Study Area

The IGWLS (lat 10°12′–10°35′N, long 76°49′–77°24′E) in the southern Western Ghats, is one of the largest protected areas in India, with an area of 958 km². Several Forest Department posts are located within its boundaries, Topslip (lat 10°28.561′N, long 76°49.992′E, 750 msl) and Varagaliar (lat 10°25.070′N, long 76°51.940′E, 600 msl) being 2 such sites. There also exist tribal settlements within the sanctuary—the Kadar settlement of Erumaparai is located 1.5 km from Topslip, and the elephant camp at Varagaliar is inhabited mainly by Malasars and Pulayars. The 2 sites are 23 km apart by road (straight-line distance about 7 km, No.58/B/15/4, 1:15,000 Tamil Nadu Coimbatore Forest Circle, Government of India). The study area comprised the region between Topslip and Varagaliar in the east and the Kerala–Tamil Nadu interstate border in the west (Fig. 2). This region receives rainfall from both the southwest and northeast monsoons; June and July are the wettest months of the year, and December to April, the driest.

View Figure

• Download as Powerpoint
• Download Figure

Five localities within the study area were searched intensively for tortoises. They included the Karian Shola National Park at Topslip, which has tropical wet evergreen forests of the Drypetes–Calophyllum–Polyalthia–Vateria type (Champion and Seth 1968; Pascal 1988); Kozhikamuthy, which has tropical moist deciduous forest of the Terminalia–Anogeissus–Tectona type (Champion and Seth 1968); the teak (Tectona grandis) plantations of Sichali and Varagaliar; and the Anaikundhy Shola—a small wet evergreen forest fragment. Small, seasonal streams run through all 5 localities, and the teak plantations also have stretches of marshy ground covered with grasses. The latter are locally known as vayal and are often the only source of water in the plantations in summer.

Methods

Based on the results of the preliminary survey (Ramesh, in press), I conducted a more intensive study from 14 December 2002 to 31 March 2003. During this period, the area-constrained visual encounter survey method was used (Crump and Scott 1994), in which 2 persons searched the banks of streams and vayals for tortoises from 1400 to 1900 hours, except for 2 days in March when 3 searchers were used. In the case of large vayals, wherever possible, the middle section was also searched for tortoises that could be resting in clumps of tall grasses. Each area was searched only once daily. Apart from direct sightings, cues indicative of tortoise presence, such as trails and foraging signs, were used to locate individuals. Search effort was used as a measure of relative abundance (Klemens and Moll 1995; Platt et al. 2003). But, unlike other studies in which search effort is presented as number of tortoises observed per man-hour of searching, i.e., sighting frequency, I converted this into its reciprocal (number of man-hours of searching required to find one tortoise) to facilitate comparisons (Freilich et al. 2000).

All tortoises found were marked with a unique combination of notches on marginal scutes (Cagle 1939) and checked for evidence of external injury and ectoparasites. Straight carapace length (SCL), straight carapace width, shell height (HT), plastron length, and plastron width were measured using Mitutoyo steel calipers (300 mm); anal fork (AF) and anal notch (AN) by dial calipers (150 mm); and body mass (WT) by Zebco Deliar weighing scales (Smith et al. 2001). Tortoises were sexed on the basis of plastral and tail-claw characteristics (Auffenberg 1964; Vijaya 1983; Das 1991), and individuals smaller than 160 mm SCL were considered to be juveniles (Ramesh, in press). Morphometric differences between male and female tortoises were tested using a 2-tailed Mann-Whitney U test.

Results

In total, 57 individuals were found. Adults of both sexes were sighted in equal numbers (27 each), but only 3 juveniles were recorded. There were no recaptures despite repeated searches of some localities (Karian Shola, Sichali, and Anaikundhy). Apart from signs such as trails and bitten leaves, the rustling sound produced when a tortoise moves on dry leaf litter also proved useful in locating individuals. In one instance, 3 tortoises were located by another auditory cue—a tortoise (male, SCL 251mm, WT 2.3 kg) was heard grunting loudly and shell-ramming another individual (male, SCL 209.5mm, WT 1.4 kg) while a third individual remained nearby (female, SCL 266 mm, WT 3.2kg). This is the first record of male combat observed in the wild. Overall, the mean search effort was 3.4 man-hours/tortoise. Search effort was lowest in Anaikundhy Shola and highest in Varagaliar (Table 1). It also varied across the 4 months, with search effort being highest in January and lowest in March (Table 2). All tortoises were found singly, except during the month of March when after convectional summer showers, 2 or more individuals were found in close spatial proximity and therefore only 1 hour of search effort was required to locate each tortoise (Table 3).

Table 1. Search effort for Indotestudo travancorica across localities in the Indira Gandhi Wildlife Sanctuary (December 2001–March 2003).

View Fullscreen

Table 2. Search effort for Indotestudo travancorica in the Indira Gandhi Wildlife Sanctuary from December 2002 to March 2003.

View Fullscreen

Table 3. Search effort for Indotestudo travancorica in the Indira Gandhi Wildlife Sanctuary from 10 March 2003 to 25 March 2003.

View Fullscreen

Morphometric data were collected from 49 adults and 3 juveniles (Table 4); 5 tortoises could not be measured because fieldwork was interrupted by elephants. There were no significant differences among sexes in SCL (z = 1.34, p = 0.18), WT (z = 0.74, p = 0.45) and AN (z = −0.46, p = 0.65), but males had significantly lower HT (z = −2.12, p = 0.034) and larger AF (z = 2.22, p = 0.026). The largest tortoise (a male) measured 307 mm (SCL) and weighed 3.6 kg.

Table 4. Morphometric data (in mm and kg) of Indotestudo travancorica from the Indira Gandhi Wildlife Sanctuary, from December 2002 to March 2003.

View Fullscreen

In addition, carapacial color appeared to vary according to the locality: 35% of tortoises found in Karian Shola (n = 20) and 33% in Anaikundhy (n = 21) were dark brown and/or had large black blotches on the scutes; whereas, tortoises from Sichali, Kozhikamuthy, and Varagaliar were a lighter, duller brown.

Most of the tortoises (51 of 57 found, 90%) had ticks (average = 3 ticks/tortoise) attached to the marginal scutes of the carapace. Occasionally, ticks were also found near the base of the tail. Some tortoises were found to shelter in thickets of weeds (Lantana camara and Cromolarium glandulosum) (Ramesh, in press; 6 of 57 found, 10.5%).

Discussion

Because of its cryptic coloration the Travancore tortoise is considered difficult to find in the wild (Frazier 1989). Even local people such as the Kadars (a hunter-gatherer tribe), who are knowledgeable about the area and familiar with this species, consider it an arduous task to find tortoises, and in surveys conducted by other biologists, the largest number observed was 7 (Moll 1989). However, the overall mean search effort reported in this study is much lower than those reported elsewhere (Table 5). Therefore, factors that are thought to have resulted in greater capture success are discussed here. Suitable localities and microhabitats (near streams and vayals) were identified and search effort was concentrated here. Signs of tortoise presence such as trails, bitten leaves, and fresh scat were used to locate individuals—trails proved to be the most reliable indicators, since they are clearly visible in wet and dry seasons and in different microhabitats. Searches were conducted only during the period when tortoises are most active, thereby improving capture probabilities (Freilich et al. 2000; Ramesh, in press).

Table 5. Comparison of search effort for Indotestudo travancorica in the southern Western Ghats, India.

View Fullscreen

Further, confining the number of searchers to 2 instead of a group improved search efficiency as tortoises were less likely to be disturbed and hide under dense cover (Ramesh, in press), and auditory cues (rustling of dry leaves caused by moving tortoises, grunting and shell-ramming by males) were more noticeable because the noise made by the movement of searchers was also reduced. Moreover, because the same 2 searchers worked in the 5 localities, it also reduced interobserver bias, which can be significant in locating tortoises and causing variations in capture probability (Freilich and LaRue 1998; Platt et al. 2003).

Environmental factors such as the availability of suitable microhabitats and forage need to be investigated further to explain variation in relative abundance across localities within IGWLS (Table 1). However, although Sichali and Varagaliar are very similar in terms of habitat type, the hours of search effort required to find a single tortoise at the latter site was almost 3 times as much as that required in the former. A probable reason is that Varagaliar has fairly heavy levels of disturbance caused by activities such as collection of fuel wood and forage for camp elephants. But more importantly, the inhabitants have dogs that are often used for hunting small animals, including tortoises. Kozhikamuthy is also used as a seasonal elephant camp, but by a smaller number of inhabitants, and this could contribute to the low encounter rates at that locality. The presence of dogs and increased movement of people may result in a greater number of tortoises being hunted and, therefore, affect relative abundance more significantly than habitat disturbance per se because tortoises are known to shelter even in thickets of weeds such as Lantana camara and Cromolarium glandulosum, which grow in disturbed areas.

The high relative abundance of I. travancorica in Anaikundhy and the variation in search effort across the 4-month study period could have been caused by seasonal changes in the microhabitat. For example, after the first 2 days of summer showers that occurred in March, 3 searchers found 14 tortoises in 15 man-hours (Table 3): individuals were found clustered together; therefore, a larger number of tortoises were found per man-hour of searching. Some chelonians are known to drink rainwater (Peterson 1996), so the sudden availability of water at the end of the dry season could have resulted in tortoises aggregating near streams and vayals. Fresh forage may also have been available in these microhabitats. Such seasonal changes can alter tortoise activity thereby affecting encounter rates (Rose and Judd 1975; Berry and Turner 1986; Geffen and Mendelssohn 1989; Freilich et al. 2000). This clustering of tortoises could also be indicative of reproductive behavior because male combat reported herein is known to occur alongside courtship in captive tortoises (Das 1991; Ramesh, pers. obs.), and in 4 instances (10 March to 17 March), a female was found with 1 or 2 males. However, no courtship behavior was observed. Further investigation is required because there is some uncertainty regarding the breeding season of I. travancorica. Auffenberg (1964) reported it to be from November to January; whereas, according to another report, it coincides with the monsoons (Das 1991) beginning in June, but only single individuals were encountered in an earlier survey conducted from May to June (Ramesh, in press). In addition, the Kadars too believe that the summer showers (called the Elavan pumari) signal the onset of the breeding season of this species.

It is interesting to note that of the 57 individuals found during this study, juveniles comprised only 5% of all captures; whereas, during the earlier survey, they comprised 33% of all captures (n = 27). With the onset of the monsoons in June, water and forage may have been easily available, and this in turn could have increased the activity levels and capture probabilities of juveniles during the earlier study. However, this needs to be verified as in general, neonates and juvenile tortoises are very difficult to find and, as a result their ecological requirements, are poorly understood (Berry and Turner 1986; Mason et al. 2000; Freilich et al. 2000).

Among adult I. travancorica, it has been reported that males are larger than females (Das 1991), but despite the relatively large sample size of this study (24 males, 25 females), there was no significant difference in carapace length (Fig. 3). The difference in shell height and anal notch has been recorded earlier, and the latter is believed to provide the tail of the male greater freedom of movement during copulation (Vijaya 1983). There is a marked difference in the shape of the carapace as well: males tended to have a more uniformly elongated carapace; whereas, females have an ovoid shape, with a broader anterior end. In addition to the characters mentioned earlier, this may also serve as useful guide to differentiate between sexes in the field. In view of the fact that both the male-to-female ratio (1:1 December–March; 2.6:1 May–June) and representation of juveniles varies according to season, sampling during both wet and dry periods may be necessary to determine size distribution of the population.

View Figure

• Download as Powerpoint
• Download Figure

The carapacial color variation seen in tortoises is believed to provide effective camouflage (Pritchard 1979), with tortoises in wetter habitats being darker than those in dry habitats (Kabigumila 2000). Therefore, the color variation observed in I. travancorica may serve a similar purpose. It should be noted that I. travancorica may have fairly large home ranges; there is a recorded instance of a tortoise travelling at least 20 km in as many years (Ramesh 2004), and given the mosaic landscape of the Western Ghats, habitat variables can vary considerably between sites.

To conclude, the search methodology described here can be used to effectively sample this rare species and assess the relative abundance of I. travancorica in different types of habitats. If surveys are conducted during March–April, the sound of shell-ramming and grunting can be heard about 50 m away and, therefore, can serve as an important auditory cue and further reduce search effort. At the same time, it could yield valuable data on the reproductive behavior of this species. Although further studies are necessary in order to estimate the size and structure of the population, I. travancorica may not be as rare as believed earlier because it has been sighted even in moderately disturbed areas during this study and others (Bhupathy and Choudhury 1995); it is known to occur in several protected areas totaling about 3900 km² (Das 1991; Bhupathy and Choudhury 1995); and so far, large-scale trade or exploitation of this species has not been recorded (Das 1991; Choudhury et al. 2000). However, it is necessary to continue with the enforcement of wildlife protection laws applicable to the species because tortoises are very vulnerable to exploitation by humans (Moll 1989) and to large-scale habitat destruction such as those caused by wildfires, conversion of forests to plantations, and construction of hydroelectric dams (Groombridge et al. 1983; Das 1991; Choudhury et al. 2000).