Population Structure and Sexual Dimorphism in a Southern Population of Tent Tortoises (Psammobates tentorius tentorius) of the Karoo, South Africa
ABSTRACT
This study examined population ecology parameters for a southern population of the tent tortoise (Psammobates tentorius tentorius) in the Karoo. We found a relatively low density of tortoises (49 tortoises/km2) skewed to adults with a 1:1 sex ratio. The tortoises' contrasting dark and light carapacial patterns create crypsis with the disruptive shadows cast at the base of vegetation (mostly shrubs) where the tortoises seek refuge. The largest female was 14.7 cm, which makes her approximately 42 yrs old. The largest male was 11.6 cm, which makes him approximately 42 yrs old. We can distinguish males and females at carapace lengths of approximately 8–10 cm, when the tortoises are approximately 8–10 yrs old. Psammobates t. tentorius is a sexually size-dimorphic species in which adult females are significantly larger than are males. Selection for body shape and size appears attributable to fitness between the sexes (e.g., egg production and size in females, and opportunities to mate by males). We characterize this first group of nominate tent tortoises for future comparisons with other population. The study provides critical baseline information for the long-term management and conservation of this species, and adds to our understanding of the fauna of the Succulent Karoo.
Southern Africa has the highest diversity of tortoises in the world. One-third of all tortoise species (i.e., 14 of 43) occur here, and 11 species are endemic to the region (Branch 1998), Here we examine fundamental population structure and sexual dimorphism of the tent tortoise (Psammobates tentorius). Tent tortoises occur throughout the Great and Little Karoo, to the Succulent Karoo of the west coast, and north into southern Namibia (Branch 1998; Hofmeyr et al. 2018). There are currently 3 recognized subspecies: P. t. tentorius, P. t. verroxii, and P. t. trimeni. These subspecies are based on plastral patterns, which is a concept first explored by Duerden in 1907 and elaborated on by Loveridge and Williams (1957). They created a key to subspecies of Psammobates that link 5 distinct plastral patterns to specific geographic locations We examined a population of the nominate southern subspecies (P. t. tentorius) near the town of Prince Albert (an area where Loveridge and Williams indicate the convergence of plastral pattern) and found they had straight carapace length (SCL) that ranged from 2.9 to 14.7 cm. This subspecies occurs in the southeast Cape from Grahamstown to Matjiesfontein, north into the central Karoo (Branch 1998; Hofmeyr et al. 2018).
Examining differences in morphology between different populations of turtles traditionally advances the discovery of new species, subspecies, or regional polymorphism (as done by Loveridge and Williams [1957]). Yet little work has been done to examine morphological differences in tent tortoises in the field. There can also be morphological differences that are seen between the sexes.
Turtles exhibit a wide variety of size differences between the sexes (Berry and Shine 1980). In particular in North America, male Graptemys spp. and Malaclemys terrapin are significantly smaller than their females (Ernst and Lovich 2009). In Africa, male Centrochelys sulcata, Chersina angulata, and Astrochelys radiata are significantly larger than females (Lambert 1993; Leuteritz and Gantz 2013; Boycott and Bourquin 2000; T. Leuteritz, pers. obs.). Berry and Shine (1980) suggest female optimal body size may depend primarily on the number of eggs she lays, and secondarily on the degree of nest predation occurring. They believe these factors to be similar between females of closely related species, but are different from the factors affecting male body size. It is possible that the mating system (i.e., whether there is strong female choice) is responsible for the variability in direction and degree of interspecific sexual size dimorphism (SSD). Although the selective advantage of small versus large male body size is not fully understood, Berry and Shine (1980) showed that in species with male combat or forcible insemination, males are often as large as or larger than females. The larger the male the stronger he may be and presumably the better his ability to fight with rival males or to overpower females during mating. Conversely, in species with female choice, males are usually smaller than females (Berry and Shine 1980; Bonnet et al. 2010; Leuteritz and Ganz 2013). Males of this type often exhibit elaborate precoital behavior. Small size may increase mobility (males can move greater distances), which may aid in the ability to search for females (adaptation for male dispersal). Smaller tortoises have greater mass-specific metabolic rates than do large tortoises. (Berry and Shine 1980).
In addition to body size, a suite of other divergent characters exist that have often been used to assess the sex of turtles, including shell morphology, claw length, eye color, tail size, and distance of cloacal vent to body (Berry and Shine 1980; McRae et al. 1981). The presence of a plastral concavity (located at the abdominal and femoral scutes) is indicative of males in many species (Carr 1952; Ernst and Lovich 2009; Leuteritz and Gantz 2013). Plastron concavity appears to aid males in mounting females for terrestrial copulation and is most pronounced in terrestrial domed genera such as Geochelone, Gopherus, and Terrapene. Males in some tortoises have elongate gular scutes, none so pronounced as the Madagascar ploughshare tortoise (Astrochelys yniphora), which are used in combat and courtship rituals (Juvik et al. 1981; Durrell et al. 1989).
For the Prince Albert area of the Great Karoo, the climate is arid with < 200 mm of rainfall per yr, and a mean annual temperature of 17.5°C. Prince Albert is a between-seasons rainfall region with the higher rainfalls occurring between February and April, and a second smaller peak between October and December, although this is often unpredictable (Milton et al. 1992; Leuteritz and Hofmeyr 2007).
As with most South African species of tortoise, very little is known about this species basic ecology (i.e., only observations on predation: Malan and Branch 1992; Lloyd and Stadler 1998). This study examines population ecology parameters for a southern population of P. t. tentorius. These parameters are critical to the long-term monitoring necessary to guide management and conservation of this species and adds to our understanding of the fauna of the Succulent Karoo, the one and only arid region in the World's 25 Biodiversity Hotspots (Myers et al. 2000).
METHODS
Study Site. — We studied P. t. tentorius from October 2002 to October 2003 at the Tierberg Karoo Research Centre (otherwise known as the study site) on the southern edge of the Great Karoo, 30 km east of Prince Albert. The 100-ha site has been fenced off from domestic livestock since 1989 and is located on the flat open plains north of the Swartberg Mountains. The site is considered typical of Acocks' (1975) Little Karoo form of Karroid Broken veld with few grasses. It is characterized by low-growing evergreen and deciduous succulents (Ruschia spp. and Pteronia pallens) with interspersed major drainage lines that have taller (up to 4 m) vegetation. The presence of mima-like mound termite hills (heuweltjies) dispersed throughout the habitat is of note because the fauna and the flora on the heuweltjies differ from that of the surrounding habitat (Milton et al. 1992).
Data Collection. — In October 2002, we used strip transects (1 km × 5 m) for an intense search of the study site to locate, and attach radiotransmitters to, 20 adult tent tortoises. During the remaining 11 mo, we located new tortoises opportunistically while tracking the transmittered individuals. All tortoises were hand-captured and given a unique number by filing notches in their marginal scutes (Cagle 1939). We recorded the time each field-worker spent in the field to calculate search effort in person-hours.
Body mass (BM) was recorded to the nearest 1 or 10 g, respectively, with Pesola spring balances (two sizes of 100 and 1000 g), and morphometric measurements were recorded to the nearest 0.1 cm with vernier callipers (two sizes of 15.0 and 50.0 cm). We measured straight carapace length (SCL) from the nuchal to supracaudal scutes; domed carapace length (DCL) along the midline over the arch of the carapace; plastron length (PL) as the straight distance across the gular and anal scutes; shell width (SW) across the sixth marginal scutes; and shell height (SH) over the third vertebral scute. The lengths of the plastral scutes, the gular (G), humeral (H), pectoral (P), abdominal (Ab), femoral (F), and anal (A) scutes were measured along the plastral midline seam. Additional measurements included the posterior shell opening (anal gap, AG) at the midline between the supracaudal and anal scutes; the distance between the tips of the two anal scutes (anal width, AW); and the width of the gular scute (GW) at the seam of the gular and humeral scutes.
After measuring a tortoise, we counted the different carapacial scute types (nuchal, vertebrals, pleurals, marginals, supracaudal), the plastral scutes at the bridge (axillaries and inguinals), and the number of toes on the front and hind feet. We examined the radiation pattern on the carapace by counting the number of rays on the third vertebral, second left pleural, and first left marginal scutes as representatives of the different scute types. We used secondary sexual characteristics, especially the shape of the supracaudal scute and tail length (i.e., we noted whether the tail extended laterally beyond the anal scutes of the plastron), to distinguish males from females and categorized each tortoise as juvenile, male, or female. The age of a tortoise was estimated by counting annuli (growth rings, GR; Zug 1991).
Data and Statistical Analyses. — Statistical tests were completed using SigmaStat 2.03 and were considered significant at α = 0.05. We summarized data as means and standard deviations. Data were first tested for normality and homoscedasticity before using parametric tests or the nonparametric equivalent. To test for differences between independent samples, we used the Student's t-test or Mann-Whitney rank sum test (U), while we used repeated measures analysis of variance (ANOVA) to compare more than two matched samples. Body parameters were scaled by linear regression. When regressions were significant, we used analysis of covariance (ANCOVA; followed by Tukey's post hoc analysis) to determine whether regression slopes and elevations differed among groups (Zar 1999). The sexual size index (SDI) for the population was calculated as the mean SCL of the larger sex divided by the mean SCL of the smaller sex (Gibbons and Lovich 1990) and as the compressed–modified SDI (Lovich and Gibbons 1992). Lovich and Gibbons (1992) propose a size dimorphism indices (SDIs) to quantify SSD that should meet 4 criteria as follows: 1) it should be properly scaled; 2) it should have high intuitive value; 3) it should produce values with one sign, (positive) when sex A is larger than sex B, and the opposite sign when sex B is larger; and 4) it should produce values that are symmetric around a central value, preferably zero. The compressed–modified SDI is an alternative SDI based on the mean size of the smaller sex, with the result arbitrarily defined as positive (minus one) when females are larger and negative (plus one) in the converse case. Careful selection of a primary size variable is crucial to meaningful interpretation of sexual size differences.
RESULTS
Tortoise Numbers, Density, and Search Effort. — From October 2002 to October 2003, we found 79 P. t. tentorius (39 males, 31 females, 9 juveniles), 6 Stigmochelys pardalis, and 4 C. angulata tortoises in or adjacent to the study site. The sex ratio for adult male to female tent tortoises (1:0.8) did not differ from the predicted 1:1 sex ratio (χ2 = 0.258, p = 0.612). Of these tortoises we found 49 P. t. tentorius tortoises (23 males, 23 females, 3 juveniles) and 3 C. angulata tortoises (2 females, 1 juvenile) within the 100-ha study site (or 49 tortoises/km2) and recaptured only 4 of these tortoises (excluding recaptures with radio) during the study period. We spent 494 man-hours searching for tortoises, but seasonal search effort was greatest in spring 2002 and winter 2003. The number of new tent tortoises found within the study site decreased through the study period, but not in direct proportion to search effort (Table 1). The lowest detection time per tortoise occurred during the wettest and warmest season, summer.
Morphological Characteristics. — The carapace scutes of tent tortoises were black based with yellow rays emanating from the birth plate (areola). The mean number of rays on the vertebral, pleural, and marginal scutes of 71 individuals was 9.5 ± 3.1 (range, 4–17), 11.3 ± 3.5 (range, 4–19), and 2.1 ± 0.6 (range, 1–4), respectively. As the animals age, the annuli become worn and the scute color changes to a duller black with amber-colored radiations. The plastron had a parchment color with a distinct brown oval patch (tapering anteriorly like a shoe print) centered on the midline. In all tortoises (n = 32) where the feet were visible and not drawn into the shell, there were always 5 and claws on the front and hind feet, respectively.
The nuchal scute was completely absent in 6%, or very small in 10%, of the tortoises. The supracaudal scute was undivided, although one male had a partially divided supracaudal scute. There were 10–13 marginal scutes on each side of the carapace; 11 per side (52%) were most common, followed by 12 per side (29%). Vertebral scutes varied from 5 to 7 per tortoise but most tortoises (70%) had 5 vertebral scutes. The vertebral, and to a lesser extent the pleural scutes, were strongly raised in a conical or pyramidal fashion. This pyramiding was more prominent in females than in males and was reflected in the underlying carapacial bones (as evidenced by dead specimens found at the site). The axillaries (1–5) and inguinals (1–7) of the bridge were highly variable and often difficult to see when appendages were in the way. The plastral formula varied among females, males, and juveniles, mainly with regard to the placement of the gular, femoral, and anal scutes. In all cohorts, the abdominal scute seam was longer than all other scutes. The humeral scute seam was longer than all remaining scute seams in male tortoises, but not in females or juveniles. In males, the gular and femoral scute seams were longer than the anal and pectoral scute seams, while in females and juveniles, the anal, gular, and femoral scute seams were of similar length. For females, but not for juveniles, the femoral scute seam was longer than the seam at the pectoral scutes. Overall, the plastral formula for adult P. t. tentorius can be summarized as Ab > H > G ≥ A > < F > P.
Body Size and Population Size Structure. — Psammobates t. tentorius is a small tortoise and the SCL ranged from 2.9 to 14.7 cm (Table 2). Based on 79 individuals found, the population size structure was skewed toward the larger individuals (Fig. 1) and 88.6% of the tortoises were adults. Few tortoises (3.8%) were larger than 14 cm, while only 6.3% of the tortoises measured less than 8 cm.
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World's Turtle and Tortoise Journal 22, 1; 10.2744/CCB-1566.1
Body mass for juveniles, males, and females increased allometrically (not in direct proportion) with SCL (all slopes < 3, p < 0.05; Fig. 2). The regression slopes of body mass on SCL were similar for the three cohorts but their elevations differed (ANCOVA: F2,74 = 19.3, p < 0.0001; with BM F > M at p < 0.0001). At a similar SCL, females had a higher body mass than males had (t66 = 4.67, p < 0.0001), while the elevations for males versus juveniles and females versus juveniles were not different.
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World's Turtle and Tortoise Journal 22, 1; 10.2744/CCB-1566.1
Growth rings of adult tortoises become worn, so we could only count growth rings for 36 tent tortoises. Straight carapace lengths of each group (F, M, J) were correlated with their growth rings (Fig. 3). The log–log regression slopes for the three cohorts were similar but elevations differed (F2,31 = 33.6, p < 0.0001). The female regression line was higher (i.e., SCL was longer per growth ring count) than those for males (t24 = 6.09, p < 0.0001) and juveniles (t19 = 6.56, p < 0.0001). The elevation of the male regression line was higher than that of the juvenile line (t18 = 3.37, p < 0.005).
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World's Turtle and Tortoise Journal 22, 1; 10.2744/CCB-1566.1
Sexual Dimorphism and Shell Allometry. — Females were bigger than males in most morphometrics (Table 2), ranging from SCL to BM. The sexual size index (SDI) for the study site population was 1.27 while the compressed-modified SDI was 0.27. The plastral concavity of males was slight and visible on the hind lobe (i.e., on the abdominal, femoral, and anal scutes) in 77% of 31 males. The supracaudal scute of females projected off the posterior carapace at a 45° angle, whereas that of males hooked down distally to give a 90° angle (exhibited by 90% of males). Males had longer tails than females had (exhibited by 67% of males) and, when withdrawn, the male's tail extended laterally beyond the tip of the anal scute.
The shape of males and females differed as shown by the relationships between SH and SW on SCL (Fig. 4). Shell height was strongly correlated to SCL for juveniles (SH = 0.589 SCL – 0.090, r2 = 0.965, F1,7 = 192, p < 0.0001), males (SH = 0.375 SCL + 1.44, r2 = 0.517, F1,37 = 39.7, p < 0.0001), and females (SH = 0.506 SCL + 0.751, r2 = 0.821, F1,29 = 133, p < 0.0001). Among the three cohorts, regression slopes just failed significance (ANCOVA: F2,73 = 3.11, p > 0.05) for differences with p = 0.05 at F2,73 = 3.12. Post hoc tests indicated that juveniles had a steeper slope than males had (p < 0.05). Regression elevations for sexes differed and the shell height of female tent tortoises was greater than for males of similar SCL (t67 = 6.25, p < 0.0001; Fig. 4A).
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World's Turtle and Tortoise Journal 22, 1; 10.2744/CCB-1566.1
There was a strong correlation between shell width and SCL for juveniles (SW = 0.679 SCL + 1.24, r2 = 0.965, F1,7 = 192, p < 0.0001), males (SW = 0.554 SCL + 2.13, r2 = 0.664, F1,37 = 73.3, p < 0.0001), and females (SW = 0.686 SCL + 1.19, r2 = 0.857, F1,29 = 174, p < 0.0001). Regressions slopes among the three cohorts were similar but regression elevations were not equal (F2,75 = 13.8, p < 0.0001). For males and females of similar SCL, the shells of females were wider than were the shells of males (t67 = 3.66, p < 0.001; Fig. 4B).
Plastron length was also strongly correlated to SCL for juveniles (PL = 0.792 SCL + 0.443, r2 = 0.991, F1, 7 = 801, p < 0.0001), males (PL = 0.796 SCL + 0.102, r2 = 0.767, F1,36 = 118, p < 0.0001), and females (PL = 0.798 SCL + 0.537, r2 = 0.924, F1,29 = 355, p < 0.0001). Regression slopes among the three cohorts were similar but elevations differed (F2,74 = 16.7, p < 0.0001). For males and females of similar SCL, plastron length was greater for females than for males (t66 = 3.70, p < 0.001; Fig. 4C). The width of the gular scutes (on the plastron) was correlated to SCL for juveniles (GW = 0.122 SCL + 0.459, r2 = 0.776, F1, 7 = 24.2, p < 0.002), males (GW = 0.253 SCL + 0.559, r2 = 0.421, F1,37 = 26.9, p < 0.0001), and females (GW = 0.102 SCL + 0.817, r2 = 0.222, F1,28 = 7.98, p < 0.009). The regression slopes for gular width on SCL differed significantly (F2,72 = 3.16, p < 0.05); the slope was higher for males than for females (t65 = 2.40, p < 0.01), but slopes were similar for juveniles and females, and for juveniles and males. Although gular length correlated with SCL in female tent tortoises, gular length was not correlated (p > 0) with SCL in either males or juveniles.
The posterior shell opening of tent tortoise cohorts differed in relation to body size. Anal width scaled to SCL for juveniles (AW = 0.286 – 0.114, r2 = 0.899, F1, 7 = 62.5, p < 0.0001), males (AW = 0.298 SCL – 0.114, r2 = 0.367, F1,37 = 21.4, p < 0.0001), and females (AW = 0.234 SCL + 0.159, r2 = 0.468, F1,29 = 25.5, p < 0.0001). The slopes were equal among the cohorts, but elevations differed (F2,75 = 13.8, p < 0.0001); anal width was larger for males than for females (t67 = 3.49, p < 0.001; Fig. 5A). Anal gap also scaled to SCL for juveniles (AG = 0.179 – 0.233, r2 = 0.964, F1,7 = 188, p < 0.0001), males (AG = 0.131 SCL + 0.283, r2 = 0.161, F1,37 = 7.07, p = 0.0115), and females (AG = 0.257 SCL – 1.22, r2 = 0.668, F1,29 = 58.3, p < 0.0001; Fig.6B). Among the three cohorts, slopes and elevations for the regression lines were equal. However, when comparing regression lines of only males and females, slopes differed with females having a slope steeper than the males slope (t66 = 2.06, p < 0.05; Fig. 5B).
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World's Turtle and Tortoise Journal 22, 1; 10.2744/CCB-1566.1
Citation: Chelonian Conservation and Biology: Celebrating 25 Years as the World's Turtle and Tortoise Journal 22, 1; 10.2744/CCB-1566.1
DISCUSSION
Population Characteristics. — Psammobates tentorius in our study site had a relatively low density (numbers relative to search effort and density per se). The density is comparable to that found for P. geometricus and in some populations of C. angulata (Hofmeyr 2009; Hofmeyr et al. 2012). However, C. angulata has very high-density populations on Dassen and Robben islands (Hofmeyr 2009).
McMaster and Down (2006) and Asefa et al. (2020) also found very low densities for S. pardalis in semi-arid Nama-Karoo habitat and in savanna or semi-arid biomes of Ethiopia, respectively. They suggest that tortoise population numbers in savanna or semi-arid biomes are expected to be lower than estimates of densities in the more mesic habitats, which would be consistent with our findings for P. t. tentorius.
Four demographic factors influence population sex ratios: 1) hatchling sex ratio, 2) differential mortality of sexes, 3) differential rate of immigration and emigration of the sexes, and 4) difference in age of maturity of the sexes (Gibbons 1990). Like P. geometricus and C. angulata, P. t. tentorius in our study was found to have an equal (1:1) adult sex ratio (Hofmeyr et al. 2012; B. Henen, pers. comm., February 2021). There have been several reviews of adult sex ratios for freshwater turtles (Bury 1979; Gibbons 1990; Ewert and Nelson 1991; Georges et al. 2006). Sex ratio data indicate most freshwater turtles tend to have a 1:1 adult sex ratio. However, given that P. t. tentorius is a sexually size dimorphic species in which males are significantly smaller than females, one may anticipate males to mature at a younger age leading to male-biased sex ratios. The discrepancy maybe attributed to sampling bias (Bury 1979; Gibbons 1990; Georges et al. 2006), with larger individuals detected more readily.
The population size structure was skewed toward adults. The low number of juveniles seen during the study may reflect a sampling bias as reported in other turtle studies (Dodd 1997). Juvenile numbers were most likely underestimated because of the difficulty in researchers seeing them amongst the Karoo vegetation (smaller size facilitated completely burying themselves in leaf litter) and their ability to hide from predators as a survival mechanism (Nafus et al. 2015).
Morphology of Tent Tortoises. — Considering that no published data on the population ecology of P. t. tentorius existed, we wanted to document this first group of animals in detail for future comparisons, especially given the need of a systematic review of this species. Here, in conjunction with this aim, we offered an extensive morphological description of the tortoises.
Tent tortoises did not burrow and would normally shelter under vegetation (i.e., small bushes found in the open fields of the Karoo). Their scutes exhibit a yellow radiation pattern on a dark scute. The contrasting dark and light patterns of the carapace scutes blend well with the disruptive patterns cast by shadows at the base of the bushes they use as refugia.
Loveridge and Williams (1957) used plastral pattern, which they considered a geographically differentiated morphological characteristic to group P. tentorius into the currently recognized 3 subspecies. For our population, this characteristic was uniform across individuals and is consistent with the description for the subspecies (i.e., “tentorius (Bell, 1828) has a sharply bounded plastral pattern which is little indented or quite intact”). Recently Zhao et al. (2020) investigated whether the microsatellite DNA-based genetic structure derived for the P. tentorius species complex corroborated the phylogeny retrieved using mitochondrial (mtDNA) and nuclear (nDNA) DNA sequence data. However, a review of the systematics of the genus Psammobates using a total evidence approach (i.e., using morphological, ecological, behavioral, and genetic characters) is still overdue and necessary.
Body Size and Age. — The smallest juvenile tortoise (SCL = 2.9 cm, BM = 8 g) was found on 5 April 2003. Psammobates t. tentorius at the study site laid eggs with an average mass of 12.87 ± 1.57 g and oviposition occurred from October through June (Leuteritz and Hofmeyr 2007). Given this individual's mass and size it probably hatched recently—most likely in March, which is the hottest month with highest rainfall at the study site. The smallest gravid female was 11.5 cm, which makes her approximately 14–15 yrs old.
The absolute data and allometric analyses indicate female shells are longer, wider, and higher than those of males, so overall females are bigger and heavier. Consequently, females must assimilate more nutrients, or allocate nutrients differently, than do males. This size pattern seems consistent with that of the congener, P. geometricus (Henen et al. 2017), which inhabits more mesic habitat. A larger female size for both species likely accommodates the much larger gametes females must create, compared with that of males, and the digestion to support the vitellogenesis, egg formation, and nesting (see also Bonnet et al. 2001, among others).
The growth rates (SCL to growth rings; Fig. 3) are high for small juveniles but plateau with age. Male size plateaus with adult growth rates being very slow. Female size plateaus at larger sizes than that of males, females seem to have slightly higher adult growth rates. A continued growth in adulthood may increase reproductive ability as females age.
The largest female was 14.7 cm, which makes her approximately 42 yrs old. The largest male was 11.6 cm, which makes him approximately 42 yrs old. The data (Figs. 1, 4, and 5) indicate that males and females can be distinguished at approximately 8–10 cm, when they are approximately 8–10 yrs old. As for other species, males may mature earlier than do females (Gibbons 1990; B. Henen, pers. obs.).
Sexual Dimorphism. — Our data confirm that the southern population of P. t. tentorius is sexually size dimorphic; that is, females are significantly larger than males. The shape or morphology of the carapace and plastron of the two sexes differ—females' being higher and wider than those of males with similar SCL. When they reach sexual maturity, shell height and shell width of females increased more relative to SCL than in males, and shell height and width of females are greater than SH and SW of males with similar SCL. Thus, the larger body size in females influences their ability to be able to carry more eggs, larger eggs, or a combination of both, resulting in larger clutch volumes of eggs. Females also had a larger anal gap (AG) to anal width (AW) opening, which allows for easier passage of the elliptical egg (egg length:width ratios of 1.32 ± 70.20) at oviposition (Leuteritz and Hofmeyr 2007). Consequently, the anal gap helps the tortoises lay big eggs, which produce larger hatchlings that are likely to have higher survival rates in their arid environments.
The gular scute of males is wider than the gular of females relative to their body size. We observed no male combat (i.e., fighting) during the study. However, males, sometimes multiple at one time, often followed females for a few days during courtship. This following behavior could facilitate many attempts to copulate and inseminate females, and serve as mate guarding from competing males. A wider gular may allow more space (width) for the head. It may be that the wider gular of males is an adaptation that allows them to sway their heads from side to side more easily as part of their courtship behavior.
We also found males have a smaller plastron than females with similar SCL (Fig. 4C), implying that males have larger shell openings. The smaller body size and larger shell openings of males may be part of an adaptation to a mating strategy employed by this species (see also Bonnet et al. 2001; Leuteritz and Gantz 2013). Small size may enable greater mobility and search distances by males, which may aid to the search for females, refugia from weather and predators (e.g., Henen et al. 2017), and ability to disperse (Berry and Shine 1980; Bonnet et al. 2010). Male P. t. tentorius would actively seek mates. In observing mating behavior we often noted that 2 or 3 males would follow the female around rocks and vegetation in a “train” waiting to mate with her or one male would spend several days following a female. Unlike turtles or tortoise in which males are the same size or bigger (e.g., Astrochelys radiata, S. pardalis, Centrochelys sulcata, and C. angulata) and exhibit forced insemination (Berry and Shine 1980; Leuteritz and Gantz 2013), P. t. tentorius males appear to wait for the female to be receptive to copulation.
Finally, the AG is similar for the two sexes, but AW differed (males being larger per unit SCL; Fig. 5A). Thus, additional space for a larger tail of the male tent tortoise is primarily in the width and not so much in the length of the posterior opening. This greater space, combined with the male's plastral concavity, facilitates copulation, especially given size discrepancies between males and female tortoises. This feature is consistent with findings for other tortoise species (e.g., Bonnet et al. 2001).
Tortoise Communities. — Although the study focused on P. t. tentorius, it should be noted that 3 other tortoise species (C. angulata, S. pardalis, and Homopus boulengeri) are historically known to occur sympatrically with P. t. tentorius in the Prince Albert area (Boycott and Bourquin 2000). Despite the overlap at our site, tortoises tended to partition their habitat. Chersina angulata utilized drainages with tall vegetation. Stigmochelys pardalis was found in the adjacent small rocky outcrop areas (koppies), whereas P. t. tentorius was found in the open fields. No H. boulengeri were ever encountered in any of the habitat types, but this is not revealing considering search effort was focused on open field habitat and not on koppies where H. boulengeri are known to occur (Branch 1998).
Size-class structure of a Psammobates tentorius tentorius population near Prince Albert, South Africa.
Relationships (power curves) of body mass (BM) to straight carapace length (SCL) for Psammobates tentorius tentorius groups (females, males, juveniles) near Prince Albert, South Africa. Log–log regressions were significant for juveniles (log BM = 2.66 log SCL – 0.301, r2 = 0.987, F1,7 = 519, p < 0.0001), males (log BM = 2.59 log SCL – 0.259, r2 = 0.850, F1,36 = 205, p < 0.0001), and females (log BM = 2.64 log SCL – 0.250, r2 = 0.936, F1,29 = 422, p < 0.0001).
Relationships (power curves) of straight carapace length (SCL) to growth rings (GR) for Psammobates tentorius tentorius cohorts near Prince Albert, South Africa. Log–log regressions were significant for juveniles (log SCL = 0.221 log GR + 0.668, r2 = 0.956, F1,7 = 152, p < 0.0001), males (log SCL = 0.171 log GR + 0.786, r2 = 0.343, F1,11 = 5.74, p = 0.0354), and females (log SCL = 0.231 log GR + 0.793, r2 = 0.459, F1,12 = 10.2, p = 0.00777).
Relationships of (A) shell height, (B) shell width, and (C) plastron length with straight carapace length for Psammobates tentorius tentorius males and females near Prince Albert, South Africa.
Relationships of (A) anal width, and (B) anal gap with straight carapace length for Psammobates tentorius tentorius males and females near Prince Albert, South Africa.
Sexual dimorphism of carapace and plastron of (a) female (top) and (b) male (bottom) for Psammobates tentorius tentorius near Prince Albert, South Africa.
Contributor Notes
†Deceased
Handling Editor: James O. Juvik
