcocki were escaped pets because they would not naturally be found in the suburbs of Durban. In addition, most had distorted carapaces (e.g., pyramiding of scutes), which is an indication of a tortoise raised on a “captive” diet (Gerlach 2004).

Selection of Suitable Release Sites

Various criteria are important for choosing release areas (Armstrong 2003). The release areas should be within the native range of S. p. babcocki and should have had a population of the same species in the past but should have few or no S. p. babcocki at the time of the release. The factors that caused the decline in the tortoise population in the release areas should be known and no longer operating or else should be under long-term control. The region surrounding each release area should have suitable habitat for dispersal of the tortoises or their offspring should the population exceed the carrying capacity of the release area. Suitable habitat for leopard tortoises is bushland, savanna, open woodland and grassland (including vleis) with relatively sparse ground cover (Greig and Burdett 1976; Rall 1985; Bourquin 1990; Boycott and Bourquin 2000). The release site should have suitable refuge sites present, which for leopard tortoises consist of dense undergrowth of trees and shrubs, thickets and vegetation clumps, grass tussocks, logs, rocks, river banks, termite mounds, and mammal burrows (Grobler 1982; Hailey and Coulson 1995; Boycott and Bourquin 2000; McMaster and Downs 2006).

Plant species known to be eaten by leopard tortoises must grow in adequate variety and abundance in the release areas (Ashton and Ashton 2009). Leopard tortoises are regarded as intermediate between generalist and specialist feeders and are known to eat a variety of plants and fruits, including grasses (e.g., Cynodon dacytlon) and succulents (e.g., Crassula sp.) (Mason et al. 1999; Boycott and Bourquin 2000). The full diet of S. p. pardalis in KZN is unknown because this has not been studied. However, a number of species of plants indigenous to KZN are known to be eaten by leopard tortoises (Branch and Braack 1987; Broadley 1989; Boycott and Bourquin 2000; Rall, unpubl. data; P. Goodman, Ezemvelo KZN Wildlife, pers. comm.). At the release site, leopard tortoises should be able to access water because they drink from open water sources (e.g., puddles) but are said to be able to survive long periods without drinking (Grobler 1982; Rall 1985; Bourquin 1990; Boycott and Bourquin 2000).

The release program should be fully understood, accepted, and supported by the neighboring landowners and local communities. Protection of the tortoise population must be assured, by ensuring that the release sites have the infrastructure to prevent wildlife poaching and interference by humans. Release areas had to be on private land in KZN because release is not permitted in state protected areas.

Study Sites

After consideration of the aforementioned criteria, 2 release sites were chosen. In January 2005, 22 S. p. babcocki were released into the 913-ha Leopard Mountain Game Reserve (GR; 27°48′S, 32°12′E). In December 2006 and February 2007, 7 S. p. babcocki were released into the 2196-ha Usuthu Gorge Community Conservation Area (CCA; 26°52′S, 32°06′E). Leopard Mountain GR had been in existence for 8 years before the release. The land use of some of the neighboring areas had changed from cattle farming to wildlife conservation at least 2 years before, and all these areas were joining up to form a much larger protected area. Usuthu Gorge CCA was a newly established protected area at the time of the initial release there.

The vegetation at Leopard Mountain GR is characterized as Zululand Lowveld and occurs between 50- and 450-m altitude, mainly on flat or slight undulating landscapes in a summer rainfall area (500–900 mm) (Mucina and Rutherford 2006). The reserve is covered by woodland, thicket, bushland, and wooded grassland, which are all suitable habitats for leopard tortoises. Some of the leopard tortoise's preferred food plants occur widely on the reserve, refuge sites are present, and water is generally accessible to these tortoises because it is present in various areas of the reserve (C. Viviers, landowner, pers. comm.). The vegetation at Usuthu Gorge CCA is characterized as Southern Lebombo Bushveld, occurring between 100- and 600-m altitude, on more undulating landscapes, including gorges and ridges, also in a summer rainfall area (550–1000 mm) (Mucina and Rutherford 2006). Some of the known food plants, as well as refuge sites and permanent water were present. We did not have the resources to undertake a survey of S. p. babcocki present on the release areas or to perform quantitative habitat and food plant analyses (as suggested by Ashton and Ashton 2008, 2009). However, because both reserves had S. p. babcocki and the reserves were within the historical range of the species (Bourquin 2004; Branch 1998), the other ecological requirements should be met.

The number of tortoises in the reserves was unknown but thought to be below carrying capacity because of a recent severe drought in the region and having recently been converted from cattle farms to wildlife conservation areas (some areas neighboring Leopard Mountain GR) and being a recently formed nature reserve (Usuthu Gorge CCA). Tortoise shells were found on Leopard Mountain GR during the drought, but it is not known whether the tortoises succumbed to the drought (severe droughts often kill leopard tortoises; Van Zyl 1966) or to other factors, such as fire (another known cause of mortality; Boycott and Bourquin 2000). Before becoming wildlife reserves, there may have been high tortoise mortalities on the release areas and surrounding land because of the use of tortoises in traditional medicine (Cunningham and Zondi 1991) and for food, and from being burnt during uncontrolled fires or during fires designed to promote livestock production as opposed to wildlife conservation (Boycott and Bourquin 2000). Because tortoises are killed by vehicles while crossing roads (Boycott and Bourquin 2000), it was important that neither release areas had tarred roads (which promote greater traffic flow and higher traffic speeds), and only Leopard Mountain GR had a district road passing through it, which was used mainly by reserve vehicles and vehicles of tourist clients entering or exiting the reserve.

Because we expected the numbers of S. p. babcocki to be below carrying capacity on the release areas and because they are not territorial with overlapping home ranges (5%–90%, average 24% for telemetered tortoises; McMaster 2001) and because of the small number released, we did not think that much social disruption in the resident population would result from the release, as indicated in a tortoise relocation study (Berry 1986).

Release

On the release days, the tortoises were transported in crates to the reserves early in the morning to minimize heat stress and were hard-released at one or more predetermined sites in each reserve. At Leopard Mountain GR, the group of 22 tortoises (including 10 with radiotelemeters) was divided and released at 2 different sites, about 2 km apart. At Usuthu Gorge CCA, the tortoises were released at the same site, but tortoises T1–T5 were released in December 2006, and T6 and T7 were released in February 2007. Tortoises were released in summer because this is when there would be the most food available for them compared with winter, and so no supplementary feeding was provided.

Postrelease Interventions

For the purposes of the study, we accepted that some of the S. p. babcocki might try to disperse from the release areas, but such dispersal was accommodated by the likelihood that the existing populations in the surrounding areas were below carrying capacity. However, we decided to return those tortoises that we detected as having moved from the fenced reserves because we wanted to ensure that we could relocate the tortoises through the study, and we did not want the tortoises to potentially be exposed to greater threats than might occur on the patrolled areas during the study. We realized that some of the tortoises might disperse again later, but we hoped that by returning them they might settle down in the release areas (because this has been done in some tortoise relocation studies, e.g., Belzer 1999; Tuberville et al. 2005), or else that by the end of the study the tortoises would be more familiar with the habitat of the region. Furthermore, if any of the released S. p. babcocki showed signs of disease, then the tortoise was taken to a veterinarian to be treated.

Monitoring

The radiotelemetered tortoises released at Leopard Mountain GR were located monthly for the first 10 months after release and sporadically (maximum 5 times) up to 25 months after release. The radiotelemetered wild tortoise was located monthly (after affixing the radiotelemeter), until the telemeter was found detached on the ground. Because of malfunctioning of some of the radiotelemetry equipment, not all radiotelemetered tortoises were found at each monitoring session. Nontelemetered tortoises were located opportunistically. Tortoises released at Usuthu Gorge CCA were located monthly for up to 13 months, when the study ended.

A 3-tier Yagi aerial and a wide-range receiver (DJ-X10; Alinco) was used to locate the radiotelemetered tortoises. Once found, their locations were obtained by using a global positioning system (12XL; Garmin), the microhabitat and their activity were noted, and the tortoises were suspended in a bag and weighed by using a scale (0.5 kg accuracy, Leopard Mountain GR) and a spring balance (0.1 kg accuracy, Usuthu Gorge CCA). General health was also noted. The global position system locations were exported into the Geographical Information System ArcMap 9.2 (Environmental Systems Research Institute Inc, Redlands, CA) for further analyses.

Data Analyses

The minimum convex polygons (MCP) encompassing the recorded locations for each tortoise were estimated by using Hawth's Analysis Tools extension (Beyer 2004) for ArcMap. We recognize that the tortoises had probably not yet developed a home range during the study periods, and standard home range analyses might not be biologically significant or appropriate (Field et al. 2007), but we considered that these MCPs would be informative as indices of the areas covered by the tortoises after release.

To determine movement of each tortoise, the distance function in ArcMap was used to measure the minimum straight-line distance travelled each month (“minimum monthly movement”). If a tortoise left the reserve and was brought back, then the distance measured for the next month was from that new location, not from where the tortoise was found outside the reserve. Minimum total distance travelled by each tortoise was the sum of these minimum monthly straight-line distances. Changes in body mass were calculated as the difference between initial and final mass and were expressed as a percentage.

Survival and Changes in Mass

After 25 months of postrelease monitoring at Leopard Mountain GR, there was only 1 tortoise confirmed to be alive (10%), 3 confirmed deaths (30%), and 6 whose fate was unknown (Table 1). Tortoise L6 was found injured 2 months after release, with a hole in its carapace likely caused by a pick-axe or similar sharp instrument, and with the radiotelemeter detached. It subsequently died. Tortoise L10 was driven over and killed by a vehicle 10 months after release, and tortoise L2 was found dead 17 months after release, lying on its side, wedged against a shrub (Table 1). In addition, one nontelemetered released tortoise was found driven over and killed 3 months after release. The radiotelemeter from tortoise LW was found detached but still functioning on the ground 21 months after it was attached, and no further sighting of the tortoise was made. Because of the known failure of 2 radiotelemeters (on L4 and L7), it was the likely cause of the disappearance of the other tortoises.

After 13 months of postrelease monitoring at Usuthu Gorge CCA, there were only 3 of 7 tortoises alive, namely T1, T4, and T5. However, T5 was listed as “dead” in terms of the study, because it was taken back to CROW 4 months after being released, because it had mucus bubbling from its nares, loose skin, sunken eyes, was losing weight (Table 2), and did not move far each month (see below). The cause of its illness was undetermined. Other than T4 being taken out to be treated for extensive skin sloughing on its front legs, diagnosed as “noncontagious dermatitis caused by a hypersensitivity response” (R. Last, VetDiagnostix, pers. comm.), both T1 and T4 were healthy and had an accumulative weight gain of more than 20% (Table 2). However, they could not be located 14 months after release.

The first tortoise to die at Usuthu Gorge CCA was T7, 4 months after being released. It was found freshly decapitated and with a laceration on one of its back legs. The edges of the wound were sharp and straight, which suggested that a person used a knife to kill it (J. Vorster, Vetdiagnostix, pers. comm.). This tortoise was the only other one in this release (besides T1 and T4 mentioned above) that had gained body mass (Table 2). The second tortoise (T6) died 6 months after release, having lost 17% of its body mass (Table 2). It was too decomposed for a necropsy, but visual inspection showed no visible marks on the carapace. It had behaved differently from most other tortoises, except T3, by always being in the open (not in a refuge) during autumn and winter. T3 was taken out of the reserve 9 months after release because it had continued to lose body mass since May, was found with loose skin and sunken eyes, and only retracted its limbs when touched. T3 had the largest body mass loss of all the released tortoises (Table 2). It was brought to a relatively large, secured garden outside of the reserve, fed, given water, and allowed to rest. Because it ate and drank, it was decided not to take it back to CROW, and it was left to recuperate. Unfortunately, it died a few days later, and no necropsy was performed. The last known death was of T2, whose body mass had decreased (Table 2) before it was found dead, lying on its back, 10 months after release. A necropsy was not done because it was too decomposed when found. There were no visible injuries on its carapace.

Minimum Straight-line Distances

At Leopard Mountain GR, most of the tortoises' first large recorded movements from the release points were either in a northeasterly (n  =  4) or southwesterly direction (n  =  3), whereas a few others moved in a northwesterly (n  =  2) or southeasterly direction (n  =  1) (Fig. 1). Similarly, those released at Usuthu Gorge CCA had minimum straight-line distances in either a northeasterly direction from the release point (n  =  5) or in a northwesterly direction (n  =  2) (Fig. 2). At both reserves, most tortoises changed directions from this initial direction, which was pronounced in L2 (Fig. 1) and T3 (Fig. 2).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Several tortoises travelled outside each of the reserves after each release. At Usuthu Gorge CCA, both T4 and T7 travelled outside the reserve, in a similar direction, in December and April, respectively (Fig. 2). When T7 was brought back to the original release site, it went in a similar direction as it had done previously (T7b, Fig. 2); whereas, T4 went roughly 180° in the other direction compared with the direction it had originally taken (T4b, Fig. 2). T4 left the reserve again a month later and was retrieved and released (T4*, Fig. 2) near where T1 was last located. At Leopard Mountain GR, 3 months after release, L5 left the confines of the reserve. It was retrieved and released within the reserve boundary (Fig. 1) but by the next month had disappeared. Tortoise L10 had also left the confines of the reserve after 3 months (Fig. 1), but the boundary fence had been removed at this time, and so it was not brought back to the reserve.

At both reserves, monthly recorded movements were variable between the tortoises and between months, but most tortoises travelled less than 400 m (minimum straight-line distance) each month. At Leopard Mountain GR, large movements (> 400 m minimum straight-line distance) were recorded in the first month after release (L5, 1419 m; L6, 891 m) and again in June (L3, 431 m; L8, 673 m), July (L10, 456 m), August (L2, 558 m), and November (L1, 434 m; L8, 567 m). The wild tortoise only exceeded 200 m (minimum straight-line distance) per month in spring (October). Conversely, all tortoises released at Usuthu Gorge CCA (except T5, because it was diseased and thus did not move further than 156 m each month), had recorded minimum straight-line movements larger than 900 m, either in the first month (T2, 928 m; T4, 1283 m), the second (T6, 1012 m), or the third month (T1, 1138 m) after release. Tortoise T3 had recorded minimum straight-line movements of 936, 1633, 990, and 976 m for the first 4 months after release; whereas, tortoise T7 had values of 1428, 1115, and 1375 m for the first 3 months after release. In addition, there was a clear decrease in recorded minimum straight-line movements of tortoises released at Usuthu Gorge CCA in winter months (June–August), these being generally < 100 m per month, which was not seen as distinctly in the tortoises released at Leopard Mountain GR. After winter, both tortoises T2 and T4 showed an increase in recorded minimum straight-line movements from September onward, but T4 especially travelled far each month and attained a recorded minimum straight-line movement of more than 2000 m in January 2008.

When excluding T5, the average monthly minimum straight-line distance for tortoises at Usuthu Gorge CCA ranged from 132 to 1045 m, total minimum straight-line distance travelled ranged from 1092 to 6144 m, and total MCP area ranged from 3.1 to 150.5 ha (Table 3). At Leopard Mountain GR, excluding L5, L6, and L7 because of the small sample size, average monthly minimum straight-line distance ranged from 83 to 283 m, total minimum straight-line distance travelled ranged from 650 to 1801 m, and total MCP area ranged from 4.6 to 48.4 ha (Table 3). The wild tortoise had the second lowest average monthly minimum straight-line distance and MCP area (Table 3). The maximum “minimum straight-line distance” from the release site for tortoises released at Usuthu Gorge CCA (excluding T5) ranged from 922 to 2585 m; whereas, those at Leopard Mountain GR (excluding L5, L6, and L7) tended to be less (449–1612 m; Table 3).

Table 3 Total and monthly (average ± SE) minimum straight-line distances travelled, maximum straight-line distance travelled from release site, and estimated area covered as recorded for Stigmochelys pardalis babcocki up to 10 mo after release (Nov 2005) at Leopard Mountain Game Reserve, and for S. p. babcocki released (Dec 2006, Feb 2007) at Usuthu Gorge Community Conservation Area, before their demise, disappearance, or the end of the study. The number of months used in calculations is specified as data were not recorded for each month during the period shown.

View Fullscreen

Management Implications

To allow for rehabilitated S. p. babcocki to become accustomed to the diseases present at the release site or for diseases to reveal themselves, they could be placed in an enclosure for a period before release (Dodd and Siegel 1991). This may also help to increase site fidelity (Pedrono and Sarovy 2000; Tuberville et al. 2005; Ashton and Ashton 2008) and to allow them to adapt to eating the indigenous vegetation in the area. Penning the animals for 12 months before release has significantly increased site fidelity in 1 study on gopher tortoises, where only 1 of 12 tortoises dispersed, compared with no penning (10 dispersing of 13) and penning for 9 months (5 dispersing of 13) (Tuberville et al. 2005). Therefore, it would be useful to examine the effects of penning in future releases of rehabilitated S. p. babcocki. However, because some of the penned gopher tortoises did disperse and had to be retrieved and 3 tortoises dispersed again in the second year after release (Tuberville et al. 2005), penning is not a foolproof solution in ensuring site fidelity. Furthermore, this intervention is likely for rehabilitated S. p. babcocki to be impractical because of costs in relation to the conservation status of this species (Boycott and Bourquin 2000). Therefore, less expensive and less time-consuming options could be attempted in future releases, such as ensuring that the release area is large enough to accommodate the tortoises' ability to disperse. Calculations were made of the reserve sizes needed to accommodate released gopher tortoises (Berry 1986) and box turtles (Cook 2004). By using the method outlined by Cook (2004), the required reserve size for future rehabilitated S. p. babcocki releases would need to be approximately 2099 ha. This size is actually smaller than the Usuthu Gorge CCA where tortoises were released and thus other preventative measures are suggested. These include ensuring the reserve fencing is properly secured to prevent tortoises from pushing through and releasing tortoises at a suitable site in the middle of the reserve to decrease the possibility of them encountering the boundary. Rehabilitated S. p. babcocki should also be released in less undulating landscapes than Usuthu Gorge CCA because even though some wild leopard tortoises were seen there, rehabilitated tortoises might not be as fit as wild tortoises because of their time in captivity, as found in a study on rehabilitated raptors (Curtis and Jenkins 2002). In addition, the quality and quantity of food available in the habitat needs to be assessed (Ashton and Ashton 2008, 2009). Future releases should be at the start of spring, just after the first rains, when tortoises generally become more active (Boycott and Bourquin 2000), so that they have more time to build up energy reserves before winter than if released later. Lastly, future postrelease monitoring could be carried out by the local residents (e.g., game rangers) because having local people involved in tracking has been shown to decrease the interest in harvesting tortoises in Egypt (Attum et al. 2008).

To assess whether these suggestions are beneficial, future releases of rehabilitated leopard tortoises should be in accordance with an amended release protocol that includes long-term postrelease monitoring. Otherwise, one has to accept that animal welfare will be compromised because there will be mortality that is human induced, caused by disease, or because of an inability of the tortoises to adapt to the environment and habitat of the release area. The alternative of keeping these tortoises in captivity should not be a standard option because of the large space requirements of the leopard tortoise, which normally cannot be supplied in captivity. Generally allowing the keeping of leopard tortoises in captivity is likely to stimulate the desire of other people to obtain leopard tortoises as pets. In both these respects, captivity may have negative welfare implications, such as the acquisition of shell deformities because of poor diet, and injuries received from vehicles, lawnmowers, dogs, etc., all of which have been noted on tortoises at CROW in Durban. An imperative is that extensive public education is carried out to dissuade the public from illegally keeping tortoises as pets. Otherwise the option of euthanasia may need to be considered.