Home Range and Philopatry in the Ornate Box Turtle, Terrapene ornata ornata, in Iowa

Am. Midl. Nat. 157:162 174 Home Range and Philopatry in the Ornate Box Turtle, Terrapene ornata ornata, in Iowa NEIL P. BERNSTEIN 1 AND REBECCA J. RICHTSMEIER Department of Biology, Mount Mercy College, Cedar Rapids, Iowa 52402 ROBERT W. BLACK Department of Biology, Cornell College, Mount Vernon, Iowa 52312 AND BENJAMIN R. MONTGOMERY Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor 48100 ABSTRACT. We quantified home range size and philopatry in ornate box turtles, Terrapene ornata ornata, with minimum convex polygons (MCP) and 95% Kernel areas (K) calculated from radio telemetry data in Iowa during 1999 2001. We also analyzed philopatry by markrecapture data during 1994 2000. Ten adult turtles were followed every other day from May to Sep. 2000 and seven of the turtles were followed in 2001. Mean monthly MCP and K home ranges of males and females did not significantly differ. Mean annual MCP and K home ranges for males were significantly larger than those of females. For an individual, monthly MCP home ranges overlapped on average 19.3% between years and K home ranges overlapped an average of 6.0%. There were no significant differences in annual or monthly home range overlap between sexes for K home ranges, but MCP monthly home range overlap was significantly larger for males. Annual MCP home ranges overlapped on average 40.9% between years and K annual home range overlap averaged 6.8%. Seven years of markrecapture data were analyzed to determine movements between habitat zones. Turtles were found in sand prairie or shrubby border to sand prairies 69% of the time. Regardless of sex, turtles were twice as likely to be recaptured in the same sub-habitat from where they initially were found compared to the adjacent sub-habitat. Turtles were 2 3 times more likely to be in the adjacent sub-habitat than two sub-habitats away. Given the degree of philopatry exhibited by ornate box turtles in this area, turtle populations faced with habitat fragmentation or destruction will most likely decline. INTRODUCTION Studies of home range in box turtles are important for not only knowledge of basic ecology (Dodd, 2001), but also for making management and conservation decisions (Swarth, 2005; Redder et al., 2006). Dodd (2001) and Redder et al. (2006) summarized some of the numerous studies documenting home range size in eastern box turtles (Terrapene carolina carolina), as well as ornate box turtles (Terrapene ornata ornata), and noted variation in reported home ranges. However, information on seasonal habitat use and gender differences in home ranges is often lacking for box turtle populations (Swarth, 2005) and data are still needed on turtle behavior in relation to movements and habitat use (Maret et al., 2005). Several multi-year studies documented home range fidelity and philopatry in individual box turtles (e.g., Nichols, 1939; Metcalf and Metcalf, 1970; Madden, 1975; Stickel, 1950, 1989; Nieuwolt, 1996; Hall et al., 1999). Although studies have documented homing ability in displaced box turtles (e.g., Legler, 1960; Metcalf and Metcalf, 1970, 1978; Schwartz and Schwartz, 1974; DeRosa and Taylor, 1980; Holy, 1 e-mail: nbernstein@mtmercy.edu 162

2007 BERNSTEIN ET AL.: ORNATE BOX TURTLE 163 1995), homing ability has not been tested rigorously in natural conditions. Recently, however, philpatry was demonstrated in the common musk turtle (Sternotherus odoratus) (Andres and Chambers, 2006). Whereas Terrapene ornata ornata once ranged over a large area of eastern Iowa (Richtsmeier et al., 2007), extant populations are now separated and surrounded by non-native habitats. To guide conservation efforts for this threatened species in Iowa, it is important to understand habitat use and home ranges in an increasingly fragmented landscape that exacerbates several potential threats to long term survivorship of T. o. ornata populations (Redder et al., 2006). Therefore, we established a means to accurately quantify year-to-year home range fidelity. Our objectives were to: (1) quantify size of home ranges and the percent overlap in home ranges within individual turtles between years and to (2) quantify philopatry as related to habitat selection. Quantification of these objectives will assist in evaluating threats to a localized population. METHODS Studies of Terrapene ornata ornata began in 1994 within a 75 ha section of the Hawkeye Wildlife Area of Johnson Co, Iowa, a complex managed by the Iowa Department of Natural Resources. The area, located on the south side of the Iowa River, contains a series of sand dunes used by the turtles. Whereas the habitat is relatively undisturbed to the north, farmland adjoins the area to the south and, in part, to the east and west. From 1994 1997 R. Rhodes searched for turtles, notched marginal carapace scutes using an identification system adapted from (Cagle, 1939) and began collecting mark-recapture data. In 1998 three of us (Bernstein, Black, Richtsmeier) began monitoring turtle movements with radio telemetry, continued the system of mark-recapture started by Rhodes and initiated Geographical Information System (GIS) analysis. As of 2006 over 600 turtles have been identified in this area, making this the second largest population in eastern Iowa (Richtsmeier et al., 2007). Rhodes identified habitats of interest by dividing the study area into seven macro-habitats based on large differences in spatial location. The macro-habitats were further divided into smaller sub-habitats based on finer habitat distinctions (e.g., prairie, wetland, forest) and spatial distinctions (e.g., slope). Although this approach resulted in units of unequal size and shape (0.2 3.66 ha, mean 5 1.33 ha), it had the advantage of being more biologically meaningful relative to a simple grid that crossed differing habitats. GIS analysis. We affixed radio transmitters (approximately 9.5 g, Advanced Telemetry Systems, Isanti, MN) to either the front or rear costal scutes of the carapace with 5-min epoxy. The transmitter was covered with an adhesive tape which secured a temperature datalogger used in another study (Bernstein and Black, 2005) and the total weight of the unit was 3 to 4.5% of adult body weight. Transmitters were placed off-center with the antenna trailing behind the turtle; observations indicated that copulation, nesting and burrowing were not hindered by this placement. Turtles were tracked with a 3-element yagi antenna and an FM100 receiver (Advanced Telemetry Systems, Isanti, MN). Transmitter loss within the winter burrow as well as during the summer activity resulted in an incomplete data set. Overall, of the10 adult turtles monitored from 1999 to 2000, seven of were also monitored from 2000 to 2001. Turtles were located every other morning from 1 May to 30 Sep. during 1999 2001. Multiple observations at constant time intervals should increase accuracy of home range estimates (DeSolla et al., 1999). Tests of Global Positioning Systems (GPS) proved too

164 THE AMERICAN MIDLAND NATURALIST 157(1) FIG. 1. May and Jun. MCP home ranges of three ornate box turtles (242, 348, 355) showing home range overlap, as well as movement between habitats (AG 5 Agriculture, F 5 forest, P 5 Prairie, PS 5 Prairie Shrub, S 5 Shrub, W 5 Wetland, WS 5 Wetland Shrub) inaccurate (610 to 30 m) to transfer locations to Geographical Information Systems (GIS). Therefore, teams of researchers carried copies of the identical aerial photo used in GIS analysis (3 m accuracy) and marked the position of each turtle based upon recognizable landmarks within 5 to 10 m of each other (e.g., trees, wetland shorelines, vegetation borders, fence lines) (Fig. 1). This method resulted in high accuracy because specific locations of turtles within the habitat were manually transferred to the computer photo in ArcView 3.2 (ESRI, Redlands, CA) rather than downloaded from a GPS. Monthly and overall yearly home ranges were calculated as Minimum Convex Polygons (MCP) and 95% Kernel (K) in ArcView 3.2 with the Animal Movement extension developed by (Hooge and Eichenlaub, 2000). MCP and K home ranges were compared with a paired t- test. Percent overlap was calculated using the Geoprocessing feature in ArcView 3.2 to calculate percent overlap of two polygons by the formula: Area Intersect AreaPolygon 1 z Area Polygon 2 { Area Intersect 100 To test for differences between individual annual home ranges and for differences based on sex, we used a MANOVA to account for the repeated-measures aspect of the data using only turtles with complete data sets for all three years. To test for differences in monthly home ranges and whether there were differences based on sex, MANOVA was conducted on the average, individual monthly home ranges for individuals with complete data sets for all 3 y. Mean male and female monthly and yearly percent home range overlaps were

2007 BERNSTEIN ET AL.: ORNATE BOX TURTLE 165 compared with a t-test following a square root, arcsine transformation of percentages. Significance was assessed at an alpha of 0.05; analyses were run using (SPSS, 2004). Philopatry to habitat zones. During 1994 1997 searches for turtles were largely conducted by Rhodes approximately once a week during the spring, summer and fall. After 1996 turtles were less visible late Jun. mid-aug. Subsequently, searches were not as frequent during those months. During 1998 2001 we continued marking and recapturing turtles in conjunction with our radio telemetry study. Several focused searches were conducted with teams of students and volunteers in spring of some years, but the majority of effort involved one to two researchers searching for turtles every other day. We also noted that turtles without radios were inconspicuous in mid-summer. For all of the above, when a marked turtle was found, the date, as well as the macro- and sub-habitat zone was noted. Unmarked turtles were notched, and the above data were recorded for each new turtle. Mark-recapture data were analyzed from 1994 to 2000 data based on four categories of recapture: (1) in the same sub-habitat, (2) a sub-habitat immediately adjacent, (3) two sub-habitats adjacent and (4) more than two sub-habitats away from sub-habitat initially marked. These categories were analyzed with a goodness of fit test in three ways: (1) total of all recaptures both within and between years, (2) only recaptures between years and (3) recaptures between years that occurred within three blocks of time representing: (a) breeding, (b) nesting/summer and (c) fall/return to overwintering sites (Apr. May, Jun. Jul., Aug. Sep., respectively). Males and females were analyzed separately because analysis of GIS data indicated statistically significant differences in male and female home ranges For the tests to be valid, turtles must have uninhibited access to all habitats. As (Dodd, 2001) noted, box turtles are not territorial; we also observed overlap in individual home ranges between (Fig. 1). Further, there were no obstacles to movement between habitats and subhabitats, and turtles frequently used multiple habitats within the same month (Fig. 1). RESULTS GIS analysis. There were no significant differences in MCP annual home range areas between years (MANOVA MCP, F 5 0.559, df 5 2, P 5 0.58), but males had significantly larger mean annual home ranges (58,006 m 2, S. E. 5 8360) than did females (26,927 m 2, S. E. 5 4471) (MANOVA MCP, F 5 12.3, df 5 1, P, 0.002) (Table 1). Similarly, there were no significant differences between annual K home ranges (MANOVA K, F 5 1.414, df 5 2, P 5 0.27), and males also had larger mean annual home ranges (10,763 m 2, S. E. 5 911) than did females (8059 m 2, S. E. 5 376) (MANOVA K, F 5 9.11, df 5 1, P, 0.07) (Table 1). Therefore, further analyses combined years and analyzed sexes separately. Mean MCP monthly home ranges were significantly larger than K (MCP 5 8638 m 2, K 5 2607 m 2 ) (paired t-test, t 5 7.93, df 5 127, P, 0.0001) as were mean annual home ranges (MCP 5 34,858 m 2, K 5 8727 m 2 ) (paired t-test, t 5 6.10, df 5 25, P, 0.0001). Therefore, MCP and K areas were analyzed separately. Mean monthly MCP areas were not significantly different between males (7812 m 2, S.E. 5 1465) and females (8062 m 2, S.E. 5 1487) (MANOVA MCP, df 5 1, F 5 1.05, P 5 0.33) (Tables 1, 2). Additionally, no significant difference between months was found (MANOVA MCP, F 5 0.32, df 5 4, P 5 0.86. (Tables 1, 2). For K, mean monthly areas were also not significantly different between males (2860 m 2, S.E. 5 128) and females (2590 m 2, S.E. 5 179) (MANOVA K, F 5 1.23, df 5 1, P 5 0.28) (Tables 1, 2); there were significant differences in the K monthly home ranges (MANOVA K, F 5 3.10, df 5 4, P, 0.05 (Tables 1, 2).

166 THE AMERICAN MIDLAND NATURALIST 157(1) TABLE 1. Individual MCP (P) and Kernel (K) mean home ranges (m 2 ) in Terrapene ornata ornata from 1999 2001 I.D./sex May Jun. Jul. Aug. Sep. Annual 32F P 20125 6 4711 1 9038 6 3489 1679 6 1080 2825 6 1353 7097 6 4427 49449 6 6505 K 3050 6 333 3788 6 294 2075 6 427 1827 6 568 2120 6 751 7890 6 221 48M P 18474 6 2686 6922 6 2762 7668 6 429 9370 6 2192 11357 6 4269 34279 6 2045 K 3055 6 334 3153 6 451 2781 6 419 2680 6 412 3283 6 412 9279 6 433 65F P 8440 6 4017 12723 6 11024 2480 6 140 4395 6 294 4231 6 3731 27056 6 2445 K 2390 6 554 2799 6 500 1580 6 190 2391 6 237 1594 6 720 7462 6 1018 66F P 8416 6 7271 11410 6 531 1538 6 403 1695 6 348 3423 6 622 48324 6 1398 3 K 1961 6 893 3746 6 18 2383 6 446 1882 6 93 1899 6 506 8931 6 890 94M P 17542 6 16665 32446 6 3144 1345 6 130 NA 2 NA NA K 2562 6 1522 2832 6 728 1395 6 354 242F P 5141 6 1048 8106 6 2216 5329 6 1126 3762 6 1961 1986 6 499 16405 6 2822 K 2999 6 324 3457 6 92 2640 6 57 2148 6 190 2406 6 517 9528 6 567 348M P 7949 6 1218 16654 6 5820 10911 6 5551 6625 6 4290 NA 25910 6 2970 3 K 2801 6 279 4092 6 581 2469 6 66 2034 6 95 7912 6 464 355F P 4838 6 3865 6558 6 2263 1836 6 445 775 6 388 3965 6 858 14800 6 999 K 2239 6 889 3063 6 335 1742 6 167 1651 6 471 2413 6 560 7800 6 262 477M P 10361 6 2497 11586 6 6416 15204 6 6226 9511 6 1851 15545 6 10597 59094 6 14398 K 2174 6 398 3177 6 332 3571 6 258 2634 6 88 3325 6 678 12933 6 260 478M P NA 5724 6 446 6681 6 4691 2018 6 563 9183 6 4707 80717 6 9350 K 3291 6 135 3107 6 603 2450 6 690 2699 6 608 10079 6 2477 Mean P 11313 6 1788 1164 6 1744 5564 6 1182 6113 6 1622 7653 6 1917 Monthly K 2556 6 162 3316 6 130 2407 6 162 2216 6 204 2481 6 204 1 One Standard Error 2 Not Available. Transmitter loss prevented multiple readings. Overall calculated from remaining months when possible 3 Based on 2 y data only

2007 BERNSTEIN ET AL.: ORNATE BOX TURTLE 167 TABLE 2. Male and female Terrapene ornata ornata mean monthly minimum convex polygon (P) and kernel (K) home ranges (m 2 ) Male Female May P 15310 6 2244 1 9817 6 2243 K 2518 6 211 2594 6 211 Jun. P 13146 6 2244 9482 6 2162 K 3286 6 211 3344 6 204 Jul. P 8705 6 2244 2708 6 2162 K 2777 6 211 2063 6 204 Aug. P 10223 6 2439 2636 6 2162 K 2524 6 230 1956 6 211 Sep. P 12219 6 2559 3962 6 2162 K 3013 6 241 2100 6 204 1 One Standard Error There were no significant differences among individual turtles for MCP or K mean monthly areas (MANOVA MCP, F 5 3.0, df 5 3, P 5 0.06; MANOVA K, F 5 1.93, df 5 3, P 5 0.06) and variation in home ranges between sexes, as well as between months for the same individual, was evident (Table 1). MCP monthly home range overlap varied 0 54.9% (mean 5 19.3%, SE 5 1.8) and annual overlap varied 20.3 60.0% (mean 5 40.9%, SE 5 3.0) (Table 3). Mean monthly overlap for males (23.7%, SE 5 2.7) differed significantly from females (16.7%, SE 5 2.3) (t 5 2.08, 68 df,, P, 0.042). However, there was no significant difference between males (45.2%, SE 5 5.3) and females (37.6%, SE 5 3.3) in annual percent overlap (t 5 1.176, df 5 14, P 5 0.50). K monthly home range overlap varied 0 28.1% (mean 5 6.0%, SE 5 0.7) and annual overlap varied,1 31.2% (mean 5 6.8%, SE 5 2.6) (Table 3). Mean monthly overlap for males (3.7%, SE 5 0.9) was not significantly different then females (4.6%, SE 5 0.9) (t 5 0.55, df 5 70, P 5 0.59). Additionally, there was no significant difference in mean annual overlap between males (5.6%, SE 5 3.0) and females (7.7%, SE 5 4.2) (t 5 0.37, df 5 14, P 5 0.70). Not apparent from these data was that turtles with smaller monthly home ranges tended to have lower percent overlap. An example would be turtle 32 which, between Jul. 1999 and 2000, had no MCP overlap, minimal K overlap (Table 3), and also had relatively small home ranges during those months (Table 1). In this case, whereas there was no overlap in home ranges, the monthly areas were within 25 m of each other (i.e., in the same subhabitat as the previous year). A similar situation occurred for another female in previous years (unpubl. data). In non-consecutive years, this female moved in an area,25 m radius from late-may to early-sep., the smallest home range we have recorded, but although these two home ranges were closer than 50 m apart, she never intersected her previous year s area. Such relatively small distances between areas used more greatly affected K overlap analysis (e.g., 48 M, Table 3) and kernel home ranges were often within meters of each other without overlapping. The MCP and K analysis resulted in four, seemingly illogical conclusions of no MCP overlap but K overlap (Jul. 99 00, 32F; May, Jul., Aug. 99 00, 355F). This occurred when MCP polygons were close, without intersecting, but the K area generated around adjacent, but separate points, intersected.

168 THE AMERICAN MIDLAND NATURALIST 157(1) TABLE 3. Minimum convex polygon (P) and Kernel (K) overlap (m 2 ) and percentage overlap in Terrapene ornata ornata home ranges during 1999 2000 and 2000 2001 May Jun. Jul. Aug. Sep. Annual I.D./sex 99 00 00 01 99 00 00 01 99 00 00 01 99 00 00 01 99 00 00 01 99 00 00 01 32F P 2755 (9.0) 9875 (24.8) 430 (25.5) 2732 (30.6) 0 (0) 377 (8.2) 62 (1.4) 1763 (26.8) 27 (,1.0) 3880 (22.4) 15686 (20.7) 31359 (40.0) K 121 (1.8) 1 (,1.0) 474 (5.8) 9 (,1.0) 12 (,1.0) 3 (,1.0) 386 (15.7) 107 (2.3) 0 (0) 0 (0) 14 (,1.0) 257 (2.3) 48M P 25584 (44.7) 8130 (24.3) 2847 (20.1) 2956 (54.9) 5331 (53.6) 3053 (26.4) 566 (32.9) 3851 (27.4) 2562 (2.0) 7918 (1.7) 24524 (60.0) 21683 (42.6) K 0 (0) 265 (4.9) 27 (,1.0) 92 (1.6) 115 (2.5) 0 (0) 321 (5.5) 6 (,1.0) 197 (3.4) 0 (0) 819 (6.6) 148 (,1.0) 65F P 6258 (36.3) 810 (4.9) 1495 (40.9) 1120 (3.1) 1495 (40.9) 457 (11.9) NA NA 0 (0) 344 (4.2) 13251 (29.6) 12992 (35.5) K 578 (12.1) 0 (0) 361 (5.8) 598 (13.7) 274 (10.9) 76 (2.8) NA NA 0 (0) 424 (12.2) 42 (,1.0) 14 (,1.0) 66F P 1103 (4.0) NA 1890 (9.0) NA 0 (0) NA 0 (0) NA 123 (1.8) NA 19209 (29.8) NA K 115 (3.0) NA 78 (1.1) NA 0 (0) NA 0 (0) NA 0 (0) NA 86 (,1.0) NA 94M P 876 (2.6) NA 13098 (25.3) NA NA NA NA NA NA NA NA NA K 0 (0) NA 0 (0) NA NA NA NA NA NA NA NA NA 242F P 1908 (24.9) 2762 (44.8) 4527 (3.5) 4997 (36.2) 1316 (11.4) 2477 (32.7) 276 (3.0) 1819 (23.7) 47 (1.4) 975 (31.8) 12486 (38.5) 11979 (48.8) K 862 (16.1) 261 (5.2) 154 (2.3) 259 (3.8) 0 (0) 285 (5.7) 29 (,1.0) 227 (5.1) 274 (8.0) 4235 (28.1) 4235 (28.1) 4735 (31.2) 348M P 4255 (36.6) NA 10834 (48.2) NA 2068 (10.5) NA 2255 (20.5) NA NA NA 19349 (59.6) NA K 77 (1.4) NA 4 (,1.0) NA 65 (1.3) NA 68 (1.7) NA NA NA 92 (,1.0) NA 355F P 0 (0) NA 3700 (27.9) 981 (13.8) 0 (0) 1038 (33.7) 0 (0) 0 (0) 901 (17.0) 2361 (36.4) 9290 (50.5) 9338 (45.2) K 133 (3.1) NA 31 (,1.0) 0 (0) 219 (6.6) 357 (12.7) 126 (5.5) 0 (0) 168 (4.7) 444 (8.6) 351 (2.3) 4 (,1.0) 477M P 6251 (39.6) 2103 (9.5) 5801 (23.6) 4236 (17.3) 5155 (15.0) 6031 (22.2) 1621 (8.2) 3411 (28.5) 9693 (19.3) 27596 (47.5) 42814 (48.5) 44181 (49.4) K 176 (3.9) 0 (0) 4 (,1.0) 77 (1.1) 322 (4.4) 130 (2.0) 217 (4.3) 393 (7.8) 198 (3.9) 449 (6.6) 9 (,1.0) 200 (,1.0) 478M P 4448 (8.8) 3320 (28.1) 1978 (3.9) 6059 (12.7) 2671 (7.8) 5007 (8.6) 1061 (35.7) NA 2330 (14.5) NA 26363 (20.3) 47267 (35.7) K 0 (0) 342 (8.4) 277 (3.8) 0 (0) 705 (11.6) 54 (,1.0) 1165 (31.2) NA 528 (10.9) NA 4234 (22.3) 1369 (6.6)

2007 BERNSTEIN ET AL.: ORNATE BOX TURTLE 169 FIG. 2. All recaptures of ornate box turtles within and between years, 1994 2000 Philopatry to habitat zones. As of summer 2005, 597 adult turtles had been marked, over 95% within 75 ha. Population density was estimated to be 8.0 turtles/ha after subtracting turtles known to be dead. More conservative estimates were calculated by also subtracting turtles only observed once, potential transients. This yielded a density estimate between 5.1 6.4 turtles/ha for turtles recaptured in at least 2 or 3 y (483 and 381, respectively). These individuals were found in sand prairie or shrubby border to sand prairie habitats 69% of the time. Analysis of all mark-recapture data, between and within years, indicated that males and females were twice as likely to be recaptured in the same sub-habitat in which they were initially found when compared to one sub-habitat adjacent (Fig. 2). In addition, males were three times more likely to be one sub-habitat away from initial capture than two blocks away, and recaptures more than two sub-habitats away were rare (Fig. 2). Similarly, females were twice as likely to be recaptured one sub-habitat away compared to two sub-habitats away, and, like males, recaptures more than two sub-habitats away were rare (Fig. 2). All of these patterns were significant (males: x 2 5 368, df 5 3, P, 0.001; females: x 2 5 340, df 5 3, P, 0.001). Analysis of recapture between years yielded similar results (Fig. 3). Both males and females were more likely to be recaptured annually in either the same sub-habitat or one sub-habitat adjacent when compared to two or more than two sub-habitats away (males: x 2 5 97, df 5 3, P, 0.001; females: x 2 5 96, df 5 3, P, 0.001). Results were similar for recapture data between years and within the same 2-mo block of time (Fig. 4). Both sexes were more likely to be recaptured in either the same sub-habitat or one sub-habitat adjacent when compared to two or more than two sub-habitats away (males: x 2 5 86, df 5 3, p, 0.001; females: x 2 5 85, df 5 3, P, 0.001). Philopatry may also extend to specific behaviors within the sub-habitats. In consecutive years, female 242 built nests within a meter of each other and male 477 overwintered almost adjacent to previous burrows in three consecutive years.

170 THE AMERICAN MIDLAND NATURALIST 157(1) FIG. 3. Recaptures of ornate box turtles between years 1994 2000 DISCUSSION GIS analysis. Mean annual MCP and K home range areas for males were larger than that of females, and this was similar to (Blair, 1976), who reported larger male home ranges in Terrapene ornata in Texas. (Nieuwolt, 1996) also noted a tendency for male T. ornata luteola to travel farther than females in New Mexico, however, no sexual differences in home range were found. Box turtles studied by (Stickel, 1950) in Maryland, (Legler, 1960) in Kansas, (Schwartz and Schwartz, 1974) in Missouri, (Doroff and Keith, 1990) in (Wisconsin and Holly, 1995) in Nebraska also did not find home range differences between sexes. FIG. 4. Recaptures or ornate box turtles between years within the same block of time (Apr. May, Jun. Jul., Aug. Sep.), 1994 2000

2007 BERNSTEIN ET AL.: ORNATE BOX TURTLE 171 Direct comparisons are difficult because researchers applied different methodologies to determine home range and ecological factors also affect home range calculations (Redder et al., 2006). As noted by (Stickel, 1950), some of our turtles exhibited a dumbbell home range, two separate areas of activity during the summer. However, we controlled for that behavior by analyzing monthly home ranges as well as overall yearly home range, and we also examined kernel home ranges. Whereas variation between home range monthly areas resulted in no statistical differences between males and females in monthly home ranges, males had MCP home ranges approximately 33 that of females during Jul., Aug. and Sep. However, gender differences in K home ranges during Jul. through Sep. were less pronounced. These patterns can partially be explained by mating and nesting behavior. Both sexes were more active in May and Jun. during mating and nesting. After nesting, females moved away from the open prairie into the shrub border, and, after establishing a home range in late Jun., they were relatively sedentary until Sep. when they moved to a location to overwinter. In contrast, males ranged more widely throughout the year, although they also were more sedentary after the mating period and more active before overwintering. The female movements were similar to those of spotted turtles (Clemmys guttata) in Ohio (Lewis and Faulhaber, 1999) in which female movements associated with mating and nesting were reflected in larger K home range, core activity area, in May and Jun. with smaller home ranges in Jul., Aug. and Sep. Although our annual home ranges are within those reported in other studies, our largest annual MCP home ranges were much lower than the maximum area reported for Terrapene ornata ornata (Doroff and Keith, 1990). Our MCP data were more comparable with MCP calculations for T. ornata ornata in Nebraska (Holy, 1995), although still much lower than the maximum that she reported. Our K home range areas were also within those reported in other studies where core area usage was estimated (e.g., Nieuwolt, 1996), although our data were again more comparable with the smaller home range estimates. Differences in these results could have several explanations: our sample size was larger (e.g., Blair, 1976; Holly, 1995; Nieuwolt, 1996), there generally was high variation in home range size between individual turtles (e.g., Doroff and Keith, 1990; Nieuwolt, 1996) which affected home range calculations and statistical comparisons, different methods of calculating home ranges yielded different results and are, therefore, not comparable (Nieuwolt, 1996; Dodd, 2001), or there were differences in our habitat relative to environments of other studies (Redder et al., 2006). Higher quality habitats with abundant food supplies could produce smaller home ranges with higher population densities (Stickel, 1989; Nieuwolt, 1996; Dodd, 2001). Although we have not quantified food resources, our study area does contain a variety of fruits, fungi, invertebrates and plants, and, therefore, food was probably not limited in our area. As noted, (Stickel, 1989) suggested that population density was inversely correlated with home ranges size and that both were related to habitat quality; however, (Dodd, 2001) stated that the relationship had not been demonstrated. Regarding the relationship between home range size and population density, our conservative estimate of turtle density was higher than estimates of (Doroff and Keith, 1990) (2.9 5.0 turtles/ha) and over twice the estimates by (Legler, 1960) (0.16 2.6 turtles/ha) and our maximum home range areas tended to be smaller than other studies. We know of no other study of turtles that attempted to quantify area of overlap of home ranges although (Dodd, 2001) noted the stability of individual home range size among years. For MCP areas, individual turtles in our study displayed philopatry with up to 54.9%

172 THE AMERICAN MIDLAND NATURALIST 157(1) overlap between months in consecutive years and up to 62.8% overlap in consecutive annual home ranges. Both maximum monthly and annual K percent overlap was 31.2%, but most K overlap calculations were much lower. A possible reason male and female monthly MCP philopatry differed relates to the monthly differences in home ranges and the tendency for females to be more sedentary for much of the active season. As noted, females did not travel far away from the open sand prairie after nesting and, in some cases, barely moved at all whereas males were less sedentary. However, K area overlap was much lower and did not differ between sexes. Whereas overall conclusions based on MCP and K home ranges were similar, why were there differences in amount of overlap between MCP and K, especially comparing home range overlap among years? As noted by (Horne and Garton, 2006), selecting an appropriate home range model involves understanding of complex ecological factors, and our turtles core areas of activity, K, were not precisely, geographically established compared to larger areas of habitat usage, as analyzed by MCP areas and the mark-recapture data. Philopatry to habitat zones. Goodness of fit analysis of mark-recapture data indicated that turtles exhibited a strong tendency to be recaptured either in the same sub-habitat or a sub-habitat immediately adjacent to where they were first encountered. As noted, philopatry has been reported in other box turtle studies, as has a tendency to overwinter and/or nest in the same location (Stickel, 1989; Doroff and Keith, 1990; Redder et al., 2006). Philopatry has also been reported in Clemmys guttata (Perillo, 1997) and Glyptemys insculpa (Tuttle and Carroll, 2003). When mark-recapture data were combined with the GIS analysis, we concluded that Terrapene ornata ornata exhibited philopatry for the habitat, but their movements within that sub-habitat and habitat may vary between years. Our observations provide information for conservation of box turtles in Iowa and perhaps elsewhere. Although we worked in a small portion of the Hawkeye Wildlife Area, farms and an increasing number of houses border the property. Soil maps indicate that suitable sandy substrate extends throughout the northern third of Johnson County, as well as through what is now a large city, but this habitat has been altered and fragmented since settlement. We likely studied a subset of a once widely dispersed population (Richtsmeier et al., 2007). We have noted injuries and mortality from farm machinery, and we have received reports of Terrapene ornata ornata found near new housing developments. As reported from Wisconsin (Doroff and Keith, 1990; Curtin, 1997), despite severe habitat alteration and disturbance, the turtles returned to the same areas to breed, nest and feed. As humans intrude into established box turtle home ranges, the turtles are unlikely to establish new areas of activity, placing the turtle at increased risk of injury and disturbance within its home range. Although relocating threatened turtles has been attempted, such efforts have met with mixed success (e.g., Doroff and Keith, 1990; Hatch, 1997; Belzer, 2002), and philopatry may be one factor limiting relocation efforts. Many of the turtles displaced to an enclosure in one Iowa relocation oriented towards the area from which they were taken, although other factors than homing instinct might have been involved (J. Christiansen, pers. comm.). To the best of our knowledge, the Hawkeye Wildlife Area Terrapene ornata ornata population is one of two in Iowa that contain over 600 individuals, and most other populations are well under 25 individuals. Because of philopatry, as well as the basic properties of habitat, we feel that protection of the existing habitat and surrounding areas for a population is imperative to the long term survival of T. o. ornata in Iowa. Acknowledgments. A. Bauch, C. Barker, B. Comito, T. Hornbeck, R. Nieuwendaal, H. Secrist, A. Shaw, L. Steiner, R. Stokes, C. DeVries helped in several aspects of the fieldwork, K. Brumm and J. Schubert

2007 BERNSTEIN ET AL.: ORNATE BOX TURTLE 173 analyzed some of the GIS data, P. Brown and P. Garvin provided GIS advice, J. L. Christiansen, C. A. Ribic and two anonymous reviewers reviewed earlier drafts and provided helpful criticisms and the Iowa DNR allowed us to conduct studies at Hawkeye Wildlife Area and work with the threatened species. Financial support was received from the Mount Mercy College Department of Biology and Faculty Development Program, the Cornell College Department of Biology and Faculty Development Program, the Associated Colleges of the Midwest, the Iowa Science Foundation of the Iowa Academy of Science, the Iowa College Foundation of the R. J. McElroy Trust, the Iowa Wildlife Diversity Program of the Iowa DNR, the Iowa Natural History Association,and R. Rhodes II, in whose memory the paper is dedicated. LITERATURE CITED ANDRES, K. M. AND R. M. CHAMBERS. 2006. A test of philopatry by common musk turtles. Am. Midl. Nat., 156:45 51. BELZER, B. 2002. A nine year study of eastern box turtle courtship with implications for reproductive success and conservation in a translocated population. Turtle and Tortoise Newsletter, 6:17 26. BERNSTEIN, N. P. AND R. W. BLACK. 2005. Thermal environment of overwintering ornate box turtles, Terrapene ornata ornata, in Iowa. Am. Midl. Nat., 153:370 377. BLAIR, W. F. 1976. Some aspects of the biology of the ornate box turtle, Terrapene ornata. Southwest. Nat., 21:89 104. CAGLE, F. R. 1939. A system of marking turtles for future identification. Copeia, 1939:170 173. CURTIN, C. G. 1997. Biophysical analysis of the impact of shifting land use on ornate box turtles, Wisconsin, USA. p. 31 36. In: J. van Abbema (ed.). Proceedings: conservation, restoration, and management of tortoises and turtles-an international conference. New York Turtle and Tortoise Society and the WCS Turtle Recovery Program, New York. DEROSA, C. T. AND D. H. TAYLOR. 1980. A comparison of compass orientation mechanisms in three turtles (Trionyx spinifer, Chrysemys picata and Terrapene carolina). Copeia, 1982:394 399. DESOLLA, S. R., R. BONDURIANSKY AND R. J. BROOKS. 1999. Eliminating autocorrelation reduces biological relevance of home range estimates. J. Anim. Ecol., 68:221 234. DODD, C. K., JR. 2001. North American box turtles. A natural history. University of Oklahoma Press, Norman. 231 p. DOROFF, A. M. AND L. B. KEITH. 1990. Demography and ecology of an ornate box turtle (Terrapene ornata) population in south-central Wisconsin. Copeia, 1990:387 399. HALL, R. J., P. F. P. HENRY AND C. M. BUNCK. 1999. Fifty year trends in a box turtle population in Maryland. Biol. Cons., 88:165 172. HATCH, K. 1997. Translocation, reintroduction, and habitat use of ornate box turtles. Report to the Zoological Society of Milwaukee County, Milwaukee, WI. HOLY, L. L. 1995. Home range, homing ability, orientation and navigational mechanisms of the western box turtle (Terrapene ornata) from western Nebraska. M.S. thesis. University of Nebraska, Lincoln. 140 p. HOOGE, P. N. AND W. M. EICHENLAUB. 2000. Animal movement extension to ArcView. Version 2. Alaska Biological Science Center, U.S. Geological Survey, Anchorage, Alaska. http://www.absc.usgs. gov/gistools/. HORNE, J. S. AND E. O. GARTON. 2006. Selecting the best home range model: an information-theoretic approach. Ecology, 87:146 1152. LEGLER, J. M. 1960. Natural history of the ornate box turtle, Terrapene ornata ornata Aggisiz. Univ. Kansas Publ. Mus. Nat. Hist., 11:527 669. LEWIS, T. L. AND C. A. FAULHABER. 1999. Home ranges of spotted turtles (Clemmys guttata) in southwestern Ohio. Chelon. Cons. Biol., 3:430 434. MADDEN, R. 1975. Home range, movements, and orientation in the eastern box turtle, Terrapene carolina carolina. Ph.D. Diss. The City University of New York, New York. 217 p. MARET, T., J. MITCHELL, JR. AND I. L. BRISBIN. 2005. Research Needs. p. 13 14. In: C. Swarth and S. Hagood (eds.). Summary of the eastern box turtle regional conservation workshop. The Humane Society of the United States, Washington, D.C.

174 THE AMERICAN MIDLAND NATURALIST 157(1) METCALF, A. L. AND E. METCALF. 1970. Observations on ornate box turtles (Terrapene ornata ornata Aggisiz). Trans. Kansas Acad. of Sci., 73:96 117. AND. 1978. An experiment with homing in ornate box turtles (Terrapene ornata ornata Aggisiz). J. Herp., 12:411 412. NICHOLS, J. T. 1939. Range and homing of individual box turtles. Copeia, 1939:125 127. NIEUWOLT, P. M. 1996. Movement, activity, and microhabitat selection in the western box turtle, Terrapene ornata luteola, in New Mexico. Herpetologica, 52:487 495. PERILLO, K. M. 1997. Seasonal movements and habitat preferences of spotted turtles (Clemmys guttata) in north central Connecticut: a status report. p. 24 25. In: T. F. Tyning (ed.). Status and conservation of turtles of the northeastern United States. Massachusetts Audubon Society, Lincoln, Massachusetts. REDDER, A. J., C. K. DODD, JR. AND D. KEINATH. 2006, May 16. Ornate box turtle (Terrapene ornata ornata): a technical conservation assessment [Online]. USDA Forest Service, Rocky Mountain Region. Available: http://fs.fed.us/r2/projects/scp/assessments/ornateboxturtle.pdf. Accessed 30 Jun. 2006. RICHTSMEIER, R. J., N. P. BERNSTEIN, J. W. DEMASTES AND R. W. BLACK. 2007. Migration, gene flow, and genetic diversity within and among populations of ornate box turtles (Terrapene ornata ornata). Chelon. Cons. and Biol., 8: In press. SCHWARTZ, C. W. AND E. R. SCHWARTZ. 1974. The three-toed box turtle in central Missouri. Missouri Department of Conservation Terrestrial Series #5:1 28, Jefferson, Missouri. SPSS. 2006. SPSS for Windows, Release 14. SPSS, Inc. Chicago, Illinois. STICKEL, L. F. 1950. Populations and home range relationships of the box turtle, Terrapene c. carolina (Linnaeus). Ecol. Monog., 20:351 378.. 1989. Home range behavior among box turtles (Terrapene c. carolina) of a bottomland in Maryland. J. Herp., 23:40 44. SWARTH, C. 2005. Home range characteristics of box turtles. p. 9 10. In: C. Swarth and S. Hagood (eds.). Summary of the eastern box turtle regional conservation workshop. The Humane Society of the United States, Washington, D.C. TUTTLE, S. E. AND D. M. CARROLL. 2003. Home range and seasonal movements of the wood turtle (Glyptemys insculpa) in southern New Hampshire. Chelon. Cons. Biol., 4:656 663. SUBMITTED 21 JANUARY 2006 ACCEPTED 21 JULY 2006