BULLETIN of the FLORIDA STATE MUSEUM Biological Sciences Volume 27 1982 Number 2 GROWTH AND MATURITY OF THE GOPHER TORTOISE IN SOUTHWESTERN GEORGIA J. LARRY LANDERS W. ALAN MCRAE AND JAMES A. GARNER , 't ' 1 , 4 502eB, - $ Iff & 9 -9/1. & 1' 4 ' .=,9 $3 /1 . 50 ...... I ./*jill ./ . I 's**s'... ....R 4 ' I %*» 0 1 8 % 1 .. 1 . 4&2%. UNIVERSITY OF FLORIDA GAINESVILLE Numbers·of the BULLETIN OF THE FLORIDA STATE MUSEUM, BIOLOGICAL SCIENCES, are published.at irregular intervals. Volumes contain about 300 pages and are not necessarily completed in any one calendar year. OLIVER L. AUSTIN, JR. , Editor RHODA J·BRYANT, Managing Editor Consultants for this issue: JOHN F. DOUGLASS JOHN B. IVERSON Communications concerning purchase or exchange of the publications and all manuscripts should be addressed to: Managing Editor, Bulletin; Florida State Museum; University of Florida; Gainesville, Florida 32611. Copyright © 1982 by The Florida State Museum of the University of Florida This public document was promulgated at an annual cost of $1,990.00 or $1.99 per copy. It makes available to libraries, scholars and all inter- ested persons the results of researches in the natural sciences, emphasiz- ing the circum-Caribbean region. Publication date: January 5, 1982 Price $2.00 GROWTH AND MATURITY OF THE GOPHER TORTOISE IN SOUTHWESTERN GEORGIA J. LARRY LANDERS, W. ALAN MCRAE, AND JAMES A. GARNERi SYNOPSIS: Growth rates of gopher tortoises (Gopherus polyphemus) were determined by mea- suring individuals during two consecutive years and by correlating plastral annuli counts and measurements with shell dimensions. Ninety-three percent of growth occurred from April- October. The annual increment in carapace length (CL) for immatures was 11-12 nim; this varied by CL class. Growth increments were greatest for specimens 100-190 mm CL (about 4-11 yr of age), and gradually decreased thereafter. However, body volume increments peaked as tortoises grew from 220-230 mm CL, shortly before attainment of somatic ma- turity. S6xual maturity typically occurred at age 16-18 years (230-240 mm CL) in males and at 19-21 years (250-265 mm CL) in females. Thereafter, growth in both sexes was reduced, but females grew at greater rates and became significantly larger. Development of tortoises living in sand ridge habitats in southwestern Georgia was much slower than in Florida: ma- turity was delayed about 7 years. Growth was influenced by habitat quality, ambient tem- perature, and length of the annual activity season. The extended period of time required for attainment of maturity is potentially a limiting factor in population growth, and becomes more important in the northern portions of the species' range. TABLE OF CONTENTS INTRODUCTION .......... 82 ACKNOWLEDGEMENTS . .... 82 METHODS ... .- 83 Study Area . .. 83 Data Collection and Analysis . 83 SEASONAL GROWTH CYCLE ... 85 ANNUAL GROWTH RATE .. 87 Carapacial Increment . .. .. 87 Allometric Relations 92 Volumetric Increment 97 SIZE IN RELATION TO LIFE HISTORY . 98 Immature Stages . .. ... 98 Size at Sexual Maturity . . 101 Age at Sexual Maturity .. .. 102 Post Maturity .... ... 103 SURVIVAL AND ADAPTIVE IMPLICATIONS . 104 APPENDIX . .. .. 108 Synopsis of Methods Used to Estimate Growth . . 108 LITERATURE CITED . 109 'The authors are Wildlife Research Biologists for International Paper Company at Southlands Experiment Forest, Bain- bridge, Georgia 31717. LANDERS, J. LARRY, W, ALAN MCRAE, AND JAMES A. GARNER. 1982. Growth and maturity of the gopher tortoise in 56uthwestern Georgia. Bull. Florida State Mus., Biol. Sci. 27(2):81-110. 82 BULLETIN FLORIDA,STATE MUSEUM Vol. 27, No. 2 INTRODUCTION The gopher tortoise is one of four extant species of the genus Gopherus and the only tortoise now indigenous to the southeastern United States. In recent years this species has drawn much attention, due both to the eco- logical significance of its burrows (Landers and Speake 1980) and its ques- tionable population status (Auffenburg and Franz 1975). Gopher tortoise populations have declined steadily over much of the range, due largely to habitat alteration and human predation (Franz and Auffenberg 1978). Few data exist on factors that influence the capacity of depleted popula- tions to return to their former levels. A major aspect of population dy- namics is the growth rate of individuals, especially in relation to sexual maturity. In reference to terrestrial turtles, Auffenberg and Iverson (1979) stated that the span of time between the reproductive period and prereproductive period is most important: it determines the amount of energy required and the length of time an individual is exposed to imping- ing factors (e.g., predation and disease) before entering the breeding population. Unfortunately, no growth data are available on known-age gopher tortoises living in the wild. Growth has been estimated in only two areas. In central Florida, Coin and Goff (1941) determined annual growth in- crements of 33 specimens captured at approximately one-year intervals. In north-central Florida, Auffenberg and Iverson (1979) correlated plas- tron lengths of 168 gopher tortoises with ages estimated from shield annuli counts. Iverson (1980) utilized these two data sets to predict the age at sex- ual maturity of necropsied females collected from north-central Florida. No comparable data have been ,contributed from other portions of the range, and environmental factors that influence growth have not been determined. This study includes samples from southwestern Georgia and north- central Florida and employs several methods commonly used for growth analysis of turtles. Major objectives were: (1) to determine annual growth increments in relation to age and size, (2) to estimate age at sexual ma- turity, and (3) to determine the influence of selected environmental fae- tors on growth. Information on geographic variations in growth parame- ters and comparisons with certain other turtle species are also included. ACKNOWLEDGMENTS This work is part of a study on gopher tortoise ecology sponsored by International Paper Company. It was partially financed witb grant-in-aid funds under Section 6 of the Endan- gered Species Act of 1973 (PL 93-205) administered through the Georgia Department of Natural Resources (Project No. 8-933). W. Auffenberg and R. Franz provided specimens for study from the Florida State Museum and contributed additional data from Marion County, Florida. Without their assistance, parts of this study would never have been completed. 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 83 J..Iverson provided data on maturity of females in the FSM sample, and L. Hunt kindly brought in many specimens from various locations in southwestern Georgia. J. Douglass and J. Iverson reviewed the manuscript and offered many helpful suggestions. We especially ap- preciate the interest of members of the Gopher Tortoise Council; their efforts to conserve the remaining populations of gopher tortoises provided the impetus for this study. METHODS STUDY AAEA Most data used in this study were collected on Silver Lake Station, a 1400-ha tract of International Paper Company's Southlands Experiment Forest, and on adjacent lands in Decatur County, Georgia. The forested area was dominated by natural stands of longleaf pine (Pinus palustris) (965 ha). Sand ridges with scrub oaks (Quercus incana, Q margaretta, and Q. laevis) and scattered longleaf pines comprised approximately 210 ha. A dense groundstory of wiregrass (Aristida stricta), bracken fern (Pteridium aquilinum), and running oak (Q. pumita) grew beneath the longleaf pine stands, but these plants were sparse on sand ridges. Open habitat conditions have been maintained by late-winter burning, and fire swept through sand ridge habitats once every 2-4 years. Large colonies of gopher tor- toises were found only on xeric sites characteri*d by Lakeland and Troup soils. Sites with pure stands of longleaf pine were more mesic (Orangeburg and Norfolk soil series) and sup- ported relatively few tortoises. On cultivated sites adjacent to the forested tract, isolated in- dividuals or small groups of tortoises were locatud along fence-rows and field borders. Elevations in the study area range from 24-31 m above mean sea level. The average maxi- mum and minimum temperatures for the summer months are 32°C and 21°C, respectively; annual rainfall averages 1270 mm. DATA COLLEZZION AND ANALYSIS From March 1978-March 1980, gopher tortoises were captured by hand or in pit-fall traps placed in front of dens. Hatchlings from nests in the area were included, and data were col- lected on tortoises found dead along roadsides. Several specimens in the Florida State Museum collection were examined, and data on tortoises from Marion County, Florida, were also utilized. When possible, the age, sex, reproductive status, shell dimensions, and body weight of tor- toises were recorded. Prior to release, tortoises were marked by notching the marginal scutes according to a numbering scheme. Age was estimated by counting annuli where they were most distinct, usually on the ab- dominal shields (Fig. 1). This technique has been used to estimate age in G. poly- phemus (Auffenberg and Iverson 1979), in Geochelone gigantea (Grubb 1971), and in several species of water turtles (Sexton 1959, Moll 1976). False rings, formed occasionally by de- creased growth during an activity season, were discounted; these were usually discontinuous and were bounded by relatively shallow marginal grooves. The ability to discern false rings was developed from observations of many specimens inspected seasonally and annually. The sex and reproductive status of tortoises found dead were determined during necropsy; otherwise, the sex of larger specimens was determined by monitoring sexual behavior in the field and by measuring dimorphic characteristics (McRae et al. 1980). Subdentary glands, which function in courtship (Auffenberg 1966), were inspected each time a tortoise was ex- amined. These glands were recorded as active if they were enlarged and contained copious fluid. Observations of courtship and mating of marked tortoises provided additional data on reproductive status. Measurements taken were straight-line dimensions of plastron length (PL), carapace 84 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 length (CL), and (at a point near the middle of the body) maximum width (W) and thickness (TH). The carapace was chosen as the primary parameter because it is most commonly used in linear growth analysis. Also, in G. polyphemus, the carapace is less variable than the plas- tron: the latter is sexually dimorphic in length and the anterior portion (gulars) becomes reduced by abrasion with age. Tortoises were weighed as soon as possible after capture on.a triple-beam balance. Many individuals excreted profusely before they could be weighed, and this loss coupled with that which occurs during winter dormancy made weight change data too variable for accurate analysis. A box-model volume (CL x W x TH) was calculated as an indicator of total size; this parameter has been used effectively as a size index in water turtles (Mosimann 1958). Three methods were employed to determine growth rates: (1) capture-recapture analysis; (2) age-dimension correlations; (3) annuli measurement-dimension predictions. These three approaches were undertaken to allow comparisons with the findings of various other studies. When possible, growth parameters were calculated by more than one method to provide comparative checks of results. Annuli 4 thC 3 rd Shield B 2 nd at birth 1 st SCALE 1.Omm 05 FIGURE 1. Measurements of abdominal annuli used to estimate carapace length in previous years. Total annuli width (A) was correlated with CL at present age (here, N=4 yr) to form regression equation (Fig. 2). CL for first previous year (N-1) was calculated using distance A minus D, for second previous year (N-2) using A-C, etc. 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 85 Capture-recapture method: A total of 106 tortoises were caught in corresponding months of two consecutive years. The annual growth increment was taken as the difference in measurements from one year to the next (cf. Goin and Goff 1941) Age-dimension correlations: These included measurements regressed against estimated age at the time of eachcapture (N = 579). Only tortoises with distinct plastral annuli were in- cluded in this analysis (cf Auffenberg and Iverson 1979). Annuli measurement-dimension predictions: These were made to discern the growth history of selected tortoises. A regression equation was developed to predict carapace length from a measurement of cummulative annuli width on abdominal scutes (Figs. 1 and 2) for all previous years of an individual's life. Growth of 72 tortoises from sand ridge habitats was compared to that of 12 others living nearby in cultivated fields. Comparisons were also made with samples from Florida (15 from Alachua County, 30 from Marion County). Studies of various species of water turtles (Sexton 1959, Gibbons 1968, Moll 1976) and a giant tortoise (Geochelone gigantea) (Grubb 1971) have shown' the utility of body dimension-annuli measurement relationships in growth analysis. Unless otherwise stated, differences in means were assessed using t-tests. SEASONAL GROWTH CYCLE During this study it was noted that tortoises of all sizes were essentially dormant during winter (December-March). In early spring they showed increased activity, especially on warm daysi they were most active during summer (late May-August). Tortoises were seen more infrequently as fall progressed. 200 C ar ap ac e Le ng th Im m l 150 ... 100 4 2.556 A 50 r lili 0 10 20 30 40 50 60 Total Annuli Width Imml FIGURE 2. Relationship between carapace length and an abdominal annuli measurement; total annuli width (A) as in Figure 1. 86 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 In an attempt to determine the relationship between seasonid activity and growth, young tortoises were chosen because they appeared# to grow at a rapid, fairly uniform rate from year to year. Mean monthly CL plotted against age indicated that growth occurred from spring through fall and ceased during winter (Fig. 3). Mean bimonthly CL increments were averaged through the first five years of life and expressed as 'a per- centage of annual growth (Fig. 4). There was slight growth during early spring, a rapid increase to a constant rate (34 % bimonthly) during late spring and summer, and a gradual decrease through fall to zero growth during winter dormancy. Ninety-three percent of total growth occurred from April-October. Growth was negligible when mean maximum tem- perature was 18-20 °C (December-February). It increased slightly at tem- peratures of 21-27 °C (March-April, October-November), peaked when the mean maximum temperature was fairly stable at 32 °C during the warmer months (May-August). Eighty percent of growth occurred during May-September, the peak growth period noted in north-central Florida by Auffenberg and Iverson (1979). 120- 3 110- 41 5 67 1 C ar ap ac e Le ng th Im m t 100- 1 3 3 2 90- ---- so- 1 2 Ar 1,3 1 /2 -Act ivity Season i5 Winter Dormancy 4,7 12 ) 50'«- 12345 Age I years l FIGURE 3. Seasonal variation in growth as indicated by CL of young tortoises during various months of the year. Numerals indicate sample sizes by month of capture. 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 87 32 PE RC EN T OF AN NU AL GR OW TH /30 0 o /28 25# /26 Bkin.LVE!3dW ,131 H IV 20/ O g ~~ r24 15-* B- 0 «- 22 u! wnlll!xe,Al uee,Al8101 /20 5 -.-18 0 JF MA MJ JA SO ND FIGURE 4. Percent of total annual growth by 2-month intervals in relation to mean maximum ambient temperature. Bars represent the mean bimonthly CL increment averaged through the first 5 years of life and expressed as percentage of annual growth. ANNUAL GROWTH RATE CARAPACIAL INCREMENT Measurements of tortoises captured during corresponding months of two consecutive years were initially used to determine growth in- crements. Capture-recapture data for tortoises of CL < 240 mm were compared with those obtained by Coin and Goff (1941) in central Florida (one of the specimens in the latter sample (i.e., number 13) was deleted because it showed no growth). For purposes of comparison, Coin and Goffs (1941) data were adjusted to a growth interval of 12 months based on the calculated percentage growth per month for each individual (Fig. 4). In both Coin and Goffs (1941) sample and our own, the mean CL in- crement was about 12 mm per year (Table 1). The Florida sample includ- ed only three individuals in smaller size classes (initial CL < 130 mm), and these represent the stages of most rapid development. Larger tortoises grew at a mean rate of 11.1 mm per year, while those in our sample grew 88 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 TABLE 1. Annual carapacial increments of gopher tortoises from southwestern Georgia and central Florida. Data were obtained by capture-recapture. First year CL Mean CL increment (mm) class (mm) Southwestern Georgia Central Floridaa Range x N Range x N 50-59 7-20 13 15 -- 10 1 60-69 - 13 1- 80-89 7-15 12 4 - 90-99 8-22 14 4 -- 13 1 100-109 -- 18 2 110-119 15-20 17 2 -- 17 1 120-129 10-16 13 4 - - - 130-139 7-9 8 3 3-19 13 4 140-149 4-19 11 2 13-60 37 2 150-159 - 11 1 - 4 1 160-169 8-19 13 2 3-18 11 4 170-179 - - - 4-16 10 4 180-189 -- 14 1 4-14 8 4 190-199 -- 7 1 1-9 6 4 200-209 -- 8 1 6-9 8 3 210-219 5-9 7 2 - - - 220-229 -- 11 1 -- 4 1 230-239 4-24 10 4 3-18 10 2 All classes 7-18 11.8 16 4-37 11.6 13 All individuals 4-24 12.5 50 1-60 11.2 32 .Data adapted from Coin and Goff (1941). 9.8 mm annually (about 12 % less). The greater potential for growth in the Florida sample was evidenced by specimen number 9, which grew 42 mm in 9 months, or about 60 mm per year (our estimate). This rate is ap- proximately twice that of our fastest-growing specimen. Douglass and Layne (1978) presented information that also indicates more rapid growth in more southerly populations: in an area near Lake Placid, Florida, they found tortoises to exceed 100 mm in PL by age 3-4 years; this size was attained (on the average) after 4-5 years in our part of Georgia. A second approach to determining relative carapacial growth was derived by estimating CL in previous years from annuli width measure- ments (see METHODS). The difference between reconstructed CL from one year to the next was plotted against CL class. Figure 5 depicts an increment curve in relation to initial size based on two methods. Small individuals with CL of about 50 mm grew rapidly, but at 60-90 mm CL experienced below average ( < 12 mm per year) de- velopment. This trough was followed by a growth surge at about 100-120 mm CL, following which tortoises grew regularly at above average rates 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 89 20- . Capture Recapture . Annuli Measurement18- · C ar ap ac e In cr em en t [ m i · Mean E 16- 14- 12- 10- 8- 6- 4- 2- O 6 ' ' ' 1 ' ' I ' I80 100 120 140 160 180 200 220 240 Carapace Length Imm I FIGURE 5. Annual carapacial increments in relation to carapace length. Data points are mean increments,of individuals in 10-mm CL classes, plotted over class midpoints. until they reached 180-190 mm CL. Growth generally decreased beyond that point, but a slight surge was noted in tortoises 220-230 mm CL. A comparison of growth increments in relation to age, as estimated by three methods (Table 2), shows most rapid development to occur in younger age classes, followed by sharply decreasing increments in later years. The similarity in results of these methods indicates that any of the three methods can be used to estimate growth. A summary of CL growth per 4-year class as estimated from annuli measurements is given in Table 3; only individuals in wiregrass habitats were used in this analysis. Of 599 annuli measurements for growing seasons 1-12, over 67 % represented CL increases of at least 10 mm. Dur- ing these years of immaturity, annual growth was characteristically 5-20 TABLE 2. Annual growth in CL in 4-year groups of G. polypheinus as determined by each of three methods. Specimens are from natural habitats in southwestern Georgia. CL-age Inter-annuli Capture-recapture correlation Measurement Age N x N -x N X 1-4 23 12.6 82 11.8 271 12.0 5-8 15 12.3 36 12.0 210 12.1 9-12 2 11.8 25 11.6 118 11.5 13-16 5 9.3 24 9.8 70 9.5 17-20 5 5.7 37 6.2 35 7.8 90 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 mm (x = 12.0 mm) (Table 3). Auffenberg and Weaver (1969) suggested that growth in CL in immature G. berlandieri averaged 16-18 mm per year. Giant tortoises grow 15-20 mm per year in CL early in life, then steadily decline in growth rate to about 5 mm annually by age 15-20 years, depending upon population density (Grubb 1971), Occasionally, annual increments of 20 mm or more were detected in our study area; only rarely did tortoises < 20 years of age grow < 5 mm per season. Growth in the 17-20 year age class was typically only 5-10 mm per year (80 %), and was occasionally 10-15 mm per year (13%). Most tortoises (55 of 69) grew in an intermittent pattern, with peak CL increases typically occurring at 3 to 6 year intervals. Grubb (1971) found that giant tortoises grew intermittently, often exhibiting sudden in- creases in growth rates after several years of very slow growth. In our study, the most rapid growth usually occurred during 1 to 3 year periods within the 6th-10th growing seasons of an animal's life. Carapace length increases of 2 15 mm occurred on the average once every 5 years in the life of an individual; increments of 10-15 mm were twice as common, oc- curring an average of once every 2.5 years in the lives of individuals. Growth surges of > 15 mm per season were manifested by individuals for periods of 1-4 years only; 44 such surges were confined to a single year, 12 occurred in two successive years, and only three each occurred during 3 and 4 consecutive years. Fourteen of 69 individuals (20.3 %) grew at fairly uniform rates throughout life (i.e., variation in CL growth was not significant [P > 0.10] among years). These individuals generally grew at rapid rates (x= 13.3 mm/yr, N=6) or at comparatively reduced rates (R = 8.8 mm/yr, N = 8). Varying nutritional regimes are possible determinants of these unusual growth patterns. Of 52 tortoises living in wiregrass habitats, 13 experienced growth of > 20 mm CL in a single season; five of these grew at least 20 mm during TABLE 3. Increases in CL, per 4-year period, of 72 tortoises living in natural habitats. Incre- ments were estimated from annuli measurements using regression equation (see METHODS). Numbers in the body of the table are percentages of total annuli (N) measured that represent an estimated CL increase falling within each increment class. Carapace growth/yr (mm classes) Year N <5 5-9.9 10-14.9 15-19.9 20-24.9 25-29.9 Z30 1-4 271 6.2 28.1 41.0 19.1 5.6 0.0 0.0 5-8 210 3.8 28.6 38.1 20.9 4.8 3.3 0.5 9-12 118 1.7 30.5 35.6 20.3 6.8 4.3 0.8 13-16 70 9.1 42.4 30.3 15.2 3.0 0.0 0.0 17-20 35 6.7 80.0 13.3 0.0 0.0 0.0 0.0 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 91 two or more years. All of these atypical growth surges were experienced during growing seasons 2, 6-11, or 13, and usually during 7-8. Only one of 25 tortoises aged 13 and older grew more than 20 mm per year in CL after the 12th growing season. Only two single-year increases were over 30 mm; one occurred in the 8th year, the other in the 9th year of life. It ap- peared that surges of above-average growth ( > 12 mm/yr) did not occur in any specific year of life, but tended to occur during the 6th-10th grow- ing seasons. To determine if there was significant variation in growth in relation to rainfall or other environmental factors, annual carapacial increments were estimated for the years 1967-1978 (Fig. 6). Two 5-year age cohorts, 2-6 and 7-11, were used in the analysis; these groups were chosen because of the insignificant (P> 0.20) variation in annual growth within each group. There were no significant differences in growth rates among the 12 years in either of the two samples (P > 0.20). Nor were annual increments in growth correlated with rainfall in the spring, summer, or preceding fall, although seasonal rainfall was quite variable among these years. An- nual growth of G. polyphemus in this area is determined to a greater ex- tent by the age and growth potential of the individual and by food quality (see below) than by seasonal rainfall. This is to be expected as rainfall in this area, especially in spring and early summer, is typically sufficient for adequate plant development. 73 74 76 16 16-- 7772 15 C ar ap ac ia l In cr em en t [m m l 15~ _67 69 71 14 13- --~ ~ 13 1212 - 78 11 11 L 1010I- m== 7th - 11 th Growing Seasons czz] 2nd - 6 th Growing Seasons 9 8 8 67 68 69 70 71 72 73 74 75 76 77 78 Growing Seasons FIGURE 6. Estimated annual growth in carapace length for two age-groups (2-6 and 7-11 yr) by calendar year: Sample sizes were 16-50 (x = 32) per year. 92 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 Annual growth rates thus did not vary significantly with rainfall; the effects of nutrition and geographic location on growth were then ex- amined. Depicted in Fig. 7 are CL growth rates estimated for years 1-10 for G. polyphemus from north-central Florida and from sand ridge and agricultural habitats in southwestern Georgia. Individuals in natural habitats in Georgia were significantly (P < 0.01) more likely to grow 10 mm per year than were individuals in Florida or in fertilized Georgia sites. The influence of improved nutrition was evident: tortoises in agricul- tural situations consistently grew about 4 mm more per year (x incre- ment = 16.2 mm) than those in natural communities (* = 12.0 mm). This was to be expected: for example, Jackson et al. (1978) reported· CLs of 236 mm at 4 years in G. agassizii captives on highly nutritious diets, while those living in natural habitats require 15-20 years to reach this size (Woodbury and Hardy 1948). Nearly 20 % of the gopher tortoises in a colony in a ruderal habitat in Florida exceeded 318 mm CL, while less than 7 % of the individuals in each of 11 other colonies were this large (Alford 1980). Significant differences in growth rates of Chrysemys picta on diets differing in nutritional quality are well documented (Cibbons 1967; Ernst 1971; Christiansen and Moll 1973). Giant tortoises in two low-density populations grew much more rapidly (Swingland 1977)and reached sizes over 2.4 times larger (Coe et al. 1979) than those in a food- limited, high-density population. That the warmer climate and longer growing season in north-central Florida led to a greater annual CL increment (x= 17.7 mm) seems clear (Fig. 7); this increment is identical to that determined from data pre- sented by Auffenberg and Iverson (1979). The growing season at the Florida locality was 295 days, approximately 12% longer than that in southwestern Georgia (259 days). Over 75 % of Georgia G. polyphemus, regardless of habitat, were in increment classes of less than 20 mm; in- dividuals in Florida were significantly (P < 0.01) more likely to grow more than 20 mm per year. In an additional sample of 30 tortoises from Marion County, Florida (provided by W. Affenberg), projected CL increase ex- ceeded 13 mm annually during the first 17 years of life and was 15-16 mm at ages 1-8 years. These values also are significantly greater than corres- ponding data in our samples (P < 0.01). Earlier maturity, faster growth rates, and larger body size of southern versus northern Kinosternon subru- brum (Ernst et al. 1973) and Chrysemys picta (Christiansen and Moll 1973) have been attributed to longer annual growing seasons in the south. Jackson et al. (1978) felt that year-round activity was a significant factor, in addition to nutrition, in the rapid growth of G. agassizii captives. ALLOMETRIC RELATIONS Proportional changes in various body dimensions are important com- 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 93 40- 40 P er ce nt ag e of A nn u M ea su re d /£ Georgia, Wiregrass 30 - -.'111 ~ Georgia, Agricultural - 30 ) Florida, WoodlandS -*20- . - 20 10 <5 5 -9.9 10-14.9 15-19.9 20-24.9 25-29.9 230 Carapace Increase / Year (mm) FIGURE 7. Comparison of mean annual carapacial increment during growing seasons 1-10 for gopher tortoises in Georgia (wiregrass and agricultural habitats) and Florida (woodland habitat). ponents of shape and development (Mosimann 1958). In this section, allometry is utilized to depict changes from the juvenile to subadult to fully mature form in G. polyphemus. The relative rates of increase in CL, width (W), and thickness (TH) are apparent in Fig. 8. As has been noted for CL increase, mean body widths and thicknesses for age cohorts 0-17 were not significantly different from values predicted from the linear regression lines (P > 0.20). Thus it is assumed that deviations about the line (Fig. 8, notably at age 9) are due to small sample sizes in some age cohorts. Mean annual increments for various dimensions are given in Table 4. The major axis of growth through age 17 was antero-posterior, width and thickness increasing at only 73 and 41 % of the CL rate of increase, respec- tively. The annual increase in CL is a larger percentage of carapace length (compared to relative growth in W and TH) for 4-5 years after hatching. Mosimann (1958) noted dissimilarity between hatchlings and later forms and suggested that this reflected selection for "form-changing" types,'i. e., the advantageous form is produced mainly just after hatching through differential growth. This also seems to characterize the relative growth relations in G. polyphemus. From about age 5 to just before maturity, growth is linear with age at the same relative rate in all 3 dimensions. Grubb (1971) reported isometry between CL and width in giant tortoises. Beginning at about age 17, as individuals reached adult form, annual increments in all body dimensions decreased sharply (Fig. 8). Growth rates in linear dimensions for the first 11 years as adults ranged from 23-54 % of rates for the younger group (Table 4). 94 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 300- CL 1.1- a LL250 - w M 200- mm 150- Th F M100 - r „fAt 50 2 . ~ CO 5% - -1 3 0 8 S*9882900 u'~M°"r)0(0'°< w q g ZP Ze N q 11 0 5 10 15 20 25 30 35 Age (years) FIGURE 8. Relative growth in carapace length (CL), body width (W), and body thickness (TH) in relation to age estimated from annuli counts for male (M) and female (F) tortoises. Symbols are mean values for tortoises in 1-year age classes; sample sizes are indicated above the age axis. Through age 15, regression lines are used; thereafter, lines are fitted to the points by hand. TABLE 4. Mean annual increments in body dimensions of young (0-17 yr) and older (18-28 yr) gopher tortoises. Increments for the first group are the slopes of regression equations (see text); for the older group, increments are values for the 26-30 agh-cohort ( x = 28) minus values,at age 18 divided by 11 yr. Sample sizes in parentheses. Group 1 Group 2 0-17 yr 18-28 yr (298) Males (63) Females (61) Dimension x x % Group 1 x % Group 1 Carapace length (mm) 11.3 2.7 23.9 5.2 46.0 Width (mm) 8.3 2.5 30.1 4.5 54.2 Thickness (mm) 4.7 1.3 27.7 1.8 38.3 Volume (cc) 210.6 181.8 86.3 345.5 164.1 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 95 Allometric relations of older tortoises were clearly sexually dimorphic (Table 4). Width increases of males were almost equal to CL increments, and growth in thickness, although greatly reduced from the rate for the smaller group, was more rapid relative to CL increase. In females, however, CL and W continued to increase at approximately half the rates in the younger cohort, contrasted to rates in males of less than 1/3 those in the younger cohort (Table 4). The thickness of females increased at a much lower rate (38 %). As a result of these allometric changes, mature females were longer and wider relative to their thickness than they had been during the latter stage of rapid growth (ages 12-17 yr). Increase in TH of both sexes and increase in CL and W of males typically was asymp- totic soon after maturity was reached; CL and W of females usually in- creased substantially for several more years. Relative growth in the three dimensions is depicted in Figure 9. Box-model volume (CL xWx TH) was a curvilinear function of both CL and age (Fig. 10). As weight data were too variable for accurate analysis, volume was employed as an index to size. Regression analysis us- ing 605 individuals demonstrated that weight and volume (V) of gopher tortoises were linearly correlated (re =0.95).It was found that patterns of development of secondary sex characters (e.g., plastral concavity) were Hatchling S W CL [IZES> <~12 F -2.5 - ~-35 »1~ 28 Years (3 17 Years 4.5 1~-5 \ 2 ~ 28 Years 9 SCALE I mml 0 10 30 FIGURE 9, Three-dimensional growth models of Gopherus poll/phemus in the hatchling, subadult, and mature stages of development. Arrows represent growth vectors of the three body dimensions, expressed in terms of mean annual increments (mm) between developmen- tad stages 96 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 -5000 9- F 8 -4000 Vo lu m e (T ho u. c m 3 ) 7 M 6 -3000 6 146!eM A pog 5 0 4 -2000 3 0 o Box Model Volume -1000 · Body Weight 1 5 10 15 20 25 30 35 Age (Years) FIGURE 10. Volumetric growth and body weight in relation to age. Data were obtained using the volume-age correlation method, and the line for volume was fitted by eye, highly correlated with body volume in linear relationships (MeRae et al. 1980). Since volume is a quadratic of three dimensions (CL, W, TH), all of which increase linearly with age (through 17 yr), its exponential increase (Fig. 10) is expected. The 90 % greater rate of volumetric increase in 18 to 28 year old females (Table 4) relative to males was due mainly to con- tinued growth in CL and W of females. Older females almost doubled in body volume during this 11-year span (Fig. 10), while males began to ap- proach terminal size. The exponential increase in total body size demonstrates a pronounced change that is not clearly reflected in linear dimensions taken separately. Volumetric change is a more meaningful indicator of growth; it best reflects the rapid development that is apparent in weight change with age (Table 5). A very close relationship exists between the box-model estimate of volume and body mass from age 0-17 years; the carapace in this age cohort is characteristically dome-shaped. In older, larger tortoises, the carapace becomes more dorsoventrally compressed and the body con- forms more closely to the box model. This change in shape has also been noted in certain aquatic turtle species (Mosimann 1958). The body mass per cubic unit of box-model volume consequently increases greatly in older individuals (Fig. 10). 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 97 VOLUMErRIC INCREMENTS Estimates of annual volumetric increments were determined by cal- culating the mean volume of tortoises in each age class (Fig. 10) and then determining differences in volume from one year to the next. Values were plotted over CL to provide a general comparison with CL increments (Fig. 5). This curve (Fig. 11) shows a slight decrease in added size at CL of 60-70 mm, after which volume increased at an increasing rate. The in- flection point occurred at 220-230 mm CL, beyond which tortoises were still growing, but at decreasing rates. This growth change occurred at the size of most rapid development of sexually dimorphic external characters (MeRae et al. 1981). TABLE 5. Body weights (g) of gopher tortoises, by age, in natural habitats in southwestern Georgia. Age N R SD 0 130 34.4 3.1 1 40 50.1 10.4 2 4 64.1 12.4 3 10 100.1 19.1 4 12 181.4 28.8 5 16 238.1 41.8 6 13 319.3 45.2 7 10 400.2 47.4 8 8 519.7 100.2 9 5 779.1 81.6 10 7 817.6 118.3 11 3 1103.7 338.4 12 9 1272.5 303.1 13 5 1427.9 220.0 14 8 1581.2 358.6 15 6 2064.0 188.6 16 5 2165.4 366.9 17 5 2278.3 325.0 18-20 M 25 3012.3 575.9 F 7 3184.2 513.7 21-25 M 44 3287.4 688.2 F 18 3959.6 1102.0 26-30 M 58 3603.2 558.7 F 56 4205.3 958.7 31 + M 32 3651.9 573.6 F 42 4911.6 775.5 98 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 500- 400 V ol um et ric In cr em en t (c m ) 0 0 0 C 300 0 200 0 100 0 lili 50 100 150 200 250 Carapace Length (m m) FIGURE 11. Annual volumetric increment in relation to CL. Increments are differences in mean volume between age classes, plotted over mean CL for each age class. Sample sizes as in Figure 8. SIZE IN RELATION TO LIFE HISTORY IMMATURE STAGES As shown in previous sections, gopher tortoises develop most rapidly in early years, though not at a uniform rate. The cumulative effect of these variable rates is an overall growth curve, which is useful in depict- ing life history stages. The progressive development of the carapace was estimated by three methods. Capture-recapture data from Table 1 were applied, starting at hatchling size (50 mm CL) and adding CL increments; theoretically, as a tortoise entered a new CL class, the mean annual increment for that class was added as the next year's growth. The second method was to plot CL over estimated age at the time of capture (data from Table 6). A curve was developed by connecting CL values obtained in September-October, when termination of yearly growth occurred. Carapacial increases estimated from annuli widths provided a third estimate (see Table 6). These approaches indicate similar, curvilinear relationships (Fig. 12), beginning with a steep linear increase extending through age 11 years 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 99 (190 mm CL). This pronounced growth coincides with the above-average growth reflected in annual increments (Fig. 5). Thereafter growth gradually decreased; it was greatly reduced beyond age 17 years, espe- cially in males. TABLE 6. Carapace lengths of young gopher tortoises, by age, in natural habitats in South- western Georgia. Values calculated by CL-age correlation are data for the middle of the growing season; annuli measurements provide values for the end of the grow- ing season (winter). Carapace length (mm) Growing CL-age correlation Annuli measurement season N x SD N R SD 0 130 50.la 2.1 40 59.4 5.5 1 52 64.6 0.6 4 69.8 7.6 2 44 76.0 1.0 10 79.4 4.2 3 44 88.2 1.4 12 97.4 6.0 4 42 100.5 1.9 16 106.4 6.9 5 31 111.6 2.7 13 122.6 5.3 6 29 122.9 3.1 10 134.1 5.5 7 23 136.2 3.5 8 144.4 7.5 8 20 148.9 4.6 5 159.6 7.6 9 17 162.6 6.4 7 163.0 5.1 10 16 177.5 7.6 4 181.8 15.0 11 16 187.5 7.9 9 191.7 11.9 12 12 197.0 10.8 5 194.8 15.6 13 11 207.1 10.9 8 206.4 14.9 14 11 216.2 10.1 6 223.2 16.1 15 7 225.0 10.2 5 231.1 13.3 16 7 228.8 10.4 5 232.1 11.8 17 7 237.9 10.4 aActual size during first day of life. 100 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 300- F -M 250 - .7- 200 C ar ap ac e Le ng th Im m l 150 100 ~ Capture - recapture • CL - age correlation 50 o Annuli measurement 1 1 1 1 0 5 10 15 20 25 30 35 Age Iyearsl FIGURE 12. Comparison of CL-age relationships estimated by three methods of analysis. The curves are fitted to the CL-age correlation data only. Tortoises passed through two general life history stages before reach- ing maturity. The juvenile stage lasted until the carapace was about 100-120 mm in length. The shells of young tortoises were very soft, and carapacial scutes usually had distinct yellow centers. Laminal spurs on the humeral and femoral shields were pronounced. During the following few years the juvenile coloration faded as tortoises entered the subadult stage. Laminal spurs became indistinct from abrasion, and the shell less compressible. During the latter portion of the subadult stage (190-220 mm CL), the shell was nearly rigid and sexual dimorphism became more apparent; body volume increased rapidly during this stage (Fig. 8). There was no evidence of social interaction (i. e., visits to dens of others) in the juvenile stage, and very little interaction was observed among subadults. Auffenberg and Iverson (1979) studied aging in relation to plastron growth in the Florida State Museum (FSM) collection. The growth curve in that sample appeared to be sigmoid, compared to the somewhat parabolic curve determined for tortoises at the southwestern Georgia locality (Fig. 13). Student's t-tests indicated that yearly means of PL 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 101 growth increments for the Florida tortoises 7-14 years of age were significantly larger (P < 0.01). Means at ages 6 and 15 were different only at the 10 % level, and at other ages the means were not significantly dif- ferent (P > 0.10). The Florida specimens were collected near Gainesville, approximately 140 km south of our study site. The differences in growth increments seem attributable to dissimilar yearly activity periods, as noted earlier. SIZE AT SEXUAL MATURITY Several criteria were used to determine size at maturity. Estimating the size at which males are functionally mature is a complex matter. Necrospy data showed that physiological maturity, as indicated by sperm production, occurs in individuals with a CL of 210 mm, and possibly smaller. Somatic maturity, including rapid development of secondary sex characters, usually occurs from that point up to 240 mm CL or more (McRae et al. 1981). Of greater importance is psychogenic maturity, which occurs when males acquire a propensity to breed. Determining that stage from field observation is difficult because large males attempt to prevent smaller ones from visiting females (Douglass 1980). Subdentary glands were active from spring-fall in 55 males over 230 mm CL; glands of 18 others (160-235 mm CL) were inactive, but those of two males (222 and 227 mm CL) were slightly active. During a study of the mating system in southern Florida, Douglass (1980) observed seven males that visited females at their dens; plastron lengths of these males 250- • NC Florida ° SW Georgia I 200- P la st ro n Le ng th Im m l 150- ... -44 100- ,A 50-0~' 01234567891611121314151617181920 Estimated Age [ Years I FIGURE 13. Plastral growth in relation to age in southwestern Georgia (this study) and north- central Florida (from Auffenberg and Iverson 1979). Symbols represent means for 0. 1-year age classes. 102 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 were 222-256 mm (227-251 mm CL, our estimate). Of 32 males seen copulating or visiting females in our study, all were at least 240 mm CL. Taking all factors into account, we estimate that males usually be- come functionally mature in this region as they reach 230-240 mm CL. The lower size limit of mature males indicated by Auffenberg and Iverson (1979) was also 230 mm CL. Females in our study area reach somatic maturity at about 250 mm CL (MeRae et al. 1981). The minimum CL of confirmed sexually mature females (N = 44) was 252 mm. Ten small females (215-261 mm CL), necropsied during April and early May, contained no oviducal eggs, enlarged follicles, or corpora lutea; reproductive activity would probably have been discernible at this season (e.g., Iverson 1980). Subdentary glands were active during spring for females 255 mm ih length and larger, and inactive in smaller individuals. From the FSM collection we also examined 23 females for which re- productive data were provided by J. Iverson. External sex characters in this sample were evident at smaller sizes than in our area (MeRae et al. 1981). The smallest mature females in the FSM collection were 242,245, and 249 mm CL. Other mature females reported from Florida have been as small as 238 mm CL (Auffenberg and Iverson 1979) and 241 mm CL (Hallinan 1923). Behavioral evidence presented by Douglass (1980) in- dicated that seven females 242-303 mm PL were frequently visited at their dens and often showed interest in male courtship; two of the larger in- dividuals (245 and 303 mm PL) apparently laid eggs. Smaller females (184-240 mm PL) were rarely visited by males, and 5howed no interest in courtship. Based on a review of the literature and examination of the FSM and other samples, Iverson (1980) concluded that females mature at 220-230 mm PL (226-236 mm CL) in northern Florida. We examined four immature females from our study area that were 243-261 mm CL: the smallest mature female was 252 mm CL, and all other confirmed mature individuals were 260 mm in CL or larger. From a collection made near Douglas, Georgia, Jackson et al. (1974) examined a female 241 mm CL that contained enlarged follicles, but this is not con- elusive evidence of maturity. In turtles ovarian follicular development can occur several years prior to ovulation; this process is demonstrated in studies of Chelydra serpentina (Christiansen and Burken 1979) and Geochelone gigantea (Swingland 1977). Therefore we estimate that females in this part of Georgia reach maturity at some point between 250-265 mm CL. AGE AT SEXUAL MATURITY Growth rates of the two sexes were very similar up to about 240 mm CL, but began to diverge beyond that size. Volumetric increments peaked 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 103 at 220-230 mm CL (Fig. 8). After divergence the reduced growth rates of males and the greater rates in females led to sexual dimorphism in size beyond age 17 years (Table 7). Males attained the size of estimated sexual maturity (230-240 mm CL) at 16-18 years while females reached mature size (255-265 mm CL) at 19-21 years. The latter period is considerably greater than the 10-15 years estimated by Iverson (1980) for attainment of maturity by females in northern Florida. As stated earlier, the Florida specimens examined became mature at smaller sizes, and the minimum age was determined from faster-growing tortoises (Auffenberg and Iver- son 1979). Estimated ages at maturity in other Gopherus species in the wild vary considerably. Gopherus agassizii is said to mature at 15-20 years (Wood- bury and Hardy 1948; Berry 1976). Burge and Bradley (1976) suggested that both sexes mature at 215-220 mm CL in that species, and data ob- tained by Medica et al. (in Auffenberg and Iverson 1979) indicate that this size is attained some time after 17 years of age. Miles (1953), however, stated that G. agassizii can reach maturity at ages as low as 10 years. TABLE 7. Mean CL (mm) of large gopher tortoises, by sex and age group, in southwestern Georgia. Males . Females Age N x SD N SD 18-20 25 248.2 10.0 7 251.8 8.7 21-25 44 254.9 12.7 18 271.6 7.7 26-30 58 261,9 10.5 56 288.6 12.6 31 + 32 263.7 16.1 42 294.3 15.9 In the Texas tortoise (G. berlandieri), attainment of breeding size has been estimated to occur at 15-20 years of age (Anonymous 1967, in Douglass 1980). However, Auffenberg and Weaver (1969) estimated that maturity in this species is reached at 105-128 mm CL, at ages of only 3-5 years. Apparently the only information on sexual maturity in G. flavomar- ginatus is that of Legler and Webb (1961). They suggested that both sexes mature at 220-300 mm CL, and that the age of one female in this size range was about 11 years. POST MATURITY After maturity is reached, growth occurs occasionally, but at reduced rates. Unusually large size in adult Chrysemys picta (Gibbons 1967) and 104 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 Geochelone gigantea (Grubb 1971) has been attributed to exceptionally rapid growth as immatures rather than to continuing growth as adults. Auffenberg and Weaver (1969) noted a sharp decrease in growth of the Texas tortoise after attainment of sexual maturity. Abrasion of the shell of gopher tortoises over the years can hinder accurate age estimation. Males in this part of Georgia typically stop growing at some point between 250-270 mm CL. Females grow to be significantly larger; mean CL for adult females in our area was 283 mm (N = 102); the largest captured was 335 mm CL. Sexual dimorphism in size may be related to differential energy intake and expenditure. Movement data from radio-instrumented tortoises in this area (unpublished) and data presented by Douglass (1980) show that males travel much more and spend more time in mate-seeking and courtship than do females. By applying the mean growth rate of adults that grew measurable amounts from one year to the next (Table 8), we estimated that the age of the largest female in our sample was at least 57 years, and of the largest male (290 mm CL) was 46 years. So about 53 % of recaptured adults did not grow during a one-year period, sporadic growth from year to year is indicated. Probably some tortoises at our locality live to be perhaps 80-100 years of age. Hubbard (1894) suggested that tortoises of this species may live more than 100 years. TABLE 8, Annual growth.of 56 adult gopher tortoises in southwestern Georgia, as determined by capture-recapture. Malesa Femalesa Growing Tortoises N 11 15 x increment (mm) 2.1 2.3 Increment range (mm) 0.2-5.2 0.4-5.8 All Tortoisesb N 24 32 R increment (mm) 0.9 1.1 Increment range (mm) 0.0-5.2 0.0-5.8 'CL >240 mm for males, > 265 mm for females. bIncludes tortoises that did not show growth (13 males, 17 females). SURVIVAL AND ADAPTIVE IMPLICATIONS Data gathered thus far suggest a latitudinal cline in body size in G. polyphemus, with a general northward increase. Notable exceptions are record-sized individuals from Florida, such as one of 368 mm CL 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 105 (Conant 1958) and another of 343 mm CL (Carr 1952). Local variations in growth due to differences in habitat quality and nutrition, documented in this study and by Gibbons (1967), Jackson et al. (1978), and Swingland (1977) for other species, may explain these exceptions. The typical upper size limit for this species, according to Cochran (1952), is about 305 mm CL in Florida, and the largest reported by Ditmars (1946) and Hutt (1967) in Florida are less than 300 mm CL. Douglass and Layne (1978) noted that adults of PL greater than 300 mm were rare in a large sample (N = 366) from south-central Florida, and Iverson (1980) reported a mean CL of 262 mm for adult females in a north-central Florida sample (N = 16). Alford (1980) reported that only 4 % of the individuals in 12 northern Florida colonies exceeded 300 mm in CL. Contrasting with these reports are: (1) in our study area in south Georgia, a mean CL for adult females of 283 mm (over 24 % with CL ex- ceeding 300 mm); and (2) observations that the largest individuals of a Florida-Georgia sample were in the fall-line hills of Georgia, at the north- ern limit of the range (D. Franz, pers. comm.). Further evidence that gopher tortoises in the south are generally smaller is implied by existing data on clutch size, which increases with the size of females (Iverson 1980; Landers et al. 1981). The mean and upper limit of clutch size in our study area (7 + and 12 eggs, respectively) exceed corresponding data from Florida (see Landers et al. 1981) with only two exceptions, both from Dade County (a clutch of 15 eggs and a group of 19 enlarged follicles in a female, Iverson 1980). Considerable evidence suggests a similar north- south cline in clutch size in some aquatic turtles (Cagle 1954; Tinkle 1961; Christiansen and Moll 1973). The higher growth rates of Florida G. polyphemus relative to those in our study area seem attributable to longer growing seasons in the south. The reasons why females in Florida attain mature form and lay eggs at smaller sizes (McRae et al. 1980; Landers et al. 1981) and are apparently smaller on the average than those in southwestern Georgia are probably more complex. Several possible explanations are presented here for con- sideration. One hypothesis is based on differential physiological age. In south- western Georgia tortoises are dormant for 3-4 months during winter. Those in the southern portion of the range are active for much longer periods during a given year, and in extreme southern Florida are active year-round (W. Auffenberg, pers. comm.). Thus a southern specimen liv- ing the same number of calendar years as a more northerly one is physio- logically older. If physiological age is a primary determinant of maturity, internal and external morphological traits accompanying maturity would be expressed in fewer calendar years and at smaller sizes in southern animals. Maturity is followed by a cessation or great reduction in growth, 106 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 and, consistent with the hypothesis, terminal sizes appear to be smaller in the south. Another possible explanation for these size differences is based on the probable tendency of gopher tortoises to modify their habitats through grazing pressure. A longer growing season (i.e., Florida vs. Georgia) results in earlier maturity and may induce more frequent additions to the population. If a high density ensues, overgrazing may reduce the more nutritious herbs in the feeding stratum and serve to limit the growth of larger individuals, which require greater quantities of food. Work on the giant tortoise (Swingland 1977; Coe et al. 1979) is interesting in this con- nection: Adults in a dense, food-limited population attained sizes only 40 % as great as those attained in two low-density populations. Additionally, environmental influences may select for terminal body size. If high-density populations have occurred more frequently and ex- erted stronger grazing pressure in southern portions of the range, selection for smaller, more energy-efficient tortoises could have resulted. Selection for thermal efficiency may also be at work. Southern specimens not only tend to be smaller, they are also lighter in color on the carapace and plastron. In southwestern Georgia the carapace is dark gray and the plastron mottled black and yellow, while at the northern limits of the range (e.g., the Georgia fall-line hills) the carapace is even darker and the plastron is often totally black (our observations and those of R. Franz, pers. comm.). It seems that smaller, lighter-colored tortoises may be bet- ter adapted to reflect heat (i. e., in the south), while larger, darker indi- viduals probably conserve heat more efficiently (i.e., northern portions of the range). Selection for larger females may be linked to fecundity. For example, in a study of geographic variation in size and reproduction of Ster- notherus odoratus, Tinkle (1961) hypothesized that selection in the south must be mainly for smaller body size, while in the north, selection for both larger size and higher reproductive potential seems to exist. These selective forces may also affect gopher tortoises. The less frequent addi- ti6n to colonies in southern Georgia, as indicated by longer periods of growth to sexual maturity, may be somewhat compensated by selection for larger females producing larger clutches. Growth is probably further reduced in more northerly populations, which may be even less produc- tive. This species is distributed patchily in the fall-line hills of Georgia, and sand ridge habitats exist in many areas north of the present and known historic range of these animals. As annual temperature cycles in- fluence growth and age at sexual maturity, it is possible that the length of the growing season is a major factor limiting the range of this species. As Auffenberg and Iverson (1979) pointed out, growth to the age of sexual maturity is a very important parameter in population dynamics. 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 107 This period is critical, as individuals are exposed to predators and diseases before becoming reproductively active. A study made in this area (Landers et al. 1981) revealed that female G. polyphemus nest at most once per year, and most nests (87 %) are destroyed by predators. Many hatchlings are killed in the nests or shortly after emerging by predatory mammals and fire ants (Solenopsis saevissima) and, for at least two years after hatching by several species of snakes and mammals. In some areas, larger tortoises are preyed upon by canids (Causey and Cude 1978) and black bears (Ursus americanus) D. Speake, pers. comm.). In addition, gopher tortoises are known hosts for several types of parasites (e.g., Petter and Douglass 1976), but the ef- fects of these and other possible agents of disease on populations are unknown. During this study, three immature tortoises (aged 2,4, and 5 yr) and four adults were found dead in or near their burrows, but the causes of death could not be determined. Low reproductive rates and the high incidence of postnatal mortality in this species no doubt greatly impede the comeback of depleted popula- tions. The extended period of time required for hatchlings to reach maturity (e.g., 10-15 yr in Florida, 16-21 yr in our study area) is an addi- tional limiting factor, especially in more northerly parts of the range. 108 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 APPENDIX SYNOPSIS OF METHODS USED TO ESTIMATE GROWTH Results of this study are based on application of three methods commonly used in testudinid growth estimation. In using these methods, it was noted that each had its merits and shortcomings. This section is a comparison of the approaches, and is presented as an aid on future studies. The capture-recapture method is most time-consuming. Due to varying rates of increase among and within size classes, a large sample is required (i.e., at least 100 specimens). It is difficult to capture gopher tortoises at the same time in two successive years because they relocate frequently and become wary of even well-camouflaged traps. Recapture success was only 27 % in this study and 25% in that of Coin and Goff (1941). If measuring seasonal growth is a study objective, this method should not be used because frequent capture may cause gopher tortoises to move to other burrows and may alter normal patterns of movement and feeding. If volumetric increments are to be studied, even larger samples are required; ex- ponential increases make inferences from small samples of little value. The advantage of the capture-recapture method is that it yields exact measurements of linear increments over time. Its effectiveness can be improved by concentrating trapping efforts on juveniles and subadults, as adults grow very little. The accuracy of age-dimension correlations is dependent upon the regularity of ring deposition in a given location and the ability of the observer to discern actual annuli. Both can be understood by close examination of specimens captured at various time intervals. To depict the entire aging process, a sufficient sample is needed for each 1-year classthrough maturity. In gopher tortoises, 5-10 per class is adequate to determine the overall pattern of growth, but not to estimate accurately increments between yearly age classes. The advantages of this method over the capture-recapture procedure are: (1) data can be gathered each time a live or dead specimen is available; (2) seasonal growth can be estimated by comparing sizes of the in- dividuals captured during various months of the year. This method is also the most practical means of discerning trends in volumetric increase, as adequate data are easily obtained. The utility of annuli measurement-prediction equations is·also dependent upon accuracy in discerning rings. Total annuli width on the abdominal scutes was considered to be the best measurement for our purposes because (1) it takes into account the breadth of corresponding (paired) abdominal annuli; (2) it provides more precise points of measurement in comparison to other dimensions of that scute; (3) rings on the other plastral scutes are'difficult to measure because of their irregular shapes and greater susceptibility to abrasion. Annuli are also deposited on the carapace, but measurements made on the convex shields are somewhat in- consistent from one tortoise to the next and false rings are not so easily detected as on the plastron. Resulting equations are valuable in estimating linear dimensions in past years, but not volume; the logarithmic relationship involvedand the variable nature of the data prevent predicting volume from annuli measurements within acceptable limits. The advantages of this method are: (1) few specimens are required to obtain an adequate data set; (2) both preserved and live specimens can be utilized. Growth histories can also be used t6 compare the effects of differing environmental conditions over long periods of time. Studies of growth and maturity of this and similar species are most thorough when all three methods can be employed. Capture-recapture data provide an essential check of results from other approaches. Seasonal growth and volumetric increase are most efficiently deter- mined by age-dimension correlation, and long-term environmental effects and individual growth histories by the annuli-prediction equations. 1982 LANDERS ET AL.: S.W. GEORGIA GOPHER TORTOISES 109 LITERATURE CITED Alford, R. A. 1980. Population structure of Copherus polyphemus in northern Florida. J. Herpetol. (14):In press. Auffenberg, W. 1966. On thecourtship of Gopherus pollmhemus..Herpetologica 22:113-117. Auffenberg, W., and R. Franz. In press. The status and distribution 6f Gopherus pot!/- phemus. In R. B. Bury (ed.), North American Tortoises: Conservation and Ecology. Wildl. Res. Rept. 12, , and J. B. Iverson. 1979. Demography of terrestrial turtles. Pages 541-569 in Turtles: Research and Perspectives. Wiley-Interscience, New York. , and W. G. Weaver, Jr. 1969. Gopherus beriandieri in southeastern Texas. Bull. Florida State Mus., Biol. Sci. 13(3):141-203. Berry, K. H. 1976. A comparison of size classes and sex ratios in four populations of the desert tortoise. Proc. 1st Ann. Symp., Desert Tortoise Council, Las Vegas, Nevada. Burge, B. L.,and W. G. Bradley. 1976. Population density, structure, and feeding habits of the desert tortoise, Gopherus agassizii, in a low desert study area in southern Nevada. Proc. 1st Ann. Symp., Desert Tortoise Council, Las Vegas, Nevada. Cagle, F. R, 1954, Observations on the life cycles of painted turtles (genus Chrysemys). Am. Midl. Nat. 52:225-235. Carr, A. 1952. Handbook of turtles. Cornell University Press, Ithaca. 542 p. Causey, M. K., and C. A. Cude. 1978. Feral dog predation of the gopher tortoise, Copherus polyphemus (Reptilia, Testudines, Testudinidae) in southeast Alabama. Herp. Review 9(3):94-95. Christiansen, J. L., and R. R. Burken. 1979. Growth and maturity of the snapping turtle , (Chelgdra serpentina) in Iowa. Herpetologica 35(3):261-266. , and E. 0. Moll. 1973. Latitudinal reproductive variation within, a single sub- species of painted turtle, Chr!/semys picta bellii. Herpetologica 29(2): 152-163. Cochran, D. M. 1952. Nature's tank, the turtle. Nat. Geogr 101:665-684. Coe, M. J., D. Bourn, and I. R. Swingland. 1979. The biomass, production and carrying capacity of giant tortoises on Aldabra. Phil. Trans. R. Soc. Lond. B. 286: 163-176. Conant, R. 1958. A field guide to reptiles and amphibians. Houghton Mifflin Co., Boston. 366 p. Ditmars, R. L. 1946. The reptiles of North America. Doubleday and Company. Inc., Garden City, New York. 476 p. Douglass, J· F. In press. The mating system of the gopher tortoise, Gopherus polyphemus, in southern Florida. Bull. Florida State Mus., Biol. Sci. , and J. N. Layne. 1978. Activity and thermoregulation of the gopher tortoise (Gopherus polyphemus) in southern Florida. Herpetologica 34(4):359-374. Ernst, C. H. 1971. Growth of the painted turtle, Chrysemys picta, in southeastern Pennsyl- vania. Herpetologica 27:135-141. , R. W. Barbour, E. M. Ernst, and J. R. Butler. 1973, Growth of the mud turtle, Kinosternon subrubrum, in Florida. Herpetologica 29(3):247-250. Franz, R., and W. Auffenberg. 1978. The gopher tortoise: a declining species. Pages 61-63 in R. R. Odum and L. Landers (eds.). Proc. Rare and Endangered Wildl. Symp., Georgia Dept. Nat. Res., Game and Fish Div. Tech. Bull. WL#, Gibbons, J. W. 1967. Variation in growth rates in three populations of the painted turtle, Chrysemus picta. Herpetologica 23(4): 296-303. . 1968. Growth rates of the common snapping turtle, Chelydra serpentina, in a polluted river. Herpetologica 24(3):266-267. Coin,CJ·, and C.C. Goff. 1941, Notes on the growth rate of the gopher turtle, Copherus polyphemus. Herpetologica 2:66-68. 110 BULLETIN FLORIDA STATE MUSEUM Vol. 27, No. 2 Grubb, P. 1971. The growth, ecology and population structure of giant tortoises on Aldabra. Phil. Trans. R. Soc. Lond. B. 260:327-372. Hallinan, T. 1923. Observations made in Duval County, northern Florida, on the gopher tortoise (Gopherus polyphemus) Copeia 1923(115):11-20. Hubbard, H. C. 1894. The insect guests of the Florida land tortoise. Insect Life 6:302-315. Hutt, A. 1967. The gopher tortoise, a versatile vegetarian. Fla. Wildl. 21(7):20-24. Iverson, J. B. 1980. The reproductive biology of Gopherus polyphemus (Chelonia:Testu- dinidae). Amer. Midl. Nat. 103(2):353-359. Jackson, C, G,, Jr,, C, M, Holcomb, and M. M. Jackson. 1974. Aortic calcification, serum calcium, magnesium, sodium and cholesterol in Gopherus polyphemus. Comp. Bio- chem. Physiol. 49A:603-605. , T. H. Trotter, J. A. Trotter, and M. W. Trotter. 1978. Further observations of growth and sexual maturity in captive desert tortoises (Reptilia:Testudines). Herpeto- logica 34(2):225,227. Landers, J. L., J. A. Garner, and W. A. McRae. 1980. Reproduction of gopher tortoises (Gopherus polyphemus) in southwestern Georgia. Herpetologica 36(4):353-361. Landers, J. L., and D. W. Speake. In press. Management needs of sandhill reptiles in south- ern Georgia. Proc. Assoc. Fish Wildl. Agencies 34 (in press). Legler, J. M., and R. C. Webb. 1961. Remarks on a collection of Bolson tortoises, Gopherus flavomarginatus. Herpetologica 17:26-37. MeRae, W. A., J. L. Landers, and C. D. Cleveland. 1981. Sexual dimorphism and develop- ment in the gopher tortoise (Gopherw polyphemus). Herpetologica 37(1):46-52. Miles, L. E. 1953. The desert tortoise. Audubon Mag. 55(4):172-175. Moll, D. 1976. Environmental influence on growth rate in the Ouachita map turtle, Graptemys pseudogeographica ouachitensis. Herpetologica 32(4):439-443. Mosimann, J. E. 1958. An analysis of allometry in the chelonian shell. Rev. Can. Biol. 17(2):137-228. Petter, A. J., and J. F. Douglass:1976. Etude des populations d'Oxyures du c61on des Gopherus (testudinidae). Bull. Mus. Nat. Hist. Nat. (Paris) SZr. 3, No. 389, Zoologie 271:731-768. Sexton, 0. J. 1959. A method of estimating the age of painted turtles for use in demographic studies. Ecology 40(4):716-718. Swingland, I. R. 1977. Reproductive effort and life history of the Aldabran giant tortoise. Nature 269:402-404. Tinkle, D. W. 1961. Geographic variation in reproduction, size, sex ratio and maturity of Sternothaerus odoratus (Testudinata:Chelydridae). Ecology 42:68-76. Woodbury, A. M., and R. Hardy. 1948. Studies of the desert tortoise, Gopherus agassizii. Ecol. Monogr. 18: 145-200. Contributions to the BULLETIN OF THE FLORIDA STATE MUSEUM, BIOLOGICAL SCI- ENCEs SERIES, may be in any field of biology. Manuscripts dealing with natural history of systematic problems involving the southeastern United States or the New World tropics are solicited especially. Manuscripts should be of medium length-circa 35 to 200 pages (10,500- 16,00() words). Examination for suitability is made by an Editorial Board. The BULLETIN is distributed worldwide through institutional subscriptions and exchanges. It is considered the responsibility of the author to distribute his paper to all interested individ- uals. To aid in this the author(s) receive(s) 50 copies free, and he may purchase additional separates at cost if ordered when page proof is returned. The author is also responsible for any charges incurred for alterations made by him on galley or page proofs. The Museum will send an invoice to the author for this amount upon completion of publication. PREPARATION OF MANUSCRIPT Contributors should consult recent numbers of the BULLETIN for preferred style and format. Highly recommended as a guide is the CBE Style Manual, 3rd Edition, 1972 (Washington, D.C., Amer. Inst. Biol. Sci.). MSS must be submitted in duplicate (please no onionskin) and satisfy the following minimal re- quirements: They should be typewritten, double-spaced (especially tables, figure captions, and "Literature Cited"), on one side of numbered sheets of standard (8-1/2 x11 in.) bond paper, with at least one-inch margins all around. Figure legends and tables should be typed on separate sheets. All illustrations are referred to as figures. They must comply with the following stan- dards: Photographs should be sharp, with good contrast, and printed on glossy paper. Drawings should be made with dense black waterproof ink on quality paper or illustration board. All il- lustrations should have a cover sheet. All lettering will be medium weight, sans-serif type (e.g. F,itura Medium, News Gothic) in cutout, dry transfer, or lettering guide letters, Make allow- anee so that after reduction no lowercase letter will be less than 1 mm high (2 mm is preferred) nor any capital letter greater than 5 Inm high. The maximum size for illustrations is 86/8 x 14 in. (twice typepage size); illustrations should not be less than typepage width (4-5/16 in.). Designate the top of each illustration and identify on the back with soft pencil by author's name, MS title, and figure number. All manuscripts not submitted in BULLETIN format will be returned to the author for retyping. Mamiscripts tind all editorial matters should be addressed to: Managing Editor of the BULLETIN Florida State Museum Museum Road University of Florida Gainesville, FL 32611