From all batches sampled, we caught a total of 425 specimens of N. atrata (males=209, females=216), with an SSD index of 0.16. The sex ratio showed no significant differences between months (χ2=15.89, P=0.145, n=425), climate season (χ2=2.30, P=0.1286, n=425), or batches sampled (χ2=0.93, P=0.61, n=425). In contrast, significant differences in snake abundance were observed between good and bad years (χ2=176.15, P<0.0001, n=425; Figure 2). Health condition was recorded as a potential variable affecting reproductive output. However, through the whole sampling period, only 11 specimens (2.6%) suffered poor condition (ticks n=1; mycoses n=8; injury=2). Therefore, these variables were excluded from analyses and we considered the N. atrata population under study to be healthy.
According to the macroscopic reproductive traits observed in 79 N. atrata females, the smallest female with vitellogenic follicles, indicating sexual maturity, was 270 mm SVL. All females larger than 270 mm SVL (48%) were in some stage of reproduction (Table 1). Females were asynchronous in reproductive stage between months or climate years (χ2 months=27.374, P=0.44; χ2 climate years=3.05, P=0.38). All reproductive stages were observed throughout the year. Oviduct width exhibited significant differences among female reproductive stages (ANOVA F=13.7, P<0.000 1), with the oviduct being less wide in previtellogenic females than the remaining reproductive stages. However, differences in oviduct width among females at ovigerous, vitellogenic, and vitellogenic-ovigerous stages were not significant (ANOVA F= 0.14, P=0.86). Thus, an oviduct width larger than 6.14±3.85 mm indicated sexual maturity. Nonetheless, a high degree of oviduct width overlap was observed among female reproductive stages, suggesting that this macroscopic character is not an accurate predictor of sexual maturity (Figure 3).
Female reproductive stage SVL (mm) Mass (g) Oviduct length (mm) Sperm in oviduct Uterine scar No. primary follicles No. secondary follicles No. eggs Previtellogenic 205.05
Absent Absent Vitellogenic 313.57
Absent Ovigerous 335.08
Table 1. Mean values of female reproductive features
In the 79 adult females examined, primary follicles were present throughout the year. However, in mid-June, most follicles began to enlarge, reaching a maximum in November to January until the beginning of the dry season (Figure 4A). Likewise, the greatest abundance of primary follicles was observed in January to March and September to November. Consequently, secondary follicles were only observed from April to November (rainy season), with abundance and size increasing gradually, reaching the greatest size in the October to November period (Figure 4B).
We observed mating signals at all female reproductive stages. However, the number of immature females exhibiting mating signals was significantly lower than that of mature females (χ2=20.14, P<0.001, n=79). As expected, the frequency of uterine scars was significantly higher in females at ovigerous and vitellogenic stages (χ2=44.32, P<0.001, n=79). Notwithstanding, uterine scars also were observed in six previtellogenic females, four of which had sperm inside the infundibulum and two of which had sperm inside the oviduct. Mature females had sperm inside their infundibulum or oviduct almost the entire year (except December), indicating that copulation is continuous even for females not ready to mate. We observed eggs throughout the year, though the greatest abundance was recorded in September to December (Figure 4C). Based on the incubation of three clutches oviposited by three gravid females, the estimated birth-time was 108±1.41 d/egg. Clutch size ranged from 1–4 eggs (2.44±1.02, n=34). We recaptured two female snakes who produced eggs twice in the same reproductive season (first capture in July and recapture in November). Also, during the dissection of snakes collected in August to November from batches 8 and 9, we recorded five females with vitellogenic follicles and oviductal eggs simultaneously, indicating that females could produce more than one clutch per reproductive season.
Neonates were also observed throughout most of the year (except July), with three abundance peaks during the recruitment season. The greatest recruitment peak occurred during the early dry season (January–February). A second moderate recruitment peak was observed in September, and a third pronounced recruitment peak occurred at the end of the rainy season (November). The greatest dearth occurred during the first half of the rainy season (June–July; Figure 4D). Based on the birth of eight neonates, birth-size was estimated to be 114.63±10.69 mm SVL and 1.91±0.74 g of body mass. Likewise, RCM and RF were highly variable and ranged from 4.32% to 8.54% (7.21±1.14%, n=16) and 2.72% to 12.74% (7.28±3.01%, n=34), respectively.
We observed a remarkable decrease in neonates between good and bad climate years. Based on the presence or absence of the ENSO, the number of neonates declined significantly, from 57 in good years to four in bad years (10.557, P=0.001, n=61). Despite this, we observed a clear synchronization between recruitment peaks and prey abundance in years without ENSO effects. Increased snail and slug abundance coincided with increased neonate abundance over the same time period (Figure 5). This relationship was confirmed by multiple regression analysis, with neonate abundance being strongly correlated with slug and snail abundance, but not with other prey (R2=0.46, P= 0.032; Figure 6). Nonetheless, when the height of piles of palm leaves was included in the analysis, it provided a better model, explaining 55.29% (P=0.011) of the neonate abundance variability observed (Table 2).
Model R2 AIC dAIC df Nor. test P value Hom. test P value Aut. test P value Clutch size versus female reproductive drivers Clutch size vs. Var2+Var3+Var4 0.70 –90.62 0.0 1 0.07 0.35 0.63 Clutch size vs. Var2+Var3+Var4+Var5+Var6 –89.16 –1.46 1 Clutch size vs. Var1+Var2+Var3+Var4+Var5+Var6 –87.28 –2.98 1 Clutch size vs. Var1+Var2+Var3+Var4+Var5+Var6+Var9 –85.35 –5.39 1 Secondary follicles versus female reproductive drivers Secondary follicles vs. Var2+Var5+Var6 0.42 11.18 0.0 1 0.17 0.36 0.87 Secondary follicles vs. Var2+Var5+Var6+Var9 11.69 0.51 1 Secondary follicles vs. Var2+Var3+Var5+Var6+Var9 13.24 2.06 1 Secondary follicles vs. Var1+Var2+Var3+Var5+Var6 15.21 4.03 1 Sperm count versus male reproductive traits and environmental variables Sperm count vs. Var2+Var6+Var9 0.198 –166.21 0.0 1 0.07 0.34 0.2 Sperm count vs. Var1+Var2+Var6+Var9 –165.20 –1.01 1 Sperm count vs. Var1+Var2+Var6+Var9+Var10 –163.32 –2.89 1 Sperm count vs. Var1+Var2+Var6+Var9+Var10+Var11 –161.38 –4.83 1 Neonates versus prey abundances Neonates vs. Slugs+Snails 0.46 –9.06 0.0 1 0.34 0.50 0.46 Neonates vs. Slugs+Snails+Leeches –7.60 1.46 1 Neonates vs. Slugs+Snails+Leeches+Earthworms –6.41 2.65 1 Neonates versus prey abundances and height of piles of palm leaves Neonates vs. Var9+Snails 0.55 –11.59 0.0 1 0.83 0.43 0.92 Neonates vs. Var9+Snails+ Slugs –11.49 –0.10 1 Neonates vs. Var9+Snails+ Slugs+Leeches –10.55 –1.04 1 Neonates vs. Var9+Snails+ Slugs+Leeches+Earthworms –8.65 –2.94 AIC: Akaike information criterion, employed to select“best model”, was used to test whether environmental factors rather than intrinsic reproductive traits are main drivers of reproductive output. Var1: Snout-vent length, Var2: Body mass, Var3: Primary follicles number, Var4: Secondary follicles number, Var5: Fat body area, and Var6: Stomach bolus volume, Var9: Height of piles of palm leaves, Var10: Testicular volume, Var11: Width of sexual segment of kidney, and Var12: Distal width of deferent duct. R2: Proportion of variance for reproductive output explained by microenvironment or reproductive intrinsic trait variables. Nor. test: Kolmogorov-Smirnov test for normality; Hom. test: Breusch-Pagan test for homoscedasticity; Aut. test: Durbin-Watson test for autocorrelation.
Table 2. Multiple regression analysis models
Figure 5. Temporal prey abundance variation and neonates of Ninia atrata during years without ENSO effects (2014−2015)
Ninia atrata males exhibited early sexual activity. The smallest male with sperm in their testes, indicating sexual maturity, was 145 mm SVL, and the largest male without sperm was 212 mm SVL. These males represent the extreme body-size limits of sexual maturity. However, 98.60% (n=71) of males larger than 187 mm SVL had sperm in their testes and deferent duct. The smallest male with metamorphosing spermatocytes was 137 mm SVL.
Based on macroscopic examination of the male gonads, testes showed noticeable size variation throughout the year. Testicular size gradually changed, decreasing from February to August and increasing from August to November. Maximum volume was attained in April (beginning of rainy season) and October-November (end of rainy season) and was 136.4% greater than the minimum testicular volume observed in August (mid rainy season; Figure 7A). Despite a lack of mature male samples in December, January, and March, our data suggests that testicular volume declines at the beginning of the dry season but increases mid-way through.
Likewise, macroscopic sexual features such as SSK and distal end of deferent duct exhibited a similar monthly variability pattern as observed for testicular volume (Figure 7C–D). Indeed, monthly variability of these traits was closely related to testis size (RSSK2 =0.68, P<0.000 1, n=71; Rdeferent duct2= 0.45, P<0.000 1, n=71). In contrast, sperm production was not correlated with monthly variability in macroscopic male sexual features (RSSK2 =0.031, P=0.15, n=69; RTestis volume2=0.028, P=0.18, n=71; Rdeferent duct2=0.033, P=0.14, n=71), indicating that testicular volume, SSK hypertrophy, and deferent duct width are not concordant with spermatogenic activity. Sperm production was present for most of the year, including the dry and rainy seasons. Specifically, production increased gradually from April to November and reached a maximum in July–August (mid rainy season) without significant decline once production began (Figure 7B).
Conversely, significant differences in the size of males with either weak or prominent chin tubercles were found (ANOVA F=37.28, P<0.000 1, n=71). The weak condition was generally associated with males within the SVL range of 135–241 mm (168.16±2.02, n=12), although two larger males (SVL=276 and 280) exhibited this condition even though they had reached the minimum size of sexual maturity. In contrast, the prominent condition was generally associated with males within the SVL range of 183–354 mm, in accordance with the results that 98.60% (n=71) of males of this size have sperm in their testes and deferent duct. Despite this, three smaller-sized males (SVL=146, 154, and 175) also had prominent chin tubercles. In fact, the SVL ranges of chin tubercle condition showed a high degree of overlap (46.53%), indicating that this secondary sexual character may not be an accurate predictor of male reproductive stage.
We found notable differences in the reproductive cycles between the sexes. First, males showed a continuous cyclical pattern, in which spermatogenesis, gonads, and SSK were active throughout the year. Although they did show reduced activity during the dry and mid rainy seasons, they never displayed total regression or quiescence. In contrast, females showed a cyclical pattern in which oogenesis, gonads, and accessory organs become inactive or absent during the dry season. Second, size at sexual maturity was significantly different between the sexes (t=9.54, P<0.000 1, n=150), with males and females attaining sexual maturity at 56% and 86% of mean adult SVL, respectively. Finally, higher sperm and vitellogenic follicle production were not synchronized. While maximum sperm abundance occurred from July to August (mid rainy season), maximum vitellogenic follicle abundance occurred from October to November (end of rainy season).
Despite the divergence in reproductive cycles between sexes, sperm production and follicle maturation patterns indicated that the reproductive cycle was seasonal at the population level. Both sexes presented a synchronized increase in reproductive output through the rainy season, with highest abundance from June to November. Even though no mating behaviors were observed among individuals of N. atrata, the high frequencies of oviductal sperm, as well as sperm production and follicle maturation, suggest two main mating pulses, one at the beginning of the rainy season (April) and one at the end of the rainy season (October–November). However, mating signals were present all year, including the dry season (Figure 8).
The seasonal reproductive cycle of N. atrata was closely correlated with the strong climate seasonality of the study area, as well as prey and hatchling abundance. For instance, the greatest recruitment peak occurred in the mid dry season, which coincided with the increase in snail abundance and greatest production of primary follicles, whereas the greatest recruitment dearth was observed in the mid rainy season, which coincided with the decline in snail abundance and maximum sperm production (Figures 7–8).
The main drivers of reproductive output in N. atrata diverged between the sexes. In females, multiple regression analysis indicated that clutch size was strongly correlated with almost all reproductive traits evaluated (Table 2). However, among these variables, the "best model" was comprised of the number of primary and secondary follicles and body mass, which explained 70.63% (P<0.000 1) of clutch size variability. Given the importance of secondary follicles in clutch size, a second multiple regression analysis was carried out exploring the relationships among secondary follicles, maternal traits, and height of piles of palm leaves (Figure 9). As a result, secondary follicle variability was highly correlated with stomach bolus volume, fat body area, and body mass. These variables composed the "best model" and explained 44% (P= 0.003) of secondary follicle variability. Similarly, female SVL, but not the remaining variables, was significantly related to egg mass (F=7.64, P=0.014, n=17), indicating that larger females produced heavier eggs.
Figure 9. Multiple regression models related to clutch size and secondary follicles with environmental variables and intrinsic reproductive traits
In males, body mass, height of piles of palm leaves, and stomach volume, rather than intrinsic reproductive traits, showed the greatest contribution to sperm production (R2= 0.198, P<0.001; Table 2). This result agrees with the discordance observed between sperm production and monthly variation in testicular volume and size of SSK (Figure 10).
Exploring the reproductive ecology of the tropical semifossorial snake Ninia atrata
- Received Date: 2019-06-21
- Publish Date: 2020-03-01
- Continuous male reproduction /
- Clutch mass /
- Income breeding /
- Iteroparity /
- Spermatogenesis /
- Oogenesis /
- Reproductive effort /
- El Niño-Southern Oscillation (ENSO)
Abstract: Based on histological analyses and field studies, this research describes the reproductive ecology of a population of Ninia atrata snakes inhabiting an oil palm plantation. Furthermore, through a multivariate approach, we explored the main drivers of reproductive output in N. atrata. Results showed that prey abundance and food intake were crucial variables contributing to reproductive output. Multiple linear regression models showed that neonates had high sensitivity (R2=55.29%) to extreme changes in climate, which was strongly related to slug and snail abundance variability and microhabitat quality. Reproductive cycles were markedly different between the sexes, being continuous in males and cyclical in females. Despite this variation, reproductive cycles at the population level were seasonal semi-synchronous. Constant recruitment of neonates all year, multiple clutches, high mating frequency, and continuous sperm production characterized the reproductive phenology of N. atrata. In addition, a significant number of previtellogenic females presented oviductal sperm as well as uterine scars, suggesting a high precocity in the species. The main drivers of reproductive output also differed between the sexes. In females, clutch size and secondary follicle variability were highly related to stomach bolus volume, fat body area, and body mass. In males, height of piles of palm leaves and body mass, rather than intrinsic reproductive traits, were the main drivers of sperm production. Nevertheless, in both cases, the relationship between body mass, prey abundance, and food intake suggests that N. atrata follows the income breeding strategy to compensate for reproductive costs and to maximize fitness.
|Citation:||Teddy Angarita-Sierra, Cesar Alejandro López-Hurtado. Exploring the reproductive ecology of the tropical semifossorial snake Ninia atrata. Zoological Research, 2020, 41(2): 157-171. doi: 10.24272/j.issn.2095-8137.2020.015|