Growth rates of black caiman (Melanosuchus niger) and spectacled caiman (Caiman crocodilus) from two different Amazonian flooded habitats

Transcription

1 Amphibia-Reptilia 34 (2013): Growth rates of black caiman (Melanosuchus niger) and spectacled caiman (Caiman crocodilus) from two different Amazonian flooded habitats Ronis Da Silveira 1,, Zilca Campos 2, John Thorbjarnarson 3,, William E. Magnusson 4 Abstract. Rates of growth and survival in wild populations are affected by the physical environment, biotic interactions, and density-dependent processes, such as growth and fecundity. However, the relative importance of these factors in longlived reptiles is poorly understood. We analyzed growth rates of Melanosuchus niger and Caiman crocodilus coexisting in two areas of the Brazilian Amazon with very different environmental characteristics. Growth rates of Caiman crocodilus at the two sites were similar, but M. niger grew more slowly in the area with higher productivity and higher density of caimans. Growth rates of the same species from other sites and of the temperate-zone Alligator mississippiensis indicate large differences among sites, but little evidence that these differences are primarily due to differences in productivity or temperature. Demographic models used to estimate sustained yields from caiman harvests should take into account the likely importance of density-dependent growth. Keywords: Alligatoridae, Crocodylia, demography, mark-recapture. Introduction Effective wildlife management requires information on the demography of the species involved, including an understanding of individual parameters, such as somatic growth (Rees and Crawley, 1989; Abercrombie, 1992; Wilkinson and Rhodes, 1997). Growth rates of crocodilians are complex, and can vary within a species according to age, size, sex, genetics, incubation conditions, season, habitat, location (Gorzula, 1978; Magnusson and Taylor, 1981; Webb et al., 1983; Hutton, 1987a; Jacobsen and Kushlan, 1989; Webb and Cooper-Preston, 1989; Booth, 2006) and could be density- 1 - Laboratório de Zoologia Aplicada à Conservação, Departamento de Biologia, Instituto de Ciências Biológicas, Universidade Federal do Amazonas, Av. Gen. Rodrigo Otávio 3000, CEP , Manaus, Amazonas, Brazil 2 - EMBRAPA Pantanal, CP 109, CEP , Corumbá, Mato Grosso do Sul, Brazil 3 - Wildlife Conservation Society, PO Box , Gainesville, Florida, USA 4 - Coordenação de Biodiversidade, Instituto Nacional de Pesquisas da Amazônia, CP 478, CEP , Manaus, Amazonas, Brazil Corresponding author; ronis@ufam.edu.br Deceased. dependent. Information on growth is important for determining age at sexual maturity, and understanding relationships between age and fecundity of females (Wilkinson, 1983; Thorbjarnarson, 1996; Campos et al., 2008). Body size also likely plays an important role in determining the reproductive success of males (Lang, 1987; Webb, Whitehead and Manolis, 1987). Most studies of crocodilian growth patterns have been based on mark-recapture studies and the relationship between growth rate and the size of individuals (Magnusson and Sanaiotti, 1995; Dalrymple, 1996; Saalfeld et al., 2008). This information is generally analyzed using sigmoidal growth models to describe the relationship between size and age of individuals (Andrews, 1982). Use of these models for crocodilians has been criticized because growth rates of individuals within the same habitat can vary greatly and this can lead to errors in predictions relating size and age (Webb, Buckworth and Manolis, 1983; Magnusson et al., 1997) and mark-recapture studies of crocodilians frequently lack data on the growth of adults (Abercrombie, 1992). Ideally, growth models would be validated with information on the size Koninklijke Brill NV, Leiden, DOI: /

2 438 R. Da Silveira et al. of known-age individuals that were marked as hatchlings and recaptured after they had reached sexual maturity. However, this has rarely been done with crocodilians as many species require a decade or more to mature (Magnusson and Sanaiotti, 1995; Thorbjarnarson, 1996). Eaton and Link (2011) used Bayesian models to include individual variation in the estimation of growth trajectories. This approach is appropriate where sufficient data are available to characterize between-individual parameter variation. However, much of the variation in growth rate within a species is because of differences among sites. For management, it is not appropriate to model between-site variation in parameters as a random variable because management generally has to be tailored to each site and no general decision should be made across all sites. External factors that could cause such variation include climate (Wilkinson and Rhodes, 1997) and ecosystem productivity (Saalfeld et al., 2008). However, densitydependent processes also affect demography (Abercrombie, 1989). In the Amazon basin, rivers have been classified broadly as either white water with high Andean sediment loads and elevated levels of aquatic primary productivity, or clear or black waters with low sediment loads and low levels of primary productivity (Sioli, 1984; Junk and Piedade, 2010). Whether or not growth rates of caimans differ among these types of river systems is unknown, as the only information comes from the Rio Tapajos, a clear-water river system (Magnusson and Sanaiotti, 1995). Understanding crocodilian growth patterns is important for the design of management programs based on selective hunting and for evaluation of the effects of illegal hunting. Hunting pressure may vary because of the value of the skin, but given similar hunting pressure, it has been postulated that the spectacled caiman (Caiman crocodilus) can support greater hunting pressure than the larger black caiman (Melanosuchus niger) because it is smaller and reaches sexual maturity earlier (Rebêlo and Magnusson, 1983; Vallejo, Ron and Asanza, 1996). However, this conclusion is difficult to evaluate with so little knowledge about the growth of either species in the Amazon. Caiman crocodilus has been relatively well studied in seasonally flooded savanna habitats (Gorzula, 1978; Ayarzagüena, 1984; Rebêlo et al., 1997), but few investigations have examined its ecology in forested habitats (Ouboter and Nanhoe, 1984; Da Silveira, Magnusson, Campos, 1997). Information on growth and other aspects of the ecology of M. niger is even more limited. In this study, we compared growth rates within and between C. crocodilus and M. niger, at two sites in Brazilian Amazonia with very different primary productivity, but similar temperature regimes: a white-water varzea with high levels of primary productivity and a lowproductivity black-water system. Our objectives were to determine whether growth rates vary as a function of site, body size, and sex, and to estimate how long caiman of each species need to reach sexual maturity or maximum size. We also identify which age classes of caiman are most commonly hunted, and compare growth rates of these species to those obtained in other areas and those of the well-studied, but more temperate dwelling A. mississippiensis. Material and methods Study sites Caiman growth was studied in two sites located in Amazonas State, Central Amazonia, Brazil. The first area was a wide fluvial archipelago (350,000 ha) located in the Anavilhanas National Park (ANP) in the lower Rio Negro, the world s largest black water river system, whose aquatic ecosystems are characterized by low levels of primary productivity (Goulding, Carvalho and Ferreira, 1988). However, the Anavilhanas Archipelago receives some sediment-laden water from the Branco River, and growth rates of trees in the archipelago are intermediate between those of most black-water rivers and whitewater rivers, such as the Amazon (Scabin, Costa and Schöngart, 2011). Capture effort in the ANP was concentrated in a region between the downstream extremity of the archipelago to 130 km upstream from the city of

3 Growth of Amazonian caimans 439 Manaus. This area included some 100 lakes and canals that encompass approximately 1000 km of island shoreline within the Anavilhanas Archipelago. Further description of the archipelago can be found at br/portal/biodiversidade/unidades-de-conservacao/biomasbrasileiros/amazonia/unidades-de-conservacao-amazonia/ 1977-parna-de-anavilhanas.html. The other study site was in the Mamirauá Sustainable Development Reserve located between the Amazon River (Solimões River) and the Japurá River. This area has forest seasonally flooded by nutrient-rich white waters (Sioli, 1984) that is known locally as varzea, and is characterized by high levels of primary aquatic productivity (Junk, 1982; Junk and Welcomme, 1990). Detailed information on the Mamirauá region and its aquatic habitats can be found in Ayres (1986, 1993), and Mamirauá (1996) as well as Data collection and analyses Caiman were captured by hand, using tongs, or with locking cable snares at night from a 4.8 m aluminum boat with a 15 Hp outboard motor. All caiman were marked by removing a unique combination of single and double caudal scutes. Recently hatched caiman (<22.3 cm of snout ventlength) were marked only by removing two caudal scutes, one indicating the year and the other the region or water body where it was captured. After 1996, caiman were also marked with numbered self-piercing metal tags placed in the interdigital membrane of a hind foot. Snout vent-length (SVL from the tip of the snout to the posterior margin of the cloaca) and total body length (TL) of caimans were measured to the nearest 0.1 cm using a flexible metallic tape. Sexing was done by examination of the cloaca (Webb et al., 1984). All caiman were released after a maximum of 20 minutes at the point of capture. TL was estimated from SVL because many caimans had injuries that resulted in loss of the tail tip (see Results). Although some caiman were recaptured throughout the first part of the study, intensive recapture efforts were carried out in both the Anavilhanas (September-October) and Mamirauá (November-December) sites in The transparency of the water in the Anavilhanas Archipelago allowed hundreds of caimans to be visually inspected to determine whether they were marked, and only marked animals were captured. The relationship between growth rate (SVL at recapture initial SVL/time between captures) and caiman body size (geometric mean of SVL at initial capture and SVL at recapture SVL m ) was analyzed using linear regression (Jacobsen and Kushlan, 1989; Brandt, 1991; Dalrymple, 1996). Because it was not possible to recognize individually captured hatchlings, we used the mean SVL of hatchlings with the same mark as an estimate of size at first capture for all individuals with known age. The Richards and von Bertalanffy sigmoidal growth models (Brisbin, 1988) were used to predict the relationship between body size (SVL) and age of recaptured caiman, excluding individuals of known age (those marked as hatchlings). The Richards model, in its complete form is: L t = [A (1 m) (A (1 m) L (1 m) 0 ) exp( 2t/ T(m + 1))] 1/1 m,wherel t and L 0 are SVL of animals at recapture and at hatchling, respectively. A is the theoretical mean maximum SVL for individuals of the species, T is the period of time necessary for the animals to approach the asymptotic value of A, t is the interval between capture and recapture, and m is a growth curve value as described by Richards (1959). When the value of m = 0, the growth curve is generally referred to as the von Bertalanffy model (Abercrombie, 1992), but may be more correctly called the monomolecular model (L. Brisbin, pers. com.). Caimans of known age were not included in these growth-model analyses but were used to validate the models. Analyses were carried out using the NONLIN module in Systat 8.0 (Systat 8.0, SPSS Inc., Chicago). In the case of M. niger, the piecewise regression identified intersection points for the relationship between SVL m and growth rate for both males and females, and growth of animals with SVL m below and above these intersection points were analyzed separately. For small M. niger, we used a Richards monomolecular growth model (Abercrombie, 1996; Wilkinson and Rhodes, 1997; Saalfeld et al., 2008). However, there was no detectable relationship between age and growth rate in larger M. niger, making use of the Richards growth model impossible. In this case we developed an empirical model that was based on the mean growth rate for each sex in specific body-size classes: SVL = k + (b c) GR, where k is the growth curve intersection point for each sex, b is the age of the caiman, c is the length of time necessary for each sex to reach k, and GR is the sex-specific mean growth rate for large caiman. The use of this empirical relationship is supported by the results of other studies on M. niger growth (see Discussion). The growth curves for C. crocodilus and M. niger were compared with those of individuals from other sites and also with those of A. mississippiensis, whose growth has been the subject of several studies. Adjustments for use of TL instead of SVL to measure animals in other studies and calculate growth, and differences in how SVL is measured are outlined in Da Silveira (2001). Growth rates of M. niger calculated by Herron (1991) in Peru and Vallejo, Ron and Asanza (1996) in Ecuador were estimated for males and females together, and in this case we used the combined data in the growth curve for large M. niger. Growth of A. mississippiensis was evaluated in many sites (e.g. Chabreck and Joanen, 1979; Brandt, 1991; Elsey et al., 1992). We used studies in Louisiana (Rootes et al., 1991) and South Carolina (Wilkinson and Rhodes, 1997) to illustrate growth in this species because they represent geographical and climatic extremes of the species range, and provided detailed data on males and females separately. Results For the species we studied, SVL is about half TL, as described by the following simple linear

4 440 R. Da Silveira et al. regression equations based on individuals with complete tails: C. crocodilus, TL = SVL; r 2 = 0.998, F 1,358 = , P<0.001, M. niger, TL = SVL; r 2 = 0.998, F 1,466 = , P< In the Anavilhanas Archipelago, between December 1990 and March 1992, we marked 372 C. crocodilus of all sizes, of which 48% were hatchlings. During the same period, we marked 120 M. niger (SVL 150 cm), of which 64% were hatchlings. From 1992 to 1998 (April or May) field work was focused entirely on capturing recently hatched caiman, and we marked 1166 C. crocodilus and 393 M. niger at this site. In the Mamirauá Reserve, we marked 1476 C. crocodilus (86% hatchlings) and 248 M. niger (14% hatchlings) between February 1994 and February Captures were carried out in 55 water bodies throughout the entire southern portion of the reserve. A subsequent capture effort conducted between August 1996 and December 1998 in 30 water bodies in the downstream end of the reserve resulted in capture of 493 C. crocodilus (43% hatchlings) and 665 M. niger (31% hatchlings). Of the 1376 caimans we approached in 88 water bodies in the Anavilhanas in 1999, only the 25 marked animals (18 C. crocodilus and 7 M. niger) were captured. Known-aged animals marked as hatchings included 14 C. crocodilus and all the M. niger recaptures. Other M. niger had been recaptured in 1992 and other C. crocodilus in In Mamirauá, the 1999 dry season was characterized by extremely low water levels and recaptures were only possible in six water bodies ( 11% of all water bodies initially visited). All caimans had to be captured (336 were M. niger and 68 were C. crocodilus) to determine if they had been marked. Of these, 45 M. niger and 23 C. crocodilus had been marked previously, including three C. crocodilus marked as hatchlings. Another 25 M. niger and 22 C. crocodilus had been recaptured between 1995 and 1998, of which, two C. crocodilus and four M. niger had been marked as hatchlings. Growth of C. crocodilus In Anavilhanas, the interval between capture and recapture of C. crocodilus ranged from one month to 8.8 years and in Mamirauá from one month to 4.3 years. Recaptures of C. crocodilus included individuals of both sexes and all size classes in both study sites. The growth rate of males (n = 46) varied from 3.6 to 18.4 cm/yr (mean ± SD = 9.2 ± 3.3 cm/yr; fig. 1A). Female growth rates (n = 17) were slightly lower and ranged from 2.3 to 14.9 cm/yr (mean ± SD = 8.5 ± 3.5 cm/yr; fig. 1C). The relationship between growth rate (GR) and SVL m in males of unknown age (fig. 1A) was GR = SVL m, r 2 = 0.50, F 1,32 = 31.9, P < The equivalent relationship for females (fig. 1C) was GR = SVL m, r 2 = 0.83, F 1,8 = 39.9, P < Linear models for the derivatives of the size-on-age curves fit the data well, indicating that they conform to the monomolecular model (von Bertalanffy model for body length). ThegrowthrateofC. crocodilus varied significantly (r 2 = 0.41, F 4,73 = 12.8, P<0.001) as a function of SVL m (P = 0.003) and sex (P < 0.001), but did not vary significantly between study sites (P = 0.491). As there was a significant interaction between SVL m and sex (P < 0.001), we analyzed males and females separately. Growth of M. niger In Anavilhanas, the interval between capture and recapture of eight M. niger ranged from 1.0 to 4.4 years, and for 70 M. niger in Mamirauá, ranged from one month to 4.3 years. Recaptures of M. niger with SVL cm at recapture included individuals of both sexes.

5 Growth of Amazonian caimans 441 Figure 1. Relationship between growth rate and geometric-mean snout vent-length (between capture and recapture) of (A) male and (C) female C. crocodilus. Growth curves are those of the Richards (dashed line) or monomolecular (solid line) models for (B) male and (D) female C. crocodilus and do not include known-age individuals. Circles and crosses represent unknown-age caimans, and stars and squares are known-age caimans from Anavilhanas and Mamirauá, respectively. We combined them because the test showed that growth did not vary between sites for the species. The curves are growth-rate on length relationships based on 45 animals varying from 17.0 to cm in length. Growth rates of male M. niger (n = 55) varied from 2.7 to 16.4 cm/yr (mean ± SD = 7.3 ± 2.7 cm/yr; fig. 2A). For females (n = 18), growth rates ranged from 1.9 to 12.5 cm/yr (mean ± SD = 6.0 ± 3.0 cm/yr; fig. 2C). For this species, an analysis of covariance (ANCOVA) indicated that growth rates varied significantly (r 2 = 0.59) as a function of sex (P < 0.001) and study site (P < 0.001), but not as a function of SVL m (P = 0.131). M. niger grew significantly faster in the Anavilhanas than in the Mamirauá. Within the size intervals (SVL cm) for which we had recaptures, growth rates of males and females was greater (>double) in Anavilhanas compared with Mamirauá (fig. 2A, 2C). Using breakpoint regression analyses, the relationship between growth rate and SVL m for M. niger was not considered linear for either males (fig. 2A) or females (fig. 2C). The intersection point was at SVL m = 41.5 cm for males and SVL m = 39.7 cm for females. The relationship between SVL m and growth rate for small male M. niger (fig. 2A) was GR = SVL m, r 2 = 0.670, F 1,9 = , P = For small females (fig. 2C), the relationship was GR = SVL m, r 2 = 0.702, F 1,5 = 11, 761, P = For individuals above the growth intersection points there was no significant relationship between SVL and growth rate (Males: n = 41, P = 0.892, mean growth rate =

6 442 R. Da Silveira et al. Figure 2. Relationship between growth rate and geometric-mean snout vent-length (between capture and recapture) of (A) male and (C) female M. niger. Growth curves are those of the Richards (dashed line) and monomolecular (solid line) models for (B) male and (D) female M. niger and do not include known-age individuals. Circles and crosses represent unknownage caimans, and stars and squares are known-age caimans from Anavilhanas and Mamirauá, respectively. The curves are growth-rate on length relationships based on 65 animals varying from 22.0 to cm in length. 6.4 cm/yr; Females: n = 13, P = 0.401), mean growth rate = 4.8 cm/yr). Growth curves and age of caimans The Richards (m = 0.101) and the monomolecular (m = 0) growth models produced similar results for male C. crocodilus (fig. 1B). Growth asymptotes (where growth rates approach zero) for this sex were similar in the two models: cm SVL in the Richards model and cm in the monomolecular. The time parameter, which reflects the time necessary to approach the asymptote for male C. crocodilus was similar in the two models: 14.3 yr and 14.2 yr, respectively (fig. 1B). This asymptote is somewhat less than the mean SVL of the five largest male C. crocodilus (mean ± SD = ± 1.9 cm) captured (not recaptured) during our study in both reserves. The two growth models also produced comparable results for female C. crocodilus, with the Richards model (m = 0.600) and the monomolecular model estimating similar asymptotic SVL (77.6 cm and 81.3 cm SVL, respectively) and time parameters (9.8 yr and 9.1 yr, respectively). The mean SVL of the five largest female C. crocodilus captured was 81.3 cm (SD =±1.7 cm). A comparison of the known-age C. crocodilus and the growth curves of the two models indicate that both the Richards and the monomolecular models adequately predict the relationship between SVL and age for both

7 Growth of Amazonian caimans 443 males (fig. 1B) and females (fig. 1D). Nevertheless, two of the three known-age males greater than 80 cm SVL (all from the same hatchling group) from Anavilhanas grew faster than the rates predicted by the models (fig. 1B). Among females, only one individual from Mamirauá grew faster than predicted by the models (fig. 1D). For small M. niger, the only data for knownage individuals was for males. Despite the fact that the monomolecular model was developed using only data from recaptures in Mamirauá, it proved to be adequate to predict the age-svl relationships for small M. niger at both study sites. However, for large males and females, the empirical growth models that were developed based on recaptures at Mamirauá underestimated the size of 2.5 to 4.5 yr old M. niger in the Anavilhanas (fig. 2B, D). We expected this result, because growth rates at comparable SVLs for males (19.9 SVL 41.6 cm) and females (33.4 SVL 44.6 cm) differed between the two reserves. Within these size intervals (SVL cm), growth rates for males were larger in Anavilhanas (fig. 2A), and growth rates for females were more than double those found in Mamirauá (fig. 2C). The empirical growth model accurately predicted the SVL of the only female M. niger recaptured in Mamirauá (fig. 2D). Comparison with other populations Methodological differences between this study and previous studies of the growth of C. crocodilus and M. niger do not allow statistical comparisons. Nevertheless, it is possible to examine general patterns and variability in growth rates of these species in our study sites and other areas. Growth rates of C. crocodilus in Anavilhanas and Mamirauá were similar to those from Suriname and the Tapajos River, Brazil, but faster than the values reported for the Llanos and Guyana regions of Venezuela (fig. 3). However, both males and females of the closely related Caiman crocodilus yacare grew at faster rates than the C. crocodilus at our study sites (fig. 3). Figure 3. Growth curves for (A) male and (B) female Caiman crocodilus yacare in the Brazilian Pantanal (Rebelo et al., 1997), and for (C) male and (D) female of C. c. crocodilus in the Anavilhanas and Mamirauá reserves. Growth curves of C. c. crocodilus from (E) Suriname (Ouboter and Nanhoe, 1984), (F) the Tapajós River Brazil (Magnusson and Sanaiotti, 1995), (G) the Venezuelan Llanos (Ayarzagüena, 1984) and (H) the Venezuelan Guayana (Gorzula, 1978). All authors, except Rebelo et al. (1997), showed the growth curves for both sexes. See Methods for a description of how the curves were constructed. Growth rates of M. niger in Mamirauá were similar to those reported from Ecuador but somewhat less than those from Peru (fig. 4). However, M. niger growth in the Anavilhanas appears to be much faster, at least for the first 3.5 yr of life. Sex-specific comparisons were not possible as growth data for Peru (Herron, 1991) and Ecuador (Vallejo, Ron and Asanza, 1996) were presented without information on the sex of the individuals. Overall growth rates of M. niger at our two study sites were similar to values reported for the American alligator (Alligator mississippiensis) in Louisiana (Rootes et al., 1991) and South Carolina (Wilkinson and Rhodes, 1997). At Mamirauá, growth rates of male M. niger were lower than those reported for South Carolina and Louisiana coastal marshes, but similar to those reported for Louisiana freshwater marshes. Growth of female M. niger from Mamirauá was, overall, somewhat less than

8 444 R. Da Silveira et al. Figure 4. Growth curves of (A) male and (B) female Melanosuchus niger in (1) Anavilhanas, (2) Peru (Herron, 1991), (3) Mamirauá, and (4) Ecuador (Vallejo et al., 1996). Growth curves of (C) male and (D) female Alligator mississippiensis in (5) a Louisiana estuary (Rootes et al., 1991), (6) South Carolina (Wilkinson and Rhodes, 1997) and (7) freshwater marshes in Louisiana (Rootes et al., 1991). See Methods for a description of how the curves were constructed. values reported for female A. mississippiensis (fig. 3). Discussion The Richards and the monomolecular model of von Bertalanffy produced similar results in describing the growth of Caiman crocodilus, as has been found for some other crocodilians (Abercrombie, 1996; Wilkinson and Rhodes, 1997). However, the Richards and other sigmoidal growth models were not able to accurately describe the growth of M. niger and we had to use different models for juveniles and adults, as has been found for several other species (Webb et al., 1978; Webb, Buckworth, Manolis, 1983; Magnusson and Sanaiotti, 1995). In the Mamirauá reserve, we were unable to find evidence of growth rate slowing in larger M. niger. Similar findings were reported in the Peruvian Amazon (Herron, 1991) and in Ecuador (Vallejo, Ron and Asanza, 1996). The relatively long period of time over which this study was conducted was still insufficient to analyze the growth of large M. niger, asthe largest animal recaptured was only cm SVL. These findings suggest that longer-term studies are required to more accurately measure the relatively slow growth rates of these large individuals, a common problem in studies of crocodilian growth (Abercrombie, 1992). Given the need for very long studies to estimate growth

9 Growth of Amazonian caimans 445 based on the recapture of marked individuals, and the increasing difficulty of recapturing large crocodilians (Bayliss, 1987; Da Silveira and Da Silveira, 1997), the use of skeletochronology methods may offer advantages (Hutton, 1987b). Growth of crocodilians is frequently associated with prey availability (Webb, Buckworth and Manolis, 1983; Ouboter and Nanhoe, 1984; Rootes et al., 1991; Wilkinson and Rhodes, 1997), so our finding of similar growth rates of C. crocodilus in the blackwater Anavilhanas system, and the more biologically productive white-water varzea of Mamirauá was not expected. Even more difficult to interpret when considering the biological productivity of these aquatic habitats are the results showing faster growth of M. niger in the Anavilhanas than in Mamirauá. Instead of mirroring resource availability, the differences in growth rates between the two study sites may reflect densitydependent compensatory mechanisms that influence somatic growth rates (Hines and Abercrombie, 1987). Cannibalism and interspecific predation between crocodilian species are common (Delany and Abercrombie, 1986; Rootes and Chabreck, 1993; Da Silveira and Magnusson, 1999). These and other types of nonlethal agonistic interactions likely increase under high-density conditions and are likely to affect growth rates. In Mamirauá, the density of caiman is the highest reported for Amazonia, with encounter rates during dry-season nocturnal spotlight counts ranging from 77 to almost 2000 caiman/km, with 53-83% of these being M. niger (Da Silveira, Magnusson and Thorbjarnarson, 2008). Melanosuchus niger and A. mississippiensis are the two largest species of Alligatorids, but one is a tropical species and the other only occurs in the temperate zone. Studies of growth rates of A. mississippiensis have found differences between habitats and geographic location (e.g. Brandt, 1991; Rootes et al., 1991; Dalrymple, 1996). Some of the slowest growth rates have been reported for sites near the northern limit of the species distribution where ambient conditions are cooler (South Carolina, Wilkinson and Rhodes, 1997), and some of the fastest, along the Gulf of Mexico where ambient conditions are much warmer (Louisiana, Rootes et al., 1991). Evidence from M. niger in our study and others suggests that the two species in general have similar growth rates and also exhibit great variability in growth rates in different parts of their ranges. In Mamirauá, the smallest female C. crocodilus we found nesting was 53.5 cm SVL and, based on our growth models, we would predict that females of this species can reach this size in 5.5 years in both reserves. In Suriname, female C. crocodilus reach this size in 4.5 years (Ouboter and Nanhoe, 1984), and in the Tapajos River in Brazil, minimum age at sexual maturity for C. crocodilus has been estimated as 5-6 years (Magnusson and Sanaiotti, 1995). The minimum size at sexual maturity for female M. niger is less well known (Thorbjarnarson and Da Silveira, 2000). Based on our studies of reproduction of this species in Mamirauá (Da Silveira, unpublished data), we estimate that females begin nesting when approximately 120 cm SVL, and our growth models predict that females are years old when they reach this size. However, our model was based on a relatively small number of animals (n = 14), the largest female recaptured was only 93.4 cm SVL, and our estimate of minimum reproductive size was also based on a relatively small number of individuals (n = 10). If we use 100 cm SVL as the minimum size of adult females (Thorbjarnarson, 1996, 1998), age at first reproduction would be more like years. This age at sexual maturity is still considerably older than has been reported for A. mississippiensis (8-13 years; Rootes et al., 1991) and Crocodylus porosus (12 years; Webb, Whitehead and Manolis, 1987). That sexual maturity in female M. niger is reached at a relatively older age compared with C. crocodilus, suggests that this species remains vulnerable to hunting for an extended period (Rebêlo and Magnusson, 1983; Vallejo, Ron

10 446 R. Da Silveira et al. and Asanza, 1996). However, time to reproduction is not the only factor that affects susceptibility to hunting. Recent studies suggest that habitat type, and particularly the density of vegetation that provides refuges from hunters, may be more important (Magnusson, 1999; Da Silveira and Thorbjarnarson, 1999). Management programs involving the wild harvest of crocodilians usually target larger individuals (males), and prohibit the harvest of females by imposing a minimum size restriction in order to reduce the impact on the population (Joanen and MacNease, 1987; Thorbjarnarson and Velasco, 1998). However, size restrictions can vary according to species, habitat, location, and the intended purpose of the harvested animals (meat or skins). Studies of illegal caiman hunting in the Mamirauá reserve have found that all C. crocodilus taken are greater than 60 cm SVL, with the majority between cm SVL. The sex ratio of C. crocodilus taken is 3.5 males/female (Da Silveira and Thorbjarnarson, 1999). Based on our growth curves, the males are 4.5 yrs of age and the females 5.5 yrs. Males over 95 cm SVL would be 12.5 yrs old. The majority of the M. niger that are illegally hunted in Mamirauá are cm SVL, with a sex ratio of 12.4 males/female (Da Silveira and Thorbjarnarson, 1999). Nevertheless, the largest individual that was recaptured in our study was cm SVL, so growth rates of animals larger than this are only an extrapolation. Of the sample of hunted animals, 49% were cm SVL and estimated to be between 7.5 and 14 years old. Despite the probable imprecision in the growth curve for large individuals, we estimate the age of animals in the cm SVL size class to be years old. However, the age of 205 cm SVL males may be considerably more based on our findings that growth rates slowed significantly in larger animals. The corresponding ages of M. niger in the Anavilhanas may be less. Illegal hunting of both species of caiman has been reported from our two study sites. While hunting has probably always been light and intermittent in the Anavilhanas Archipelago, there have been frequent reports of hunting in the Mamirauá SDR (Da Silveira and Thorbjarnarson, 1999). However, this probably reflects the higher number of caimans that inhabit the Mamirauá due to greater primary productivity. Therefore, it is difficult to evaluate the effect of hunting in these localities, and there is no evidence that populations in either area are currently much below carrying capacity. Ecology and social behavior of crocodilians is complex and varies between species (Lang, 1987; Webb, 1987), and it has been suggested that recruitment of breeding females of A. mississippiensis is density-dependent (Hines and Abercrombie, 1987). Density, nesting sites and intensity of illegal hunting of Amazonian caimans are highly related to productivity of the flooded forest, especially in varzea habitats (Da Silveira, 2002; Da Silveira, Magnusson Thorbjarnarson, 2008; Villamarin et al., 2011). It is therefore possible that growth rates also will be density-dependent. Our study provides evidence that suggests there is a density-dependent mechanism controlling growth rates of M. niger. Ifthisisthe case, a reduction in population density could potentially increase the growth rates of juveniles and allow animals to reach sexual maturity at a younger age. For C. crocodilus, if there is a density dependent mechanism at work it is more likely to be acting on survival rates and not growth. In the case of the Mamirauá reserve, one of the most important questions to evaluate will be how growth rates vary as a result of a reduction in the population density of the two species. The liberation of caiman harvesting in sustainable use reserves in Amazonas State in Brazilian Amazonia may allow these predictions to be tested if monitoring is continued for the next few decades. Acknowledgements. This research was financed or supported by a group of Brazilian and international agencies, including the Ministério da Ciência, Tecnologia e Inovação,

11 Growth of Amazonian caimans 447 the Conselho Nacional de Desenvolvimento Científico e Tecnológico, the Academia Brasileira de Ciências, the Instituto Brasileiro do Meio Ambiente e Recursos Naturais Renováveis, the Instituto de Proteção Ambiental do Amazonas (IPAAM), the Fundação Vitória Amazônica, the Fundação O Boticário de Proteção à Natureza, the Direction de Environment, Securitie Nucleaire et Protection Civile of the European Commission, the Wildlife Conservation Society, and the World Wide Fund for Nature/UK. We thank all inhabitants of the Mamirauá Reserve and especially our field staff: J. Carvalho, J. Tapioca, E. Martins, H. Carvalho, Dalvino, A. Cardoso, J. Carvalho, Macedo, A. Carvalho and N. Carvalho. In Anavilhanas, we thank Olegário, Maracanã, Adenilson, Marli, Maria and especially M.P. Pontes. J.M. Ayres (in memorian) and A.R. Alves gave us the unique opportunity to work in the Mamirauá Reserve. References Abercrombie, C.L. (1989): Population dynamics of the American alligator. In: Crocodiles: Their Ecology, Management, and Conservation, p Hall, H., Bryant, R., Eds, Gland. Abercrombie, C.L. (1992): Fitting curves to crocodilian age-size data: some hesitant recommendations. In: Proc. of the 11th Working Meeting of the IUCN/SSC Crocodile Specialist Group, p Zimbabwe. IUCN Publ., Gland, Switzerland. Abercrombie, C.L. (1996): Summarizing croc growth using von Bertalanffy curve. In: Proc. of the 13th Working Meeting of the IUCN/SSC Crocodile Specialist Group, Santa Fe, p IUCN Publ., Gland, Switzerland. Andrews, R.M. (1982): Patterns of growth in reptiles. In: Biology of the Reptilia, p Gans, C., Pough, F.H., Eds, Vol. 12, Academic Press, New York. Ayarzagüena, J. (1984): Biologia y ecologia de la baba (Caiman crocodilus L.) en los Llanos de Venezuela. Doñana Act. Vert., Ser. Monog. 136 pp. Ayres, J.M. (1986): Uacaris and the Amazon flooded forest. Ph.D. Thesis, Subdepartament of Veterinary Anatomy, Cambridge University. Unpublished. Ayres, J.M. (1993): As matas de várzea do Mamirauá. CNPq Sociedade Civil Mamirauá, Brasília, DF. Bayliss, P. (1987): Survey methods and monitoring within crocodile management programmes. In: Wildlife and Management: Crocodiles and Alligators, p Webb, G.J.W., Manolis, S.C., Whitehead, P.J., Eds, Surrey Beatty & Sons Pty Limited, Chipping Norton. Booth, D.T. (2006): Influence of incubation temperature on hatchling phenotype in reptiles. Physiol. Biochem. Zool. 79: Brandt, L.A. (1991): Growth of juvenile alligators in Par Pond, Savannah River Site, South Carolina. Copeia 1991: Brisbin, I.L. (1988): Growth curve analyses and their application to the conservation and captive management of crocodilians. In: Proc. of the 9th Working Meeting of the IUCN/SSC Crocodile Specialist Group, p Gainesville. IUCN Publ., Gland, Switzerland. Campos, Z., Magnusson, W., Sanaiotti, T., Coutinho, M. (2008): Reproductive trade-offs in Caiman crocodilus crocodilus and Caiman crocodilus yacare: implications for size-related management strategies. Herpetol. J. 18: Chabreck, R.H., Joanen, T. (1979): Growth rates of American alligators in Louisiana. Herpetologica 35: Dalrymple, G.H. (1996): Growth of American alligators in the Shark Valley Region of Everglades National Park. Copeia 1996: Da Silveira, R. (2001): Monitoramento, Crescimento e Caça de jacaré-açu (Melanosuchus niger) e de jacarétinga (Caiman crocodilus crocodilus). Ph.D. Thesis, INPA/UFAM, Manaus, Brazil. Da Silveira, R. (2002): Conservação e manejo do jacaré-açu (Melanosuchus niger) na Amazônia brasileira. In: La Conservación y el Manejo de Caimanes e Cocodrilos de América Latina, p Larriera, A., Verdade, L.M., Eds, C.N. Editoria, Piracicaba, São Paulo, Brazil. Da Silveira, R., Da Silveira, B.B. (1997): Capture and recapture of an adult black caiman in Mamirauá. Newsletter Crocodile Specialist Group IUCN/SSC. 16: Da Silveira, R., Thorbjarnarson, J.B. (1999): Conservation implications of commercial hunting of black and spectacled caiman in the Mamirauá Sustainable Development Reserve, Brazil. Biol. Cons. 88: Da Silveira, R., Magnusson, W.E., Campos, Z. (1997): Monitoring the distribution, abundance and breeding areas of Caiman crocodilus crocodilus and Melanosuchus niger in the Anavilhanas Archipelago, Central Amazonia, Brazil. J. Herpetol. 31: Da Silveira, R., Magnusson, W.E. (1999): Diets of spectacled and black caiman in the Anavilhanas Archipelago, Central Amazonia, Brazil. J. Herpetol. 33: Da Silveira, R., Magnusson, W.E., Thorbjarnarson, J.B. (2008): Factors affecting the number of caimans seen during spotlight surveys in the Mamirauá Reserve, Brazilian Amazonia. Copeia 2008: Delany, M.F., Abercrombie, C.L. (1986): American alligator food habits in Northcentral Florida. J. Wildl. Manage. 50: Eaton, M.J., Link, W.A. (2011): Estimating age from recapture data: integrating incremental growth measures with ancillary data to infer age-at-length. Ecol. Applications 21: Elsey, R.M., Joanen, T., McNease, L., Kinler, N. (1992): Growth rates and body condition factors of Alligator mississippiensis in coastal Louisiana wetlands: a comparison of wild and farm-released juveniles. Comp. Biochem. Physiol. 103A: Gorzula, S.J. (1978): An ecological study of Caiman crocodilus crocodilus inhabiting savanna lagoons in Venezuelan Guayana. Oecologia 35: Goulding, A., Carvalho, M.L., Ferreira, E.G. (1988): Rio Negro: Rich Life in Poor Water. SPB Academic Publishing, Amsterdam. Herron, J.C. (1991): Growth rates of black caiman Melanosuchus niger and black caiman Caiman crocodilus, and the recruitment of breeders in hunted caiman population. Biol. Cons. 55:

12 448 R. Da Silveira et al. Hines, T.C., Abercrombie, C.L. (1987): The management of alligators in Florida, USA. In: Wildlife and Management: Crocodiles and Alligators, p Webb, G.J.W., Manolis, S.C., Whitehead, P.J., Eds, Surrey Beatty & Sons Pty Limited, Chipping Norton. Hutton, J.M. (1987a): Growth and feeding ecology of the Nile crocodile Crocodylus niloticus at Ngezi, Zimbabwe. J. Anim. Ecol. 56: Hutton, J.M. (1987b): Techniques for ageing wild crocodilians. In: Wildlife and Management: Crocodiles and Alligators, p Webb, G.J.W., Manolis, S.C., Whitehead, P.J., Eds, Surrey Beatty & Sons Pty Limited, Chipping Norton. Joanen, T., McNease, L. (1987): The management of alligators in Louisiana, USA. In: Wildlife and Management: Crocodiles and Alligators, p Webb, G.J.W., Manolis, S.C., Whitehead, P.J., Eds, Surrey Beatty & Sons Pty Limited, Chipping Norton. Jacobsen, T., Kushlan, J.A. (1989): Growth dynamics in the American alligator (Alligator mississippiensis). J. Zool. 219: Junk, W.J. (1982): Amazonian floodplains: their ecology, present and potential use. Rev. Hydrobiol. Trop. 15: Junk, W.J., Welcomme, R.L. (1990): Floodplains. In: Wetlands and Shallow Continental Water Bodies, p Patten, B.C., Jorgensen, S.E., Dumont, H., Eds, SPB Academic Publishing, The Hague, Amsterdam. Junk, W.J., Piedade, M.T.F. (2010): An introduction to South American wetlands forests: distribution, definitions and general characterization. In: Amazonian Floodplain Forests: Ecophysiology, Biodiversity and Sustainable Management, p Junk, W.J. et al., Eds, Ecol. Stud., Springer Science + Media B.V. Lang, J. (1987): Crocodilian behavior: implications for management. In: Wildlife and Management: Crocodiles and Alligators, p Webb, G.J.W., Manolis, S.C., Whitehead, P.J., Eds, Surrey Beatty & Sons Pty Limited, Chipping Norton. Mamirauá (1996): Mamirauá: Plano de Manejo. Brasília: SCM; CNPq/MCT; Manaus: IPAAM. Magnusson, W.E. (1999): Hábitat de suporte e estratégias de exploração de crocodilianos. In: Manejo y Conservación de Fauna Silvestre em América Latina, p Fang, T.G., Montenegro, O.L., Bodmer, R.E., Eds, La Paz. Magnusson, W.E., Taylor, J.A. (1981): Growth of juvenile Crocodylus porosus as effected by season of hatching. J. Herpetol. 15: Magnusson, W.E., Mourão, G. (1997): Manejo extensivo de jacarés no Brasil. In: Manejo e Conservação de Vida Silvestre no Brasil, p Valladares-Padua, C., Bodmer, R.E., Cullen, Jr. L., Eds, CNPq/Belém, PA: Sociedade Civil Mamirauá, Brasília, DF. Magnusson, W.E., Sanaiotti, T.M. (1995): Growth of Caiman crocodilus in Central Amazonia, Brazil. Copeia 2: Magnusson, W.E., Lima, A.P., Costa, V.L., Lima, A., Araújo, M.C. (1997): Growth during middle age in Schneider s Dwarf caiman, Paleosuchus trigonatus. Herpetol. Rev. 28: 183. Ouboter, P.E., Nanhoe, L.M.R. (1984): An ecological study of Caiman crocodilus crocodilus in northern Suriname. Unpubl. MS Thesis, Report no. 233, Zoöl. Lab. Afd. Dieroecologie, Kath. Univ. Nijmegen. Richards, C.H. (1959): A flexible growth function for empirical use. J. Exp. Bot. 10: Rebêlo, G.H., Magnusson, W.E. (1983): An analysis of the effect of hunting on Caiman crocodilus and Melanosuchus niger base on the sizes of confiscated skins. Biol. Cons. 26: Rebêlo, G.H., Borges, G.A.N., Yamashita, C., Filho, A.G. (1997): Growth, sex ratio, population structure, and hunting mortality of Caiman yacare in the Pantanal, Brazil. Vida Silvestre Neotropical. 6: Rees, M., Crawley, M.J. (1989): Growth, reproduction and population dynamics. Funct. Ecol. 3: Rootes, W.L., Chabreck, R.H. (1993): Cannibalism in the American alligator. Herpetologica 49: Rootes, W.L., Wright, V.L., Brown, B.W., Hess, T.J. (1991): Growth rates of American alligator in estuarine and palustrine wetlands in Louisiana. Estuaries 14: Saalfeld, D.T., Webb, K.K., Conway, W.C., Calkins, G.E., Duguay, J.P. (2008): Growth and condition of American alligators (Alligator mississippiensis) in and inland wetland of East Texas. Southeast. Nat. 7: Scabin, A.B., Costa, F.R.C., Schöngart, J. (2011): The spatial distribution of illegal logging in the Anavilhanas Archipelago (Central Amazonia) and logging impacts on species. Environ. Conserv. 39: Sioli, H. (1984): The Amazon and its main affluents: hydrography, morphology of the river courses, and river types. In: The Amazon: Limnology and Landscape Ecology of a Mighty Tropical River and its Basin Pages, p Sioli, H., Ed., Dordrecht, Monogr. Biol. Thorbjarnarson, J.B. (1996): Reproductive characteristics of the Order Crocodylia. Herpetologica 52: Thorbjarnarson, J., Velasco, A. (1998): Venezuela s Caiman Harvest Program. WCS. Working Paper n. 11, New York, USA. Thorbjarnarson, J.B., Da Silveira, R. (2000): Secrets of the flooded forest. Nat. Hist. 3: Vallejo, A., Ron, S., Asanza, E. (1996): Growth in Melanosuchus niger and Caiman crocodilus at Zancudococha and Cuyabeno, Ecuadorian Amazon. In: Proc. of the 13th Working Meeting of the IUCN/SSC Crocodile Specialist Group, p Santa Fe. IUCN Publ., Gland, Switzerland. Webb, G.J.W., Smith, A.M.A. (1987): Life history parameters, population dynamics and the management of crocodilians. In: Wildlife and Management: Crocodiles and Alligators, p Webb, G.J.W., Manolis, S.C., Whitehead, P.J., Eds, Surrey Beatty & Sons Pty Limited, Chipping Norton. Webb, G.J.W., Messel, H., Crawford, J., Yerbury, M.J. (1978): Growth rates of Crocodylus porosus (Reptilia: Crocodylia) from Arnhem Land, Northern Australia. Aust. Wildl. Res. 5: Webb, G.J.W., Buckworth, R., Manolis, S.C. (1983): Crocodylus johnstoni in the McKinlay river area, N. T. III. Growth, movement and population age structure. Aust. Wildl. Res. 10:

13 Growth of Amazonian caimans 449 Webb, G.J.W., Manolis, S.C., Sack, G.C. (1984): Cloacal sexing of hatchling crocodiles. Aust. Wildl. Res. 11: Webb, G.J.W., Whitehead, P.J., Manolis, S.C. (1987): Crocodile management in the Northern Territory of Australia. In: Wildlife and Management: Crocodiles and Alligators, p Webb, G.J.W., Manolis, S.C., Whitehead, P.J., Eds, Surrey Beatty & Sons Pty Limited, Chipping Norton. Webb, G.J.W., Cooper-Preston, H. (1989): Effects of incubation temperature on crocodiles and the evolution of reptilian oviparity. Am. Zool. 29: Wilkinson, P.M. (1983): Nesting ecology of the American alligator in coastal South Carolina. S. C. Wildl. Mar. Resour. Dep., Columbia. Wilkinson, P.M., Rhodes, W.E. (1997): Growth rates of American alligators in coastal South Carolina. J. Wildl. Manage. 61: Submitted: December 12, Final revision received: June 17, Accepted: June 17, Associated Editor: Sylvain Ursenbacher.