Amazonian chelonians, mainly species of the genus Podocnemis, are one of the richest biological resources of Brazil. For centuries, the high quality of their meat and eggs has been essential to support the people who live in the region (Veríssimo 1895; Smith 1979; Cantarelli 1997). The rational use of these resources may offer alternatives for the social and economical development of the region. Nevertheless, to manage turtle populations and produce appropriate conservation plans, it is necessary to document aspects of their breeding biology. Although the distribution of species of Podocnemis covers a large area of the Amazon Basin, very little is known about their reproductive strategies.

Six species of Podocnemis are currently recognized, all restricted to South America: P. expansa and P. unifilis are widely distributed in the Amazon and Orinoco basins; P. sextuberculata is restricted to the Amazon basin; P. erythrocephala occurs mainly in the Rio Negro basin; P. vogli in the Orinoco; and P. lewyana in the Rio Magdalena (Pritchard and Trebbau 1984; Iverson 1992). Studies of reproductive biology have been published on P. expansa (Vanzolini, 1977; Alho et al., 1979; Alho and Pádua, 1982a,b; Pádua and Alho 1982; Valenzuela 2001), P. unifilis (Foote 1978; Pritchard and Trebbau 1984; Thorbjarnarson et al. 1993; Souza and Vogt 1994; Thorbjarnarson and Silveira 1996; Escalona and Fa 1998), P. sextuberculata (Vanzolini and Gomes 1979; Pezzuti and Vogt 1999), and P. erythrocephala (Castano-Mora et al. 2003).

During the beginning of the dry season, when water levels begin to recede in Amazonian rivers, podocnemid turtles migrate to nesting areas—which may be the shores of streams or rivers, river islands with sandy beaches, or riverine sandbanks.

This paper investigates the breeding biology of P. sextuberculata at Reserva Biológica do Rio Trombetas, state of Pará, Brazil. We studied nesting behavior, biometry of reproductive females, eggs and hatchlings, morphometry of nesting locations, and the relation between incubation periods and hatching success in transferred and natural nests.

Methods

We obtained data from September to December during 2000 and 2001 at Reserva Biológica do Rio Trombetas (Fig. 1). The reserve has an area of 385,000 ha and is located in northwestern Pará, municipality of Oriximiná, on the left shoreline of the Rio Trombetas (1°46′S, 55°52′W).

View Figure

• Download as Powerpoint
• Download Figure

In the reserve, nesting takes place mainly on sandbanks exposed by decreasing water levels. Observations and data were collected at Tabuleiro do Farias and Tabuleiro do Jacaré (Fig. 1). See Pádua and Alho (1982) for additional descriptions of the Reserve.

During September–October 2000–2001, nesting was monitored on the beaches at night (1900 to 0600 hours). Using a caliper and a balance (± 100 g), we measured 24 nesting females (maximum straight-line length and width of carapaces and plastrons, maximum height, head width, tail length – from posterior edge of plastron to tip of tail) and weight; of these, only 9 females oviposited. An additional 187 clutches were located by following female tracks; of these, 141 clutches were measured (clutch size and mass, egg length, width, and mass). Egg volume was estimated based on the method of Vanzolini (1977). Dimensions of 59 nests (total depth, depth of the top egg in relation to the surface, diameter of the nest mouth, length and width of the egg chamber) were measured.

Eggs were removed from the nest and marked with a pencil on the day following oviposition so that their original orientation could be maintained, and 129 nests were transferred and reburied in a safer location within a few hours of oviposition; 55 nests were left in situ to incubate without disturbance. A Mann-Whitney U-test was used to compare the distribution of incubation times and the percentage of live hatchlings for transferred and natural nests.

Twenty-four nests were enclosed with wire; hatchlings captured after emergence from these nests were measured (maximum straight-line length and width of carapaces and plastrons, maximum height, head width, tail length – from posterior edge of plastron to tip of tail) and weighed immediately after hatching, using a caliper (± 0.05 mm) and Pesola balance to the nearest 1 g.

Results

During 2000, P. sextuberculata nested between 27 September and 3 November, and during 2001, from 13 September to 15 October. The nesting behavior of 7 P. sextuberculata females was observed. Females were found wandering on the beaches from 2030 to 0430 hours. Nests were located at distances of 1–620 m from the water. Actual oviposition times varied from 45 to 90 minutes.

Measures of carapace length (CL) and width (CW) and plastron length and width, height, head width, tail length, and weight were taken from 24 females captured on the beach after nesting (Table 1). Clutch size varied from 8 to 24 eggs (mean: 15 eggs) in both years (Table 2). Eggs of P. sextuberculata are ellipsoid, and the pergaminous shell is clearly flexible. Table 3 presents biometric data obtained from hatchlings. Carapacial deformities precluded data collection for a large number of hatchlings.

Table 1. Biometry of female Podocnemis sextuberculata from Reserva Biológica do Rio Trombetas, Pará, Brazil, September–October 2000–2001.

View Fullscreen

Table 2. Podocnemis sextuberculata egg and hatchling morphometrics from Reserva Biológica do Rio Trombetas, Pará, Brazil, September–October 2000–2001.

View Fullscreen

Table 3. Biometry of hatchlings of Podocnemis sextuberculata from Reserva Biológica do Rio Trombetas, Pará, Brazil, September–October 2000–2001.

View Fullscreen

Females of P. sextuberculata always nest in sandbanks totally exposed to the sun. The data for 59 nests are given in Table 4. Although many nests had asymmetrical shapes, the egg chamber always was larger than the mouth.

Table 4. Summary of data from 59 nests of Podocnemis sextuberculata from Reserva Biológica do Rio Trombetas, Pará, Brazil, September–October 2000–2001.

View Fullscreen

Linear regressions between the number of eggs and female CL (n = 20; F = 10.32; R² adjusted = 32.92%; p < 0.01) and between the number of eggs and female CW (n = 20; F = 7.43; R² adjusted = 25.29%; p = 0.01) were positive and significant. There also was a significant regression between female CL and the mean egg weight (n = 12; F = 6.87; R² adjusted = 34.81%; p = 0.02); between female CL and the mean egg volume (n = 12; F = 7.52; R² adjusted = 37.23%; p = 0.02); and between female CL and the clutch mass (n = 12; F = 28.88; R² adjusted = 67.53%; p < 0.01).

Comparison of transferred and natural nests indicated that manipulated nests had a longer incubation time (U = 2819.5; p = 0.02), with mean incubation times 57 days for the natural nests (range 48–64 days). The percentage of live hatchlings produced from transferred nests (49%) was significantly lower (U = 2192.5; p < 0.01) than from natural nests (73%).

Eggs were consumed by the lizard (Tupinambis teguixin), and some were infested with insect larvae during several stages of the incubation period. High levels of egg collection by local inhabitants also occurred during this study. During the nesting period, almost every beach on the Rio Trombetas (about 14) were monitored daily by local people who were looking for eggs and females. Adults also were captured from the river with nets, harpoons, and hooks.

Discussion

Nesting of P. sextuberculata resembles that described for P. unifilis (Foote 1978) with some differences: P. sextuberculata always nests in sandbanks totally exposed to the sun, while P. unifilis nests mainly in beaches located near water edges. Nests of P. unifilis also were found in the sandbanks in Reserva Biológica do Rio Trombetas, but they were located closer to or covered by vegetation. Unlike P. expansa, P. sextuberculata does not nest in aggregations. Pezzuti and Vogt (1999) commented that P. expansa and P. sextuberculata, which specifically nest only on exposed river sand beaches, have serious conservation problems because nests are easily detectable by local inhabitants.

Vanzolini and Gomes (1979) concluded that clutch investment of P. sextuberculata is not proportional to the individual's size within its age class, but to the mean body size of the age class. Their data on clutch size for 17 females, ranging from 219 to 280 mm in plastron length (15 smaller than 236 mm), ranged from 7 to 21 eggs. Our data on female and clutch size agree with these results, but the linear regressions we obtained between female CL and egg and clutch sizes suggest that larger females produce heavier and larger clutches. Congdon and Gibbons (1985) commented that both within and among species, the size of nests and the clutch size are in general positively related to the body size of the female. Pezzuti and Vogt (1999) found clutches in the Japurá River, Amazonas, Brazil, ranging from 6 to 25 eggs (n = 24); mean egg length was 41.4 mm, mean egg width was 28.9 mm, and mean egg mass was 19.6 g.

We verified that embryo survival is affected by handling of the eggs, but other factors might be influencing survival rates. Nest-site selection by females, type and soil texture, moisture, and temperature are important factors for embryo survival (Congdon and Gibbons 1990). Packard et al. (1991) also noted that soil humidity of nests affects hatching success, and mortality may be higher among embryos developing in nests placed in dry soils. Bonach et al. (2001) reported that although egg viability of P. expansa decreases naturally during incubation period, early translocation increases embryo mortality and later translocation during the second third of the incubation period does not harm eggs and may be an effective tool for ranching programs of the species in the Brazilian Amazon. They recommended further research on the possible influences of egg translocation on sex determination of embryos and fitness of hatchlings.

Predation affected the survival of eggs of P. sextuberculata throughout the area. Tupinambis detected fresh nests, principally near vegetation, and were responsible for destruction of approximately 30 nests (completely or partially destroyed). Soini (1995) commented that Tupinambis was the most important predator and had a great ability to detect nests of P. expansa, P. uniflis, and P. sextuberculata—even older ones—on the Pacaya-Samíria in Peru. Vogt (1981) noted egg predation by fly larvae, and Pezzuti and Vogt (1999) found Sarcophagidae larvae causing the death of hatchlings from infested eggs.

The main exploitation pressure on turtles and their eggs at Reserva Biológica do Rio Trombetas is due to local inhabitants. Nests and adults in the process of egg laying are easily detectable and collected by locals. The high level of collection may result in changes in nesting behavior, because many people walk on the beaches looking for nesting females. Furthermore, it can also cause a serious decrease in nesting densities because nests are placed in different atypical places, which may affect recruitment and also change hatchling sex ratios. Most importantly, this exploitation is intense enough to raise concerns over its potential sustainability, and management efforts to control the local harvest of P. sextuberculata should be considered.