In reptiles, hatching and emergence within a single egg clutch may either be synchronized or separated by days, weeks, or even months (Andrews 2004). A number of different factors cause variation in the duration of the incubation period among eggs in the same clutch, including the size of the clutch and the order in which the eggs were laid. The synchronization of hatching has evolved in several animal species (O'Donoghue and Boutin 1995), although development in oviparous species, such as chelonians, depends on microenvironmental conditions (Colbert et al. 2010). These conditions may affect not only the timing of hatching and emergence, but also the interval between the two events, an important stage in embryonic development (Godfrey and Mrosovsky 1997). The implications of this delay in the emergence of hatchlings are poorly understood, but they may have a direct influence on the probability of predation, energy loss, and infestation of the eggshell by dipterans (Gibbons and Nelson 1978; Hays et al. 1992; Godley and Kelly 1996; McGowan et al. 2001).

Chelonian hatchlings usually emerge during the early evening (Limpus 1985). They normally emerge simultaneously, which probably reflects the more or less constant rate of incubation within the nest and may be advantageous due to the reduction in the probability of predation or the more efficient excavation of the exit hole (Carr and Hirth 1961; Andrews 2004). However, environmental conditions may include gradients of humidity or temperature, for example, which may cause differences in the timing of hatching by a number of days or even longer (Andrews 2004).

The timing of hatching and the hatching and emergence patterns of neonates have not been studied previously in the yellow-spotted river turtle, Podocnemis unifilis, a species found throughout the Amazon and Tocantins-Araguaia basins, Brazil. In this study, we evaluated the hatching and emergence patterns of P. unifilis hatchlings in the várzea floodplains of the lower Amazon River at Santarém, in the Brazilian state of Pará. We tested the hypothesis that, because the physical and environmental microhabitat is particular to each nest and these characteristics may influence the hatchling hatching and emergence, there is variation in the hatching and emergence patterns of P. unifilis hatchlings within and between nests.

METHODS

Study Area

The Amazonian várzea is composed primarily of seasonally flooded marshes interspersed with areas of forest (Ribeiro 2007), where recurrent floods cause erosion of riverbanks and formation of sandbanks (Irion et al. 1997). The peak of the rainy season typically occurs between April and June and the dry season between October and December (CPRM 2010). The soils are alluvial and rich in nutrients and organic material deposited by annual floods (Ribeiro 2007).

The present study was conducted in a várzea forest of the lower Amazon River, known as Taboleiro da Água Preta (lat 2°09′12.5′S, long 54°37′51.7′W), located near the confluence of the Amazon and Tapajós rivers (Fig. 1). The Taboleiro da Água Preta is an important nesting site for chelonians located in the municipality of Santarém, eastern Pará (Brazil), and has been protected by the local community since 1991.

View Figure

• Download as Powerpoint
• Download Figure

Monitoring

We monitored nests of P. unifilis in a 25.60-ha section of the Taboleiro da Água Preta in 2007 and a 29.68-ha section in 2009. Nests were located by following fresh female tracks early in the morning after egg-laying had taken place. The location of each nest was recorded using a hand-held global positioning system (Etrex H, Garmin®) and marked with a numbered stake. The following data were collected for each nest on the day it was found: date, number of days elapsed since the beginning of the nesting season, distance (m) to the edge of the river channel, distance (m) to the nearest vegetation area, vegetation cover over the nest (%), and height (cm) of the nest in relation to the level of the water in the Lago da Água Preta channel. The vegetation cover was measured using a model C sphere densiometer (Lemmon 1956) and the angle of the substrate surface was measured using a clinometer.

All clutches were included in unpaired groups for monitoring of either hatching or emergence. Eggs in the clutches of the hatching group were counted at initial inspection, whereas those in the emergence group were counted after the emergence of hatchlings. We selected groups of very close and similar nests, monitoring in each pair one nest for hatching and the other for emergence, to ensure a comparison of nests with similar characteristics in each group (e.g., height above the river, distance from vegetation, type of cover, and granulometry of the substrate).

In each group of nests, the incubation period (days between oviposition and hatching of first hatchling, according to Godfrey et al. 1996) was monitored. From the 40th d onward, these nests were checked daily to verify whether hatching had taken place.

In 2009, the hatching sequence was also monitored in order to establish whether hatching in a given nest took place simultaneously or over a longer period and to verify the interval (days) between the hatching of the first and last hatchlings. This procedure was not considered for 2007 due to the small sample size. The nests were opened for inspection each day during the early hours of the morning starting on the 40th d of incubation, based on the known mean incubation period of 60 d for the species (Souza and Vogt 1994). Eggs were pulled out, put back in the same order inside a nylon net, and then put back in egg chamber. There was no noninvasive method to check hatching chronology in natural nests without manipulation. During each inspection, the numbers of unhatched, ruptured, and hatched eggs were recorded.

In nests monitored for emergence, the time in days between oviposition and first hatchling emergence (emergence period) was measured. From the 50th d of incubation onward, the nests were externally checked daily, without any interference or manipulation (no opening of the nest chamber), until the day when the first hatchlings emerged. Emergence was confirmed by the observation of hatchling tracks. When emergence was confirmed, the nest was examined carefully in order to verify whether all the hatchlings had emerged or not. Emergence success of a nest was determined by subtracting the number of hatchlings still in the nest from the number of eggshells found in the nest and dividing the result by the original clutch size.

To estimate the difference between the hatching period and the emergence period (first hatching–first emergence interval), pairs of hatching-monitored and emergence-monitored nests were used. The hatching–emergence interval (days) was determined by subtracting the hatching period from the emergence period. Means of hatching period, emergence period, and hatching–emergence interval are reported with standard deviations.

The vegetation cover variable was transformed by arcsine square root to achieve a normal distribution. The relationship between clutch size and days before first hatching was tested using simple linear regression. The relationship between the interval (in days) to hatching of the first hatchlings and days before first hatching was tested using Spearman's correlation coefficient. The relationship of emergence period (days) with the independent variables distance to the river and vegetation, height, inclination, plant cover, and the number of days from oviposition was tested using multiple regression, analyzed separately for each year.

RESULTS

In 2007, 29 nests were monitored for hatching and 17 for emergence. In 2009, 78 nests were monitored for hatching, with the timing of hatching being measured in 69 of these, while 27 were monitored for emergence.

In 2007, females nested between 4 October and 1 November and hatching was observed from 26 November through 26 December. The mean incubation period was 58.9 ± 3.64 d (n  =  29), with a range of 54–66 d. Hatchlings from all nests emerged between 2 December 2007 and 7 January 2008, i.e., over a period of 37 d (Fig.  2, top). In 2009, the nesting period lasted from 21 September through 20 October and hatching occurred from 12 November to 25 December. The mean incubation period was 60.4 ± 5.22 d (n  =  78, range 51–74 d) and the hatchlings emerged over 32 d between 25 November and 27 December (Fig. 2, bottom).

View Figure

• Download as Powerpoint
• Download Figure

In 2009, eggs in the upper part of the nests hatched before those in the lower part of the nest. The first hatchlings took an average of 1.5 ± 1.40 d (range 0–9 d) to hatch completely after eggshell pipping and mean hatching interval was 4.0 ± 2.40 d (range 0–10 d, n  =  69). Most eggs (52%) hatched on the first day (Fig. 3). There was a significant positive relationship in 2009 between clutch size and the interval in days between first and last hatching (simple linear regression: F  =  4.80, n  =  69, p  =  0.030; Fig. 4). By contrast, the number of days passed since the beginning of the nesting season did not have a significant effect on hatching interval (Spearman's r_s  =  0.020, n  =  69, p  =  0.87).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

In 2007, the mean emergence period was 69.9 ± 6.44 d (range 55–83 d, n  =  17) and mean proportional emergence success was 0.75 ± 0.239 (range 0.19–1.00, n  =  17). Hatchlings spent a variable number of days in the nest before emerging. The mean hatching–emergence interval was 10.5 ± 5.63 d in 2007 (n  =  17) and 7.1 ± 6.77 d in 2009 (n  =  27). In most cases (89%), the hatchlings did not all emerge on the same night, with some hatchlings remaining in the nest and emerging on 1 to 3 subsequent nights. In 2009, the mean emergence period was 68.1 ± 4.97 d (range 58–78 d, n  =  27) and mean proportional emergence success was 0.82 ± 0.172 (range 0.32–1.00, n  =  27). Most of the time all the hatchlings emerged together on the same night, but in 2 cases (7.4% of the total), some hatchlings emerged on the following night.

In 2007, emergence period was significantly negatively related to the distance of the nest to the nearest vegetation, whereas there was a significant positive relationship with the number of days passed since the beginning of the nesting season in both years (Table 1). Nests laid later and closer to vegetation had longer emergence periods in both years (Figs. 5 and 6). No other variables were significantly related to emergence period (Table 1).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Table 1.  Influence of environmental variables on the emergence period (days) of Podocnemis unifilis hatchlings in the várzea floodplains of the lower Amazon River in Santarém (Brazil), in 2007 and 2009. Significant p-values for model variables are in bold type. R² was 0.777 for 2007 and 0.458 for 2009.

View Fullscreen

DISCUSSION

In reptiles, the period of transition from the live embryo inside the egg to an emerged hatchling may vary from a few hours to several days, with the hatchling remaining within the ruptured eggshell (Ferguson 1985). Colbert et al. (2010) suggested that synchronous hatching may be impeded in aquatic chelonians by the difference in the incubation period between the eggs located on the top and bottom of the egg chamber. In a study of the freshwater species Emydura macquarii in Australia, Thompson (1988) observed that the eggs located in the upper part of the nest experienced higher temperatures, resulting in an increase in developmental rates and a reduction in the incubation period in comparison with the eggs located in the lower portion of the nest, and thus, in asynchronous hatching.

The majority of P. unifilis eggs in each nest hatched synchronously in Colombian Amazonia (Páez and Bock 1998). Studies of the nests of Chrysemys picta in Illinois (Colbert et al. 2010), and the nests of Emydura macquarii in Australia (McGlashan et al. 2012) showed that less-developed embryos undergo accelerated development in order to hatch together with the more developed embryos in the same clutch. McGlashan et al. (2012) suggested that less-developed embryos responded to the presence of more-developed embryos in a clutch by increasing both metabolic and heart rates. Colbert et al. (2010) suggest that the synchronized incubation of the eggs is probably an ancestral trait in chelonians, which functions primarily to reduce the mean predation risk for the clutch.

After hatching, hatchlings of Podocnemis expansa move to the upper part of the egg chamber to await environmental conditions favorable for their exit from the nest (Alho et al. 1979; Alho and Pádua 1982; Soini and Soini 1995a). This is an important stage for the hatchlings, considering that it allows more time for the plastron to set and for the residues of the yolk sac to be absorbed (Godfrey and Mrosovsky 1997).

Data on the hatching–emergence interval in chelonians are scarce. In general, hatchlings tend to emerge at the surface at least 2 d after hatching and the interval may be affected by the depth of the nest and the compression of the sediment (Miller et al. 2003). Godfrey and Mrosovsky (1997) estimated a mean interval of 4.1 d in Caretta caretta, whereas Soini and Soini (1995b) observed that P. unifilis hatchlings in the Pacaya-Samiria National Reserve remained in the egg for 2 to 7 d following the rupture of the eggshell and waited in the nest a further 2 wk before emerging. We found that P. unifilis hatchlings remained buried for at least a week after hatching and emerged only after the complete absorption of the yolk, consistent with the observations of Soini and Soini (1995b) on this species and the findings of Alho et al. (1979) and Alho and Pádua (1982) for P. expansa.

According to Glen et al. (2005), the timing and emergence pattern of chelonian hatchlings is a crucial factor for the survival of a clutch. For example, by emerging at night, a hatchling will avoid potential diurnal predators and lethal daytime temperatures (Glen et al. 2005). Gibbons and Nelson (1978) suggested that before emergence the hatchlings wait inside the egg chamber for some environmental sign that indicates a higher probability of favorable conditions to emerge.

In the present study, the nests under influence of shading (closer to the vegetation) and nests deposited late (in the period of more rainfall) exhibited longer emergence periods in both years. With this it can be inferred that under certain conditions the emergence period is longer. The synchronous emergence including hatchlings that hatched even at different times (different developmental rates and different metabolic rates) is the strongest evidence that there is an important external factor that triggers the process, as suggested by Gibbons and Nelson (1978). Additionally, the nests presented lower temperatures compared to 2009 (Pignati 2011), which likely contributed to the longer hatching–emergence interval in 2007.

The synchronous emergence of the hatchlings on a single night, as observed in 2009, is a pattern also observed in other taxonomically distant species, such as Phrynops hilarii in Brazil (Bujes and Verrastro 2009) and Natator depressus in Australia (Koch et al. 2008). Nocturnal emergence presumably avoids the threat of potentially lethal temperatures and the presence of more effective diurnal predators (Miller et al. 2003). In addition to synchrony, the emergence of P. expansa occurs primarily on rainy nights (Alho et al. 1979; Alho and Pádua 1982; Soini and Soini 1995a). In Arkansas, 24 of 26 clutches of Apalone mutica emerged on single nights, while the other 2 emerged over 2 consecutive nights (Plummer 2007). Another clear advantage of synchronized hatching is the more efficient excavation of the exit hole, especially in the case of relatively deep nests or those in compacted substrate (Carr and Hirth 1961; Andrews 2004).

The less-developed embryos in a nest may either accelerate their development or hatch prematurely in order to emerge simultaneously with their clutch mates, diluting the risk of predation (Spencer et al. 2001). Although synchronous emergence may have the advantage of reduced energetic expenditure due to social facilitation, there may also be disadvantages, such as the progressive reduction in the energy stocks of hatchlings that remain in the nest for long periods (Carr and Hirth 1961). In Michigan, Congdon et al. (1983) found that 45% of Emydoidea blandingi hatchlings emerged synchronously on a single night, whereas the remaining 55% emerged over more than one night. Glen et al. (2005) recorded asynchronous emergence in the hatchlings of marine species in Cyprus, where hatchlings of Chelonia mydas emerged over 1–7 nights and hatchlings of Caretta caretta emerged over 1–7 nights. In Greece, hatchlings of C. caretta took 1–12 nights to emerge, with a mean of 6 nights (Houghton and Hays 2001). Overall, there is no consistent universal pattern of emergence in either marine or freshwater chelonians, despite its apparent benefits such as social facilitation and predator avoidance.

In the present study, we observed patterns that showed some differences between years. Further studies should address the probable link between hatching and emergence processes and environmental conditions such as temperature, humidity, and precipitation levels. Given the overall paucity of data on hatching and emergence patterns in chelonians, however, further, more detailed studies will be needed for a more systematic understanding of the process, particularly the influence of external factors such as ambient and incubation temperatures.