<%BANNER%>

Body Condition Factor Analysis for the American Alligator (Alligator Mississippiensis)


PAGE 1

BODY CONDITION INDEX ANALYSIS FOR THE AMERICAN ALLIGATOR (A lligator mississippiensis) By CHRISTA L. ZWEIG A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Christa L. Zweig

PAGE 3

ACKNOWLEDGMENTS I thank my family for all of the incredible support they gave me while still letting me make my own mistakes. I am grateful to Zach Welch, who had the demanding duty of keeping me in high spirits. I also thank the members of my committee: Frank Mazzotti for offering his support and giving me the opportunity to pursue this degree, Ken Rice for providing me with all sorts of guidance, Laura Brandt for her thoughtful comments, and Franklin Percival for his excellent, practical opinions. There are many people to thank, and I wish I could enumerate the countless reasons why they are special to me, but then this section would be longer than my thesis. Everyone knows why they are on this list: Ab Abercrombie, Lindsey Hord, Mike Cherkiss, Ashley Traut, Gayle Martin, Becky Hylton, Diana Swan, Jamie Duberstein, Jocie Graham, Phil Wilkinson, Woody Woodward, Chris Tucker, Jason Williams, Geoff Cook, Phil George, Matt Chopp, Adam Finger, Stan Howarter, Cathy Ritchie, Cherie Cook, Sarah Kern, Monica Lindberg, Caprice McRae, Debra Hughes, Barbara Fesler, Marion Bailey, Mary Hudson-Kelley, Lauren J. Chapman, Peggy VanArman, the Palm Beach Atlantic College Science Club, IFAS statistics, Marinella Capanu, and Brenny. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.....................................................................................................................viii CHAPTER 1 INTRODUCTION.........................................................................................................1 2 EVALUATION OF FIELD MEASUREMENTS OF THE AMERICAN ALLIGATOR FOR USE IN MORPHOMETRIC STUDIES.......................................3 Introduction.....................................................................................................................3 Methods...........................................................................................................................3 Results.............................................................................................................................6 Discussion.......................................................................................................................6 3 CONDITION INDEX ANALYSIS OF THE AMERICAN ALLIGATOR..................8 Introduction.....................................................................................................................8 Study Area....................................................................................................................12 Materials and Methods..................................................................................................13 Capture Methods....................................................................................................13 Condition Indices Analysis....................................................................................14 Simulated Data.......................................................................................................15 Everglades Condition Analysis..............................................................................16 Results...........................................................................................................................16 Condition Indices Analysis....................................................................................16 ANOVA/ANCOVA Analysis by Capture Event...................................................17 Simulated Data.......................................................................................................18 Everglades Condition Analysis..............................................................................18 Discussion.....................................................................................................................19 4 CONCLUSION............................................................................................................37 iv

PAGE 5

APPENDIX A DATA SHEETS AND MEASUREMENT INSTRUCTIONS GIVEN TO EXPERIMENT PARTICIPANTS...............................................................................38 B GENERAL PICTORIAL MORPHOMETRIC MEASUREMENT INSTRUCTIONS.........................................................................................................41 LIST OF REFERENCES...................................................................................................45 BIOGRAPHICAL SKETCH.............................................................................................49 v

PAGE 6

LIST OF TABLES Table page 1: T-test results for coefficient of variation of experienced individuals versus inexperienced groups of volunteers measuring American alligators at A.R.M. Loxahatchee National Wildlife Refuge...................................................................7 2: Within alligator variance and covariance components for length and volumetric measures...................................................................................................................7 3: Descriptive statistics of indices used in condition analysis for the American alligator..................................................................................................................24 4: Condition indices calculated for the American alligator whose Pearson r-values are less than 30%...................................................................................................25 5: Results of ANOVA/Least Significant Difference and ANCOVA/ Least Significant Difference analyses of condition indices for the American alligator..26 6: ANOVA/ANCOVA analyses of condition indices by capture event from October 1999 to Spring 2002 (n = 395). Order of condition index is from highest to lowest mean. There were no ENP-Est captures for March/April 00.................................27 7: Ability of ANOVA/LSD analysis to detect differences in condition from Relative K simulated data....................................................................................................28 8: Ability of ANOVA/LSD analysis to detect differences in condition from Fultons K simulated data....................................................................................................28 vi

PAGE 7

LIST OF FIGURES Figure page 1: Flowchart of condition indices analysis for the American alligator in the Everglades..............................................................................................................29 2: Capture areas in south Florida within Everglades wetlands core..........................30 3: Linear regression of tail girth vs. snout-vent length of the American alligator of captures in south Florida from October 1999 to March 2002................................31 4: Range of Fultons K index values by capture area in south Florida. Captures are from October 1999 to March 2002........................................................................32 5: Hierarchy of Everglades alligator condition from October 1999 to March 2002 using Fultons K analysis.......................................................................................33 6: Sample condition range of alligators for Fultons K from Everglades data from October 1999 to March 2002.................................................................................34 7: SVL/Mass Fultons K analysis for the American alligator in areas of South Carolina, north Florida and the Everglades...........................................................35 vii

PAGE 8

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science BODY CONDITION FACTOR ANALYSIS FOR THE AMERICAN ALLIGATOR (Alligator mississippiensis) By Christa Zweig May 2003 Chair: Frank Mazzotti Major Department: Wildlife Ecology and Conservation The American alligator (Alligator mississippiensis) is an integral part of the Everglades ecosystem. They affect and are affected by the landscape and changes in hydrology, which makes them an excellent organism to use in evaluating current Everglades restoration efforts. Information on alligatorscensus data, capture or morphometric data, blood chemistry, and reproduction statisticshas been collected in the Everglades since the 1950s. Historical information provides a suite of useful life history characteristics or population parameters (i.e., health and condition, nesting effort, growth rate and survival, and density) that can be used for evaluating restoration. However, some life history traits (e.g., absolute density and survival) are difficult to accurately measure and may take decades of data to detect trends. Body condition can be measured using indices and requires less data to begin an analysis. If used properly, condition can be a useful monitoring tool to assess the health of a population and thus the health of its ecosystem. This project evaluated morphometric measurements taken for the viii

PAGE 9

American alligator, available condition indices, and, using a stepwise process, recommends the appropriate index for use in ecological applications. We analyzed morphometric measurements of captured animals to determine which are measured most accurately and are appropriate for condition analyses. Condition indices are functions of a body length indicator and a volumetric measurement and are only as accurate as the measurements used. Head length, snout-vent length and total length are suitable for body length indicators, and tail girth, neck girth, chest girth, and mass can all be used as volumetric measurements. We then compared four condition indices and two models of volume/length relationships for their ability to distinguish between populations with known qualitative condition differences. It was determined through ANOVA/LSD analysis of the condition indices that the HL/Mass combination of Fultons K and the SVL/Mass combination of ANCOVA were best able to distinguish differences in condition between areas of the Everglades; however ANCOVA can not be used to compare across populations, unless strict assumptions about the two populations can be met. HL/Mass Fultons K can be used to spatially and temporally compare populations of the American alligator, and is suggested by this study as the best condition factor to use for these purposes. The primary objective of this study was to determine the appropriate condition index for the American alligator with the intention that it will be used as a monitoring tool of population health for the current restoration process. Because the alligators life history is so closely linked to hydropattern, body condition can reflect the impacts of changes in hydrology. ix

PAGE 10

CHAPTER 1 INTRODUCTION The American alligator (Alligator mississippiensis) ranges from eastern Texas to North Carolina and south to the tip of Florida. Alligators are vital components of wetland ecosystems, serving as keystone predators (Mazzotti and Brandt 1994) and exhibiting the qualities of ecosystem engineers (Jones et al. 1994), creating critical landscape features. In the Everglades, they create topography in a system where small changes can be very important. They construct alligator holes for thermal refuge, feeding, and reproduction by making depressions in the muck soil and keeping them clear of surrounding vegetation. These areas hold water during dry-down periods and provide refugia for aquatic and amphibious animals during the dry season (Craighead 1968; Kushlan 1974; Campbell 1999). Alligators also create dry land in the marsh by mounding vegetation to build their nests. Habitat for less flood-tolerant plants is created. Further, it has been documented that other reptiles use alligator nest mounds to deposit their eggs (Kushlan and Kushlan 1980; Hall and Meier 1993; Enge et al. 2000). The natural history and reproductive patterns of the American alligator are closely tied to its physical environment. There is a relationship between spring water levels and late summer water levels in a premanaged landscape (Kushlan and Jacobsen 1990; Mazzotti and Brandt 1994). Based on spring water levels, alligators construct nests that will be above historical average high water in the summer, wet season. However, because water levels in the Everglades have been altered from historical seasonal fluctuations, unnatural water regimes may cause nest flooding (Kushlan and Jacobsen 1

PAGE 11

2 1990). Further, water levels can affect body condition of Everglades alligators (Dalyrmple 1996). The American alligator is an integral part of the Everglades ecosystem. They effect and are affected by the landscape and changes in hydrology, which makes them an excellent organism to use in evaluating current Everglades restoration efforts. Information on alligatorscensus data, capture or morphometric data, blood chemistry, and reproduction statistics, has been collected in the Everglades since the 1950s (Rice et al. 2002). Historical information provides a suite of useful life history characteristics or population parameters (i.e., health and condition, nesting effort, growth rate and survival, and density) that can be used for evaluating restoration. However, some life history traits (e.g., absolute density and survival) are difficult to accurately measure and may take decades of data to detect trends. Condition is the relative fatness of [an animal]. . a measure of how well that animal is coping with its environment (Taylor 1979, p. 349), and can be measured using indices and requires less data to begin an analysis. If used properly, condition can be a useful monitoring tool to assess the health of a population and thus the health of its ecosystem. This project evaluates morphometric measurements taken for the American alligator (Chapter 2) and available condition indices, and using a stepwise process (Figure 1), recommends the appropriate index for use in ecological applications (Chapter 3).

PAGE 12

CHAPTER 2 EVALUATION OF FIELD MEASUREMENTS OF THE AMERICAN ALLIGATOR FOR USE IN MORPHOMETRIC STUDIES Introduction Long-term data are typically collected by different people under varying field conditions, resulting in data sets that are difficult to standardize. Several researchers have examined this problem in varied taxa, such as birds (Grant 1979; Zink 1983; Arendt and Faaborg 1989; Yezerinac et al. 1992), bats (Palmeirim 1998), and snails (Bailey and Byrnes 1990). However, they have used museum specimens or shells. Live animals are even more difficult to standardize. The most common data collected on crocodilians are morphometric measurements (e.g., head length, snout-vent length, total length, and weight). These measurements are used in a wide range of analyses, from ecological to taxonomic to evolutionary (Chabreck and Joanen 1979; Hutton 1986; Hall 1991). However, these analyses are only as accurate as the data from which they are derived. Methods We performed morphometric measurement trials on the American alligator to quantify inter-measurer error and determine which morphometric measurement had the least measurer error associated with it. This is necessary to increase confidence levels in future analyses. Two trials were performed: 1) measurers were inexperienced volunteers; 2) measurers were experienced alligator biologists. The inexperienced trial was performed Chapter reprinted with permission from the editors of Herpetological Review (in press) 3

PAGE 13

4 in the fall of 2000 using ten alligators captured at Arthur R. Marshall Loxahatchee National Wildlife Refuge (Loxahatchee) located in Boynton Beach, FL, with six groups of volunteers. Alligators were located by observing eye shines from an airboat in the marsh interior and captured using a wire snare. They were secured in the boat and brought back to a covered area for measurement. Alligator size varied from 108-248 cm total length. The volunteers consisted of Loxahatchee staff, University of Florida employees, U.S. Geological Survey employees, and students from Palm Beach Atlantic Community College. Each group consisted of 4-6 people and they were given both verbal and written measurement instructions (Appendix A). Each group measured every alligator. Different individuals within the group took measurements of each alligator, but measurements were agreed upon within the group before being recorded. Participants were allowed to collaborate within but not between groups. The following were measured by each group on every alligator: head length (HL), snout-vent length (SVL), total length (TL), right hind foot length (HFL), neck girth (NG), tail girth (TG), chest girth (CG), and mass. HL was measured dorsally, while SVL and TL were measured ventrally. A measurement kit was provided with every alligator that contained the following: a clipboard, pencil, string for measuring tail girth between scutes, a Pesola scale, and a flexible centimeter sewing tape. These kits stayed with the alligator so that the same equipment was used. Measurements were made with the flexible sewing tape to the nearest 0.1 cm. Mass was measured with 10-50 kg Pesola scales to the nearest 0.1 kg. The scales were calibrated before use using a weight of known mass. Alligators also were marked with individually numbered size 3 Monel tags, provided by the Florida Fish

PAGE 14

5 and Wildlife Conservation Commission, in the webbing of the back left foot and scute clipped for permanent identification. The second trial was performed in fall 2001 with experienced alligator biologists. Nine alligators ranging from 156 cm to 255 TL were captured at Loxahatchee and used for the trial. Ten biologists were present and measured each alligator using the same procedures and instructions as in the previous experiment with three exceptions: 1) each individual measured every alligator and was not allowed to discuss their findings; 2) the participants worked in groups of two to facilitate measuring; and 3) each person recorded their own measurements to reduce bias in and among groups. Data were analyzed to determine which measurements were most reliable, or contained the least percent measurement error (%ME). Reliability for between experience and inexperienced groups was evaluated using coefficient of variation. Standard deviation for each measurement was calculated and divided by the mean of that measurement for that alligator. A T-test was performed on the coefficients of variation to determine significant differences in the reliability of the measurements between experienced and inexperienced measurers. Six of the experienced measurers were randomly chosen so that the N in the T-test would be equal to the six groups of inexperienced volunteers. Percent measurement error was calculated for the experienced group to determine which measurements are more reliable for morphometric studies. Bailey and Byrnes (1990) pioneered the use of Model II ANOVA and ANCOVA to estimate within-individual and among-individual components of covariance and variance to predict percent measurement error. A Model II ANOVA and ANCOVA was run as a part of the

PAGE 15

6 SAS NESTED procedure (SAS Institute, Inc. 1988), according to the Bailey and Byrnes procedure. The among-individual variance was calculated by the following equation (Yezerinac, et al. 1992): s 2 among = (MS among MS within ) / m, where MS is the mean squared deviation and m is the number of repeated measurements. Covariances were calculated as follows: r x.y among = (cov x.y among / s x among s y among ). %ME was then determined by: %ME = 100% (s 2 within / s 2 within + s 2 among ). Results The results of the T-test (Table 1) suggest that experienced individuals measure more accurately for HL, CG, TG, and mass than groups of inexperienced volunteers. There is no statistical difference for SVL, TL, HFL, and NG. Percent measurement error ranged from 0.50 to 49.53% (Table 2). HFL had the highest %ME because the landmarks used to measure it are not easily located. TL was also relatively high. It should be one of the most reliable measurements, as it is the largest and allows for greater margin of error. Its interesting to note the high covariance that NG and CG share with the other volumetric measurements. Discussion This assessment can be used as a guide for future crocodilian studies that use morphometric measurements, such as the analysis of growth rate and condition. If performing morphometric analysis using data collected from inexperienced groups, it would be more accurate to use TL, SVL, or NG. For experienced individuals, HL, SVL, TG, or weight should be used. For example, condition factor analysis requires a skeletal

PAGE 16

7 measurement and a volumetric measurement. SVL would be adequate for the skeletal measurement regardless of who collected the data, but more care should be taken in choosing the volumetric measurement. If these trials were to be replicated, it would be useful to modify the design so that each individual would take three or more measurements of the same measurement (e.g. three measurements of HL, three of SVL, etc.). This would provide data to quantify intra-observer error. Table 1: T-test results for coefficient of variation of experienced individuals versus inexperienced groups of volunteers measuring American alligators at A.R.M. Loxahatchee National Wildlife Refuge Measurement P-value HL 0.046* SVL 0.098 TL 0.129 HFL 0.758 CG 0.010* NG 0.284 TG 0.015* Mass 0.010* Table 2: Within alligator variance and covariance components for length and volumetric measures. Diagonal elements are variance components represented by percent measurement error (%ME). The off-diagonal elements are covariance components represented by within-gator correlations (r within ). HL SVL TL HFL NG CG TG Weight HL 0.79 SVL -0.07 1.52 TL 0.02 0.01 3.88 HFL 0.09 -0.15 0.01 49.53 NG -0.04 0.16 -0.15 0.14 7.64 CG -0.03 0.10 0.16 -0.22 0.31 4.89 TG -0.06 0.09 0.10 0.10 0.63 0.40 2.93 Weight -0.08 -0.08 0.09 0.04 -0.25 -0.10 -0.13 0.50

PAGE 17

CHAPTER 3 CONDITION INDEX ANALYSIS OF THE AMERICAN ALLIGATOR Introduction Condition as defined in chapter one is the relative fatness of [an animal]. . a measure of how well that animal is coping with its environment(Taylor 1979, p. 349). This relationship and its relative ease of measurement is the key to using the American alligator (Alligator mississippiensis) as an indicator of the health of its environment. Condition indices are often used to quantify body condition over space and time (LeCren 1951, Taylor 1979, Springer 1990, Brandt 1991, Elsey 1992, Krebs and Singleton 1993, Dalrymple 1996, Jakob 1996, Barr 1997, Leslie 1997, Delany 1999). Biologists often note that an alligator is too skinny or a healthy size, but those observations are qualitative. For example, this project was developed because of observed condition differences between alligators within parts of the Everglades system. The difference was apparent, but not quantified. Quantitative condition indices can compare condition within and among populations, and if used carefully, can provide insights into the status of ecosystems (Murphy et al. 1990, p. 86) Condition indices can be calculated in several different ways, but is always a function of skeletal length and a volumetric measurement. They have been used in fisheries literature since the 1940s, where they have become deeply entrenched as a research paradigm (Cone 1989). LeCren first reviewed this process in 1951. He described one relationship on which the condition indices are based: (1) W = aL n 8

PAGE 18

9 where W = weight, L = length, a is a constant and n is an exponent that has been calculated in past literature (LeCren 1951) to be between 2.5 and 4.0 for fish body forms. In most fisheries literature, n = 3.0, the number used when growth is isometric. Isometric growth occurs when both factors, i.e. length and weight, grow at the same rate. If it is accepted that growth is isometric and n equals 3.0, then a condition index equation can be written as: (2) K = W/L 3 x 10 n where n = 2,3,4, or 5 and is commonly chosen so that the mean of K is larger than one (Cone 1989) and K is known as Fultons condition factor. This equation is quite limited by its assumption of isometric growth and that n is equal to 3.0. To solve the problem of the strict assumptions of K, LeCren (1951) formulated the relative condition factor, or K n Instead of using an isometric length/weight relationship (3.0), he used empirical data and a least squares regression to formulate n from (1): (3) K n = W/aL n The primary difference between the two Ks is that Fultons condition factor, K, measures the deviation of an individual from an ideal, theoretical organism while the relative condition factor, K n, measures the deviation of an individual from the average of actual population data. K n is often used outside of fisheries studies where growth is usually allometric and n is estimated from field data. Relative condition factor is also limited in its use. Because you must calculate the ideal length/weight standard separately for each population, it is impossible to compare the condition factor value across populations (Murphy et al. 1990).

PAGE 19

10 A third and more recent condition index is relative weight, W r This was postulated by Wege and Anderson (1978) and is the ratio of an individuals weight (W) divided by a standard weight (W s ) for that length times 100, or (4) W r = (W/W s ) x 100. W s is an optimal standard which establishes the index value for a species. It is regularly defined in fisheries literature as the 75 th percentile (Murphy et al. 1990), which is considered optimal growth. Relative weight avoids the drawbacks of Fultons condition factor and relative condition factor, because it doesnt vary with changing species or mean sample size (Springer and Murphy 1990; Quist et al. 1998). The residual index is another method of calculating condition. The residual index is calculated by regressing the volumetric measurement on body length. The residual distance of the individual points to the regression line functions as the condition index. Some data has to be transformed so that it meets the assumptions of regression. A condition index also must be independent of body length, which can limit morphometric measurements that can be used (Green 2001). The residual index possesses the same flaw as relative condition factor in that it is not comparable across populations, because the regression line is only valid for the current population. Researchers also have used ANCOVA to evaluate condition (LeCren 1951; Packard and Boardman 1988; Garcia-Berthou and Moreno-Amich 1993; Garcia-Berthou 2001). ANCOVA allows the length-weight relationship to be examined without the confounding effects of the covariance of body length. Packard (1988) and Garcia-Berthou (2001) rule out other ratio indices for condition and endorse ANCOVA as the best method for calculating condition. However, it suffers from the same fault as relative condition and

PAGE 20

11 the residual index in that it is not possible to compare the condition of two populations with this method, unless the n in equation 1 is identical for the two populations being compared. While condition index analysis is best known in the fisheries literature, it also has been used for crocodilians. Condition index analyses (using relative condition factor) have been used during studies on the food habits of alligators in Everglades National Park (Barr 1997); to compare condition and growth of juvenile alligators in Par Pond, South Carolina (Brandt 1991); and to compare condition index for Everglades alligators in Shark Valley (Dalrymple 1996). Elsey et al. (1992) used growth rates and body condition factors to compare wild and farm-released juveniles in Louisiana; and Santos, et al. (1994) compared different habitats of the Pantanal in Brazil using condition factor of Caiman crocodilus yacare. Leslie (1997) used relative condition factor to determine the condition of a population of Crocodylus niloticus in South Africa. Australian researchers have also explored condition relating to crocodiles (Garnett and Murray 1986; Grigg et al. 1986). Many condition index analyses include an actual measure of condition by sacrificing animals and extracting total body fat (Krebs and Singleton 1993; Jakob et al. 1996; Quist et al. 1998; Sutton et al. 2000; Vila-Gispert and Moreno-Amich 2001). The body fat percentage is then regressed on the condition index to obtain a relationship. None have been done with as large an animal as an alligator. It was not feasible to digest whole alligators for this study. Alligators are qualitatively fatter or skinnier in different areas of the Everglades, and some differences in time have been noted in the literature (Dalrymple 1996, Barr

PAGE 21

12 1997). Therefore a goal of this study was to find a condition index that is able to distinguish the greatest number of differences between these Everglades areas. The main objective was to determine the appropriate condition index for the American alligator with the intention that it will used as a tool to monitor population health. Because the alligators life history is so closely linked to the water levels of their environment, condition is a way to determine how changes in their environment are affecting alligators. With this monitoring tool, condition could be used as a performance measure for the current Everglades restoration effort. Study Area Data for this study were collected from six areas within the Greater Everglades Ecosystem in south Florida (Figure 2). The northernmost area, Water Conservation Area 1, which was designated as part of the Arthur R. Marshall Loxahatchee National Wildlife Refuge (LOX) in 1951, is located in Boynton Beach, FL. It is a 572 square kilometer area bounded by the West Palm Beach Canal on the north and the Hillsboro Canal to the south. Alligators also were captured in Water Conservation Areas 2 and 3 (WCA2 and WCA3), which are operated by the Florida Fish and Wildlife Conservation Commission (FWC) and the South Florida Water Management District (SFWMD). WCA2 is a 448 square kilometer pool used for excess water from LOX, supplies water to urban areas of southeast Florida, and wildlife conservation. Two locations were sampled for alligators in WCA3, one at the southern end and other in the middle. WCA3 is 1948 square kilometers of storage for excess water from WCA2 and also is designated a water supply area for urban areas in south Florida and for wildlife conservation.

PAGE 22

13 Two sites within Everglades National Park were used: Shark Slough and the estuarine areas located near Flamingo. Shark Slough extends from the northern border of Everglades National Park to the headwaters of the Shark and Harney Rivers and is the largest drainage in the Park (Olmstead and Armentano 1997). All estuarine alligators were caught near Flamingo in Everglades National Park. Materials and Methods Capture Methods Alligators were captured during the period from October of 1999 to March/April 2002 by a multi-agency team that consisted of members from U.S. Fish and Wildlife Service (USFWS), U.S. Geological Survey (USGS), University of Florida (UF), and the FWC. Animals were captured from all study areas in marsh habitats only, excluding canal alligators, as their condition may differ from marsh alligators because of the altered habitat. Due to a concurrent aging study, only 1.22-1.83 meter alligators were caught during the first two catches. However, the size range restriction was eliminated in October of 2000 in include a larger size range for an improved analysis. Fifteen alligators were captured from each area using airboats in marshes and using motorboats in estuaries during each of six capture events. Alligators were located by spotlight and captured with a noose or dart. Head length (HL), snout-vent length (SVL), total length (TL), right hind foot length (HFL), neck girth (NG), tail girth (TG) and chest girth (CG) were measured to the nearest 0.1 cm (Appendix B). Mass was measured with a Pesola spring scale to the nearest 0.1 kg. Alligators were sexed and blood was drawn for a concurrent contaminants study and for hematocrit analyses. The alligators were then released at their original capture site.

PAGE 23

14 Measurements from several animals were eliminated from this analysis due to discrepancies. For example, captures with bobtails (an alligator missing part of its tail), missing data, or known anomalies such as an alligator that is fed by people and is heavier than a typical Everglades alligator. Condition Indices Analysis A flowchart of the process used to choose the most appropriate length and volume and condition index for this analysis is provided in Figure 1. Several length and volumetric measurements were eliminated from consideration due to measurement error (Chapter 1). Condition indices were located in the literature (see Introduction for discussion of each index) and calculated using the four combinations from Chapter 1: HL/TG, HL/Mass, SVL/TG, SVL/Mass. The slash used in this notation represents the combination of two morphometric variables, not the division of one by the other. The data were pooled (n = 395, max TL = 264 cm, min TL = 100.4 cm) for all areas when calculating all indices to provide the largest sample size, since they could be considered members of the Greater Everglades population. The regression for Relative K was run using PROC NLIN in SAS and the regression for the residual index was run using PROC REG in SAS (SAS Institute 1988). A Pearsons correlation was run on the results of the five indices (Fultons K, Relative K, Relative Mass, Residuals, and Ratio index), using the four body length and volume combinations, to determine whether body length was correlated to condition. Combinations with correlations greater than 30% were eliminated from further consideration, as this suggested lack of independence. Condition index should be independent of the body size indicator or straight measurements for the assumptions of the ordinary least squares regression used in these analyses (Green 2001).

PAGE 24

15 An ANVOA was used (SAS PROC GLM [SAS Institute, Inc.1988]) to determine whether condition differed (p < 0.05) by area for all catch events combined for the indices. Least Significant Difference (LSD) and Tukeys post-hoc tests were run using SAS to establish which areas differed significantly. Both were calculated to establish which tests were sensitive enough for our analysis. Both tests calculate differences between areas, but Tukeys is more conservative than LSD. The condition model using ANCOVA was run using PROC GLM in SAS (SAS Institute 1988) and results were analyzed with the LSD post-hoc test to determine differences among areas. Area differences were counted if an area was significantly different than another according to the LSD analysis. Among six areas there are fifteen possible differences and they are written as follows: 1|2, 1|3, 1|4, 1|5, 1|6, 2|3, 2|4, 2|5, 2|6, 3|4, 3|5, 3|6, 4|5, 4|6, 5|6, where area one is different than two, area one is different than three, etc. Condition indices combinations that delineated significantly fewer populations were eliminated. Body length and volumetric combinations of indices that displayed the maximum amount of differences were used for temporal analysis. Those data were divided by capture event and the analysis was rerun using the capture event as an added parameter for the ANOVA analysis. The same was also performed with body length and volumetric combinations of ANCOVA. Capture events involved were October 1999, March/April 2000, October 2000, March/April 2001, October 2001, and March/April 2002. There were no ENPEst alligators for the March/April 2000 capture event. Simulated Data To address the question of whether the differences found in the ANOVA and ANCOVA analysis were Type I or Type II errors, simulated data sets were constructed to

PAGE 25

16 test the analysis. Simulated data sets consisted of 15 populationsfive populations of 100 individuals with the same condition index mean and standard deviation as the original data set, five populations of 100 individuals with a 5% increase in the mean and the same standard deviation as the original data set, and five populations of 100 individuals with a 10% increase in the mean and the same standard deviation as the original data set. The populations were created using the RV.NORMAL syntax in SPSS (SPSS Inc. 2001). ANOVA and LSD were used to analyze the simulated data. ANCOVA data were created in a similar manner. Random populations of HL and SVL were also generated with the RV.NORMAL syntax in SPSS (SPSS Inc. 2001). Mass was then regressed on HL and SVL using PROC REG in SAS (SAS Insitute 1988) and the resulting equation was used to generate mass. Everglades Condition Analysis A Fultons K analysis was performed using HL and Mass. K was calculated for all captures and was analyzed using an ANOVA and LSD post-hoc test to detect differences between the areas in the Everglades. Areas were then grouped by significant differences in mean condition. Results Condition Indices Analysis The mean, minimum, and maximum values for the indices were variable, even within indices (Table 3). The b value calculated for Relative K HL/Mass, HL/TG, SVL/TG combinations is significantly different than 3.0 (2.83 0.05 S.E., 1.03 0.02 S.E., and 1.05 0.02, respectively). The SVL/Mass combination was not significantly different than 3.0 (3.01 0.04).

PAGE 26

17 For the ratio index, both mass combinations (HL/Mass and SVL/Mass) were eliminated from the analysis because they were correlated with the body length by more than 30% (Table 4). However, Fultons K mass combinations were included while those calculated from TG were excluded. All Relative K combinations were incorporated into the analysis, as were all residual indices. HL/Mass was excluded from the analysis for Relative Mass because of its high correlation value (30%). ANOVA analyses were performed on the remaining condition indices to determine if there were significant condition differences between areas (Table 5). All were significant except the residual index HL combinations. Several indices; ratio, Relative K, and relative mass HL/TG combinations, had no more than three differences between areas, according to the LSD post-hoc test. Tukeys test was deemed too conservative for the analysis, because it did not show differences for a length and volumetric combination that had a highly significant ANOVA p-value (p < 0.01). Two condition indices; Relative K and Fultons K HL/Mass combination, display seven differences between areas. The ANCOVA analysis only includes two significant differences out of four combinations (Table 5). HL/Mass and SVL/Mass found seven and eight LSD post-hoc area differences, respectively. HL/TG and SVL/TG display higher p-values (by a factor of 10) than any of the condition indices. ANOVA/ANCOVA Analysis by Capture Event Fultons K, Relative K HL/Mass, and ANCOVA SVL/Mass and HL/Mass were chosen for further analysis by capture event because of the high number of differences they exhibited in the first ANOVA analysis. Sample size for capture events ranged from 44 to 83 and had no effect on the significant differences. There were no patterns of

PAGE 27

18 significance due to sample sizes. Results for the capture event ANOVA and LSD tests were similar for the Fultons and Relative K HL/Mass combinations (Table 6). There was no difference in the order of means (from highest to lowest). Fultons K only found one additional condition difference among LOX and WCA2 in the first capture event that Relative K did not. They found significant condition differences in the same three capture events and their p-values were similar. The HL/Mass ANCOVA analysis showed differences for the first and last capture event. No area differences were found in the other capture events. The SVL/Mass combination was not able to differentiate between areas for any capture events (p> 0.10). Simulated Data The simulated data analysis confirmed that the procedure of evaluating the condition indices does not commit any Type I errors. For both Relative and Fultons K, the analysis never found differences that were not there. The populations with the exact mean and standard deviation as the original data were always significantly different than the populations with the 10% increase in the mean. However, this did not hold true for the 5% difference. Percent differences ranged from 64 to 100 percent (Tables 7 and 8). The method for creating simulated data for the ANCOVA analysis was not successful. The random HL and SVL data were correct for mean and standard deviation, however the method of creating mass for the HL and SVL did not provide results. The simulated mass did not conform to the bounds of the original data and could not be used in a simulated data analysis. Everglades Condition Analysis In this study, during the 1999-2002 sampling period, areas within the Everglades do show significant differences in condition (Figure 4). ENP-SS exhibited the highest

PAGE 28

19 mean condition and is significantly higher than any other area (Figure 5). The two areas of WCA3 display lower condition than ENP-SS, but are still significantly higher than WCA2, the lowest mean of the groups. The LOX and ENP-Est condition means are significantly lower than ENP-SS and are not higher than any other group. Discussion Condition indices are a controversial way of estimating condition for populations (Packard and Boardman 1988; Cone 1989; Krebs and Singleton 1993; Jakob et al. 1996; Kotiaho 1999; Garcia-Berthou 2001; Green 2001; Vila-Gispert and Moreno-Amich 2001). They arent appropriate for every situation and there are strict assumptions that must be met for them to be used correctly. The main assumptions are that condition is independent of body length, that the measure used for body length is an accurate measure of structural size, that body length measurements are not subject to error, and that the relationship between body size and mass is linear (Green 2001). However, they can be a good, non-destructive way to rapidly evaluate the health of populations of concern. This analysis was designed to evaluate the use of condition indices for alligators. The most accurate measurement of an alligators volume is mass. However, because large alligators often can not be weighed in the field, TG was proposed as an easily measured alternative. It is highly correlated to SVL (Figure 3), and is where alligators store fat, so it should be a good measure of fat stores. However, the b that was calculated for TG combinations for Relative K were calculated to be approximately 1.0, instead of the 3.0 that Fultons K supposes for isometric growth. SVL/Mass and HL/Mass were approximately 3.0. The TG combinations for Fultons K were too highly correlated with their straight lengths to be included in the analysis, but it would have been

PAGE 29

20 incorrect to use them because their b value is too far away from 3.0 for it to be a valid use of Fultons K. The ANOVA/ANCOVA analysis provided a simple way to narrow the condition indices down to a manageable number. Fifteen of 19 index combinations had four significant condition differences or less, as shown by the LSD post-hoc test. They are still valid ways of comparing condition, especially where HL was not measured, but were deemed not sensitive enough for this analysis. The simulated data set allayed concerns of artifacts in the data allowing differences to be found by the LSD post-hoc tests that were not legitimate. It also allowed us to determine the range of this analysis. It could accurately detect condition differences of 10% or more. It also can detect differences of 5%, but only at a range of 64-96% (Tables 7 and 8). One can be safe, however, in supposing that a difference found is truly a difference in the data. The condition indices included in the final analysis produced different results. The HL/Mass combinations of Relative K and Fultons K were very similar in the ordering of means and number of differences in condition between areas. This is most likely because they are equations that were derived from the same weight/length relationship and only vary by the distance of b from 3.0, the number used when growth is isometric. They also were similar in the pooled data analysis. The ANCOVA analysis also provides adequate differentiation and is used most recently in the literature because of its ability to remove the covariance of size with condition. However, it falls victim to the same shortcoming as many of the other indicesyou can not use it to compare populations, but only groups within a population,

PAGE 30

21 such as across capture events in this study. The SVL/Mass ANCOVA combination detected the maximum number of area differences in this study, but it did not detect any differences in the temporal analysis. It would have been useful to use an ANCOVA for temporal analyses of the same population if the sample size was sufficient.. This analysis has shown Fultons K and Relative K to be nearly equal in their ability to distinguish condition differences among areas, even though the b value calculated is significantly different than 3.0. If Fultons K and Relative K are so similar for alligator populations, it would be wise to use Fultons K because condition scores calculated using Fultons K can be compared across populations. A downside to both of these indices is that they do not provide one with an easily interpreted biological number. The Fultons condition score for this analysis ranged from 1.4 to 3.2, unlike Relative Mass, where 1.0 is an alligator in normal condition and anything less than 1.0 means the alligator is in poorer condition. However, Relative Mass does not have the ability to distinguish as finely between areas as Fultons or Relative K did for our alligator data or between populations. To overcome the problem of a lack of biological meaning in Fulton K numbers, one could identify quartiles of the data and label the bottom quartile as low condition, the middle two as average condition and the top quartile as good condition (Figure 6). One must be careful when separating condition into high and low that instant judgments are not attached to those categories. When using a condition index, results must be taken in the context of nearby populations. Fatter alligators are not necessarily living in a better environment. Like with humans, there are upper and lower limits to fat stores as they relate to condition. It is not minor fluctuations that we are interested in, but

PAGE 31

22 those extremes that could give clues to population and ecosystem health. As an example, I offer an analysis of SVL/Mass Fultons K for areas in South Carolina (P. Wilkinson, unpublished data), north Florida (A. R. Woodward, Florida Fish and Wildlife Conservation Commission, unpublished data), and the Everglades. South Carolina data is from Santee captures in 1980 (n = 100). North Florida captures are from Lake Griffin in 1987, 1989, and 1990 (n = 30), Lochloosa Lake in 1985 (n = 57), Newnans Lake in 1985 (n = 40), Orange Lake in 1985, 1987, 1989, and 1990 (n = 142), and Lake Woodruff in 1987, 1989, and 1990 (n = 33). SVL/Mass was used because HL data were not available. Lake Griffin would be classified in the high condition quartile (Figure 7) and the Everglades to be in the low condition quartile. It is not necessarily true that the alligators in Lake Griffin are in better health than those in the Everglades. If you look at Lake Griffin in context with other north Florida lakes, it looks like its condition is high and that perhaps something is wrong with that ecosystem, such as the fact that Lake Griffin is a highly eutrophic lake from agricultural run off. The recent die-off of adult and subadult alligators in Lake Griffin (Schoeb et al., 2002) corroborates the fact that Griffin is not a healthy ecosystem. Everglades population is in lower condition as compared to north Florida and South Carolina, but if data is analyzed over time for only the Everglades, it would reveal a pattern of consistent low condition and should not signal a crisis for Everglades alligators, especially when compared with other long-term Everglades alligator data. Condition is a very fluid measurement. Water management practices and rainfall can dramatically change condition of animals in a relatively short amount of time, because so many aspects of their life history (feeding, courtship, and nesting) depend on

PAGE 32

23 seasonally fluctuating water levels. In this study, we found that ENP-SS alligators had the highest condition of Everglades alligators, but Dalrymple (1996) and Barr (1997) observed Shark Slough alligators to be in very poor condition only five or six years earlier. Water levels were high during Dalrymple and Barrs period of capture, possibly affecting food availability. Hyrdologic conditions are important to the American alligator, but are not the only factor that affects condition. Disease, climatic change, nutrient input, and contaminants all contribute to condition and should be considered when using a condition index to analyze population health. While condition is a useful way to evaluate current Everglades restoration, it should not be taken out of context of events in the greater ecosystem.

PAGE 33

24 Table 3: Descriptive statistics of indices used in condition analysis for the American alligator, south Florida (n = 395) with captures from October 1999 to Spring 2002. Index Equation Factors Minimum Maximum Mean Std. Deviation Ratio Volumetric/Length HL/TG 0.944 1.725 1.447 0.103 SVL/TG 0.288 0.455 0.393 0.025 Fulton's K Volumetric/(Length^3) HL/Mass 2.572 16.389 10.849 1.556 SVL/Mass 1.430 3.200 2.162 0.253 Relative K Volumetric/(Length^b) HL/TG 8.454 15.612 13.118 0.915 HL/Mass 0.470 2.842 1.848 0.281 SVL/TG 2.327 3.719 3.204 0.203 SVL/Mass 1.381 3.089 2.087 0.243 Relative Mass Observed mass/Expected mass HL/TG 0.648 1.185 0.992 0.070 SVL/TG 0.727 1.156 0.997 0.064 SVL/Mass 0.656 1.473 0.989 0.115 Residuals Actual value Expected value HL/TG -18.833 6.967 8.314 2.623 HL/Mass -29.728 18.748 3.294 3.685 SVL/TG -11.633 6.506 -2.569 2.393 SVL/Mass -7.744 19.405 -1.476 3.466 HL = Head Length, SVL = Snout-Vent Length, TG = Tail Girth

PAGE 34

25 Table 4: Condition indices calculated for the American alligator in south Florida from October 1999 to Spring 2002 (n = 395) whose Pearson r-values are less than 30%. Condition Method Factors Pearson 'r'(p-value) Ratio HL/TG 0.122(0.015) SVL/TG 0.132(0.009) Fulton's K HL/Mass 0.059(0.239) SVL/Mass 0.065(0.196) Relative K HL/Mass 0.277(0.000) HL/TG 0.046(0.359) SVL/Mass 0.051(0.356) SVL/TG -0.007(0.888) RelativeMass SVL/TG 0.116(0.021) HL/TG 0.170(0.001) SVL/Mass 0.043(0.399) Residuals HL/Mass <0.001 HL/TG <0.001 SVL/Mass <0.001 SVL/TG <0.001

PAGE 35

Table 5: Results of ANOVA/Least Significant Difference and ANCOVA/ Least Significant Difference analyses of condition indices for the American alligator in south Florida from October 1999 to Spring 2002 (n = 395). Order of condition index is from highest to lowest mean. 26 Condition Index Factors p-value # groups LSD Order Differences (LSD) Ratio HL/TG 0.007 3 362154 2|4,3|4,4|6 SVL/TG 0.037 3 365124 2|3,3|4,4|6 Fulton's K HL/Mass <0.001 7 263514 1|2,2|3,2|4,2|5,2|6,4|3,4|6 SVL/Mass 0.028 3 265314 1|2,2|3,2|4 Relative K HL/Mass <0.001 7 263154 1|2,2|3,2|4,2|5,2|6,3|4,4|6 HL/TG 0.005 3 362154 2|4,3|4,4|6 SVL/Mass 0.025 3 265341 1|2,2|3,2|4 SVL/TG 0.028 4 365124 1|3,2|3,3|4,4|6 RelativeMass SVL/TG 0.027 3 365124 2|3,3|4,4|6 HL/TG 0.007 3 362154 2|4,3|4,4|6 SVL/Mass 0.014 3 265341 1|2,2|3,2|4 ANCOVA HL/Mass 0.004 7 136524 1|3,1|6,1|5,1|2,1|4,3|2,3|4 HL/TG 0.709 -SVL/Mass 0.005 8 136524 1|3,1|6,1|5,1|2,1|4,3|5,3|2,3|4 SVL/TG 0.436 -Residuals HL/Mass 0.0164 3 362514 2|4,3|4,6|4 HL/TG 0.0247 3 362514 2|4,3|4,6|4 SVL/Mass 0.0927 -SVL/TG 0.0747 -1 = ENPEstuaries, 2 = ENPShark Slough, 3 = A.R.M Loxahatchee NWR, 4 = WCA2A, 5 = WCA3AHoliday Park, 6 = WCA3ATamiami

PAGE 36

Table 6: ANOVA/ANCOVA analyses of condition indices by capture event from October 1999 to Spring 2002 (n = 395). Order of condition index is from highest to lowest mean. There were no ENP-Est captures for March/April 00. 27 Condition Index Factors Capture Event p-value # groups LSD Order Differences (LSD) Fulton's K HL/Mass October 99 <0.001 9 123654 1/3,1/6,1/5, 1/4, 2/3, 2/6, 2/5, 2/4, 3/4 March/April 00 0.0487 2 26534 2/3, 2/4 October 00 0.004 5 261534 2/6, 2/1 2/5, 2/3, 2/4 March/April01 0.6061 ---October01 0.5458 ---March/April02 0.0536 ---Relative K HL/Mass October 99 0.0012 8 123654 1/3, 1/6, 1/5, 1/4, 2/3, 2/6, 2/5, 2/4 March/April 00 0.0485 2 26354 2/3, 2/4 October 00 0.0056 5 261534 2/6, 2/1, 2/5, 2/3, 2/4 March/April01 0.5352 ---October01 0.5841 ---March/April02 0.0878 ---ANCOVA HL/Mass October 99 0.0299 8 236415 2/3, 2/6, 2/4, 2/1, 2/5, 3/5, 5/6, 4/5 March/April00 0.1268 ---October00 0.6431 ---March/April01 0.393 ---October01 0.7944 ---March/April02 0.0203 12 312465 2/3, 4/3, 6/3, 5/3, 1/2, 1/4, 1/6, 1/5, 2/4, 2/6, 2/5, 4/5 SVL/Mass October99 0.9094 ---March/April00 0.3708 ---October00 0.8468 ---March/April01 0.2913 ---October01 0.4194 ---March/April02 0.1403 ---1 = ENP--Estuaries, 2 = ENPShark Slough, 3 = Loxahatchee NWR, 4 = WCA2A, 5 = WCA3AHoliday Park, 6 = WCA3ATamiami

PAGE 37

28 Table 7: Ability of ANOVA/LSD analysis to detect differences in condition from Relative K simulated data. Numbers are the percentage of total populations for which the analysis found a significant difference. Type 1 is mean and standard deviation from original data, Type 2 is 5% increase of mean, and Type 3 is 10% increase of mean. Population Type 1 2 3 1 0 2 64 0 3 100 96 0 Table 8: Ability of ANOVA/LSD analysis to detect differences in condition from Fultons K simulated data. Numbers are the percentage of total populations for which the analysis found a significant difference. Type 1 is mean and standard deviation from original data, Type 2 is 5% increase of mean, and Type 3 is 10% increase of mean. Population Type 1 2 3 1 0 2 100 0 3 100 84 0

PAGE 38

% ME N arrowing of meas. to HL, SVL, TG, Mass Measurement experiment Choice of indices Lit Search Calculate indices for total data set (n=395) Examine assumptions: Post-hoc tests (LSD) Significant area differences Reduced set of indices ANOVAs 29 correlations Post-hoc by time period Ordering of indices and differences Analysis by capture event Choice of index Simulated data Figure 1: Flowchart of condition indices analysis for the American alligator in the Everglades.

PAGE 39

30 Figure 2: Capture areas in south Florida within Everglades wetlands core. Black dots represent alligator captures from October 1999 to March 2002 (n = 395). LOX = A.R.M. Loxahatchee National Wildlife Refuge, WCA2 = Water Conservation Area 2, WCA3 = Water Conservation Area 3, ENP = Everglades National Park.

PAGE 40

y = 0.4132x 1.7386R2 = 0.9164010203040506070020406080100120140160Snout-Vent Length (cm)Tail Girth (cm) 31 Figure 3: Linear regression of tail girth vs. snout-vent length of the American alligator of captures in south Florida from October 1999 to March 2002 (n = 395).

PAGE 41

02468101214161801234567Area*Fulton's K 32 Figure 4: Range of Fultons K index values by capture area in south Florida. Captures are from October 1999 to March 2002 (n = 395). *1 = ENP-Est 2 = ENP-SS 3 = LOX 4 = WCA2 5 = WCA3-HP 6 = WCA3-T

PAGE 42

33 ENP-SS WCA3-Both areas LOX ENP-Est WCA2 Mid to Low Mid Highest Figure 5: Hierarchy of Everglades alligator condition from October 1999 to March 2002 using Fultons K analysis (n = 395). Brackets signify that condition does not differ significantly within that group.

PAGE 43

34 1 st quartile: 2.57 4.13 low condition 2 nd quartile: 4.14 10.84 low to average condition 3 rd quartile: 10.85 13.61 average to high condition 4 th quartile: 13.62 16.39 high condition Figure 6: Sample condition range of alligators for Fultons K from Everglades data from October 1999 to March 2002 (n = 395). This shows division of condition by quartiles

PAGE 44

35 40395331001425730N =Area7654321Fulton's K3.63.43.23.02.82.62.42.22.0 Figure 7: SVL/Mass Fultons K analysis for the American alligator in areas of South Carolina, north Florida and the Everglades. 1 = Lake Griffin, FL (1987, 1989, 1990) 2 = Lochloosa Lake, FL (1985) 3 = Orange Lake, FL (1985, 1987, 1989, 1990) 4 = Santee, SC (1980) 5 = Lake Woodruff, FL (1987, 1989, 1990) 7 = Everglades, FL (1999, 2000, 2001, 2002) 8 = Newnans Lake, FL (1985)

PAGE 45

CHAPTER 4 CONCLUSION Condition factor can be a useful tool for evaluating the health of a population if used appropriately. This analysis examined the strict assumptions of properly calculating condition indices, beginning with body measurements that reduced measurement error. Head length (HL), snout-vent length (SVL), tail girth (TG), and mass exhibited the least measurement error and were used in the subsequent condition analysis. Five condition factors and one model of condition (ANCOVA) were calculated to establish which index or model could differentiate between Everglades areas. It was determined through ANOVA/LSD analysis of the condition indices that the HL/Mass combination of Fultons K and the SVL/Mass combination of ANCOVA were best able to distinguish differences in condition between areas of the Everglades. The condition index used depends on the question asked. For example, ANCOVA is best way to determine temporal differences between populations. It can not be used to compare across populations, unless strict assumptions about the two populations can be met. HL/Mass Fultons K can be used to spatially compare populations of the American alligator, and is suggested by this study as the best condition factor to use for that purpose. In the future, more data needs to be collected to develop a model so that condition of the American alligator can be predicted by water levels. While other aspects of condition are important to the overall model, hydrologic conditions are the focus of the current restoration and a monitoring tool needs to be sensitive to that feature. Also, before a condition index of the American alligator can used, management goals must be 36

PAGE 46

37 made to choose the power of the analysis and then determine an appropriate sample size from that power analysis.

PAGE 47

APPENDIX A DATA SHEETS AND MEASUREMENT INSTRUCTIONS GIVEN TO EXPERIMENT PARTICIPANTS Web Tag# Scute Clip # Area Crew GFC (N/A for most) GPS LOCATION Capture Date Capture Time UTM Northing UTM Easting General Location / Describe Vegetation Temperature Habitat Type Water Depth Muck Depth Air C Water C cm Total Water = cm Cap Method Capture Status Time Blood 1 (Gross) Time Blood 2 (Cardeilhac) Scute? 38

PAGE 48

39 Head Length SV Length Total Length Hind Foot L cm cm cm cm Tail Girth Neck Girth Chest Girth Weight Sex cm cm cm kg Deformities Release Status Notes Description of Measures and Coding Web Tag # -Number engraved on tag (GFC 37201). Scute Clip # -Animal number derived from scute clipping (used only on LOX). Area -Study area Everglades Nat Park Shark Slough (ENP-SS), Estuarine (ENP-EST); Loxahatchee NWR (LOX); Water Conservation Area 3A South (WCA3AS, also WCA3AN, WCA2); Big Cypress National Preserve (BCY). Crew -First initial and surname of boat crew (K. Rice). Capture Date -Date in 1 Oct 99 format. Capture Time All times are in 24 hr (military) format (0215, 1622). GPS Location -UTM coordinates of capture site (Northing 548515, Easting 2891857). General Location/Vegetation -Describe capture site and dominant vegetation (1 mile N of ..Sawgrass) Water Temp / Air Temp -Temperature (in C) of Air and Water (~ 6 below surface).

PAGE 49

40 Habitat Type -Specific habitat type: Open water = 1 Forested Wetlands = 2 Shrubs/Shrub Islands = 3 Mixed Emergents = 4 Sawgrass Marsh = 5 Spikerush Marsh = 6 Cattail Marsh = 7 Water Lily/Floating Leaved Veg = 8 Canal = 9 Alligator Hole = 10 Levee Break = 11 Water Depth -Water depth at capture site in cm. Muck Depth -Measure from surface of water to rock and subtract water depth in cm. Cap Method -Capture Method (Snare, Hand, Harpoon). Capture Status -Animal status at capture (Vigorous). Time Blood 1 -Time (24 hr) blood kit 1, Scute? Scute collected (y or n). Time Blood 2 -Time (24 hr) blood kit 2 used. Head Length -Measure (dorsal) from tip of snout to center of posterior end of skull in cm. SV Length -Measure (ventral) from tip of snout to posterior end of vent in cm. Total Length -Measure (ventral) from tip of snout to tip of tail in cm. Note if tip of tail missing, etc. Hind Foot L -Measure (ventral) from 1 st single extended scute posterior to heel to the anterior end of middle toe, not including the nail in cm. Tail Girth Measure circumference of tail at 3 rd scute row posterior of rear legs in cm. Neck Girth -Measure circumference of neck between head and shoulders in cm. Chest Girth -Measure circumference of chest just posterior to front legs in cm. Weight -Total weight of animal in kg. Sex -M or F. Deformities -Note any physical deformities or prominent scars (missing LR foot). Release Status -Animal status at release (Vigorous). Notes -Any additional information (found 1 dead 6 gator).

PAGE 50

APPENDIX B GENERAL PICTORIAL MORPHOMETRIC MEASUREMENT INSTRUCTIONS Head LengthMeasure dorsally from tip of snout to center of posterior end of skull. Dashed line indicates measurement. Snout-Vent LengthMeasure ventrally from tip of snout to back of vent or dorsally to the back of the second tail scute row. The zero tail scute is the last short tail scutethe last one from the body that doesnt wrap all the way around the tail. Detail of vent for snout-vent length: 41 Modified from King and Brazaitis (1971).

PAGE 51

42 Detail of dorsal side for snout-vent length: Zero tail scute F i r s tt a i ls c u t e Second tail scute Zero tail scute F i r s tt a i ls c u t e Second tail scute Total Length: Measure straight distance dorsally or ventrally from tip of snout to tip of tail. Note if tip of tail is missing. Make sure gator and tail are in a straight line, not like this one.

PAGE 52

43 Tail Girth: Measure the circumference of tail at 3 rd scute row posterior to rear legs. Zero tail scute F i r s tt a i ls c u t e Second tail scute Third tail scute Neck Girth: Measure circumference of neck between head and shouldersbetween two giant scutes. In this pictures, the scutes look a little farther back than they should.

PAGE 53

44 Chest Girth: Measure circumference of chest just posterior to front legs.

PAGE 54

LIST OF REFERENCES Arendt, W.J. and J. Faaborg. 1989. Sources of variation in measurements of birds in a Puerto Rican dry forest. Journal of Field Ornithology 60(1): 1-11. Bailey, R.C. and J. Byrnes. 1990. A new, old method for assessing measurement error in both univariate and multivariate morphometric studies. Systematic Zoology 39(2): 124-130. Barr, B. 1997. Food habits of the American alligator, Alligator mississippiensis, in the southern Everglades. Ph.D. dissertation. University of Miami, Miami, FL. Brandt, L.A. 1991. Growth of juvenile alligators in Par Pond, Savannah River site, South Carolina. Copeia 4: 1123-1129. Campbell, M. 1999. Everglades alligator holes: distribution and ecology. M.S. thesis. University of Florida, Gainesville. Chabreck, R. and T. Joanen. 1979. Growth-rates of American alligators in Louisiana. Herpetologica 35(1): 51-57. Cone, R.S. 1989. The need to reconsider the use of condition indexes in fishery science. Transactions of the American Fisheries Society 118(5): 510-514. Craighead, F.C. 1968. The role of the alligator in shaping plant communities and maintaining wildlife in the southern Everglades. Florida Naturalist 41: 2-7, 69-74, 94. Dalrymple, G.H. 1996. Growth of American alligators in the Shark Valley region of Everglades National Park. Copeia(1): 212-216. Delany, M.F., S.B. Linda, and C.T. Moore. 1999. Diet and condition of American alligators in 4 Florida lakes. Proceedings of the Annual Conference of Southeastern Association of Fish and Wildlife Agencies. 53: 375-389. Enge, K.M., H.F. Percival, K.G. Rice, M.L. Jennings, G.R. Masson, A.R. Woodward. 2000. Summer nesting of turtles in alligator nests in Florida. Journal of Herpetology. 34 (4): 497-503. 45

PAGE 55

46 Elsey, R.M., T. Joanen, L. McNease and N. Kinler. 1992. Growth-rates and body condition factors of Alligator mississippiensis in coastal Louisiana wetlandsa comparison of wild and farm-released juveniles. Comparative Biochemistry and Physiology A: Physiology 103(4): 667-672. Garcia-Berthou, E. 2001. On the misuse of residuals in ecology: testing regression residuals vs. the analysis of covariance. Journal of Animal Ecology 70: 708-711. Garcia-Berthou, E. and R. Moreno-Amich. 1993. Multivariate analysis of covariance in morphometric studies of the reproductive cycle. Canadian Journal of Fisheries and Aquatic Sciences 50: 1394-1395. Garnett, S.T. and R.M. Murray. 1986. Parameters affecting the growth of estuarine crocodiles, Crocodylus porosus, in captivity. Australian Journal of Zoology 34(2): 211-223. Grant, P.R. 1979. Ecological and morphological variation of Canary Island Blue tits, Parus caeruleus (Aves, Paridae). Biological Journal of the Linnean Society 11(2): 103-129. Green, A.J. 2001. Mass/length residuals: measures of body condition or generators of spurious results? Ecology 82(5): 1473-1483. Grigg, G.C., L.E. Taplin, B. Green and P. Harlow 1986. Sodium and water fluxes in free-living Crocodylus porosus in marine and brackish conditions. Physiological Zoology 59(2): 240-253. Hall, P.M. 1991. Estimation of nesting female crocodilian size from clutch characteristscorrelates of reproductive mode, and harvest implications. Journal of Herpetology 25(2): 133-141. Hall, P.M. and A.J. Meier. 1993. Reproductive behavior of Western mud snakes in American alligator nests. Copeia 1: 219-222. Hutton, J.M. 1986. Age-determination of living Nile crocodiles from the cortical stratification of bone. Copeia(1): 332-341. Jakob, E.M., J.D. Marshall and G.W. Uetz. 1996. Estimating fitness: a comparison of body condition indices. Oikos 77: 61-67. Jones, C., J. Lawton and M. Shachak 1994. Organisms as ecosystem engineers. Oikos 69(3): 373-386. King, F.W. and P. Brazaitis. 1971. Species identification of commercial crocodile skins. Zoologica. 56:15-70.

PAGE 56

47 Kotiaho, J.S. 1999. Estimating fitness: comparison of body condition indices revisited. Oikos 87(2): 399-402. Krebs, C.J. and G.R. Singleton. 1993. Indexes of condition for small mammals. Australian Journal of Zoology 41(4): 317-323. Kushlan, J.A. 1974. Observations on the role of the American alligator (Alligator mississippiensis) in the southern Florida wetlands. Copeia 1974: 993-996. Kushlan, J.A. and T. Jacobsen 1990. Environmental variability and reproductive success of Everglades alligators. Journal of Herpetology 24(2): 176-184. Kushlan, J.A. and M.S. Kushlan. 1980. Everglades alligator nests: nesting sites for marsh reptiles. Copeia. 4: 930-932. LeCren, E.D. 1951. The length-weight relationship and seasonal cycle in gonad weight and condition in the perch (Perca fluviatilis). Journal of Animal Ecology 20(2): 201-219. Leslie, A.J. 1997. The ecology and physiology of the Nile crocodile, Crocodylus niloticus, in Lake St. Lucia, Kwazulu/Natal, South Africa. Ph.D. dissertation. Drexel University, Philadelphia, PA. Mazzotti, F.J. and L.A. Brandt. 1994. Ecology of the American alligator in a seasonally fluctuating environment. Everglades: the ecosystem and its restoration. S.M. Davis and J. C. Ogden eds. Delray Beach, FL, St. Lucie Press: 485-505. Murphy, B.R., M.L. Brown and T.A. Springer. 1990. Evaluation of the relative weight (Wr) index, with new applications to walleye. North American Journal of Fisheries Management 10: 85-97. Olmstead, I. and T.V. Armentano. 1997. Vegetation of Shark Slough, Everglades National Park, SFNRC Technical Report 97-001: 43 pp. Packard, G.C. and T.J. Boardman. 1988. The misuse of ratios, indexes, and percentages in ecophysiological research. Physiological Zoology 61(1): 1-9. Palmeirim, J.M. 1998. Analysis of skull measurements and measurers: Can we use data obtained by various observers? Journal of Mammalogy 79(3): 1021-1028. Quist, M.C., C.S. Guy and P.J. Braaten. 1998. Standard weight (Ws) equation and length categories for Shovelnose sturgeon. North American Journal of Fisheries Management 18: 992-997. Rice, K.G., F.J. Mazzotti and C.L. Zweig. 2002. Annual report--compilation of alligator datasets in South Florida for restoration needs, USGS.

PAGE 57

48 Santos, S.A., M.J.S. Nogueria, M.S. Pinheiro, G.M. Mourao and Z.M.S. Campos. 1994. Condition factor of Caiman crocodiles in different habitats of Pantanal Mato-Grassense. Proceedings of the 12th Working Meeting of Crocodile Specialist Group of the Species Survival Commission of the IUCN. SAS Institute, Inc. 1988. SAS users guide: Statistics. Version 6.03 edition. Cary, NC. Schoeb, T.R., T.G. Heaton-Jones, R.M. Clemmons, D.A. Carbonneau, A.R. Woodward, D. Shelton, and R.H. Poppenga. 2002. Clinical and necropsy findings associated with increased mortality among American alligators of Lake Griffin, Florida. Journal of Wildlife Diseases. 38(2): 320-327. Springer, T.A. and B.R. Murphy. 1990. Properties of relative weight and other condition indices. Transactions of the American Fisheries Society 119: 1048. SPSS Inc. 2000. SPSS Base 11.0 for Windows users guide. SPSS Inc. Chicago, IL. Sutton, S.G., T.P. Bult and R.L. Haedrich. 2000. Relationships among fat weight, body weight, water weight, and condition factors in wild Atlantic salmon parr. Transactions of the American Fisheries Society 129(2): 527-538. Taylor, J.A. 1979. The foods and feeding habits of subadult Crocodylus porosus Schneider in Northern Australia. Australian Wildlife Research 5: 347-359. Vila-Gispert, A. and R. Moreno-Amich. 2001. Fish condition analysis by a weighted least squares procedure: testing geographical differences of an endangered Iberian cyprinodontid. Journal of Fish Biology 58: 1658-1666. Wege G.J. and R.O. Anderson. 1978. Relative weight (Wr): a new index of condition for largemouth bass. New approaches to the management of small impoundments. G.D. Novinger and J.G. Dillard, eds. American Fisheries Society, North Central Division, Special Publication 5, Bethesda, MD. Yezerinac, S.M., S.C. Lougheed and P. Handford. 1992. Measurement error and morphometric studies--statistical power and observer experience. Systematic Biology 41(4): 471-482. Zink, R.M. 1983. Evolutionary and systematic significance of temporal variation in the Fox sparrow. Systematic Zoology 32(3): 223-238.

PAGE 58

BIOGRAPHICAL SKETCH Christa Zweig was born in Lexington, Kentucky, on December 22, 1974. Her family moved to Springfield, Missouri, in 1977 where she attended elementary through high school. She originally wanted to be a brain surgeon, but soon decided that she did not like people enough to deal with them on a day-to-day basis. That and her frequent trips to Cumberland Island National Seashore combined to shape her current career path. Christa received her undergraduate degree in biology from the University of Richmond, Virginia. She then spent the next three years working temporary federal jobs for Canaveral National Seashore, Padre Island National Seashore, and the Red Bluff U.S. Fish and Wildlife Fisheries Office in northern California. In 2000, she began her work in south Florida with the Florida Alligator Research Team. Her plans for the future are vague and varied, but she would like to continue a career in the field of conservation. 49


Permanent Link: http://ufdc.ufl.edu/UFE0000836/00001

Material Information

Title: Body Condition Factor Analysis for the American Alligator (Alligator Mississippiensis)
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000836:00001

Permanent Link: http://ufdc.ufl.edu/UFE0000836/00001

Material Information

Title: Body Condition Factor Analysis for the American Alligator (Alligator Mississippiensis)
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000836:00001


This item has the following downloads:


Full Text












BODY CONDITION INDEX ANALYSIS FOR THE AMERICAN ALLIGATOR
(Alligator mississippiensis)













By

CHRISTA L. ZWEIG


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2003




























Copyright 2003

by

Christa L. Zweig















ACKNOWLEDGMENTS

I thank my family for all of the incredible support they gave me while still letting

me make my own mistakes. I am grateful to Zach Welch, who had the demanding duty

of keeping me in high spirits. I also thank the members of my committee: Frank

Mazzotti for offering his support and giving me the opportunity to pursue this degree,

Ken Rice for providing me with all sorts of guidance, Laura Brandt for her thoughtful

comments, and Franklin Percival for his excellent, practical opinions. There are many

people to thank, and I wish I could enumerate the countless reasons why they are special

to me, but then this section would be longer than my thesis. Everyone knows why they

are on this list: Ab Abercrombie, Lindsey Hord, Mike Cherkiss, Ashley Traut, Gayle

Martin, Becky Hylton, Diana Swan, Jamie Duberstein, Jocie Graham, Phil Wilkinson,

Woody Woodward, Chris Tucker, Jason Williams, Geoff Cook, Phil George, Matt

Chopp, Adam Finger, Stan Howarter, Cathy Ritchie, Cherie Cook, Sarah Kern, Monica

Lindberg, Caprice McRae, Debra Hughes, Barbara Fesler, Marion Bailey, Mary Hudson-

Kelley, Lauren J. Chapman, Peggy VanArman, the Palm Beach Atlantic College Science

Club, IFAS statistics, Marinella Capanu, and Brenny.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ................................................................................................. iii

LIST OF TABLES ................... ......... ......... .......... ............ vi

L IST O F FIG U R E S .... ....... ................................................ .... ..... .. ............. vii

A B STR A C T ...................... .................................. ........... ... ....... ....... viii

CHAPTER

1 IN T R O D U C T IO N ................................................................................ ............ .. .. 1

2 EVALUATION OF FIELD MEASUREMENTS OF THE AMERICAN
ALLIGATOR FOR USE IN MORPHOMETRIC STUDIES .............. ..................3

Introduction ......... ...... ............ ........................................................... . 3
M eth o d s ........................................................................................................................... 3
R e su lts .............. ...... ........... ................. ................................................... . 6
D isc u ssio n .............. ...... ........... ................. .......................... .................. 6

3 CONDITION INDEX ANALYSIS OF THE AMERICAN ALLIGATOR ..................8

In tro d u ctio n ....................................................................... 8
S tu d y A re a ............................................................................ 12
M materials and M ethods................................... .............. 13
C capture M methods ....................................................................................... ........ 13
Condition Indices Analysis .................................. ...................... .... ...... 14
Sim ulated D ata ..................................................... 15
Everglades Condition A analysis .................................................................... 16
R esu lts .................. .................... ................................ 16
Condition Indices Analysis .............................. .................. .. .......... 16
ANOVA/ANCOVA Analysis by Capture Event ............................................ 17
Sim ulated D ata ....................................................... 18
Everglades Condition A analysis ..................................................................... 18
D iscu ssio n ..................................................... 19

4 CON CLU SION ............... ................ ...................................... 37









APPENDIX

A DATA SHEETS AND MEASUREMENT INSTRUCTIONS GIVEN TO
EXPERIM ENT PARTICIPANTS ........................................ .......................... 38

B GENERAL PICTORIAL MORPHOMETRIC MEASUREMENT
IN STRU CTION S .................. ........................................ ................. 41

L IST O F R E FE R E N C E S ............................................................................. ............. 45

B IO G R A PH IC A L SK E TCH ..................................................................... ..................49
















LIST OF TABLES


Table p

1: T-test results for coefficient of variation of experienced individuals versus
inexperienced groups of volunteers measuring American alligators at A.R.M.
Loxahatchee National W wildlife Refuge ....................................... ............... 7

2: Within alligator variance and covariance components for length and volumetric
m e a su re s ...................... .. .. ......... .. .. ......... ....................................... 7

3: Descriptive statistics of indices used in condition analysis for the American
allig ato r ...................... .. .. ......... .. .. ................................................ 2 4

4: Condition indices calculated for the American alligator whose Pearson r-values
are less than 30% ................................................................................2 5

5: Results of ANOVA/Least Significant Difference and ANCOVA/ Least
Significant Difference analyses of condition indices for the American alligator..26

6: ANOVA/ANCOVA analyses of condition indices by capture event from October
1999 to Spring 2002 (n = 395). Order of condition index is from highest to lowest
mean. There were no ENP-Est captures for March/April 00..............................27

7: Ability of ANOVA/LSD analysis to detect differences in condition from Relative
K sim ulated data .....................................................................28

8: Ability of ANOVA/LSD analysis to detect differences in condition from Fulton's
K sim ulated data .....................................................................28
















LIST OF FIGURES


Figure page

1: Flowchart of condition indices analysis for the American alligator in the
Everglades ..................................... ................................ ......... 29

2: Capture areas in south Florida within Everglades wetlands core ........................30

3: Linear regression of tail girth vs. snout-vent length of the American alligator of
captures in south Florida from October 1999 to March 2002.............................31

4: Range of Fulton's K index values by capture area in south Florida. Captures are
from October 1999 to March 2002 .............................................32

5: Hierarchy of Everglades alligator condition from October 1999 to March 2002
u sing F u lton 's K an aly sis............................................................ .....................33

6: Sample condition range of alligators for Fulton's K from Everglades data from
October 1999 to March 2002 .................. ......... ................. 34

7: SVL/Mass Fulton's K analysis for the American alligator in areas of South
Carolina, north Florida and the Everglades ................................. ............... 35















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

BODY CONDITION FACTOR ANALYSIS FOR THE AMERICAN ALLIGATOR
(Alligator mississippiensis)

By

Christa Zweig

May 2003


Chair: Frank Mazzotti
Major Department: Wildlife Ecology and Conservation

The American alligator (Alligator mississippiensis) is an integral part of the

Everglades ecosystem. They affect and are affected by the landscape and changes in

hydrology, which makes them an excellent organism to use in evaluating current

Everglades restoration efforts. Information on alligators-census data, capture or

morphometric data, blood chemistry, and reproduction statistics-has been collected in

the Everglades since the 1950's. Historical information provides a suite of useful life

history characteristics or population parameters (i.e., health and condition, nesting effort,

growth rate and survival, and density) that can be used for evaluating restoration.

However, some life history traits (e.g., absolute density and survival) are difficult to

accurately measure and may take decades of data to detect trends. Body condition can be

measured using indices and requires less data to begin an analysis. If used properly,

condition can be a useful monitoring tool to assess the health of a population and thus the

health of its ecosystem. This project evaluated morphometric measurements taken for the









American alligator, available condition indices, and, using a stepwise process,

recommends the appropriate index for use in ecological applications.

We analyzed morphometric measurements of captured animals to determine which

are measured most accurately and are appropriate for condition analyses. Condition

indices are functions of a body length indicator and a volumetric measurement and are

only as accurate as the measurements used. Head length, snout-vent length and total

length are suitable for body length indicators, and tail girth, neck girth, chest girth, and

mass can all be used as volumetric measurements. We then compared four condition

indices and two models of volume/length relationships for their ability to distinguish

between populations with known qualitative condition differences.

It was determined through ANOVA/LSD analysis of the condition indices that the

HL/Mass combination of Fulton's K and the SVL/Mass combination of ANCOVA were

best able to distinguish differences in condition between areas of the Everglades;

however ANCOVA can not be used to compare across populations, unless strict

assumptions about the two populations can be met. HL/Mass Fulton's K can be used to

spatially and temporally compare populations of the American alligator, and is suggested

by this study as the best condition factor to use for these purposes.

The primary objective of this study was to determine the appropriate condition

index for the American alligator with the intention that it will be used as a monitoring

tool of population health for the current restoration process. Because the alligator's life

history is so closely linked to hydropattern, body condition can reflect the impacts of

changes in hydrology.














CHAPTER 1
INTRODUCTION

The American alligator (Alligator mississippiensis) ranges from eastern Texas to

North Carolina and south to the tip of Florida. Alligators are vital components of wetland

ecosystems, serving as keystone predators (Mazzotti and Brandt 1994) and exhibiting the

qualities of ecosystem engineers (Jones et al. 1994), creating critical landscape features.

In the Everglades, they create topography in a system where small changes can be very

important. They construct alligator holes for thermal refuge, feeding, and reproduction

by making depressions in the muck soil and keeping them clear of surrounding

vegetation. These areas hold water during dry-down periods and provide refugia for

aquatic and amphibious animals during the dry season (Craighead 1968; Kushlan 1974;

Campbell 1999). Alligators also create dry land in the marsh by mounding vegetation to

build their nests. Habitat for less flood-tolerant plants is created. Further, it has been

documented that other reptiles use alligator nest mounds to deposit their eggs (Kushlan

and Kushlan 1980; Hall and Meier 1993; Enge et al. 2000).

The natural history and reproductive patterns of the American alligator are closely

tied to its physical environment. There is a relationship between spring water levels and

late summer water levels in a premanaged landscape (Kushlan and Jacobsen 1990;

Mazzotti and Brandt 1994). Based on spring water levels, alligators construct nests that

will be above historical average high water in the summer, wet season. However,

because water levels in the Everglades have been altered from historical seasonal

fluctuations, unnatural water regimes may cause nest flooding (Kushlan and Jacobsen









1990). Further, water levels can affect body condition of Everglades alligators

(Dalyrmple 1996).

The American alligator is an integral part of the Everglades ecosystem. They effect

and are affected by the landscape and changes in hydrology, which makes them an

excellent organism to use in evaluating current Everglades restoration efforts.

Information on alligators-census data, capture or morphometric data, blood chemistry,

and reproduction statistics, has been collected in the Everglades since the 1950's (Rice et

al. 2002). Historical information provides a suite of useful life history characteristics or

population parameters (i.e., health and condition, nesting effort, growth rate and survival,

and density) that can be used for evaluating restoration. However, some life history traits

(e.g., absolute density and survival) are difficult to accurately measure and may take

decades of data to detect trends. Condition is the "relative fatness of [an animal]. ... a

measure of how well that animal is coping with its environment" (Taylor 1979, p. 349),

and can be measured using indices and requires less data to begin an analysis. If used

properly, condition can be a useful monitoring tool to assess the health of a population

and thus the health of its ecosystem. This project evaluates morphometric measurements

taken for the American alligator (Chapter 2) and available condition indices, and using a

stepwise process (Figure 1), recommends the appropriate index for use in ecological

applications (Chapter 3).














CHAPTER 2*
EVALUATION OF FIELD MEASUREMENTS OF THE AMERICAN ALLIGATOR
FOR USE IN MORPHOMETRIC STUDIES

Introduction

Long-term data are typically collected by different people under varying field

conditions, resulting in data sets that are difficult to standardize. Several researchers

have examined this problem in varied taxa, such as birds (Grant 1979; Zink 1983; Arendt

and Faaborg 1989; Yezerinac et al. 1992), bats (Palmeirim 1998), and snails (Bailey and

Byrnes 1990). However, they have used museum specimens or shells. Live animals are

even more difficult to standardize. The most common data collected on crocodilians are

morphometric measurements (e.g., head length, snout-vent length, total length, and

weight). These measurements are used in a wide range of analyses, from ecological to

taxonomic to evolutionary (Chabreck and Joanen 1979; Hutton 1986; Hall 1991).

However, these analyses are only as accurate as the data from which they are derived.

Methods

We performed morphometric measurement trials on the American alligator to

quantify inter-measurer error and determine which morphometric measurement had the

least measure error associated with it. This is necessary to increase confidence levels in

future analyses.

Two trials were performed: 1) measurers were inexperienced volunteers; 2)

measurers were experienced alligator biologists. The inexperienced trial was performed


*Chapter reprinted with permission from the editors ofHerpetological Review (in press)









in the fall of 2000 using ten alligators captured at Arthur R. Marshall Loxahatchee

National Wildlife Refuge (Loxahatchee) located in Boynton Beach, FL, with six groups

of volunteers. Alligators were located by observing eye shines from an airboat in the

marsh interior and captured using a wire snare. They were secured in the boat and

brought back to a covered area for measurement. Alligator size varied from 108-248 cm

total length. The volunteers consisted of Loxahatchee staff, University of Florida

employees, U.S. Geological Survey employees, and students from Palm Beach Atlantic

Community College. Each group consisted of 4-6 people and they were given both

verbal and written measurement instructions (Appendix A). Each group measured every

alligator. Different individuals within the group took measurements of each alligator, but

measurements were agreed upon within the group before being recorded. Participants

were allowed to collaborate within but not between groups. The following were

measured by each group on every alligator: head length (HL), snout-vent length (SVL),

total length (TL), right hind foot length (HFL), neck girth (NG), tail girth (TG), chest

girth (CG), and mass. HL was measured dorsally, while SVL and TL were measured

ventrally. A measurement kit was provided with every alligator that contained the

following: a clipboard, pencil, string for measuring tail girth between scutes, a Pesola

scale, and a flexible centimeter sewing tape. These kits stayed with the alligator so that

the same equipment was used. Measurements were made with the flexible sewing tape to

the nearest 0.1 cm. Mass was measured with 10-50 kg Pesola scales to the nearest 0.1

kg. The scales were calibrated before use using a weight of known mass. Alligators also

were marked with individually numbered size 3 Monel tags, provided by the Florida Fish









and Wildlife Conservation Commission, in the webbing of the back left foot and scute

clipped for permanent identification.

The second trial was performed in fall 2001 with experienced alligator biologists.

Nine alligators ranging from 156 cm to 255 TL were captured at Loxahatchee and used

for the trial. Ten biologists were present and measured each alligator using the same

procedures and instructions as in the previous experiment with three exceptions: 1) each

individual measured every alligator and was not allowed to discuss their findings; 2) the

participants worked in groups of two to facilitate measuring; and 3) each person recorded

their own measurements to reduce bias in and among groups.

Data were analyzed to determine which measurements were most reliable, or

contained the least percent measurement error (%ME). Reliability for between

experience and inexperienced groups was evaluated using coefficient of variation.

Standard deviation for each measurement was calculated and divided by the mean of that

measurement for that alligator. A T-test was performed on the coefficients of variation to

determine significant differences in the reliability of the measurements between

experienced and inexperienced measurers. Six of the experienced measurers were

randomly chosen so that the N in the T-test would be equal to the six groups of

inexperienced volunteers.

Percent measurement error was calculated for the experienced group to determine

which measurements are more reliable for morphometric studies. Bailey and Byrnes

(1990) pioneered the use of Model II ANOVA and ANCOVA to estimate within-

individual and among-individual components of covariance and variance to predict

percent measurement error. A Model II ANOVA and ANCOVA was run as a part of the









SAS NESTED procedure (SAS Institute, Inc. 1988), according to the Bailey and Byrnes

procedure. The among-individual variance was calculated by the following equation

(Yezerinac, et al. 1992):

2 among = (MSamong MSwithin) / m,

where MS is the mean squared deviation and m is the number of repeated

measurements. Covariances were calculated as follows:

rx.y among = (COVx.y among / Sx among Sy among).

%ME was then determined by:

%ME = 100% (S2within / S2within + S2among).

Results

The results of the T-test (Table 1) suggest that experienced individuals measure

more accurately for HL, CG, TG, and mass than groups of inexperienced volunteers.

There is no statistical difference for SVL, TL, HFL, and NG.

Percent measurement error ranged from 0.50 to 49.53% (Table 2). HFL had the

highest %ME because the landmarks used to measure it are not easily located. TL was

also relatively high. It should be one of the most reliable measurements, as it is the

largest and allows for greater margin of error. It's interesting to note the high covariance

that NG and CG share with the other volumetric measurements.

Discussion

This assessment can be used as a guide for future crocodilian studies that use

morphometric measurements, such as the analysis of growth rate and condition. If

performing morphometric analysis using data collected from inexperienced groups, it

would be more accurate to use TL, SVL, or NG. For experienced individuals, HL, SVL,

TG, or weight should be used. For example, condition factor analysis requires a skeletal









measurement and a volumetric measurement. SVL would be adequate for the skeletal

measurement regardless of who collected the data, but more care should be taken in

choosing the volumetric measurement. If these trials were to be replicated, it would be

useful to modify the design so that each individual would take three or more

measurements of the same measurement (e.g. three measurements of HL, three of SVL,

etc.). This would provide data to quantify intra-observer error.

Table 1: T-test results for coefficient of variation of experienced individuals versus
inexperienced groups of volunteers measuring American alligators at A.R.M.
Loxahatchee National Wildlife Refuge


Measurement P-value
HL 0.046*
SVL 0.098
TL 0.129
HFL 0.758
CG 0.010*
NG 0.284
TG 0.015*
Mass 0.010*



Table 2: Within alligator variance and covariance components for length and volumetric
measures. Diagonal elements are variance components represented by percent
measurement error (%ME). The off-diagonal elements are covariance
components represented by within-gator correlations withinin.

HL SVL TL HFL NG CG TG Weight
HL 0.79
SVL -0.07 1.52
TL 0.02 0.01 3.88
HFL 0.09 -0.15 0.01 49.53
NG -0.04 0.16 -0.15 0.14 7.64
CG -0.03 0.10 0.16 -0.22 0.31 4.89
TG -0.06 0.09 0.10 0.10 0.63 0.40 2.93
Weight -0.08 -0.08 0.09 0.04 -0.25 -0.10 -0.13 0.50














CHAPTER 3
CONDITION INDEX ANALYSIS OF THE AMERICAN ALLIGATOR

Introduction

Condition as defined in chapter one is the "relative fatness of [an animal]. ... a

measure of how well that animal is coping with its environment"(Taylor 1979, p. 349).

This relationship and its relative ease of measurement is the key to using the American

alligator (Alligator mississippiensis) as an indicator of the health of its environment.

Condition indices are often used to quantify body condition over space and time

(LeCren 1951, Taylor 1979, Springer 1990, Brandt 1991, Elsey 1992, Krebs and

Singleton 1993, Dalrymple 1996, Jakob 1996, Barr 1997, Leslie 1997, Delany 1999).

Biologists often note that an alligator is too skinny or a "healthy" size, but those

observations are qualitative. For example, this project was developed because of

observed condition differences between alligators within parts of the Everglades system.

The difference was apparent, but not quantified. Quantitative condition indices can

compare condition within and among populations, and if used carefully, "can provide

insights into the status of ecosystems" (Murphy et al. 1990, p. 86)

Condition indices can be calculated in several different ways, but is always a function

of skeletal length and a volumetric measurement. They have been used in fisheries

literature since the 1940's, where they have become deeply entrenched as a research

paradigm (Cone 1989). LeCren first reviewed this process in 1951. He described one

relationship on which the condition indices are based:

(1) W= aLn,









where W = weight, L = length, "a" is a constant and "n" is an exponent that has been

calculated in past literature (LeCren 1951) to be between 2.5 and 4.0 for fish body forms.

In most fisheries literature, "n" = 3.0, the number used when growth is isometric.

Isometric growth occurs when both factors, i.e. length and weight, grow at the same rate.

If it is accepted that growth is isometric and "n" equals 3.0, then a condition index

equation can be written as:

(2) K = W/L3 x 10n,

where "n" = 2,3,4, or 5 and is commonly chosen so that the mean ofK is larger than one

(Cone 1989) and K is known as "Fulton's condition factor". This equation is quite

limited by its assumption of isometric growth and that "n" is equal to 3.0.

To solve the problem of the strict assumptions of K, LeCren (1951) formulated the

relative condition factor, or Kn. Instead of using an isometric length/weight relationship

(3.0), he used empirical data and a least squares regression to formulate "n" from (1):

(3) Kn=W/aL".

The primary difference between the two K's is that Fulton's condition factor, K,

measures the deviation of an individual from an ideal, theoretical organism while the

relative condition factor, Kn, measures the deviation of an individual from the average of

actual population data. Kn is often used outside of fisheries studies where growth is

usually allometric and "n" is estimated from field data.

Relative condition factor is also limited in its use. Because you must calculate the

ideal length/weight standard separately for each population, it is impossible to compare

the condition factor value across populations (Murphy et al. 1990).









A third and more recent condition index is relative weight, Wr. This was postulated

by Wege and Anderson (1978) and is the ratio of an individual's weight (W) divided by a

standard weight (Ws) for that length times 100, or

(4) Wr = (W/W) x 100.

Ws is an optimal standard which establishes the index value for a species. It is regularly

defined in fisheries literature as the 75th percentile (Murphy et al. 1990), which is

considered optimal growth. Relative weight avoids the drawbacks of Fulton's condition

factor and relative condition factor, because it doesn't vary with changing species or

mean sample size (Springer and Murphy 1990; Quist et al. 1998).

The residual index is another method of calculating condition. The residual index is

calculated by regressing the volumetric measurement on body length. The residual

distance of the individual points to the regression line functions as the condition index.

Some data has to be transformed so that it meets the assumptions of regression. A

condition index also must be independent of body length, which can limit morphometric

measurements that can be used (Green 2001). The residual index possesses the same

flaw as relative condition factor in that it is not comparable across populations, because

the regression line is only valid for the current population.

Researchers also have used ANCOVA to evaluate condition (LeCren 1951; Packard

and Boardman 1988; Garcia-Berthou and Moreno-Amich 1993; Garcia-Berthou 2001).

ANCOVA allows the length-weight relationship to be examined without the confounding

effects of the covariance of body length. Packard (1988) and Garcia-Berthou (2001) rule

out other ratio indices for condition and endorse ANCOVA as the best method for

calculating condition. However, it suffers from the same fault as relative condition and









the residual index in that it is not possible to compare the condition of two populations

with this method, unless the "n" in equation 1 is identical for the two populations being

compared.

While condition index analysis is best known in the fisheries literature, it also has

been used for crocodilians. Condition index analyses (using relative condition factor)

have been used during studies on the food habits of alligators in Everglades National Park

(Barr 1997); to compare condition and growth of juvenile alligators in Par Pond, South

Carolina (Brandt 1991); and to compare condition index for Everglades alligators in

Shark Valley (Dalrymple 1996). Elsey et al. (1992) used growth rates and body

condition factors to compare wild and farm-released juveniles in Louisiana; and Santos,

et al. (1994) compared different habitats of the Pantanal in Brazil using condition factor

of Caiman crocodilusyacare. Leslie (1997) used relative condition factor to determine

the condition of a population of Crocodylus niloticus in South Africa. Australian

researchers have also explored condition relating to crocodiles (Garnett and Murray

1986; Grigg et al. 1986).

Many condition index analyses include an actual measure of condition by

sacrificing animals and extracting total body fat (Krebs and Singleton 1993; Jakob et al.

1996; Quist et al. 1998; Sutton et al. 2000; Vila-Gispert and Moreno-Amich 2001). The

body fat percentage is then regressed on the condition index to obtain a relationship.

None have been done with as large an animal as an alligator. It was not feasible to digest

whole alligators for this study.

Alligators are qualitatively fatter or skinnier in different areas of the Everglades,

and some differences in time have been noted in the literature (Dalrymple 1996, Barr









1997). Therefore a goal of this study was to find a condition index that is able to

distinguish the greatest number of differences between these Everglades areas. The main

objective was to determine the appropriate condition index for the American alligator

with the intention that it will used as a tool to monitor population health. Because the

alligator's life history is so closely linked to the water levels of their environment,

condition is a way to determine how changes in their environment are affecting alligators.

With this monitoring tool, condition could be used as a performance measure for the

current Everglades restoration effort.

Study Area

Data for this study were collected from six areas within the Greater Everglades

Ecosystem in south Florida (Figure 2). The northernmost area, Water Conservation Area

1, which was designated as part of the Arthur R. Marshall Loxahatchee National Wildlife

Refuge (LOX) in 1951, is located in Boynton Beach, FL. It is a 572 square kilometer

area bounded by the West Palm Beach Canal on the north and the Hillsboro Canal to the

south.

Alligators also were captured in Water Conservation Areas 2 and 3 (WCA2 and

WCA3), which are operated by the Florida Fish and Wildlife Conservation Commission

(FWC) and the South Florida Water Management District (SFWMD). WCA2 is a 448

square kilometer pool used for excess water from LOX, supplies water to urban areas of

southeast Florida, and wildlife conservation. Two locations were sampled for alligators

in WCA3, one at the southern end and other in the middle. WCA3 is 1948 square

kilometers of storage for excess water from WCA2 and also is designated a water supply

area for urban areas in south Florida and for wildlife conservation.









Two sites within Everglades National Park were used: Shark Slough and the

estuarine areas located near Flamingo. Shark Slough extends from the northern border of

Everglades National Park to the headwaters of the Shark and Harney Rivers and is the

largest drainage in the Park (Olmstead and Armentano 1997). All estuarine alligators

were caught near Flamingo in Everglades National Park.

Materials and Methods

Capture Methods

Alligators were captured during the period from October of 1999 to March/April

2002 by a multi-agency team that consisted of members from U.S. Fish and Wildlife

Service (USFWS), U.S. Geological Survey (USGS), University of Florida (UF), and the

FWC. Animals were captured from all study areas in marsh habitats only, excluding

canal alligators, as their condition may differ from marsh alligators because of the altered

habitat. Due to a concurrent aging study, only 1.22-1.83 meter alligators were caught

during the first two catches. However, the size range restriction was eliminated in

October of 2000 in include a larger size range for an improved analysis.

Fifteen alligators were captured from each area using airboats in marshes and using

motorboats in estuaries during each of six capture events. Alligators were located by

spotlight and captured with a noose or dart. Head length (HL), snout-vent length (SVL),

total length (TL), right hind foot length (HFL), neck girth (NG), tail girth (TG) and chest

girth (CG) were measured to the nearest 0.1 cm (Appendix B). Mass was measured with

a Pesola spring scale to the nearest 0.1 kg. Alligators were sexed and blood was drawn

for a concurrent contaminants study and for hematocrit analyses. The alligators were

then released at their original capture site.









Measurements from several animals were eliminated from this analysis due to

discrepancies. For example, captures with bobtails (an alligator missing part of its tail),

missing data, or known anomalies such as an alligator that is fed by people and is heavier

than a typical Everglades alligator.

Condition Indices Analysis

A flowchart of the process used to choose the most appropriate length and volume

and condition index for this analysis is provided in Figure 1. Several length and

volumetric measurements were eliminated from consideration due to measurement error

(Chapter 1). Condition indices were located in the literature (see Introduction for

discussion of each index) and calculated using the four combinations from Chapter 1:

HL/TG, HL/Mass, SVL/TG, SVL/Mass. The slash used in this notation represents the

combination of two morphometric variables, not the division of one by the other. The

data were pooled (n = 395, max TL = 264 cm, min TL = 100.4 cm) for all areas when

calculating all indices to provide the largest sample size, since they could be considered

members of the Greater Everglades population. The regression for Relative K was run

using PROC NLIN in SAS and the regression for the residual index was run using PROC

REG in SAS (SAS Institute 1988).

A Pearson's correlation was run on the results of the five indices (Fulton's K,

Relative K, Relative Mass, Residuals, and Ratio index), using the four body length and

volume combinations, to determine whether body length was correlated to condition.

Combinations with correlations greater than 30% were eliminated from further

consideration, as this suggested lack of independence. Condition index should be

independent of the body size indicator or straight measurements for the assumptions of

the ordinary least squares regression used in these analyses (Green 2001).









An ANVOA was used (SAS PROC GLM [SAS Institute, Inc. 1988]) to determine

whether condition differed (p < 0.05) by area for all catch events combined for the

indices. Least Significant Difference (LSD) and Tukey's post-hoc tests were run using

SAS to establish which areas differed significantly. Both were calculated to establish

which tests were sensitive enough for our analysis. Both tests calculate differences

between areas, but Tukey's is more conservative than LSD. The condition model using

ANCOVA was run using PROC GLM in SAS (SAS Institute 1988) and results were

analyzed with the LSD post-hoc test to determine differences among areas.

Area differences were counted if an area was significantly different than another

according to the LSD analysis. Among six areas there are fifteen possible differences

and they are written as follows: 112, 113, 114, 115, 116, 213, 214, 215, 216, 314, 315, 316, 415,

416, 516, where area one is different than two, area one is different than three, etc.

Condition indices combinations that delineated significantly fewer populations were

eliminated.

Body length and volumetric combinations of indices that displayed the maximum

amount of differences were used for temporal analysis. Those data were divided by

capture event and the analysis was rerun using the capture event as an added parameter

for the ANOVA analysis. The same was also performed with body length and volumetric

combinations of ANCOVA. Capture events involved were October 1999, March/April

2000, October 2000, March/April 2001, October 2001, and March/April 2002. There

were no ENP-Est alligators for the March/April 2000 capture event.

Simulated Data

To address the question of whether the differences found in the ANOVA and

ANCOVA analysis were Type I or Type II errors, simulated data sets were constructed to









test the analysis. Simulated data sets consisted of 15 populations-five populations of

100 individuals with the same condition index mean and standard deviation as the

original data set, five populations of 100 individuals with a 5% increase in the mean and

the same standard deviation as the original data set, and five populations of 100

individuals with a 10% increase in the mean and the same standard deviation as the

original data set. The populations were created using the RV.NORMAL syntax in SPSS

(SPSS Inc. 2001). ANOVA and LSD were used to analyze the simulated data.

ANCOVA data were created in a similar manner. Random populations of HL and

SVL were also generated with the RV.NORMAL syntax in SPSS (SPSS Inc. 2001).

Mass was then regressed on HL and SVL using PROC REG in SAS (SAS Insitute 1988)

and the resulting equation was used to generate mass.

Everglades Condition Analysis

A Fulton's K analysis was performed using HL and Mass. K was calculated for all

captures and was analyzed using an ANOVA and LSD post-hoc test to detect differences

between the areas in the Everglades. Areas were then grouped by significant differences

in mean condition.

Results

Condition Indices Analysis

The mean, minimum, and maximum values for the indices were variable, even

within indices (Table 3). The "b" value calculated for Relative K HL/Mass, HL/TG,

SVL/TG combinations is significantly different than 3.0 (2.83 0.05 S.E., 1.03 0.02

S.E., and 1.05 0.02, respectively). The SVL/Mass combination was not significantly

different than 3.0 (3.01 0.04).









For the ratio index, both mass combinations (HL/Mass and SVL/Mass) were

eliminated from the analysis because they were correlated with the body length by more

than 30% (Table 4). However, Fulton's K mass combinations were included while those

calculated from TG were excluded. All Relative K combinations were incorporated into

the analysis, as were all residual indices. HL/Mass was excluded from the analysis for

Relative Mass because of its high correlation value (30%).

ANOVA analyses were performed on the remaining condition indices to determine

if there were significant condition differences between areas (Table 5). All were

significant except the residual index HL combinations. Several indices; ratio, Relative K,

and relative mass HL/TG combinations, had no more than three differences between

areas, according to the LSD post-hoc test. Tukey's test was deemed too conservative for

the analysis, because it did not show differences for a length and volumetric combination

that had a highly significant ANOVA p-value (p < 0.01). Two condition indices;

Relative K and Fulton's K HL/Mass combination, display seven differences between

areas.

The ANCOVA analysis only includes two significant differences out of four

combinations (Table 5). HL/Mass and SVL/Mass found seven and eight LSD post-hoc

area differences, respectively. HL/TG and SVL/TG display higher p-values (by a factor

of 10) than any of the condition indices.

ANOVA/ANCOVA Analysis by Capture Event

Fulton's K, Relative K HL/Mass, and ANCOVA SVL/Mass and HL/Mass were

chosen for further analysis by capture event because of the high number of differences

they exhibited in the first ANOVA analysis. Sample size for capture events ranged from

44 to 83 and had no effect on the significant differences. There were no patterns of









significance due to sample sizes. Results for the capture event ANOVA and LSD tests

were similar for the Fulton's and Relative K HL/Mass combinations (Table 6). There

was no difference in the order of means (from highest to lowest). Fulton's K only found

one additional condition difference among LOX and WCA2 in the first capture event that

Relative K did not. They found significant condition differences in the same three

capture events and their p-values were similar.

The HL/Mass ANCOVA analysis showed differences for the first and last capture

event. No area differences were found in the other capture events. The SVL/Mass

combination was not able to differentiate between areas for any capture events (p> 0.10).

Simulated Data

The simulated data analysis confirmed that the procedure of evaluating the

condition indices does not commit any Type I errors. For both Relative and Fulton's K,

the analysis never found differences that were not there. The populations with the exact

mean and standard deviation as the original data were always significantly different than

the populations with the 10% increase in the mean. However, this did not hold true for

the 5% difference. Percent differences ranged from 64 to 100 percent (Tables 7 and 8).

The method for creating simulated data for the ANCOVA analysis was not

successful. The random HL and SVL data were correct for mean and standard deviation,

however the method of creating mass for the HL and SVL did not provide results. The

simulated mass did not conform to the bounds of the original data and could not be used

in a simulated data analysis.

Everglades Condition Analysis

In this study, during the 1999-2002 sampling period, areas within the Everglades

do show significant differences in condition (Figure 4). ENP-SS exhibited the highest









mean condition and is significantly higher than any other area (Figure 5). The two areas

of WCA3 display lower condition than ENP-SS, but are still significantly higher than

WCA2, the lowest mean of the groups. The LOX and ENP-Est condition means are

significantly lower than ENP-SS and are not higher than any other group.

Discussion

Condition indices are a controversial way of estimating condition for populations

(Packard and Boardman 1988; Cone 1989; Krebs and Singleton 1993; Jakob et al. 1996;

Kotiaho 1999; Garcia-Berthou 2001; Green 2001; Vila-Gispert and Moreno-Amich

2001). They aren't appropriate for every situation and there are strict assumptions that

must be met for them to be used correctly. The main assumptions are that condition is

independent of body length, that the measure used for body length is an accurate measure

of structural size, that body length measurements are not subject to error, and that the

relationship between body size and mass is linear (Green 2001). However, they can be a

good, non-destructive way to rapidly evaluate the health of populations of concern. This

analysis was designed to evaluate the use of condition indices for alligators.

The most accurate measurement of an alligator's volume is mass. However,

because large alligators often can not be weighed in the field, TG was proposed as an

easily measured alternative. It is highly correlated to SVL (Figure 3), and is where

alligators store fat, so it should be a good measure of fat stores. However, the "b" that

was calculated for TG combinations for Relative K were calculated to be approximately

1.0, instead of the 3.0 that Fulton's K supposes for isometric growth. SVL/Mass and

HL/Mass were approximately 3.0. The TG combinations for Fulton's K were too highly

correlated with their straight lengths to be included in the analysis, but it would have been









incorrect to use them because their "b" value is too far away from 3.0 for it to be a valid

use of Fulton's K.

The ANOVA/ANCOVA analysis provided a simple way to narrow the condition

indices down to a manageable number. Fifteen of 19 index combinations had four

significant condition differences or less, as shown by the LSD post-hoc test. They are

still valid ways of comparing condition, especially where HL was not measured, but were

deemed not sensitive enough for this analysis.

The simulated data set allayed concerns of artifacts in the data allowing differences

to be found by the LSD post-hoc tests that were not legitimate. It also allowed us to

determine the range of this analysis. It could accurately detect condition differences of

10% or more. It also can detect differences of 5%, but only at a range of 64-96% (Tables

7 and 8). One can be safe, however, in supposing that a difference found is truly a

difference in the data.

The condition indices included in the final analysis produced different results. The

HL/Mass combinations of Relative K and Fulton's K were very similar in the ordering of

means and number of differences in condition between areas. This is most likely because

they are equations that were derived from the same weight/length relationship and only

vary by the distance of"b" from 3.0, the number used when growth is isometric. They

also were similar in the pooled data analysis.

The ANCOVA analysis also provides adequate differentiation and is used most

recently in the literature because of its ability to remove the covariance of size with

condition. However, it falls victim to the same shortcoming as many of the other

indices-you can not use it to compare populations, but only groups within a population,









such as across capture events in this study. The SVL/Mass ANCOVA combination

detected the maximum number of area differences in this study, but it did not detect any

differences in the temporal analysis. It would have been useful to use an ANCOVA for

temporal analyses of the same population if the sample size was sufficient..

This analysis has shown Fulton's K and Relative K to be nearly equal in their

ability to distinguish condition differences among areas, even though the "b" value

calculated is significantly different than 3.0. If Fulton's K and Relative K are so similar

for alligator populations, it would be wise to use Fulton's K because condition scores

calculated using Fulton's K can be compared across populations. A downside to both of

these indices is that they do not provide one with an easily interpreted biological number.

The Fulton's condition score for this analysis ranged from 1.4 to 3.2, unlike Relative

Mass, where 1.0 is an alligator in normal condition and anything less than 1.0 means the

alligator is in poorer condition. However, Relative Mass does not have the ability to

distinguish as finely between areas as Fulton's or Relative K did for our alligator data or

between populations. To overcome the problem of a lack of biological meaning in Fulton

K numbers, one could identify quartiles of the data and label the bottom quartile as low

condition, the middle two as average condition and the top quartile as good condition

(Figure 6).

One must be careful when separating condition into high and low that instant

judgments are not attached to those categories. When using a condition index, results

must be taken in the context of nearby populations. Fatter alligators are not necessarily

living in a better environment. Like with humans, there are upper and lower limits to fat

stores as they relate to condition. It is not minor fluctuations that we are interested in, but









those extremes that could give clues to population and ecosystem health. As an example,

I offer an analysis of SVL/Mass Fulton's K for areas in South Carolina (P. Wilkinson,

unpublished data), north Florida (A. R. Woodward, Florida Fish and Wildlife

Conservation Commission, unpublished data), and the Everglades. South Carolina data is

from Santee captures in 1980 (n = 100). North Florida captures are from Lake Griffin in

1987, 1989, and 1990 (n = 30), Lochloosa Lake in 1985 (n = 57), Newnan's Lake in 1985

(n = 40), Orange Lake in 1985, 1987, 1989, and 1990 (n = 142), and Lake Woodruff in

1987, 1989, and 1990 (n = 33). SVL/Mass was used because HL data were not available.

Lake Griffin would be classified in the high condition quartile (Figure 7) and the

Everglades to be in the low condition quartile. It is not necessarily true that the alligators

in Lake Griffin are in better health than those in the Everglades. If you look at Lake

Griffin in context with other north Florida lakes, it looks like its condition is high and that

perhaps something is wrong with that ecosystem, such as the fact that Lake Griffin is a

highly eutrophic lake from agricultural run off. The recent die-off of adult and subadult

alligators in Lake Griffin (Schoeb et al., 2002) corroborates the fact that Griffin is not a

healthy ecosystem. Everglades population is in lower condition as compared to north

Florida and South Carolina, but if data is analyzed over time for only the Everglades, it

would reveal a pattern of consistent low condition and should not signal a crisis for

Everglades alligators, especially when compared with other long-term Everglades

alligator data.

Condition is a very fluid measurement. Water management practices and rainfall

can dramatically change condition of animals in a relatively short amount of time,

because so many aspects of their life history (feeding, courtship, and nesting) depend on









seasonally fluctuating water levels. In this study, we found that ENP-SS alligators had

the highest condition of Everglades alligators, but Dalrymple (1996) and Barr (1997)

observed Shark Slough alligators to be in very poor condition only five or six years

earlier. Water levels were high during Dalrymple and Barr's period of capture, possibly

affecting food availability.

Hyrdologic conditions are important to the American alligator, but are not the only

factor that affects condition. Disease, climatic change, nutrient input, and contaminants

all contribute to condition and should be considered when using a condition index to

analyze population health. While condition is a useful way to evaluate current

Everglades restoration, it should not be taken out of context of events in the greater

ecosystem.











Table 3: Descriptive statistics of indices used in condition analysis for the American
alligator, south Florida (n = 395) with captures from October 1999 to Spring
2002.

Index Equation Factors MinimumMaximum Mean Std. Deviation
Ratio Volumetric/Length HL/TG 0.944 1.725 1.447 0.103
SVL/TG 0.288 0.455 0.393 0.025
Fulton's K Volumetric/(Length^3) HL/Mass 2.572 16.389 10.849 1.556
SVL/Mass 1.430 3.200 2.162 0.253
Relative K Volumetric/(Length^b) HL/TG 8.454 15.612 13.118 0.915
HL/Mass 0.470 2.842 1.848 0.281
SVL/TG 2.327 3.719 3.204 0.203
SVL/Mass 1.381 3.089 2.087 0.243
Relative MassObserved mass/Expected mass HL/TG 0.648 1.185 0.992 0.070
SVL/TG 0.727 1.156 0.997 0.064
SVL/Mass 0.656 1.473 0.989 0.115
Residuals Actual value Expected value HL/TG -18.833 6.967 8.314 2.623
HL/Mass -29.728 18.748 3.294 3.685
SVL/TG -11.633 6.506 -2.569 2.393
SVL/Mass -7.744 19.405 -1.476 3.466
HL = Head Length, SVL = Snout-Vent Length, TG = Tail Girth










Table 4: Condition indices calculated for the American alligator in south Florida from
October 1999 to Spring 2002 (n = 395) whose Pearson r-values are less than
30%.


Condition Method Factors
Ratio HL/TG


Fulton's K


Relative K





RelativeMass




Residuals


SVL/TG
HL/Mass
SVL/Mass
HL/Mass
HL/TG
SVL/Mass
SVL/TG
SVL/TG
HL/TG
SVL/Mass
HL/Mass
HL/TG
SVL/Mass
SVL/TG


Pearson 'r'(p-value)
0.122(0.015)
0.132(0.009)
0.059(0.239)
0.065(0.196)
0.277(0.000)
0.046(0.359)
0.051(0.356)
-0.007(0.888)
0.116(0.021)
0.170(0.001)
0.043(0.399)
<0.001
<0.001
<0.001
<0.001












Table 5: Results of ANOVA/Least Significant Difference and ANCOVA/ Least Significant Difference analyses of condition indices
for the American alligator in south Florida from October 1999 to Spring 2002 (n = 395). Order of condition index is from


highest to lowest mean.
Condition Index Factors p-value
Ratio HL/TG 0.007
SVL/TG 0.037
Fulton's K HL/Mass <0.001
SVL/Mass 0.028
Relative K HL/Mass <0.001
HL/TG 0.005
SVL/Mass 0.025
SVL/TG 0.028
RelativeMass SVL/TG 0.027
HL/TG 0.007
SVL/Mass 0.014
ANCOVA HL/Mass 0.004
HL/TG 0.709
SVL/Mass 0.005
SVL/TG 0.436
Residuals HL/Mass 0.0164
HL/TG 0.0247
SVL/Mass 0.0927
SVL/TG 0.0747
1 = ENP-Estuaries, 2 = ENP-Shark Slough, 3


# groups LSD


Order
362154
365124
263514
265314
263154
362154
265341
365124
365124
362154
265341
136524


Differences (LSD)
214,314,416
213,314,416
112,213,214,215,216,413,41|6
112,213,214
112,213,214,215,216,314,41|6
214,314,416
112,213,214
113,213,314,416
213,314,416
214,314,416
112,213,214
113,116,115,112,114,312,31|4


136524 113,116,115,112,114,315,312,314


362514
362514


214,314,614
214,314,614


A.R.M Loxahatchee NWR, 4 = WCA2A, 5 = WCA3A-Holiday Park, 6 = WCA3A-Tamiami











Table 6: ANOVA/ANCOVA analyses of condition indices by capture event from October 1999 to Spring 2002 (n = 395). Order of
condition index is from highest to lowest mean. There were no ENP-Est captures for March/April 00.


Condition Index
Fulton's K





Relative K





ANCOVA


Factors Capture Event
HL/Mass October 99
March/April 00
October 00
March/April 01
October 01
March/April 02
HL/Mass October 99
March/April 00
October 00
March/April 01
October 01
March/April 02
HL/Mass October 99
March/April 00
October 00
March/April 01
October 01

March/April 02
SVL/Mass October 99
March/April 00
October 00
March/April 01
October 01
March/April 02


# groups LSD


p-value
<0.001
0.0487
0.004
0.6061
0.5458
0.0536
0.0012
0.0485
0.0056
0.5352
0.5841
0.0878
0.0299
0.1268
0.6431
0.393
0.7944

0.0203
0.9094
0.3708
0.8468
0.2913
0.4194
0.1403


Order
123654
26534
261534



123654
26354
261534



236415





312465


1 = ENP--Estuaries, 2 = ENP-Shark Slough, 3 = Loxahatchee NWR, 4


Differences (LSD)
1/3,1/6,1/5, 1/4, 2/3, 2/6, 2/5, 2/4, 3/4
2/3, 2/4
2/6, 2/1 2/5, 2/3, 2/4



1/3, 1/6, 1/5, 1/4, 2/3, 2/6, 2/5, 2/4
2/3, 2/4
2/6, 2/1, 2/5, 2/3, 2/4



2/3, 2/6, 2/4, 2/1, 2/5, 3/5, 5/6, 4/5




2/3, 4/3, 6/3, 5/3, 1/2, 1/4, 1/6, 1/5, 2/4, 2/6, 2/5,
4/5


WCA2A, 5 = WCA3A-Holiday Park, 6 = WCA3A-Tamiami










Table 7: Ability of ANOVA/LSD analysis to detect differences in condition from
Relative K simulated data. Numbers are the percentage of total populations for
which the analysis found a significant difference. Type 1 is mean and
standard deviation from original data, Type 2 is 5% increase of mean, and Type
3 is 10% increase of mean.


Population Type


1 2 3
1 0


2 64
3 100


0
96 0


Table 8: Ability of ANOVA/LSD analysis to detect differences in condition from
Fulton's K simulated data. Numbers are the percentage of total populations for
which the analysis found a significant difference. Type 1 is mean and standard
deviation from original data, Type 2 is 5% increase of mean, and Type 3 is 10%
increase of mean.


Population Type
1
2
3


2 3















Measurement
experiment


% ME


Narrowing of
meas. to HL,
SVL, TG, Mass


Lit Search
--


Choice of
indices


Calculate
indices for
total data set
(n=395)


Examine
assumptions:

correlations


Reduced set of ANOVAs
indices


Significant area
differences


Post-hoc tests
(LSD)


Ordering of indices
and differences


SPost-hoc by
time period
---


Analysis by
capture event


Simulated Choice of index
data


Figure 1: Flowchart of condition indices analysis for the American alligator in the Everglades.

















LOX


-i \ ^ WCA2



WCA3 *







ENP
\ iit
4.























Figure 2: Capture areas in south Florida within Everglades wetlands core. Black dots
represent alligator captures from October 1999 to March 2002 (n = 395). LOX
= A.R.M. Loxahatchee National Wildlife Refuge, WCA2 = Water Conservation
Area 2, WCA3 = Water Conservation Area 3, ENP = Everglades National Park.

















60



50



240 **








20



10



0-
0 20 40 60 80 100 120 140 160

Snout-Vent Length (cm)


Figure 3: Linear regression of tail girth vs. snout-vent length of the American alligator of captures in south Florida from October 1999
to March 2002 (n = 395).











16

14

12

10



6

4

2

0
0 1 2 3 4 5 6 7

Area*


Figure 4: Range of Fulton's K index values by capture area in south Florida. Captures are from October 1999 to March 2002 (n
395). 1 = ENP-Est 2 = ENP-SS 3 = LOX 4 = WCA2 5 = WCA3-HP 6 = WCA3-T









ENP-SS


WCA3-Both areas

LOX
ENP-Est

WCA2


Highest


Mid



Mid to Low


Figure 5: Hierarchy of Everglades alligator condition from October 1999 to March 2002
using Fulton's K analysis (n = 395). Brackets signify that condition does not
differ significantly within that group.










1st quartile: 2.57 -

2nd quartile: 4.14 -

3rd quartile: 10.85

4th quartile: 13.62


4.13 low condition

10.84 low to average condition

- 13.61 -average to high condition

- 16.39 high condition


Figure 6: Sample condition range of alligators for Fulton's K from Everglades data from
October 1999 to March 2002 (n = 395). This shows division of condition by
quartiles











3.6

3.4

3.2 *

3.0

S2.8 -

2.6

2.4 I I I

2.2

2.0
N= 30 57 142 100 33 395 40
1 2 3 4 5 6 7

Area


Figure 7: SVL/Mass Fulton's K analysis for the American alligator in areas of South
Carolina, north Florida and the Everglades. 1 = Lake Griffin, FL (1987, 1989,
1990) 2 = Lochloosa Lake, FL (1985) 3 = Orange Lake, FL (1985, 1987, 1989,
1990) 4 = Santee, SC (1980) 5 = Lake Woodruff, FL (1987, 1989, 1990) 7=
Everglades, FL (1999, 2000, 2001, 2002) 8 = Newnan's Lake, FL (1985)














CHAPTER 4
CONCLUSION

Condition factor can be a useful tool for evaluating the health of a population if

used appropriately. This analysis examined the strict assumptions of properly calculating

condition indices, beginning with body measurements that reduced measurement error.

Head length (HL), snout-vent length (SVL), tail girth (TG), and mass exhibited the least

measurement error and were used in the subsequent condition analysis. Five condition

factors and one model of condition (ANCOVA) were calculated to establish which index

or model could differentiate between Everglades areas. It was determined through

ANOVA/LSD analysis of the condition indices that the HL/Mass combination of

Fulton's K and the SVL/Mass combination of ANCOVA were best able to distinguish

differences in condition between areas of the Everglades. The condition index used

depends on the question asked. For example, ANCOVA is best way to determine

temporal differences between populations. It can not be used to compare across

populations, unless strict assumptions about the two populations can be met. HL/Mass

Fulton's K can be used to spatially compare populations of the American alligator, and is

suggested by this study as the best condition factor to use for that purpose.

In the future, more data needs to be collected to develop a model so that condition

of the American alligator can be predicted by water levels. While other aspects of

condition are important to the overall model, hydrologic conditions are the focus of the

current restoration and a monitoring tool needs to be sensitive to that feature. Also,

before a condition index of the American alligator can used, management goals must be






37


made to choose the power of the analysis and then determine an appropriate sample size

from that power analysis.













APPENDIX A
DATA SHEETS AND MEASUREMENT INSTRUCTIONS GIVEN TO EXPERIMENT
PARTICIPANTS


Web Tag# Scute Clip # Area

GFC
(N/A for most)


Crew


Capture Date Capture Time




General Location / Describe Vegetation


UTM Northi


GPS LOCATION
UTM
ing Easting


Lwi~


Temperature


Water


Habitat Type


Water Depth





cm


Time Blood 1
Cap Method Capture Status (Gross)


SScute?


Time Blood 2
(Cardeilhac)


Muck
Depth
Total

Water


cm










Head Length SV Length Total Length Hind Foot L


cm cm cm cm
S
e
Tail Girth Neck Girth Chest Girth Weight x


cm cm cm kg

Deformities






Release Status Notes






Description of Measures and Coding


Web Tag # -- Number engraved on tag (GFC 37201).

Scute Clip # -- Animal number derived from scute clipping (used only on LOX).

Area -- Study area Everglades Nat Park Shark Slough (ENP-SS), Estuarine (ENP-EST); Loxahatchee
NWR (LOX); Water Conservation Area 3A South (WCA3AS, also WCA3AN, WCA2); Big
Cypress National Preserve (BCY).

Crew -- First initial and surname of boat crew (K. Rice).

Capture Date -- Date in 1 Oct 99 format.

Capture Time All times are in 24 hr (military) format (0215, 1622).

GPS Location -- UTM coordinates of capture site (Northing 548515, Easting 2891857).

General Location/egetation -- Describe capture site and dominant vegetation (1 mile N of ..Sawgrass).

Water Temp / Air Temp -- Temperature (in C) of Air and Water (- 6" below surface).











Habitat Type -- Specific habitat type:
Open water = 1 Cattail Marsh = 7
Forested Wetlands = 2 Water Lily/Floating Leaved Veg = 8
Shrubs/Shrub Islands 3 Canal 9
Mixed Emergents = 4 Alligator Hole = 10
Sawgrass Marsh =5 Levee Break 11
Spikerush Marsh = 6

Water Depth -- Water depth at capture site in cm.

Muck Depth -- Measure from surface of water to rock and subtract water depth in cm.

Cap Method -- Capture Method (Snare, Hand, Harpoon).

Capture Status -- Animal status at capture (Vigorous).

Time Blood 1 -- Time (24 hr) blood kit 1, Scute? Scute collected (y or n).

Time Blood 2 -- Time (24 hr) blood kit 2 used.

Head Length -- Measure (dorsal) from tip of snout to center of posterior end of skull in cm.

SV Length -- Measure (ventral) from tip of snout to posterior end of vent in cm.

Total Length -- Measure (ventral) from tip of snout to tip of tail in cm. Note if tip of tail missing, etc.

Hind Foot L -- Measure (ventral) from 1st single extended scute posterior to heel to the anterior end of
middle toe, not including the nail in cm.

Tail Girth Measure circumference of tail at 3rd scute row posterior of rear legs in cm.

Neck Girth -- Measure circumference of neck between head and shoulders in cm.

Chest Girth -- Measure circumference of chest just posterior to front legs in cm.

Weight -- Total weight of animal in kg.

Sex -- M or F.

Deformities -- Note any physical deformities or prominent scars (missing LR foot).

Release Status -- Animal status at release (Vigorous).

Notes -- Any additional information (found 1 dead 6' gator).















APPENDIX B
GENERAL PICTORIAL MORPHOMETRIC MEASUREMENT INSTRUCTIONS

Head Length-
Measure dorsally from tip of snout to center of posterior end of skull. Dashed line
indicates measurement.










Snout-Vent Length-
Measure ventrally from tip of snout to back of vent or dorsally to the back of the
second tail scute row. The zero tail scute is the last short tail scute-the last one from the
body that doesn't wrap all the way around the tail.







Silk I ^^--i I ^9 \ ~ Si Sam




Detail of vent for snout-vent length:












*Modified from King and Brazaitis (1971).
















































Total Length:
Measure straight distance dorsally or ventrally from tip of snout to tip of tail. Note
if tip of tail is missing. Make sure gator and tail are in a straight line, not like this one.


















Lqj i


















Tail Girth:
Measure the circumference of tail at 3rd scut ar legs.
; / Zorn tail -,riite

Fir-t tail ite

'Z Roe,_n and tnil -nllte]










Third tail nmit"


Neck Girth:
Measure circumference of neck between head and shoulders-between two giant
scutes. In this pictures, the scutes look a little farther back than they should.

























Chest Girth:
Measure circumference


to front legs.















LIST OF REFERENCES


Arendt, W.J. and J. Faaborg. 1989. Sources of variation in measurements of birds in a
Puerto Rican dry forest. Journal of Field Ornithology 60(1): 1-11.

Bailey, R.C. and J. Byrnes. 1990. A new, old method for assessing measurement error in
both univariate and multivariate morphometric studies. Systematic Zoology 39(2):
124-130.

Barr, B. 1997. Food habits of the American alligator, Alligator mississippiensis, in the
southern Everglades. Ph.D. dissertation. University of Miami, Miami, FL.

Brandt, L.A. 1991. Growth of juvenile alligators in Par Pond, Savannah River site, South
Carolina. Copeia 4: 1123-1129.

Campbell, M. 1999. Everglades alligator holes: distribution and ecology. M.S. thesis.
University of Florida, Gainesville.

Chabreck, R. and T. Joanen. 1979. Growth-rates of American alligators in Louisiana.
Herpetologica 35(1): 51-57.

Cone, R.S. 1989. The need to reconsider the use of condition indexes in fishery science.
Transactions of the American Fisheries Society 118(5): 510-514.

Craighead, F.C. 1968. The role of the alligator in shaping plant communities and
maintaining wildlife in the southern Everglades. Florida Naturalist 41: 2-7, 69-74,
94.

Dalrymple, G.H. 1996. Growth of American alligators in the Shark Valley region of
Everglades National Park. Copeia(1): 212-216.

Delany, M.F., S.B. Linda, and C.T. Moore. 1999. Diet and condition of American
alligators in 4 Florida lakes. Proceedings of the Annual Conference of
Southeastern Association of Fish and Wildlife Agencies. 53: 375-389.

Enge, K.M., H.F. Percival, K.G. Rice, M.L. Jennings, G.R. Masson, A.R. Woodward.
2000. Summer nesting of turtles in alligator nests in Florida. Journal of
Herpetology. 34 (4): 497-503.









Elsey, R.M., T. Joanen, L. McNease and N. Kinler. 1992. Growth-rates and body
condition factors of Alligator mississippiensis in coastal Louisiana wetlands-a
comparison of wild and farm-released juveniles. Comparative Biochemistry and
Physiology A: Physiology 103(4): 667-672.

Garcia-Berthou, E. 2001. On the misuse of residuals in ecology: testing regression
residuals vs. the analysis of covariance. Journal of Animal Ecology 70: 708-711.

Garcia-Berthou, E. and R. Moreno-Amich. 1993. Multivariate analysis of covariance in
morphometric studies of the reproductive cycle. Canadian Journal of Fisheries
and Aquatic Sciences 50: 1394-1395.

Garnett, S.T. and R.M. Murray. 1986. Parameters affecting the growth of estuarine
crocodiles, Crocodylusporosus, in captivity. Australian Journal of Zoology 34(2):
211-223.

Grant, P.R. 1979. Ecological and morphological variation of Canary Island Blue tits,
Parus caeruleus (Aves, Paridae). Biological Journal of the Linnean Society 11(2):
103-129.

Green, A.J. 2001. Mass/length residuals: measures of body condition or generators of
spurious results? Ecology 82(5): 1473-1483.

Grigg, G.C., L.E. Taplin, B. Green and P. Harlow 1986. Sodium and water fluxes in free-
living Crocodylusporosus in marine and brackish conditions. Physiological
Zoology 59(2): 240-253.

Hall, P.M. 1991. Estimation of nesting female crocodilian size from clutch
characterists-correlates of reproductive mode, and harvest implications. Journal
of Herpetology 25(2): 133-141.

Hall, P.M. and A.J. Meier. 1993. Reproductive behavior of Western mud snakes in
American alligator nests. Copeia 1: 219-222.

Hutton, J.M. 1986. Age-determination of living Nile crocodiles from the cortical
stratification of bone. Copeia(1): 332-341.

Jakob, E.M., J.D. Marshall and G.W. Uetz. 1996. Estimating fitness: a comparison of
body condition indices. Oikos 77: 61-67.

Jones, C., J. Lawton and M. Shachak 1994. Organisms as ecosystem engineers. Oikos
69(3): 373-386.

King, F.W. and P. Brazaitis. 1971. Species identification of commercial crocodile skins.
Zoologica. 56:15-70.









Kotiaho, J.S. 1999. Estimating fitness: comparison of body condition indices revisited.
Oikos 87(2): 399-402.

Krebs, C.J. and G.R. Singleton. 1993. Indexes of condition for small mammals.
Australian Journal of Zoology 41(4): 317-323.

Kushlan, J.A. 1974. Observations on the role of the American alligator (Alligator
mississippiensis) in the southern Florida wetlands. Copeia 1974: 993-996.

Kushlan, J.A. and T. Jacobsen 1990. Environmental variability and reproductive success
of Everglades alligators. Journal of Herpetology 24(2): 176-184.

Kushlan, J.A. and M.S. Kushlan. 1980. Everglades alligator nests: nesting sites for marsh
reptiles. Copeia. 4: 930-932.

LeCren, E.D. 1951. The length-weight relationship and seasonal cycle in gonad weight
and condition in the perch (Percafluviatilis). Journal of Animal Ecology 20(2):
201-219.

Leslie, A.J. 1997. The ecology and physiology of the Nile crocodile, Crocodylus
niloticus, in Lake St. Lucia, Kwazulu/Natal, South Africa. Ph.D. dissertation.
Drexel University, Philadelphia, PA.

Mazzotti, F.J. and L.A. Brandt. 1994. Ecology of the American alligator in a seasonally
fluctuating environment. Everglades: the ecosystem and its restoration. S.M.
Davis and J. C. Ogden eds. Delray Beach, FL, St. Lucie Press: 485-505.

Murphy, B.R., M.L. Brown and T.A. Springer. 1990. Evaluation of the relative weight
(Wr) index, with new applications to walleye. North American Journal of
Fisheries Management 10: 85-97.

Olmstead, I. and T.V. Armentano. 1997. Vegetation of Shark Slough, Everglades
National Park, SFNRC Technical Report 97-001: 43 pp.

Packard, G.C. and T.J. Boardman. 1988. The misuse of ratios, indexes, and percentages
in ecophysiological research. Physiological Zoology 61(1): 1-9.

Palmeirim, J.M. 1998. Analysis of skull measurements and measurers: Can we use data
obtained by various observers? Journal of Mammalogy 79(3): 1021-1028.

Quist, M.C., C.S. Guy and P.J. Braaten. 1998. Standard weight (Ws) equation and length
categories for Shovelnose sturgeon. North American Journal of Fisheries
Management 18: 992-997.

Rice, K.G., F.J. Mazzotti and C.L. Zweig. 2002. Annual report--compilation of alligator
datasets in South Florida for restoration needs, USGS.










Santos, S.A., M.J.S. Nogueria, M.S. Pinheiro, G.M. Mourao and Z.M.S. Campos. 1994.
Condition factor of Caiman crocodiles in different habitats of Pantanal Mato-
Grassense. Proceedings of the 12th Working Meeting of Crocodile Specialist
Group of the Species Survival Commission of the IUCN.

SAS Institute, Inc. 1988. SAS user's guide: Statistics. Version 6.03 edition. Cary, NC.

Schoeb, T.R., T.G. Heaton-Jones, R.M. Clemmons, D.A. Carbonneau, A.R. Woodward,
D. Shelton, and R.H. Poppenga. 2002. Clinical and necropsy findings associated
with increased mortality among American alligators of Lake Griffin, Florida.
Journal of Wildlife Diseases. 38(2): 320-327.

Springer, T.A. and B.R. Murphy. 1990. Properties of relative weight and other condition
indices. Transactions of the American Fisheries Society 119: 1048.

SPSS Inc. 2000. SPSS Base 11.0 for Windows user's guide. SPSS Inc. Chicago, IL.

Sutton, S.G., T.P. Bult and R.L. Haedrich. 2000. Relationships among fat weight, body
weight, water weight, and condition factors in wild Atlantic salmon parr.
Transactions of the American Fisheries Society 129(2): 527-538.

Taylor, J.A. 1979. The foods and feeding habits of subadult Crocodylusporosus
Schneider in Northern Australia. Australian Wildlife Research 5: 347-359.

Vila-Gispert, A. and R. Moreno-Amich. 2001. Fish condition analysis by a weighted least
squares procedure: testing geographical differences of an endangered Iberian
cyprinodontid. Journal of Fish Biology 58: 1658-1666.

Wege G.J. and R.O. Anderson. 1978. Relative weight (Wr): a new index of condition for
largemouth bass. New approaches to the management of small impoundments.
G.D. Novinger and J.G. Dillard, eds. American Fisheries Society, North Central
Division, Special Publication 5, Bethesda, MD.

Yezerinac, S.M., S.C. Lougheed and P. Handford. 1992. Measurement error and
morphometric studies--statistical power and observer experience. Systematic
Biology 41(4): 471-482.

Zink, R.M. 1983. Evolutionary and systematic significance of temporal variation in the
Fox sparrow. Systematic Zoology 32(3): 223-238.















BIOGRAPHICAL SKETCH

Christa Zweig was born in Lexington, Kentucky, on December 22, 1974. Her

family moved to Springfield, Missouri, in 1977 where she attended elementary through

high school. She originally wanted to be a brain surgeon, but soon decided that she did

not like people enough to deal with them on a day-to-day basis. That and her frequent

trips to Cumberland Island National Seashore combined to shape her current career path.

Christa received her undergraduate degree in biology from the University of Richmond,

Virginia. She then spent the next three years working temporary federal jobs for

Canaveral National Seashore, Padre Island National Seashore, and the Red Bluff U.S.

Fish and Wildlife Fisheries Office in northern California. In 2000, she began her work in

south Florida with the Florida Alligator Research Team. Her plans for the future are

vague and varied, but she would like to continue a career in the field of conservation.