Southeastern Naturalist 705 B.B. Rothermel and T.D. Castellón 22001144 SOUTHEASTERN NATURALIST 1V3o(4l.) :1730,5 N–7o2. 04 Factors Influencing Reproductive Output and Egg Size in a Southern Population of Gopher Tortoises Betsie B. Rothermel1,* and Traci D. Castellón1 Abstract - Comparative life-history data are needed to develop effective conservation plans for Gopherus polyphemus (Gopher Tortoise), a threatened species that inhabits diverse ecosystems throughout its range. In 2010–2011, we measured and radiographed 27 female Gopher Tortoises occupying Florida scrub and mesic flatwoods habitats at Avon Park Air Force Range in south-central Florida. Counter to predictions of optimal egg size theory, both clutch size and mean egg width (measured via x-rays) increased significantly with body size. Furthermore, our data suggest the presence of a non-pelvic constraint on egg size in this species. Despite greater cover of grasses and forbs in flatwoods, clutch size, egg width, and female body condition were similar in flatwoods and scrub. Thus, the relatively low density of juvenile-sized burrows in flatwoods is not a result of low fecundity. Body condition tended to be higher in the wetter spring of 2010, although seasonal differences were not statistically significant. Clutch sizes at Avon Park Air Force Range (range = 4–9 eggs; overall mean = 5.8 ± 1.2) were comparable to other populations, but lower than reported for some peninsular Florida populations. Further research is needed to explain variation in reproductive output among individuals and populations in the southern part of the species’ range. Introduction Gopherus polyphemus (Daudin) (Gopher Tortoise) has been described as both a keystone species (Eisenberg 1983) and an ecosystem engineer (Kinlaw and Grasmueck 2012) because of its burrowing habits and the importance of its burrows for diverse assemblages of invertebrates and other vertebrates in upland communities of the southeastern US (Jackson and Milstrey 1989, Witz et al. 1991). Unfortunately, populations of Gopher Tortoises have been declining for decades (Mushinsky et al. 2006) to the point where the species is now a candidate for federal protection throughout its range (USFWS 2011). In Florida, Gopher Tortoises occupy an impressive array of inland, coastal, and island ecosystems (Mushinsky et al. 2006). Across the species’ range, clutch size exhibits a latitudinal increase from north to south, whereas female body size exhibits a polynomial relationship with latitude, which may be attributed to differences in productivity and other environmental gradients (Ashton et al. 2007). Gopher Tortoises produce a single clutch of eggs annually (Germano 1994). Clutch sizes typically range from 1 to 15 eggs (Ashton et al. 2007, Demuth 2001, Moore et al. 2009), with a maximum clutch size of 25 eggs reported from large females relocated to reclaimed phosphate-mined land in central Florida (Mushinsky et al. 2006). The main nesting period is from mid-April 1Archbold Biological Station, Venus, FL 33960. *Corresponding author - brothermel@ archbold-station.org. Manuscript Editor: Will Selman Southeastern Naturalist B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 706 to mid-June throughout most of Florida (Mushinsky et al. 2006); however, there is evidence that mating and nesting occur in almost every month of the year in extreme southern populations (i.e., latitude < 27°N; Moore et al. 2009). As in many other chelonians (Gibbons and Greene 1990, Hailey and Loumbourdis 1988, Wallis et al. 1999), clutch sizes increase linearly with maternal size in most Gopher Tortoise populations that have been studied (e.g., Colson-Moon 2003, Diemer and Moore 1994, Iverson 1980, Landers et al. 1980, Rostal and Jones 2002, Smith 1995, Smith et al. 1997). However, Ashton et al. (2007) reported exceptions to this pattern. In the two southernmost mainland populations for which data were available, there was either no significant relationship (at Archbold Biological Station, Highlands County, FL) or a polynomial relationship between clutch size and maternal size (Okeeheelee County Park, Palm Beach County, FL), suggesting a need for data from additional populations in southern Florida to clarify effects of maternal size and age (Ashton et al. 2007). Comparative data on life-history traits from multiple populations in different habitats are needed for detailed demographic analyses and development of effective conservation strategies. According to general theory regarding optimal egg size (OES), once each egg has been optimally provisioned to ensure good offspring survivorship, females with access to more resources should allocate the additional energy to production of more offspring by producing larger clutches (Brockelman 1975, Smith and Fretwell 1974). Thus, in relatively stable environments, egg size is expected to vary less than clutch size across a range of female body sizes. In turtles, however, egg-size optimization may not be possible because of anatomical constraints, including the size of the pelvic canal through which eggs must pass (Congdon and Gibbons 1987). The evidence for pelvic aperture width as a constraint on egg size is inconsistent, even among small-bodied aquatic species (Macip-Ríos et al. 2013, Wilkinson and Gibbons 2005). Furthermore, in some species or populations, both clutch size and egg size increase with female body size (Naimi et al. 2012, van Loben Sels et al. 1997, Wilkinson and Gibbons 2005). Lovich et al. (2012) identified five potential patterns of egg-size variation in turtles resulting from interactions of morphological and non-morphological factors, but these relationships have never been thoroughly examined in any species of Gopherus. To investigate the relationship between egg size and morphological and nonmorphological factors, we acquired data on egg and clutch sizes for a previously unstudied population of Gopher Tortoises at Avon Park Air Force Range (APAFR), a military training installation in south-central Florida. While monitoring Gopher Tortoise habitat use and population characteristics at this site for a related study (Castellón and Rothermel, unpublished data), we captured and xrayed female Gopher Tortoises to determine gravidity and number of eggs. Using these data, the primary objectives of our study were to 1) examine the relationship of female body size and clutch size, 2) compare reproductive parameters between two habitats (scrub versus mesic flatwoods), and 3) determine if Gopher Tortoise egg width is constrained by the pelvic aperture or is optimized in accordance with OES predictions. Southeastern Naturalist 707 B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 Field-site Description The APAFR is a 42,900-ha US Air Force installation located in Polk and Highlands counties in south-central Florida (27°35'N, 81°16'W), which has a seasonal subtropical climate with pronounced wet and dry seasons (Slocum et al. 2010). Used for military training since World War II, the installation is centered on Bombing Range Ridge, a relict sand dune that is disjunct from the larger Lake Wales Ridge to the west (Branch and Hokit 2000). The fire-dependent natural habitats of APAFR support populations of many threatened and endangered species, including the Gopher Tortoise, Drymarchon couperi Holbrook (Eastern Indigo Snake), and Picoides borealis Vielliot (Red-cockaded Woodpecker). To manage natural resources and ensure sustainability of the Air Force’s air–ground training mission at APAFR, current management includes controlled burning of scrub habitat every 7–20 y and burning of flatwoods and Pinus (pine) plantations every 2–3 y (USAF 2000). At APAFR, the highest densities of Gopher Tortoises are found in Florida scrub habitat associated with the Bombing Range Ridge (Castellón et al. 2012). For the purposes of this study, scrub habitat refers to sand pine scrub, oak scrub, mixed scrub, and scrubby flatwoods vegetation communities, which together comprise approximately 6% (2470 ha) of APAFR. The fire-maintained Florida scrub community occurs in a naturally patchy distribution (Branch and Hokit 2000) on droughty, infertile soils, and is characterized by a dense shrub layer of Quercus spp. (oaks), ericaceous species, Sabal etonia Swingle ex Nash (Scrub Palmetto), and Serenoa repens W. Bartram (Saw Palmetto), with sparse groundcover and scattered Pinus clausa (Chapm. ex Engelm.) Vasey ex Sarg. (Sand Pine) or Pinus elliottii Engelm. (Slash Pine) (Abrahamson et al. 1984, Myers 1990). The remaining suitable habitat for Gopher Tortoises at APAFR consists primarily of mesic or dry-mesic flatwoods (7421 ha) or Slash Pine plantations on mesic and dry-mesic flatwoods soils (5442 ha; Castellón et al. 2012). Although the pine flatwoods and plantations at APAFR are burned frequently and have relatively open canopies and dense groundcover, they generally have much lower densities of Gopher Tortoises than the Bombing Range Ridge sites, presumably as a result of the poorly drained soils (Castellón et al. 2012). In mesic flatwoods, groundwater is close to the surface (usually ≤1.2 m deep) for most of the year (Abrahamson and Hartnett 1990), with standing water or sheet flow common following heavy rain. Most Gopher Tortoises measured for this study were captured in 1 of 2 sites within APAFR chosen for intensive population monitoring in 2009–2011. Mean percent cover of broadleaf grasses was 7.8% in the 33.3-ha oak scrub/scrubby flatwoods monitoring site, compared to 17.3% in the 69.6-ha mesic flatwoods site (T.D. Castellón and B.B. Rothermel, unpubl. data). The flatwoods site also had higher cover of Aristida spp. (wiregrass; 12.3%) and forbs (15.3%) than the scrub site (9.0% and 6.8%, respectively). Southeastern Naturalist B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 708 Methods Data collection Beginning in December 2009, we used pitfall traps to capture adult Gopher Tortoises at burrow entrances or captured them by hand when they were opportunistically encountered outside their burrows. We measured the following parameters to 1.0 mm with calipers: straight-line carapace length (CL), plastron length (PL; from anterior edge of gular to posterior tip of anal scute), maximum body width, and maximum shell height. After weighing each Gopher Tortoise to the nearest 0.05 kg, we marked it by filing unique combinations of notches into the marginal scutes. We attached radiotransmitters (model RI-2B; Holohil Systems, Ltd., ON, Canada) to a subset of adult females >230 mm in CL (n = 23), most of which were then tracked for at least a year (range = 3–20 months duration). We used the ratio of mass per unit volume (the product of CL, maximum body width, and maximum shell height) as an index of body condition because it allowed for direct comparison with the only published data on body condition of Gopher Tortoises in Florida (McCoy et al. 2011). Because body condition might be affected by weather conditions, we examined precipitation data for 2010–2011 compared to 30-year (1981–2010) seasonal averages recorded at two NOAA weather stations located within 35 km of APAFR (NCDC 2014). During the nesting season from mid-April through mid-June in 2010 and 2011, we transported captured females to a local veterinary hospital for radiography to determine clutch sizes (300 mA, 1/60 sec, and 82 kV peak; Gibbons and Greene 1979). If no eggs were detected on the first x-ray, we attempted to recapture the tortoise again in 2–3 weeks to allow more time for eggs to calcify and become detectable on x-ray films; in a few cases, we radiographed the same tortoise 3 times within a nesting season. Each animal was released within 24 hr at the point of capture. We calculated the mean diameter of each egg by averaging the minimum and maximum egg diameters as measured from x-ray films with dial calipers to the nearest 0.01 mm (Diemer and Moore 1994). The mean diameters of all eggs within a clutch were then averaged to obtain the mean x-ray egg width (EW) for each female Gopher Tortoise. We also measured the shortest distance between the ilia (Congdon and Gibbons 1987) to obtain the x-ray pelvic aperture width (PAW). The focus-to-film distance (Graham and Petokas 1989) was kept constant across x-rays to help control for magnification effects. Statistical analyses We used SPSS (PASW Statistics 17, IBM) for all statistical analyses and report means ± 1 SD. We used both polynomial and linear regression to examine associations between clutch size and female body size (n = 27). For linear regressions, we followed King’s (2000) recommendation to log-transform data to improve linearity, reduce heteroscedasticity of variances, and facilitate comparisons with other studies of reptiles. Six females were x-rayed and found to be gravid in both years. For 3 of these 6 individuals, the clutch size was the same in both years, so we simply used measurements from the first year. For the remaining 3, we used the observation Southeastern Naturalist 709 B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 that had the largest number of eggs, to represent the maximum reproductive output for a given size. We used analyses of covariance (ANCOVA) with PL as a covariate to test for effects of habitat on number of eggs and EW. As above, we log-transformed data prior to analysis. For the habitat analysis, we only used data from 25 individuals that were known to be residing in a particular habitat based on radiotelemetry data or bucket trapping at a burrow within that habitat. We excluded two Gopher Tortoises from habitat comparisons because they were captured by hand away from burrows on or near paved roads and could not be assigned with certainty to either scrub or flatwoods. We also examined differences in clutch size between years using ANCOVA with PL as the covariate. This comparison included 10 females x-rayed in 2010 and 17 x-rayed in 2011, excluding repeat measurements of individuals (n = 6) that were x-rayed both years. Prior to running the above ANCOVAs, we tested the assumption of homogeneity of regression slopes by examining models that included the interaction between the covariate and independent variable. In every case, the interaction was not significant (P > 0.20), indicating that this assumption was met and it was appropriate to proceed with the test of the main effect. Thus, we removed the interaction from the models and report F-values derived from the full additive models. To test whether regressions of PAW and EW on PL had similar slopes, we performed an ANCOVA using log-transformed data, with PL as the covariate and including the PL x treatment interaction (Congdon and Gibbons 1987, van Loben Sels et al. 1997). If egg width is constrained by the size of the pelvic aperture, the slopes of the regressions should be equal, whereas a significant interaction would indicate heterogeneity of slopes and lack of pelvic constraint. For the 6 gravid females that were x-rayed in both years, we used the highest EW from either year. We had to omit 1 tortoise (PL = 283 mm) because the x-ray image was too poor to accurately measure PAW, leaving a sample size of 26 for this analysis. To examine the potential trade-off between egg size and egg number, as expected when resources available for reproduction are limited (Smith and Fretwell 1974), we used multiple regression to test the independent effects of log-transformed PL and clutch size on EW (Wallis et al. 1999). We compared body condition between habitats using data for all female Gopher Tortoises assignable to either scrub or flatwoods habitat, regardless of whether they were subsequently recaptured or x-rayed. However, we had to omit data for 1 tortoise that was an extreme low outlier, making us question the accuracy of the recorded measurements; the data were normally distributed once we excluded this 2011 observation. We disregarded whether the tortoise urinated or defecated prior to weighing, because results of a preliminary analysis indicated that females that excreted prior to weighing did not have significantly lower condition indices (t-test, P > 0.33). To avoid pseudoreplication because some females were measured in both years, we ran a separate ANOVA for each year. For 2010 (n = 19), we used a two-way ANOVA to test the effects of habitat and season (March–June versus July–October) and their interaction on body condition. Southeastern Naturalist B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 710 Because we only had measurements of Gopher Tortoises from a single season in 2011 (March–June; n = 18), we used a one-way ANOVA to test the effect of habitat on 2011 body condition. We pooled data across habitats and used a t-test to compare female body condition in March–June (the drier part of the activity season) between years. The timing of body size measurements relative to the nesting season was too variable among individuals to examine possible relationships between body condition and clutch or egg sizes. Results The proportion of adult females found to be gravid was 0.769 in 2010 (n = 13) and 0.885 in 2011 (n = 26); this result was similar in scrub (0.882) and flatwoods (0.800) habitats. Because most of the females lacking eggs were x-rayed only one time, it is possible that the x-rays occurred too early or too late relative to egg shelling or oviposition, respectively. Therefore, the values we calculated should be considered minimum estimates of the proportion of females that reproduced in a given year. The 6 females that were recaptured and x-rayed in 2010 and 2011 were gravid in both years. The earliest and latest dates calcified eggs were detected on x-rays were 19 April and 2 June, respectively. The smallest gravid female was 254 mm CL (Table 1). A polynomial model using either CL (r2 = 0.187, P = 0.084) or PL (r2 = 0.204, P = 0.065) to explain clutch size was not significant. Following log-transformation, linear regressions indicated that clutch size was positively associated with female body size (Fig. 1). More of the variation in clutch size was explained by PL (r2 = 0.211, P = 0.016) than by CL (r2 = 0.181, P = 0.027). Mean clutch sizes were 6.2 eggs (range = 4–9) in scrub habitat and 5.7 eggs (range = 4–7) in mesic flatwoods (Table 1). Clutch size was significantly related to PL (F1,22 = 4.687, P = 0.042), but it did not differ between habitats (F1,22 = 1.447, P = 0.242). After controlling for PL (F1,24 = 6.666, P = 0.016), clutch size also did not differ between years (F1,24 = 0.181, P = 0.675). Table 1. Carapace length (CL), plastron length (PL), and clutch size of female Gopher Tortoises in scrub and flatwoods habitats and both habitats combined at Avon Park Air Force Range, FL, 2010– 2011. The overall sample includes two tortoises that could not be assigned to a particular habitat because they were captured away from burrows on roads. Errors for means are ± 1 SD. Variable Scrub (n = 13) Flatwoods (n = 12) Overall (n = 27) CL (mm) Mean 280.5 ± 15.3 280.1 ± 14.0 278.4 ± 15.3 Range 254–307 260–301 254–307 PL (mm) Mean 274.5 ± 14.9 278.0 ± 15.9 274.7 ± 15.5 Range 253–297 251–299 251–299 # of eggs Mean 6.2 ± 1.4 5.7 ± 0.9 5.8 ± 1.2 Range 4–9 4–7 4–9 Southeastern Naturalist 711 B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 Figure 1. Relationships between clutch size versus carapace length (top) and plastron length (bottom) for Gopherus polyphemus in different habitats at Avon Park Air Force Range, FL, including scrub (●), flatwoods (○), and other (Δ; tortoises capt ured on or near roads). Southeastern Naturalist B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 712 Although EW did not differ significantly between habitats (F1,22 = 0.348, P = 0.561), egg size increased significantly with increases in PL (F1,22 = 15.014, P = 0.001; Fig. 2). PAW also increased as female body size increased (Fig. 2). Although the regression slope of PAW was steeper than that of EW, the slopes were not significantly different according to ANCOVA (PL x Treatment: F1,48 = 2.543, P = 0.117). Figure 2. Relationship between plastron length (PL) and mean egg width (EW; ●) of each clutch and pelvic aperture width (PAW; ○) in Gopherus polyphemus (n = 26) from Avon Park Air Force Range, FL. EW and PAW were measured from x-ray images. Log EW = 0.492(log PL) + 0.501 (r2 = 0.418, P < 0.0001). Log PAW = 0.953(log PL) - 0.511 (r2 = 0.353, P = 0.001). Southeastern Naturalist 713 B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 After removing the significant effect of body size by multiple regression, the slope of clutch size was negative, but clutch size was not a si gnificant predictor of egg size (P = 0.154; log EW = 0.297 + 0.593[log PL] – 0.055[log number of eggs]; r2 = 0.472). Other considerations for evaluating egg-size optimization are the actual clearance for egg passage through the pelvic canal, as well as the coefficients of variation (CV) of egg size relative to clutch size (Lovich et al. 2012). Only 1 Gopher Tortoise (PL = 255 mm) had an aperture width smaller than the largest egg in our entire sample (PAW = 54.3 mm versus maximum EW = 56.2 mm). The mean clearance, i.e., the difference between the maximum EW within a clutch and the corresponding PAW of the female, was 13.6 mm (range = 4.9–23.4 mm). The only result clearly consistent with OES theory was that variation in clutch size among females was much greater than variation in EW per clutch (CV for log-transformed data = 0.114 and 0.011, respectively). Mean body condition ranged from 0.557 to 0.593 for females residing in scrub habitat, and from 0.574 to 0.588 in flatwoods, across the 3 seasons encompassed by this study (Fig. 3). In 2010, body condition did not differ between habitats (F1,15 = 0.609, P = 0.447) or between seasons (F1,15 = 0.148, P = 0.706), and there was not a significant interaction (F1,15 = 0.514, P = 0.485). In 2011, there was also no difference in body condition between habitats (F1,16 = 0.926, P = 0.350). In both habitats, female body condition was slightly higher in March–June of 2010 than March–June of 2011, but the difference was not significant (t26 = 1.930, P = 0.065; Fig. 3). Figure 3. Mean (± 1 SD) body condition index of female Gopherus polyphemus in scrub and flatwoods habitats at Avon Park Air Force Range, FL, 2010–2011. Although March–May is the late dry season, El Niño conditions resulted in higher-than-normal rainfall in winter (December–February) and spring (March–May) of 2010 (see text). Southeastern Naturalist B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 714 Discussion An extensive survey of the Gopher Tortoise population at APAFR revealed a much smaller proportion of juvenile-sized burrows in flatwoods compared to scrub habitat (Castellón et al. 2012). Although this difference in juvenile density could result from differences in female fecundity, we expected that the lower availability of food plants in more xeric scrub habitat would restrict reproductive output. During our population survey of suitable Gopher Tortoise habitats throughout APAFR, we sampled vegetation and found significantly greater cover of food plants (i.e., grasses and forbs; Garner and Landers 1981, Mushinsky et al. 2006) in flatwoods compared to scrub habitats (Castellón et al. 2012). Despite these differences in ground-level herbaceous food resources, neither clutch size nor egg size differed significantly between habitats when we controlled for female body size. Our results indicate that female Gopher Tortoises in scrub habitats at APAFR gain access to sufficient food resources to produce clutch sizes comparable to those reported for populations in north-central Florida (mean = 5.8; Diemer and Moore 1994, Smith 1995) and a population on the adjacent Lake Wales Ridge, approximately 50 km south of APAFR (mean = 6.5; Ashton et al. 2007). From radiotelemetry monitoring, we know that 6 female tortoises in scrub and 4 females in flatwoods occasionally used burrows in habitats adjacent to their primary habitats. These movements by a subset of females may have contributed to the lack of a habitat effect on clutch and egg sizes, and suggest a need for detailed investigation of complex habitat-use patterns in these naturally patchy settings. A potential explanation for the lower density of juvenile tortoises in flatwoods is higher mortality of egg and hatchling stages. A likely source of nest failure in flatwoods is inundation of nests due to rising water tables with the onset of the wet season in May and subsequent flooding of large areas during major storms (Castellón et al. 2012). Eggs and hatchlings may also suffer predation by Solenopsis invicta Buren (Red Imported Fire Ant; Epperson and Heise 2003). Red Imported Fire Ant is adapted to flooding and may reach higher abundances in more mesic habitats (Deyrup et al. 2000, Tschinkel 1988). Anecdotally, the only 3 nests we found in our mesic flatwoods site were inundated during a high-rainfall event, and one was also disturbed and depredated, most likely by a mammal. The rigid shell of turtles imposes an upper limit on clutch sizes; thus, increases in resource availability do not necessarily translate into greater reproductive output beyond that associated with achieving larger body size (Gibbons and Greene 1990). This limit may partly explain similarities in clutch size between habitats at APAFR and in other studies. For example, Smith et al. (1997) found no difference in mean clutch size of Gopher Tortoises inhabiting 2 sites in the western part of the species’ range that differed in vegetation structure and management history. Likewise, Diemer and Moore (1994) found no difference in mean clutch size of Gopher Tortoises inhabiting 3 contrasting sites in north-central Florida. It is worth noting, however, that clutch sizes at APAFR are smaller than those reported for other populations studied at similar latitudes but divergent habitats, such as coastal strand (mean = 7.46 eggs; Demuth 2001) and sandhill (mean = 7.29 eggs; Colson-Moon 2003). Southeastern Naturalist 715 B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 Reduced clutch sizes and egg masses in a southeast Georgia population were attributed to poor habitat quality resulting from fire suppression (Rostal and Jones 2002). Thus, studies controlling for body-size effects are still needed to examine whether more extreme differences in productivity or food quality among vegetation communities, perhaps related to habitat management, also contribute to amongpopulation variation in clutch sizes. Because Gopher Tortoises skip reproduction in some years, clutch frequency is another potential source of variation in fecundity. Smith et al. (1997) estimated that 80% to 85% of female Gopher Tortoises in their study populations in Mississippi and Louisiana reproduce in a given year. Although all 6 of the APAFR females x-rayed in 2010 and again in 2011 were gravid in both years, a longer study encompassing more environmental variability is needed to adequately assess variation in clutch size and frequency within individuals and at the population level (Gibbons and Greene 1990). Nevertheless, our results from repeated x-rays combined with the high percentage of females that were gravid (77% in 2010 and 88% in 2011) suggest high reproductive frequency in the APAFR population. By comparison, the mean annual percentage of females gravid in 3 populations studied by Diemer and Moore (1994) was 73% (range = 40%–89%). As expected based on studies of other turtles and most other populations of Gopher Tortoises that have been examined, we found a significant positive linear relationship between body size and clutch size. However, body size explained only 21% of the variation in clutch size at APAFR. Similar significant but weak relationships between maternal body size and clutch size appear common in Gopher Tortoises (e.g., Diemer and Moore 1994, Iverson 1980, Smith 1995, Smith et al. 1997) as well as other species of Gopherus (Hellgren et al. 2000, Wallis et al. 1999). Although short-term seasonal variation in body condition of Gopher Tortoises appears to be slight, extended periods of low rainfall can lead to reduced body condition (McCoy et al. 2011), with potential consequences for reproduction. Smith (1995), for example, found reduced clutch sizes during an extended drought in north-central Florida. Both years of our study had belownormal rainfall (by approximately 11 cm) during autumn when vitellogenesis starts (Iverson 1980). However, El Niño conditions in 2010 resulted in rainfall exceeding 30-year averages for winter and spring by 11 cm and 16 cm, respectively (NCDC 2014). Drier conditions in autumn 2010 persisted through the winter, but rainfall in March–May 2011 exceeded the long-term average by 7 cm (NCDC 2014). Thus, drought conditions did not prevail in either year and clutch size did not differ between years, despite a trend toward higher body condition in the wetter spring of 2010 (Fig. 3). The mean body condition of females was similar in scrub (0.557–0.593) and flatwoods (0.574–0.588) and if anything, slightly higher than observed in other central Florida populations (mean = 0.537–0.561; McCoy et al. 2011). Long-term studies examining individual variation in clutch sizes of Gopher Tortoises are needed to account for the potentially important effects of interannual variation in resource availability and maternal body condition. Southeastern Naturalist B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 716 The extreme southern populations of Gopher Tortoises in Palm Beach County are noteworthy for their unusually large sizes (Ashton et al. 2007, Moore et al. 2009). The population in Palm Beach Gardens studied by Ashton et al. (2007) also exhibits an anomalous polynomial relationship between body size and clutch size. Interestingly, the 2 largest females x-rayed at APAFR measured 320 mm and 311 mm CL but were not gravid; the mean CL of non-gravid adult females (292 mm, n = 6) was slightly larger than that of gravid females (mean 278 mm, n = 27). However, we hesitate to draw any conclusions without additional observations of larger and potentially older females from this population. As in other studies of Gopher Tortoises to date, the age of tortoises in our sample is unknown and the question of reproductive senescence (Congdon et al. 2001) awaits empirical data on reproductive output versus age. Results of this first analysis of egg-size optimization in Gopher Tortoises are mostly inconsistent with OES predictions. When reproduction is constrained primarily by resource availability, OES theory predicts a trade-off between number and size of offspring, with greater variation in clutch size than in egg size (Congdon and Gibbons 1990, Smith and Fretwell 1974). In Gopherus agassizii (Cooper) (Desert Tortoise; Wallis et al. 1999) and in the population of Gopher Tortoises we studied, clutch size varied more than egg size. In contrast to Desert Tortoises, however, larger female Gopher Tortoises at APAFR produced both larger clutches and larger eggs, and we found no evidence of a trade-off between clutch size and egg size after accounting for differences in body size. Similarly, Colson-Moon (2003) found that both number of eggs and mean egg mass per clutch increased significantly with body size of Gopher Tortoises at another site in south-central Florida. A similar finding for Testudo hermanni Gmelin (Hermann’s Tortoise) led Bertolero et al. (2007) to suggest there must be strong selection to increase both egg number and offspring size, if, for example, larger offspring have greater survivorship. Although hatchling size is positively correlated with egg mass in Gopher Tortoises (Rostal and Jones 2002), the effect of body size on hatchling survivorship is largely unknown. Pike and Seigel (2006) found no effect of body size on hatchling longevity at their study site in Florida, where predation by mammals was the major cause of hatchling mortality. Because Gopher Tortoises have brittle-shelled eggs, the potential exists for the pelvic aperture to constrain egg size. We assume the significant, positive relationship between EW and PL was not an artifact of increased magnification on x-rays. In some aquatic turtles, egg height above the x-ray film (i.e., object-to-film distance) is greater in larger females (Graham and Petokas 1989), however, Wallis et al. (1999) did not find a significant relationship between body size and object-tofilm distance in G. agassizii. Based on ANCOVA, the regression slopes of EW and PAW were similar for APAFR Gopher Tortoises (Fig. 2), as expected if there is a pelvic constraint on egg size. However, in every female in our sample, the largest egg diameter was smaller than the pelvic aperture, by an average of 14 mm. Thus, the significant relationship between EW and PL suggests presence of a different body size-related constraint. The posterior space between the plastron and carapace Southeastern Naturalist 717 B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 (i.e., the caudal gap) is a constraint on egg size in some chelonians (Clark et al. 2001). In G. berlandieri (Texas Tortoise), Rose and Judd (1991) found that the most posterior plastral and carapacial bones are kinetic and can move to allow egg passage, but this has not been examined in Gopher Tortoise. Future evaluations of egg-size optimization in this species might also benefit from inclusion of smaller mature females (~230–250 mm CL) in the sample. In summary, our results for Gopher Tortoises conform with a previously reported pattern of egg-size variation, in which egg width is unconstrained by the pelvic aperture, but is still not optimized (Lovich et al. 2012). A parsimonious explanation is that the fitness advantages of larger eggs and larger clutches are nearly equal, so females simply divide extra resources between producing both more and larger eggs as body size allows (Naimi et al. 2012). Clarifying within-population variation in clutch sizes and observed differences in mean clutch sizes among southern populations of Gopher Tortoises will require more research on the relative roles of fluctuating environmental conditions and food resources on female body condition and reproductive parameters. Acknowledgments We thank research assistants J. Lopez, Z. Forsburg, D. Rankin, L. Rankin, A. Johnson, J. Ross, T. Demers, and K. Powers, as well as volunteers R. Percino-Daniel, K. Foley, A. Harrar, and G. Kamener, for their invaluable assistance in the field. We are also grateful to L. Cedola and other staff of Lake Forest Veterinary Clinic in Avon Park, S. Caster and R. Tucker for measuring eggs on x-rays, and J. Lovich for tips on x-ray measurements. M. Fredlake and other APAFR personnel provided valuable guidance and support throughout this study, and the suggestions of two reviewers and guest editor W. Selman greatly improved the manuscript. Funding was provided by the US Air Force under Cooperative Agreement W81XWH-06-2-0026, and research was conducted under Florida scientific collecting permit LSSC10-00043. Literature Cited Abrahamson, W.G., and D.C. Hartnett. 1990. Pine flatwoods and dry prairies. Pp. 103–149, In R.L. Myers and J.J. Ewel (Eds.). Ecosystems of Florida. University of Central Florida Press, Orlando, FL. 765 pp. Abrahamson, W.G., A.F. Johnson, J.N. Layne, and P.A. Peroni. 1984. Vegetation of the Archbold Biological Station, Florida: An example of the southern Lake Wales Ridge. Florida Scientist 47:209–250 Ashton, K.G., R.L. Burke, and J.N. Layne. 2007. Geographic variation in body and clutch size of Gopher Tortoises. Copeia 2007(2):355–363. Bertolero, A., J.P. Nougarède, M. Cheylan, A. Marín. 2007. Breeding traits of Hermann’s tortoise, Testudo hermanni hermanni, in two western populations. Amphibia–Reptilia 28:77–85. Branch, L.C., and D.G. Hokit. 2000. A comparison of scrub herpetofauna on two central Florida sand ridges. Florida Scientist 63:108–117. Brockelman, W.Y. 1975. Competition, the fitness of offspring, and optimal clutch size. The American Naturalist 109:677–699. Southeastern Naturalist B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 718 Castellón, T.D., B.B. Rothermel, and S.Z. Nomani. 2012. Gopher Tortoise (Gopherus polyphemus) burrow densities in scrub and flatwoods habitats of peninsular Florida. Chelonian Conservation and Biology 11:153–161. Clark, P.J., M.A. Ewert, and C.E. Nelson. 2001. Physical apertures as constraints on egg size and shape in the Common Musk Turtle, Sternotherus odoratus. Functional Ecology 15:70–77. Colson-Moon, J.C. 2003. Reproductive characteristics, multiple paternity, and mating system in a central Florida population of the Gopher Tortoise, Gopherus polyphemus. M.Sc. Thesis. University of South Florida, Tampa, FL. 61 pp. Congdon, J.D., and J.W. Gibbons. 1987. Morphological constraint on egg size: A challenge to optimal egg size theory? Proceedings of the National Academy of Sciences USA 84:4145–4147. Congdon, J.D., and J.W. Gibbons. 1990. Turtle eggs: Their ecology and evolution. Pp. 109–123, In J.W. Gibbons (Ed.). Life History and Ecology of the Slider Turtle, Smithsonian Institution Press, Washington, DC. 368 pp. Congdon, J.D., R.D. Nagle, O.M. Kinney, and R.C. van Loben Sels. 2001. Hypotheses of aging in a long-lived vertebrate, Blanding’s Turtle (Emydoidea blandingii). Experimental Gerontology 36:813–827. Demuth, J.P. 2001. The effects of constant and fluctuating incubation temperatures on sex determination, growth, and performance in the tortoise Gopherus polyphemus. Canadian Journal of Zoology 79:1609–1620. Deyrup, M., L. Davis, and S. Cover. 2000. Exotic ants in Florida. Transactions of the American Entomological Society 126:293–326. Diemer, J.E., and C.T. Moore. 1994. Reproduction of Gopher Tortoises in north-central Florida. Pp. 129–137, In R.B. Bury and D.J. Germano (Eds.). Biology of North American Tortoises. Fish and Wildlife Research 13 US Department of the Interior, National Biological Survey, Washington, DC. 204 pp. Eisenberg, J.F. 1983. The Gopher Tortoise as a keystone species. Pp. 1–4, In R.J. Bryant and R. Franz (Eds.). The Gopher Tortoise: A Keystone Species. Proceedings of the 4th Annual Meeting of the Gopher Tortoise Council, Valdosta, GA. 46 pp. Epperson, D.M., and C.D. Heise. 2003. Nesting and hatchling ecology of Gopher Tortoises (Gopherus polyphemus) in southern Mississippi. Journal of Herpetology 37:315–324. Garner, J.A., and J.L. Landers. 1981. Foods and habitat of the Gopher Tortoise in southwestern Georgia. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 35:120–134. Germano, D.J. 1994. Comparative life histories of North American tortoises. Pp. 175–185, In R.B. Bury and D.J. Germano (Eds.). Biology of North American Tortoises. Fish and Wildlife Research 13 US Department of the Interior, National Biological Survey, Washington, DC. 204 pp. Gibbons, J.W., and J.L. Greene. 1979. X-ray photography: A technique to determine reproductive patterns of freshwater turtles. Herpetologica 35:86–89. Gibbons, J.W., and J.L. Greene. 1990. Reproduction in the Slider and other species of turtles. Pp. 124–134, In J.W. Gibbons (Ed.). Life History and Ecology of the Slider Turtle. Smithsonian Institution Press, Washington, DC. 384 pp. Graham, T.E., and P.J. Petokas. 1989. Correcting for magnification when taking measurements directly from radiographs. Herpetological Review 20:46–47. Hailey, A., and N.S. Loumbourdis. 1988. Egg size and shape, clutch dynamics, and reproductive effort in European tortoises. Canadian Journal of Zoology 66:1527–1536. Southeastern Naturalist 719 B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 Hellgren, E.C., R.T. Kazmaier, D.C. Ruthven III, and D.R. Synatzske. 2000. Variation in tortoise life history: Demography of Gopherus berlandieri. Ecology 81:1297–1310. Iverson, J.B. 1980. The reproductive biology of Gopherus polyphemus. American Midland Naturalist 103:353–359. Jackson, D.R., and E.G. Milstrey. 1989. The fauna of Gopher Tortoise burrows. Pp. 86–98, In J.E. Diemer, D.R. Jackson, J.L. Landers, J.N. Layne, and D.A. Wood (Eds.). Proceedings of the Gopher Tortoise Relocation Symposium, Florida Game and Fresh Water Fish Commission, Nongame Wildlife Program Technical Report No. 5. 109 pp. King, R.B. 2000. Analyzing the relationship between clutch size and female body size in reptiles. Journal of Herpetology 34:148–150. Kinlaw, A., and M. Grasmueck. 2012. Evidence for and geomorphologic consequences of a reptilian ecosystem engineer: The burrowing cascade initiated by the Gopher Tortoise. Geomorphology 157–158:108–121. Landers, J.L., J.A. Garner, and W.A. McRae. 1980. Reproduction of the Gopher Tortoise (Gopherus polyphemus). American Midland Naturalist 103:353–359. Lovich, J.E., S.V. Madrak, C.A. Drost, A.J. Monatesti, D. Casper, and M. Znari. 2012. Optimal egg size in a suboptimal environment: Reproductive ecology of female Sonora Mud Turtles (Kinosternon sonoriense) in central Arizona, USA. Amphibia–Reptilia 33:161–170. Macip-Ríos, R., V.H. Sustaita-Rodriguez, and G. Casas-Andreu. 2013. Evidence of pelvic and nonpelvic constraint on egg size in two species of Kinosternon from Mexico. Chelonian Conservation and Biology 12:218–226. McCoy, E.D., R.D. Moore, H.R. Mushinsky, and S.C. Popa. 2011. Effects of rainfall and the potential influence of climate change on two congeneric tortoise species. Chelonian Conservation and Biology 10:34–41. Moore, J.A., M. Strattan, and V. Szabo. 2009. Evidence for year-round reproduction in the Gopher Tortoise (Gopherus polyphemus) in southeastern Florida. Bulletin of the Peabody Museum of Natural History 50:387–392. Mushinsky, H.R., E.D. McCoy, J.E. Berish, R.E. Ashton, Jr., and D.S. Wilson. 2006. Gopherus polyphemus: Gopher Tortoise. Pp. 350–375, In P.A. Meylan (Ed.). Biology and Conservation of Florida Turtles. Chelonian Research Monographs 3. 376 pp. Myers, R.L. 1990. Scrub and high pine. Pp. 150–193, In R.L. Myers and J.J. Ewel (Eds.). Ecosystems of Florida. University of Central Florida Press, Orlando, FL. 765 pp. Naimi, M., M. Znari, J.E. Lovich, Y. Feddadi, and M.A. Ait Baamrane. 2012. Clutch and egg allometry of the turtle Mauremys leprosa (Chelonia: Geoemydidae) from a polluted peri-urban river in west-central Morocco. Herpetological Journal 22:43–49. National Climatic Data Center (NCDC). 2014. Climate data online. Available online at http://www.ncdc.noaa.gov/cdo-web/. Accessed 29 January 2014. Pike, D.A., and R.A. Seigel. 2006. Variation in hatchling tortoise survivorship at three geographic localities. Herpetologica 62:125–131. Rose, F.L., and F.W. Judd. 1991. Egg size versus carapace-xiphiplastron aperture size in Gopherus berlandieri. Journal of Herpetology 25:248–250. Rostal, D.C., and D.N. Jones, Jr. 2002. Population biology of the Gopher Tortoise (Gopherus polyphemus) in southeast Georgia. Chelonian Conservation and Biology 4:479–487. Slocum, M.G., W.J. Platt, B. Beckage, S.L. Orzell, and W. Taylor. 2010. Accurate quantification of seasonal rainfall and associated climate–wildfire relationships. Journal of Applied Meteorology and Climatology 49:2559–2573. Smith, C.C., and S.D. Fretwell. 1974. The optimal balance between size and number of offspring. The American Naturalist 108:499–506. Southeastern Naturalist B.B. Rothermel and T.D. Castellón 2014 Vol. 13, No. 4 720 Smith, K.R., J.A. Hurley, and R.A. Seigel. 1997. Reproductive biology and demography of Gopher Tortoises (Gopherus polyphemus) from the western portion of their range. Chelonian Conservation and Biology 2:596–600. Smith, L.L. 1995. Nesting ecology, female home range and activity, and population sizeclass structure of the Gopher Tortoise, Gopherus polyphemus, on the Katharine Ordway Preserve, Putnam County, Florida. Bulletin of the Florida Museum of Natural History 37, Pt. I(4):97–126. Tschinkel, W.R. 1988. Distribution of the fire ants Solenopsis invicta and S. geminata (Hymenoptera: Formicidae) in northern Florida in relation to habitat and disturbance. Annals of the Entomological Society of America 81:76–81. United States Air Force (USAF). 2000. Plan for Management of the Florida Grasshopper Sparrow, Florida Scrub-jay, and Red-cockaded Woodpecker at Avon Park Air Force Range, Florida. Avon Park Air Force Range Miscellaneous Publication. Avon Park, FL. 136 pp. United States Fish and Wildlife Service (USFWS). 2011. 12-month finding on a petition to list the Gopher Tortoise (Gopherus polyphemus) as threatened in the eastern portion of its range. Federal Register 76:45-130–45-162. van Loben Sels, R.C., J.D. Congdon, and J.T. Austin. 1997. Life history and ecology of the Sonoran Mud Turtle (Kinosternon sonoriense) in southeastern Arizona. Chelonian Conservation and Biology 2:338–344. Wallis, I.R., B.T. Henen, and K.A. Nagy. 1999. Egg size and annual egg production by female Desert Tortoises (Gopherus agassizii): The importance of food abundance, body size, and date of egg shelling. Journal of Herpetology 33:394–408. Wilkinson, L.R., and J.W. Gibbons. 2005. Patterns of reproductive allocation: Clutch and egg size variation in three freshwater turtles. Copeia 2005:868–879. Witz, B.W., D.S. Wilson, and M.D. Palmer. 1991. Distribution of Gopherus polyphemus and its vertebrate symbionts in three burrow categories. American Midland Naturalist 126:152–158.