Can Reproductive Allometry Assess Population Marginality in Crocodilians? A Comparative Analysis of Gulf Coast American Alligator (Alligator mississippiensis) Populations Author(s): Christopher M. Murray, Michael Easter, Mark Merchant, Amos Cooper, and Brian I. Crother Source: Copeia, 2013(2):268-276. 2013. Published By: The American Society of Ichthyologists and Herpetologists DOI: http://dx.doi.org/10.1643/ch-11-136 URL: http://www.bioone.org/doi/full/10.1643/ch-11-136 BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder. BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.
Copeia 2013, No. 2, 268 276 Can Reproductive Allometry Assess Population Marginality in Crocodilians? A Comparative Analysis of Gulf Coast American Alligator (Alligator mississippiensis) Populations Christopher M. Murray 1, Michael Easter 2, Mark Merchant 3, Amos Cooper 4, and Brian I. Crother 5 This study uses the American Alligator (Alligator mississippiensis) to assess the use of reproductive allometry as a tool to infer crocodilian population marginality based on conformation to advantageous life-history strategies. It is hypothesized that reproductive allometry, a morphometric relationship between mother s size and her reproductive output, varies intraspecifically between populations and that this variation reflects population marginality based on size, stress, temporal exploitation, habitat fragmentation, and/or the presence of social hierarchy. This hypothesis is tested using relative comparisons of allometric correlation between a marginal population inundated with saline storm surge from Hurricane Ike in southeastern Texas and a hypothesized unstressed core population in southeastern Louisiana. Heterophil to lymphocyte ratios fail to falsify the hypothesis of a saline stressor. The number of significant morphometric correlations between various parameters, degree of correlation (R 2 ), and slope of correlation between mother and her respective nest and clutch varied greatly between study sites. Reproductive allometry, as a measure of relative population marginality, may provide a cost effective way to prioritize management with local support for crocodilian taxa. REPRODUCTIVE allometry is the relationship between a mother s morphometrics and the morphometrics of her respective reproductive output including nest, clutch and/or hatchlings. This relationship has been used in various ways to assess crocodilian life-history variation, demographics, and even stability with regard to selection and fitness (Thorbjarnarson, 1996). An intriguing hypothesis was suggested by Verdade (2001:435; parenthetic notation ours): It is tempting to believe that this (strong allometric correlation between mother, clutch, and hatchling) happens in well established populations with healthy structured age-pyramids and a well established hierarchy among reproductive individuals. However, this may not be the case of small populations inhabiting fragmented and anthropogenic landscapes on a metapopulation context, in which social structure and behavioral ecology may be completely altered. Kawecki (2008) argued that the demography of marginal habitats requires the alteration of life-history strategies to adjust to such habitats. Marginal populations are characterized by stressors that result in low density and fewer total offspring. Immigration into marginal populations from core populations is responsible for their existence and genetic input. Constant impact of stressors ensures that even small changes in a marginal habitat can severely impact the persistence of marginal populations (Kawecki, 2008). The American Alligator (Alligator mississippiensis), found only in the southeastern United States, has undergone decline, protection, and increase since the mid-1900s across the entire range of the species (Wood et al., 1985). Since the recovery of the species, biologists and managers have explored density and abundance monitoring techniques to ensure that population fluctuations can be accurately monitored. Taylor and Neal (1984), Woodward et al. (1984), Taylor et al. (1991), and Lance (2003) have investigated the reproductive ecology of the American Alligator, including density-dependent breeding patterns. At present and in reference to all extant crocodilians the susceptibility to contaminants, overharvesting, eradication of rogue crocodilians, habitat destruction, and other tribulations continue to impact populations. As of 1992, 11 of the 23 crocodilian species held an endangered status, with only five species secure enough to be harvested (Thorbjarnarson, 1992). In 2010, six were critically endangered, one was endangered, three were vulnerable, and one was labeled data deficient (Manolis and Stevenson, 2010). This situation warrants further ecological assessment of crocodilian populations. Newly proposed intraspecific reproductive allometry methods may make it possible to assess crocodilian populations with respect to dynamic life-history strategies. While intraspecific reproductive allometry has been extensively examined, no comparative intraspecific work has been done in a conservation or management context. This study tests the hypothesis that the degree of reproductive allometric correlation in a population reflects the marginality of that population. Statistically, this study hypothesizes that the strength of allometric correlation is greater in an unstressed population than a stressed population and such is evident in populational comparisons between allometric relationships. This hypothesis is directly based on Verdade s (2001) hypothesis that strong positive allometric relationships are degraded by stress, low density, and lack of an intact hierarchy as a function of size. It is also hypothesized, although not directly tested here, that with low density and constant stress, a social hierarchy based on size class will be ignored in crocodilians and food resource 1 Department of Biological Sciences, Auburn University, Auburn, Alabama 36849; E-mail: murracm04@gmail.com. Send reprint requests to this address. 2 Everglades Holiday Park, Fort Lauderdale, Florida 33332; E-mail: meast001@fiu.edu. 3 Department of Chemistry, McNeese State University, Lake Charles, Louisiana 70609; E-mail: mmerchant@mcneese.edu. 4 J. D. Murphree Wildlife Management Area, Texas Parks and Wildlife Department, Port Arthur, Texas 77640; E-mail: amos.cooper@tpwd. state.tx.us. 5 Department of Biological Sciences, Southeastern Louisiana University, Hammond, Louisiana 70402; E-mail: bcrother@selu.edu. Submitted: 22 September 2011. Accepted: 20 November 2012. Associate Editor: B. Stuart. F 2013 by the American Society of Ichthyologists and Herpetologists DOI: 10.1643/CH-11-136
Murray et al. Alligator allometry 269 consumption will not follow size-based partitioning (Verdade, 2001). Thorbjarnarson (1996) reviewed the life-history strategies of crocodilians as a subset of an examination of general crocodilian reproductive characteristics. Among other assessments, Thorbjarnarson (1996) examined allometric relationships and concluded that egg size, clutch size, and clutch mass were positively correlated with female body size both intra- and interspecifically. He noted the important role of size in life-history variation and demonstrated that female body size can predict clutch and egg size (Thorbjarnarson, 1996). When discussing intraspecific reproductive allometry among crocodilians, Thorbjarnarson (1996:20) noted that because crocodilians are long-lived, iteroparous species, they adjust short-term reproductive output such that lifetime reproductive success is maximized. Hall (1991) and Verdade (1997, 2001) also assessed intraspecific reproductive allometry in crocodilians. Hall (1991) demonstrated that demographic profiles of the female portions of populations could be estimated by clutch mass or clutch size in Alligator mississippiensis and Crocodylus novaeguineae (Hall, 1991:133). That study demonstrated the ability to estimate population demographics and stability based on clutch parameters and hatchling morphometrics. Hall (1991) suggested that this method could help to avoid the error associated with observation-based adult surveys. Verdade (1997) demonstrated the capabilities of intraspecific reproductive allometry to accurately model populations of the Broad-Snouted Caiman (Caiman latirostris). Regression analysis and multivariate statistics were used to relate maternal-clutch morphology as well as cranial morphology. This study found a positive correlation between female body size and numerous clutch morphometrics (egg size, clutch size, and hatchling body size). Morphometric models were described to potentially be an inexpensive and effective way to model and conserve metapopulations of Broad-Snouted Caiman (Verdade, 1997:145). In another study, Verdade (2001) compared morphometrics between captive Broad- Snouted Caiman adult females and their clutch and hatchlings. The most significant correlations were between female snout vent length and body mass and clutch/ hatchling mass. These data were interpreted to indicate that significant correlations between female morphometrics and clutch/hatchling morphometrics represent a stabilized conformation to selective pressures and thus maximum female fitness. With larger clutch sizes being restricted to larger females it is more unlikely that in stable, hierarchical populations reproductive allometric correlations will stray from a relatively high R 2 value. This study compares reproductive allometric correlations between two wild Gulf Coast alligator populations as an index of populational stress. Should this method prove effective, management can be prioritized for alligator populations, marginal habitat can be identified, progress toward or away from marginality can be monitored, and this method can be applied to the conservation efforts of other crocodilian taxa. MATERIALS AND METHODS This study used two intermediate salinity marsh sites (3 10 ppt) for relative comparison (Fig. 1A). Site habitat was categorized using Chabreck (1972) and Louisiana Department of Wildlife and Fisheries Conservation Habitats and Species Assessments. J. D. Murphree Wildlife Management Area (JDM WMA) is located to the southwest of Port Arthur, TX (Fig. 1B). This marsh is a 9,814 ha property that was impacted by Hurricane Ike in September 2008. This hurricane resulted in extreme salt-water inundation that has remained in the habitat. Occasional waterway openings also periodically allow salt-water inundation, but to a lesser extent than the Hurricane Ike storm surge. One nestmaternal combination in Anahuac National Wildlife Refuge (Anahuac NWR; west of JDM WMA) was utilized to increase sample size in 2009 as a result of extremely low nest density in southeastern Texas. Hurricane Ike also impacted Anahuac NWR in 2008. Las Conchas Marsh (LCM) is a 1,619 ha private marsh east of Slidell, LA (Fig. 1C) and is characterized by freshwater input from the West Pearl River and outflow into the Rigolets, a pass between Lake Pontchartrain and the Gulf of Mexico. This habitat has been unaffected by anthropogenic events since Hurricane Katrina in 2005 and was unaffected by anthropogenic perturbations during the course of this study. As a result LCM was designated as the hypothesized core population. Data were collected from nine nest-maternal combinations (three in 2009, four in 2010, and two in 2011) in JDM WMA and nine nestmaternal combinations (five in 2009 and four in 2010) in LCM. Nest-maternal combinations were identified by the presence of an attending female at the nest site. The identified maternal individuals of nests were captured and data were collected from her. Maternal individuals were captured using snare poles and/or nest site traps similar to the design of Elsey and Trosclair (2004). Maternal individuals were released and data were collected from nests. Each egg was removed from the nest and marked to avoid lethal rotation. Processed nests were revisited periodically later in the season to acquire hatchling data if hatchlings could be found and matched to the correct nest. Every attempt was made to collect egg and nest data within a small relative time frame to avoid developmental confounds on morphometrics, particularly egg mass change over time. No differences in egg mass were evident, indicating that the time between sampling had no effect on allometric analyses. Morphometric data (in this case in reference to variation in size) were collected from maternal individuals and their respective nests, clutch, and hatchlings. These data were then examined to assess the degree of reproductive allometric correlation in each population over two years. Maternal morphometrics included snout vent length (SVL), body mass (BM), and total length (TL), as well as morphometrics indicative of health: nape circumference (NC) and tail girth (TG). Nape circumference and tail girth were measured by quantifying the circumference of the widest portion of the neck between the posterior margin of the skull and the insertion of the anterior limbs, and the widest portion of the body posterior to the hind limbs, respectively. Nest parameters included nest diameter (ND) and nest height (NH). Egg morphometrics included egg length (EL), egg width (EW), egg mass (EM), clutch mass (CM), and band width (BW). Hatchling morphometrics included hatchling SVL, hatchling TL, and hatchling BM. Salt-water inundation from Hurricane Ike into the JDM WMA habitat represented a stressor for alligators and alligator nesting (Lance et al., 2010). Salinity data were taken at each nest site at the initial opening, using a Quanta HydroLab, at the waters edge closest to the nest. Turbidity, dissolved oxygen, percent dissolved oxygen, ph, water temperature, and specific conductivity were also recorded
270 Copeia 2013, No. 2 Fig. 1. (A) Google Earth satellite image showing relative location of study sites highlighted in white boxes along the Gulf Coast. J. D. Murphree WMA to the left (west) and Las Conchas Marsh to the right (east). (B) Google Earth satellite image of J. D. Murphree WMA relative to Port Arthur, TX. (C) Google Earth satellite image of Las Conchas Marsh east of Slidell, LA. at each nest. Blood samples were taken from each captured alligator to quantify individual stress levels via heterophil to lymphocyte ratios (H/L, Gross and Siegel, 1983; Davis et al., 2008; Lance et al., 2010). Ratios above about 1:1 indicate stress (Lance et al., 2010). Blood was drawn immediately after capture in the field to avoid evidence of capture stress in the samples. Up to 5 ml of whole blood were extracted from the spinal vein (Zippel et al., 2003) using 5 ml heparinized syringes with 21-gauge, 1.5-inch needles. Five ml of whole blood were smeared on slides, and were stained for leukocyte types using Giemsa-Wright stain. The first 200 leukocytes were counted at 40X magnification and the ratios of H/L were obtained, and these ratios were examined to determine relative differences in stress between maternal alligators. Two-sample t-tests were used to assess differences in salinity and H/L ratio between populations. A Pearson correlation matrix was used to indicate which maternal morphometrics and nest morphometrics were independent of other variables within their respective allometric category. Representative variables were used to compare the strength of allometric correlation between the stressed and unstressed populations. Furthermore, statistical differences in the slope and elevation between the allometric correlations of the stressed and unstressed population were assessed using standardized major axis regression (SMA). This was
Murray et al. Alligator allometry 271 done separately for each comparison with the use of the SMATR package (Warton et al., 2006; Warton, 2007). For variables that were not comparable between populations because of missing data (adult nape and adult tail girth) Spearman correlation assessments were used to test the correlation between representative variables from the Pearson correlation matrices and these variables with missing data. This was done to further determine representative variables for entry in SMA. All above-mentioned statistical analyses were performed in R (R Development Core Team, 2011). All data were log transformed to account for large variations in scale between compared morphometrics, variation in measurement unit, and to account for expected curvilinear relationships (King, 2000). A size-corrected clutch mass and size-corrected clutch size was obtained for each location. This was calculated by dividing clutch mass (g) and egg number by the body mass of their respective mother, resulting in number of eggs per one kilogram mother and grams of eggs deposited per one kilogram mother. The non-parametric statistics program Primer 6 (Clarke and Gorley, 2001) was used to assess the strength of correlation between maternal morphometrics and their respective reproductive output for both populations. This was performed using the RELATE test which compares stepwise change between Euclidean distance resemblance matrices and assigns a Rho statistic and significance level to correlations within study populations. Each RELATE test was run for 999 permutations using only those samples that lacked missing variable data. Data were normalized to correct for the use of multiple units of measurement for morphometric variables prior to non-parametric analysis (King, 2000). RESULTS Between the 2009 to 2011 reproductive seasons, data from 18 maternal-nest combinations were obtained between both sites. Five hundred ninety-six eggs were analyzed between both populations and 47 hatchlings were obtained and analyzed from four nests in the hypothesized core population at LCM. Nest site salinity at LCM ranged from 0.09 ppt to 3.88 ppt while heterophil to lymphocyte ratios ranged from 1.24 to 1.80. Nest site salinity in southeast Texas, including JDM WMA and Anahuac NWR, ranged from 3.00 ppt to 14.00 ppt while heterophil to lymphocyte ratios ranged from 1.33 to 3.70. Two sample t-test reveal significantly higher salinity levels in the JDM stressed population than the LCM unstressed population (t 5 24.32, df 5 9.06, P 5 0.002). The same directionality is true for heterophil to lymphocyte ratios (t 523.35, df 5 12.45, P 5 0.036). A Pearson correlation matrix for maternal morphometrics revealed a high degree of correlation between all maternal morphometric variables (correlation coefficients 5 0.71 between adult BM and adult SVL, 0.77 between adult BM and adult TL, and 0.92 between adult SVL and adult TL; Fig. 2). Spearman rank correlation assessments showed a high degree of correlation between those variables with missing data and adult SVL: adult NC (Rho 5 0.910, P, 0.001) and adult TG (Rho 5 0.899, P, 0.001). Adult SVL was used as the representative maternal morphometric in further SMA analyses. A Pearson correlation matrix for nest variables revealed a high degree of correlation between three morphometric parameters (correlation coefficients 5 0.56 between EL and EW, 0.80 between ND and NH, and 0.65 Fig. 2. Pearson correlation matrix indicating strong correlation between maternal morphometrics: adult snout vent length (AdultSVL), adult body mass (AdultBM), and adult total length (AdultTL). Strong relationships indicate that one morphometric is informative for all other morphometrics in further analyses. between EW and CM; Fig. 3). EM, CM, EL, and ND were used as representative nest morphometric variables in further SMA analyses. These comparisons excluded morphometric parameters that are not comparable between populations. For instance, hatchling morphometrics were excluded because hatchlings were only recovered from the core population. Complete morphometric correlations from each study site per year are not discussed here because total population correlations maximize sample size. Standard major axis regression revealed significantly different allometric relationships between stressed and unstressed populations based on slope, elevation, and correlation coefficients (R 2 ; Fig. 4). Slopes of allometric relationships in the stressed population were either negative or less than 1. The unstressed population revealed consistently positive slopes with two being greater than 1. Shift and elevation significantly differed in all allometric relationships between the stressed and unstressed populations except adult SVL and CM (Table 1). Correlation coefficients of allometric relationships in the unstressed population were consistently nine times higher than those of the stressed population (Table 1). The average size-corrected CM and clutch size in the LCM population was 67.21 g/kg and 0.97 eggs/kg, respectively. The JDM WMA revealed a sizecorrected CM of 81.65 g/kg and a size-corrected clutch size of 1.35 eggs/kg. Additionally, general linear model regression in SYSTAT 10.2 (SYSTAT Software, Inc., Richmond, CA, 2002) revealed significant correlation between bandwidth and all maternal morphometrics (all P, 0.001) expect NC and TG in the marginal JDM WMA. The non-parametric RELATE test revealed very different levels of correlation for the two study populations (Fig. 5). The recovered significance level for the LCM population was 0.5% (0.005 as a theoretical P-value), with only four permuted statistics greater than or equal to a Rho value of
272 Copeia 2013, No. 2 Fig. 3. Pearson correlation matrix indicating strong correlation between egg width and egg length, nest diameter and nest height, and clutch mass and egg width. Egg mass, egg length, clutch mass, and nest diameter were used as representative variables in subsequent analyses. 0.507. The JDM WMA data recovered a 69.9% significance level (0.699 as a theoretical P-value). Six hundred ninetyeight permuted statistics were greater than or equal to a Rho value of 20.145. These values suggest that the morphometrics between maternal individual and clutch are significantly correlated in the LCM population and are not in the JDM WMA population. DISCUSSION Multiple studies have examined reproductive allometric correlations in reptiles. For instance, Warne and Charnov (2008) examined classic life-history trade-offs comparatively between 15 lizard families. Ford and Seigel (1989) assessed maternal-offspring size relationships in three oviparous snake taxa. Much allometric discussion has been presented in a theoretical and statistical context (Smith, 1980; Harvey, 1982; King, 2000; Bonnet et al., 2003) and some in the context of comparative physiological constraint (Sinervo, 1993; Bowden et al., 2004). This study tests the use of reproductive allometry as a relative indicator of population marginality or stress and as a novel method in an intraspecific comparative framework. Heterophil to lymphocyte ratios have been utilized as a measure of stress in many vertebrates by indirectly profiling glucocorticoid levels (Davis et al., 2008). Salinity has been documented as a stressor in American Alligators (Lance et al., 2010), and this study also fails to falsify that hypothesis.
Murray et al. Alligator allometry 273 Fig. 4. Log-transformed comparison of allometric correlations between a stressed and unstressed population of Alligator mississippiensis. (A) Relationship between egg mass and maternal snout vent length (SVL) in stressed (gray) and unstressed (black) populations. (B) Egg length vs. SVL. (C) Clutch mass vs. SVL. (D) Nest diameter vs. SVL. Nest-site salinity ranges in the marginal population were considerably higher than the core population, and heterophil to lymphocyte ratios reflected this difference. Alligator nesting, reproductive activity, and physiological performance are negatively impacted by increased salinity (Lance et al., 2010). Salt-water inundation from Hurricane Ike in JDM WMA presented a stressor to induce negative impacts, and stress levels were high enough to reflect such negative effects. McNease and Joanen (1978) and Chabreck (1988) noted that the greatest alligator densities are found in intermediate salinity marshes. This observation coupled with results presented here indicates that stochastic salinity stress from salt-water inundation can have a significant population level effect on alligator physiology and resource acquisition, and potentially social hierarchy and density, even in habitat that supports the largest abundances. Before assessing the use of reproductive allometry as an indicator of marginality it is critical to assess which morphometric parameters are useful comparisons that exhibit allometric correlation between maternal individuals and their respective reproductive output. Suitable morphometric comparisons are based on the allometric relationships of the core population at LCM that revealed statistically significant correlations between all five maternal morphometrics and all seven nest and clutch morphometrics. Egg bandwidth, an indicator of embryonic developmental progress (Masser, 1993), was also significantly correlated with all maternal morphometrics in the core population and some in the marginal population. These results potentially indicate a relationship between maternal size and the timing of egg deposition or egg developmental rate. Nests in both study sites were not sampled separately in time; in fact, nest sampling alternated between study sites such that egg bandwidth was not confounded by sampling and developmental time. This finding lends credence to the presence of a social hierarchy based on size and suggests that larger Table 1. Standard Major Axis Regression Analysis Testing Effects of Stress on the Allometric Relationships between Maternal Morphometric Parameters and the Morphometrics of Their Respective Clutch and Nest. Slope estimates, elevation estimates, R squared correlation coefficients, and P-values for difference between populations are reported for stressed (marginal) and unstressed (core) populations. Slope (Marginal) Elevation (Marginal) Slope (Core) Elevation (Core) P-values Slope Elevation R 2 -value Marginal ASVL (cm) 3 EL (mm) 23.37 21.62 1.37 21.57,0.001,0.001 0.03 0.022 ASVL (cm) 3 ND (cm) 20.86 1.14 0.34 1.18,0.001,0.001 0.01 0.23 ASVL (cm) 3 CM (g) 0.39 0.75 0.77 0.288,0.001,0.01 0.12 ASVL (cm) 3 EM (g) 0.61 0.41 1.51 0.47,0.001,0.001 0.02 0.19 Core
274 Copeia 2013, No. 2 Fig. 5. (A) RELATE analysis assessing the level of correlation between mother and nest/clutch morphometrics from Las Conchas Marsh using 999 permutations and Euclidean distance resemblance matrices (Rho 5 0.507, significance level 5 0.5%). (B) RELATE analysis assessing the level of correlation between mother and nest/clutch morphometrics from J. D. Murphree WMA using 999 permutations and Euclidean distance resemblance matrices (Rho 520.145, significance level 5 69.9%). females nest earlier than smaller individuals. This indicates that larger females have first priority for nesting site choice and hatchlings have the advantage of increased developmental time prior to fall torpor. Of further interest is the significant correlation between actual nest morphometrics (diameter and height) and female size. Nesting material included Spartina patens and Phragmites in both populations. This indicates that larger females are capable of manipulating larger amounts of nest vegetation. Joanen (1969) notes that old fully grown females will build larger than average size nests. To our knowledge no other study empirically quantifies this relationship. When physically stressed, however, the energetic costs of building large nests may outweigh the potential benefits. Further research is warranted to test allometric hypothesis regarding females and nest dimension as well as bandwidth variation. The hypothesized marginal population had weaker allometric correlations, statistically different elevations, and lower slopes of lines best fit relative to the hypothesized core population. Biologically, the unstressed population conformed to a big mom big offspring hypothesis yielding strong positive relationships between maternal morphometrics and nest and clutch morphometrics. In contrast, the stressed population yielded relatively uncorrelated comparisons between mother and nest/clutch parameters or negative relationships. Negative correlations may be a product of low sample size and these relationships may in fact be random, although Thorbjarnarson (1996) also suffered from low intraspecific sample sizes and found no negative relationships. Elevationally, mothers in the unstressed population deposited seemingly larger outputs than the stressed populations among all tested parameters. However, analysis of elevation using the dataset presented here does not hold biologically meaningful results based on our interpretation. Elevation, based on centroid for data from each group (SL and TX), is confounded by nonlinearity in the data with both population in some comparisons and TX allometric relationships in others (Fig. 4). While all data were log transformed, such a transformation did not compress the data enough to make linear allometric interpretations in all presented instances (Fig. 4). Because elevation is computed using centroids for each population, relative spread and non-linearity confound interpretation. Elevational differences were also not large enough hold much biological meaning in terms of relative reproductive output sizes. Interestingly, the retention of multiple slightly concave-up parabolas using logtransformed data suggests uniquely large reproductive outputs that may be attributed to a reproductive immaturity of sorts at smaller maternal sizes. In these situations mothers may over-invest in their reproductive output as a result of little reproductive experience or at least produce larger clutches than one would expect from their size. Additionally, the large difference in correlation coefficients between the stressed and unstressed populations suggests that the big mom big offspring hypothesis breaks down in stressed populations and variance in allometric relationships increases with stress. This alone fails to falsify Verdade s (2001) hypothesis that the strength of reproductive allometry is higher in unstressed populations. Non-parametric RELATE analysis yields similar results and recovers significant correlation using normalized raw data between all maternal morphometrics and all nest and clutch morphometrics in the core population. The marginal population yielded no significant correlation between maternal and clutch morphometrics. These data and all subsequent analyses indicate a more random size-based reproductive output in the marginal population than the core. Size-corrected clutch size and clutch mass variation also indicates a difference in overall reproductive energy allocation and life-history strategies between the marginal and core populations. The intraspecific allometric relationships between mother and clutch in the core population are relatively consistent with previous intraspecific allometric comparisons in the literature. Verdade (2001) and Larriera et al. (2004) found positive significant relationships between female size (SVL and BM) and EM, EL, and EW in wild and captive Caiman latirostris. Larriera et al. (2004) also found positive significant relationships between female size, egg volume and clutch size, and SVL and CM, which Verdade (2001) did not. Hall (1991) found positive significant relationships between female size (SVL, BM, TG, and TL) and CM as well as clutch size. Such data corroborates our findings for CM, as clutch size was not assessed. Further, Thorbjarnarson (1996) found positive significant relationships between
Murray et al. Alligator allometry 275 female size (TL) and egg mass in Caiman crocodilus, and CM in Caiman crocodilus, Crocodylus porosus, and Crocodylus acutus. Our results further corroborate those data using Alligator mississippiensis. This study fails to falsify the hypothesis put forth by Verdade (2001) that suggests that strong allometric correlations break down in fragmented populations. While this study only directly tests for differences in allometric relationships between a stressed and unstressed population, it is necessary to propose an explanation as to why this may be. In agreement with Verdade (2001), stress-induced fragmentation of social structure is the most likely posit; however, additional steps are necessary. Potentially, stressful conditions are resulting in social fragmentation through subsequent decrease in population density or resource availability. These conditions then result in partitioning of resources and reproductive allocation regardless of individual size. In addition, this case study supports the use of reproductive allometry as a method to intraspecifically assess the marginality of crocodilian populations relative to others and prioritize conservation efforts accordingly. Reproductive allometry analyses may be more useful in locations in which the analysis for stress hormone levels, leukocyte profiling, or other stress measures are too costly or otherwise unattainable. ACKNOWLEDGMENTS We express gratitude to K. Piller, D. Sever, J. Sutherlin, G. Gambel,D.Llewellyn,D.Kelley,R.Moreau,andH.Renoof Turtle Cove Biological Station and Wall Alligator Services for administrative assistance and guidance. We thank C. McMahan,R.Ackery,J.Lee,J.Price,J.Rheubert,andD. Steen for field and editorial help and O. Gaoue and M. Wooten for statistical advice. All research for this paper was conducted per the guidelines created by Southeastern Louisiana University s IACUC committee (protocol #0016), and all specimens were collected with an appropriate permit from the Louisiana Department of Wildlife and Fisheries (permit #LDWF Special Alligator Permit, NA). LITERATURE CITED Bonnet, X., R. Shine, O. Lourdais, and G. Naulleau. 2003. Measures of reproductive allometry are sensitive to sampling bias. Functional Ecology 17:39 49. Bowden, R. M., H. K. Harms, R. T. Paitz, and F. J. Janzen. 2004. Does optimal egg size vary with demographic stage because of a physiological constraint? Functional Ecology 18:522 529. Chabreck, R. H. 1972. Vegetation, water, and soil characteristic of the Louisiana coastal region. Louisiana Agricultural Experiment Station Bulletin 664. Chabreck, R. H. 1988. Coastal Marshes: Ecology and Wildlife Management. University of Minnesota Press, Minneapolis, Minnesota. Clarke, K. R., and R. N. Gorley. 2001. PRIMER v6. PRIMER-E Ltd., Plymouth, U.K. Davis, A. K., D. L. Maney, and J. C. Maerz. 2008. The use of leukocyte profiles to measure stress in vertebrates: a review for ecologists. Functional Ecology 22:760 772. Elsey, R. M., and P. L. Trosclair. 2004. A new live trap for capturing alligators. Herpetological Review 35:253 255. Ford, N. B., and R. Seigel. 1989. Relationships among body size, clutch size, and egg size in three species of oviparous snakes. Herpetologica 45:75 83. Gross, W. B., and H. S. Siegel. 1983. Evaluation of the heterophil/lymphocyte ratio as a measure of stress in chickens. Avian Disease 27:972 979. Hall, P. M. 1991. Estimation of nesting female crocodilian size from clutch characteristics: correlates of reproductive mode, and harvest implications. Journal of Herpetology 25:133 141. Harvey, P. H. 1982. On rethinking allometry. Journal of Theoretical Biology 95:37 41. Joanen, T. 1969. Nesting ecology of alligators in Louisiana. Proceedings of the Annual Conference of Southeastern Association of Game and Fish Commission 23:141 151. Kawecki, T. J. 2008. Adaptation to marginal habitats. Annual Review of Ecology, Evolution, and Systematics 39:321 342. King, R. 2000. Analyzing the relationship between clutch size and female body size in reptiles. Journal of Herpetology 34:148 150. Lance, L. A. 2003. Alligator physiology and life history: the importance of temperature. Experimental Gerontology 38:801 805. Lance, L. A., R. M. Elsey, G. Butterstein, P. L. Trosclair, III, and M. Merchant. 2010. The effect of hurricane Rita and subsequent drought on alligators in southwest Louisiana. Journal of Experimental Zoology 313A:106 113. Larriera, A., C. I. Piña, P. Siroski, and L. M. Verdade. 2004. Allometry of reproduction in wild broad-snouted caiman (Caiman latirostris). Journal of Herpetology 38:301 304. Manolis, C., and C. Stevenson. 2010. Crocodiles. Status survey and conservation action plan. IUCN/SSC Crocodile Specialist Group. Available via http:// http://www.iucncsg. org/pages/publications.html. Accessed 25 March 2012. Masser, M. P. 1993. Alligator production: breeding and incubation. Southern Regional Aquaculture Center, Publication No. 231. McNease, L., and T. Joanen. 1978. Distribution and relative abundance of the alligator in Louisiana coastal marshes. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 32:182 196. R Development Core Team. 2011. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Sinervo, B. 1993. The effect of offspring size on physiology and life history: manipulation of size using allometric engineering. BioScience 43:210 218. Smith, R. J. 1980. Rethinking allometry. Journal of Theoretical Biology 87:97 111. Systat Software. 2002. Systat 10.2. Systat Software, Inc., Richmond, California. Taylor, D., N. Kinler, and G. Linscombe. 1991. Female alligator reproduction and associated population estimates. Journal of Wildlife Management 55:682 688. Taylor, D. T., and W. Neal. 1984. Management implications of size-class frequency distributions in Louisiana alligator populations. Wildlife Society Bulletin 12:312 319. Thorbjarnarson, J. B. 1992. Crocodiles: An Action Plan for Their Conservation. IUCN, Gland, Switzerland. Thorbjarnarson, J. B. 1996. Reproductive characteristics of the order Crocodylia. Herpetologica 52:8 24.
276 Copeia 2013, No. 2 Verdade, L. M. 1997. Morphometric analysis of the broadsnouted caiman (Caiman latirostris): an assessment of individuals clutch, body size, sex, age, and area of origin. Unpubl. Ph.D. diss., University of Florida, Gainesville. Verdade, L. M. 2001. Allometry of reproduction in broadsnouted caiman (Caiman latirostris). Brazilian Journal of Biology 61:431 435. Warne, R. W., and E. L. Charnov. 2008. Reproductive allometry and the size-number trade-off for lizards. The American Naturalist 172:E80 E98. Warton, D. 2007. The smatr package for R. Version 2.1. http://cran.r-project.org. Warton, D. I., I. J. Wright, D. S. Falster, and M. Westoby. 2006. Bivariate line-fitting methods for allometry. Biological Reviews 81:259 291. Wood,J.M.,A.R.Woodward,S.R.Humphrey,andT.C. Hines. 1985. Night counts as an index of American alligator population trends. Wildlife Society Bulletin 13:262 272. Woodward, A. R., T. Hines, C. Abercrombie, and C. Hope. 1984. Spacing patterns in alligator nests. Journal of Herpetology 18:8 12. Zippel, K. C., H. B. Lilywhite, and C. R. J. Mladnich. 2003. Anatomy of the crocodilian spinal vein. Journal of Morphology 258:327 335.