Reproductive Traits of a High Elevation Viviparous Lizard Sceloporus bicanthalis (Lacertilia: Phrynosomatidae) from Mexico

Journal of Herpetology, Vol. 38, No. 3, pp. 438 443, 2004 Copyright 2004 Society for the Study of Amphibians and Reptiles SHORTER COMMUNICATIONS Reproductive Traits of a High Elevation Viviparous Lizard Sceloporus bicanthalis (Lacertilia: Phrynosomatidae) from Mexico FELIPE RODRÍGUEZ-ROMERO, 1,2 GEOFFREY R. SMITH, 3 ORLANDO CUELLAR, 4 AND FAUSTO R. MÉNDEZ DE LA CRUZ 1 1 Departamento de Zoología, Instituto de Biología, Universidad Nacional Autónoma de México, Ciudad Universitaria, Circuito exterior A. P. 70-153, C. P. 04510, México, D. F. México 3 Department of Biology, Denison University, Granville, Ohio 43023, USA 4 Department of Biology, University of Utah, Salt Lake City, Utah 84112, USA; and P.O. Box 17074, Salt Lake City, Utah 84127, USA ABSTRACT. Several species of lizards exhibit significant annual variation in reproductive traits; however, most work in this area focused on populations from temperate latitudes or low to medium elevations. We examined annual variation in litter size, neonate size, and relative litter mass in a high elevation (4200 m) population of the viviparous lizard, Sceloporus bicanthalis from the Volcano Nevado de Toluca, México. We found little evidence for annual variation in reproduction in this population. Female body size influenced litter size and litter mass. Relative litter mass in this population (0.47) was among the highest reported for any Sceloporus and may be a consequence of a nearly annual life cycle. Mean neonate size was not affected by female SVL or litter size, suggesting it may be optimized in this population. RESUMEN. En algunos lacertilios se ha registrado una significativa variación anual en las características reproductoras, sin embargo, la mayoría de estos estudios se han efectuado en poblaciones de latitudes templadas y elevaciones bajas o medias. En el presente estudio analizamos la variación anual del tamaño de la camada, tamaño de las crías y masa relativa de la camada en una población de lagartijas vivíparas (Sceloporus bicanthalis), que habitan a 4200 m en el volcán Nevado de Toluca, México. Se registró una mínima evidencia de variación interanual en la reproducción. La talla de las hembras demuestra una gran influencia tanto en el tamaño de la camada como en el peso de la camada. La masa relativa de la camada se encuentra entre los valores más altos para el género Sceloporus (5 0.47), quizá como una consecuencia de un ciclo de vida de tipo anual. Finalmente, el tamaño promedio de las crías no se encuentra moldeado por la talla de las hembras o el tamaño de la camada, sugiriéndose por lo tanto, una optimización de esta característica reproductora. Reproductive traits in lizards are often thought to be under proximate or environmental control to some extent (e.g., Ballinger, 1983). For example, in several species of lizards litter size, litter mass, neonate size, and eggs size have been shown to vary from year to year, often in relation to the amount of rainfall or precipitation (e.g., Smith et al., 1995; Abell, 1999; Wapstra and Swain, 2001 and references therein). Most of these studies have been conducted on low to middle elevation (,3000 m) populations and species. Few, if any, studies are available examining annual reproductive variation in lizards from high elevations. Indeed we know little about reproduction in general in high elevation lizard populations (e.g., Eumeces copei, Guillette, 1983; Sceloporus mucronatus, Méndez-de la Cruz et al., 1988, 1993; Estrada-Flores et al., 1990; Sceloporus grammicus, Lemos-Espinal et al., 1998; Sceloporus bicanthalis, Hernández-Gallegos et al., 2002). Comparing the expression of reproductive traits among environments that are at the limits of the potential distribution of lizards might contribute to our understanding of the proximate causes of variation in reproduction. One prediction for relatively harsh environments is that little variation in the expression of reproductive traits will be observed between years 2 Corresponding Author. E-mail: feliper@ibiologia. unam.mx since there may be little scope for lizards to respond to any variation in environmental variation. Here, we examine annual variation in reproductive traits (litter size, neonate size, relative litter mass) of the viviparous lizard, Sceloporus bicanthalis, from a high elevation (4200 m) population on the Volcano Nevado de Toluca, México. In this population, S. bicanthalis are active throughout the year (unpubl.) and maintain elevated body temperatures over during the colder months of the year (Andrews et al., 1999). Thus, this population allows us to consider annual variation in a tropical latitude lizard from high elevation. MATERIALS AND METHODS Our study area was near the top of the Volcano Nevado de Toluca in México (198 079 300N, 998469150W; 4200 m above sea level). The environment in springsummer when maximum activity occurs in this population (FR-R, unpubl.) is marked by low temperatures (4.68C monthly average) and mean monthly precipitations of 171.6 mm (García, 1981). Figure 1 displays precipitation and temperature profiles collected by the Servicio Metereológico Mexicano at Nevado de Toluca (198 079N, 998 469W; 4120 m above sea level) for the duration of the study, as well as historical means. Vegetation includes the alpine bunchgrasses Festuca tolucencis, Calamagrostis tolucencis, and Eryngium protiflorum (Rzedowski, 1981).

SHORTER COMMUNICATIONS 439 FIG. 1. Temperature (A) and precipitation (B) patterns for Nevado de Toluca, México from 1960 1993 (multiyear data), 1994, 1995, 1996, and 1998. A total of 68 pregnant female S. bicanthalis were collected by hand or with a noose during four years: 1994 (N 5 19), 1995 (N 5 9), 1996 (N 515), and 1998 (N 5 25). Females in this population rarely live for longer than 12 months (mean lifespan 5 8 months; unpubl.); thus no female was sampled in more than one year. Only females approaching parturition (April to September), as determined by maximum distension of their abdomens and reduced mobility (Rodríguez-Romero et al., 2002), were captured and kept in the laboratory until parturition (for approximately one month). Females were housed individually in rectangular cages (60 3 40 cm) with sandy substrate, rocks and tree bark that provided refuges for lizards. Food (moth larvae) and water were provided ad libitum. Females were kept in a room maintained at 248C and on a 9-h day:15-h night photoperiod. A 45-watt Vita-Light bulb suspended 20 cm above each terrarium provided heat and light during daylight hours. These conditions allowed females to maintain their preferred body temperature (298C; Andrews et al., 1999) during the day. For each female, we measured snout vent length (SVL; to the nearest 1 mm with a plastic ruler) and body mass (to the nearest mg with an analytical balance). Female body mass was measured both before (female total mass, FTM; measured upon capture, generally one month before parturition) and immediately after parturition (female mass after parturition, FMAP). Relative litter mass (RLM) was calculated by dividing total litter mass by FMAP. We also recorded litter size (LS) as the number of neonates produced by a female and weighed the mass of the litter immediately following parturition. To compare variation in the expression of traits among years we used ANOVA. We used a one-way ANOVA for SVL, and an ANCOVA (with SVL as covariate) for FTM, FMAP, litter size (LS), litter mass (LM), and RLM. We used linear regressions to investigate the relationships between aspects of female body size and reproductive output. We also report the results of a ln-ln regression for litter size and SVL on the recommendation of King (2000), who suggested that such a regression often provides a better fit for linear regression and also provides allometric coefficients. RESULTS Annual Variation. Female SVL did not differ significantly among years (Table 1; ANOVA: F 3,64 5 0.04, P 5 0.99). Female total mass also did not vary significantly among years (Table 1; ANCOVA: F 3,63 5 0.22, P 5 0.88). Female mass after parturition varied significantly among years, with highest values in 1994 and lowest values in 1996 (Table 1; ANCOVA: F 3,63 5 5.89, P 5 0.0013). None of the litter or neonate traits (LS, RLM, LM, neonate mass) significantly differed among the years of our study (Table 1; P. 0.34 in all cases). Reproductive Traits. Because of the general lack of annual variation in reproductive traits, the following analyses were done on data pooled across the four years of the study. Overall means for reproductive traits are given in Table 1. Female total mass increased with increasing SVL (N 5 68, r 2 5 0.72, P, 0.0001; FTM 5 10.07 þ 0.31SVL). TABLE 1. Means (6 SE) for reproductive characteristics of a population of Sceloporus bicanthalis from Volcano Nevado de Toluca, México. Trait Year 1994 (N 5 20) 1995 (N 5 9) 1996 (N 5 15) 1998 (N 5 24) Overall (N 5 68) SVL (mm) 52.1 6 0.9 52.1 6 1.5 52.5 6 1.0 52.0 6 0.9 52.1 6 0.5 Female total mass (g) 6.05 6 0.23 5.83 6 0.53 6.03 6 0.38 5.91 6 0.29 5.97 6 0.18 Female mass after parturition (g) 3.39 6 0.23 3.21 6 0.34 2.78 6 0.17 2.91 6 0.16 3.06 6 0.11 Litter size 7.35 6 0.47 7.78 6 1.0 6.67 6 0.63 7.21 6 0.54 7.21 6 0.30 Total litter mass (g) 1.47 6 0.10 1.42 6 0.21 1.29 6 0.12 1.41 6 0.14 1.41 6 0.07 Mean individual neonate mass (g) 0.20 6 0.004 0.18 6 0.006 0.19 6 0.004 0.19 6 0.008 0.19 6 0.003 RLM 0.46 6 0.029 0.43 6 0.034 0.47 6 0.041 0.49 6 0.043 0.47 6 0.020

440 SHORTER COMMUNICATIONS Female mass after parturition also increased with increasing SVL (N 5 68, r 2 5 0.60, P, 0.0001; FMAP 5 5.68 þ 0.17SVL). Litter size increased with increasing female body size (N 5 68, r 2 5 0.35, P, 0.0001; LS 5 11.54 þ 0.36 SVL; ln-ln transformed regression: N 5 68, r 2 5 0.311, P, 0.0001; ln LS 5 7.5 þ 2.39 ln SVL), as did total litter mass (N 5 68, r 2 5 0.34, P, 0.0001; LS 5 2.76 þ 0.08SVL). In contrast, RLM was not significantly related to SVL (N 5 68, r 2 5 0.01, P 5 0.42). Mean individual neonate mass was not affected by maternal SVL (Fig. 2A; N 5 68, r 2 5 0.03, P 5 0.16). Litter size was positively related to female mass after parturition (N 5 68, r 2 5 0.20, P, 0.0001; LS 5 3.38 þ 1.25FMAP). Total litter mass also increased with increasing female mass after parturition (N 5 68, r 2 5 0.17, P 5 0.0005; TLM 5 0.62 þ 0.26FMAP). Mean individual neonate mass was not related to female mass after parturition (N 5 68, r 2 5 0.002, P 5 0.74). Neonate mass was not influenced by litter size and was relatively constant across a wide range of litter sizes (Fig. 2B; N 5 68, r 2 5 0.018, P 5 0.27). Neonate mass was unaffected by litter size after effects of female size were removed using partial regression (N 5 68, r 2 5 0.003, P 5 0.74). Relative litter mass increased with litter size (N 5 68, r 2 5 0.35, P, 0.0001; RLM 5 0.187 þ 0.039LS). DISCUSSION We found little evidence for annual variation in reproduction in S. bicanthalis in this population. The only trait that varied significantly among the four years of our study was female mass after parturition. Other studies on Sceloporus have found little or no annual variation in reproductive traits, often in spite of variation in precipitation (e.g., Sceloporus graciosus, Tinkle et al., 1993; Sceloporus jarrovi, Ballinger, 1979; Sceloporus scalaris, Ballinger and Congdon, 1981; Sceloporus undulatus, Jones and Ballinger, 1987; Jones et al., 1987; Parker, 1994; Sceloporus variabilis, Benabib, 1994). In contrast, several other studies on Sceloporus show significant annual variation in reproductive traits (e.g., Sceloporus merriami, Dunham, 1981; Sceloporus undulatus (clutch frequency), Jones et al., 1987; Sceloporus virgatus, Smith et al., 1995; Abell, 1999). It is not immediately apparent why some species of Sceloporus show annual variation in reproductive traits and others do not. It may be that the conditions at high elevations preclude or limit the scope or range of potential variation in reproductive traits in this population. Alternatively, the amount of variation in climatic variables, such as precipitation and temperature, was not great over most of the years of this study (e.g., 1995, 1996, 1998; see Fig. 1), although 1994 appeared to be a drought year. Thus, lack of variation in reproductive traits may be a result of the lack of variation in environmental conditions among most years; although it is interesting that the drought year of 1994 did not have an effect on the expression of reproductive traits. This suggests environmental fluctuations in this population may have little influence on reproductive output. Such a result would be consistent with the high reproductive effort expected of an annual species (see below). Further information from a broader range of Sceloporus are needed to determine FIG. 2. Relationship between mean individual neonate mass and (A) maternal SVL, and (B) litter size for Sceloporus bicanthalis from the Volcano Nevado de Toluca, México, at 4200 m. whether there are any habitat, latitudinal, elevational, methodological (e.g., number of years studied), or reproductive mode (viviparity vs oviparity) correlates of reproductive variation within the genus Sceloporus and lizards in general. Body size influenced litter size and litter mass in our population. Such an observation has been demonstrated numerous times in Sceloporus (Ballinger, 1973; Ballinger and Congdon, 1981; Dunham, 1981; Benabib, 1994; Ballinger and Lemos-Espinal, 1995; Smith et al., 1995; Abell, 1999) but is not always present (e.g., Sceloporus gadoviae; Lemos-Espinal et al., 1999). The relative litter mass of Sceloporus bicanthalis in our population averaged 0.47 for females sampled across the four years of our study. This average is among the highest RLM or relative clutch mass (RCM) for any Sceloporus (Table 2). It is interesting to note that other populations of S. bicanthalis also have very high RLMs (Table 2). It is not clear why S. bicanthalis should have such a large RLM relative to other Sceloporus. One possible explanation is that in our population, S. bicanthalis females rarely live more than one year (unpubl.) and, thus, may not reproduce more than once in their lifetime. In other words, they approximate semelparous annual species, which often have high reproductive efforts (see Roff, 1992). Interestingly, the members of the S. scalaris clade (sensu Wiens and Reeder, 1997), Sceloporus aeneus, S. bicanthalis, and S. scalaris all have the highest RLM or RCM. These species tend to come from relatively high elevations, tend to be small, and tend to have shorter life spans than many of the other Sceloporus with lower RCMs or RLMs (see citations in Table 2). This combination of traits may select for high reproductive effort. In this case though, it

SHORTER COMMUNICATIONS 441 TABLE 2. Relative clutch and litter masses of various species of Sceloporus lizards. In some cases we indicate the particular study zone in Mexican populations. Species RCM or RLM Source S. aeneus, Milpa Alta, México 0.34 Rodríguez-Romero (1996) S. aeneus, San Cayetano, México 0.44 Rodríguez-Romero (unpubl.) S. arenicolus, New Mexico 0.20 Greenwald and West (2002) S. bicanthalis, Nevado de Toluca, Mexico 0.47 This study S. bicanthalis, Zoquiapan, México 0.52 Rodríguez-Romero (1996) S. bicanthalis, Nopalillo, México 0.40 Rodríguez-Romero (unpubl.) S. bicanthalis, Paso de Cortez, México 0.40 Rodríguez-Romero (unpubl.) S. bicanthalis, Las Vigas, México 0.43 Rodríguez-Romero (unpubl.) S. clarkii, Arizona 0.32 Vitt and Congdon (1978) S. graciosus, Utah 0.20 Tinkle (1973); Tinkle et al. (1993) S. grammicus, Zoquiapan, México 0.36 Cuellar et al. (1996) S. jarrovii, Arizona 0.15 Ballinger (1981) S. magister, Utah 0.19 Tinkle (1976) S. malachiticus, Costa Rica 0.34 Marion and Sexton (1972) S. merriami, New Mexico 0.27 Greenwald and West (2002) S. mucronatus, Hidalgo, México 0.26 Méndez de la Cruz et al. (1988) S. mucronatus, Zoquiapan, México 0.20 Rodriguez-Romero (1999) S. poinsetti, Mapimí, México 0.33 F. Rodríguez-Romero (unpubl.) S. poinsetti, Texas 0.32 Ballinger (1973) S. scalaris, Arizona 0.39 Vitt and Congdon (1978) S. serrifer, Yucatán, México 0.21 Rodríguez-Romero (unpubl.) S. torquatus, México 0.20 0.21 Méndez de la Cruz et al. (1992) S. undulatus consobrinus, Mapimí, México 0.20 Gadsden-Esparza and Aguirre-León (1993) S. undulatus consobrinus, New Mexico 0.21 Vinegar (1975b) S. undulatus consobrinus, Texas 0.27 Tinkle and Ballinger (1972) S. undulatus elongates, Utah 0.21 Tinkle (1972) S. undulatus erythrocheilus, Colorado 0.23 Tinkle and Ballinger (1972) S. undulatus tristichus, Arizona 0.22 Marion and Sexton (1972) S. variabilis, Veracruz, México 0.20 Benabib (1994) S. virgatus, Arizona 0.29 Vinegar (1975a) is difficult to separate the potentially confounded effects of ecology and phylogeny. Mean neonate size was unaffected by either female body size or litter size (see Fig. 2). The constancy of neonate size in our study over the range of SVLs and litter sizes we observed suggests that neonate size in these lizards may be optimized. Other Sceloporus also show no relationship between egg size or offspring size and clutch or litter size (S. variabilis, Benabib, 1994; S. gadovae, Lemos-Espinal et al., 1999; S. undulatus, Angilletta et al., 2001; S. virgatus, Smith et al., 1995), but some show a negative relationship (e.g., S. virgatus, Abell, 1999; Sceloporus poinsetti, Ballinger, 1973). In summary, S. bicanthalis has a relatively high mean RLM for Sceloporus and shows little annual variation in reproductive traits. The high elevation environment and the short lifespan of S. bicanthalis in this population may help explain this constellation of traits. However, additional study on this species and other high elevation Sceloporus are needed to more fully understand what factors may be influencing these traits and their plasticity. Acknowledgments. We thank L. López González and O. Hernández for field assistance, and P. Doughty, and an anonymous reviewer for their comments and suggestions. Support was provided by CONACYT (400355-5-2155) and DGAPA (IN210594 and IN232398). LITERATURE CITED ABELL, A. J. 1999. Variation in clutch size and offspring size relative to environmental conditions in the lizard Sceloporus virgatus. Journal of Herpetology 33:173 180. ANDREWS, R. M., F. R. MÉNDEZ-DE LA CRUZ, M. VILLAGRÁN-SANTA CRUZ, AND F. RODRÍGUEZ-ROMERO. 1999. Field and selected body temperatures of the lizards Sceloporus aeneus and Sceloporus bicanthalis. Journal of Herpetology 33:93 100. ANGILLETTA, M. J., M. W. SEARS, AND R. S. WINTERS. 2001. Seasonal variation in reproductive effort and its effect on offspring size in the lizard Sceloporus undulatus. Herpetologica 57:365 375. BALLINGER, R. E. 1973. Comparative demography of two viviparous iguanid lizards (Sceloporus jarrovi and Sceloporus poinsetti). Ecology 54:269 283.. 1979. Intraspecific variation in demography and life history of the lizard, Sceloporus jarrovi, along an altitudinal gradient in southeastern Arizona. Ecology 60:901 909.. 1981. Food limiting in populations of Sceloporus jarrovi (Iguanidae). Southwestern Naturalist 25: 554 557. 1983. Life-history variations. In R. B. Huey, E. R. Pianka, and T. W. Schoener (eds.),lizard Ecology: Studies of a Model Organism, pp. 241 260. Harvard Univ. Press, Cambridge, MA.

442 SHORTER COMMUNICATIONS BALLINGER, R. E., AND J. D. CONGDON. 1981. Population ecology and life history strategy of a montane lizard (Sceloporus scalaris) in southeastern Arizona. Journal of Natural History 15:213 222. BALLINGER, R. E., AND J. A. LEMOS-ESPINAL. 1995. Reproduction in the Mexican lizard, Sceloporus torquatus. Revista Ciencia Forestal en México 20:143 148. BENABIB, M. 1994. Reproduction and lipid utilization of tropical populations of Sceloporus variabilis. Herpetological Monographs 8:160 180. CUELLAR, O., F. R. MÉNDEZ DE LA CRUZ., M. VILLAGRÁN- SANTA CRUZ, AND R. S. TREJO. 1996. Pregnancy does not increase the risk of mortality in wild viviparous lizards (Sceloporus grammicus). Amphibia-Reptilia 17:77 80. DUNHAM, A. E. 1981. Populations in a fluctuating environment: the comparative population ecology of the iguanid lizard Sceloporus merriami and Urosaurus ornatus. Miscellaneous Publications Museum of Zoology, Univ. of Michigan 158:1 62. ESTRADA-FLORES, E., M. VILLAGRAN-SANTA CRUZ, F. R. MÉNDEZ-DE LA CRUZ, G. CASAS-ANDREU. 1990. Gonadal changes throughout the reproductive cycle of the viviparous lizard Sceloporus mucronatus (Sauria: Iguanidae). Herpetologica 46:43 50. GADSDEN-ESPARZA, H., AND G. AGUIRRE-LEÓN. 1993. Historia de vida comparada en una población de Sceloporus undulatus (Sauria: Iguanidae) del Bolsón de Mapimí. Boletín de la Sociedad Herpetologica Mexicana 5:21 41. GARCÍA, E. 1981. Modificaciones al sistema de clasificación climática de Köppen. Offset Larios., D.F., México. GREENWALD N. D., AND S. WEST. 2002. Petition to list the sand dune lizard Sceloporus arenicolus as a threatened or endangered species under the U.S. Endangered Species Act. Submitted to Endangered Species Program, New Mexico. Department of Game and Fish, Albuquerque. GUILLETTE JR. L. J. 1983. Notes concerning reproduction of the montane skink, Eumeces copei. Journal of Herpetology 17:144 148. HERNÁNDEZ-GALLEGOS, O., F. R. MÉNDEZ-DE LA CRUZ, M. VILLAGRÁN-SANTA CRUZ, AND R. M. ANDREWS. 2002. Continuous spermatogenesis in the lizard Sceloporus bicanthalis (Sauria: Phrynosomatidae) from high elevation habitat of central Mexico. Herpetologica 58:415 421. JONES, S. M., AND R. E. BALLINGER. 1987. Comparative life histories of Holbrookia maculata and Sceloporus undulatus in western Nebraska. Ecology 68:1828 1838. JONES, S. M., R. E. BALLINGER, AND W. P. PORTER. 1987. Physiological and environmental sources of variation in reproduction: prairie lizards in a food rich environment. Oikos 48:325 335. KING, R. B. 2000. Analyzing the relationship between clutch size and female body size in reptiles. Journal of Herpetology 34:148 150. LEMOS-ESPINAL, J. A., R. E. BALLINGER, AND G. R. SMITH. 1998. Comparative demography of the high-altitude lizard, Sceloporus grammicus (Phrynosomatidae), on the Iztaccihuatl Volcano, Puebla, Mexico. Great Basin Naturalist 58:375 379. LEMOS-ESPINAL, J. A., G. R. SMITH, AND R. E. BALLINGER. 1999. Reproduction in Gadow s Spiny Lizard, Sceloporus gadovae (Phrynosomatidae), from arid tropical México. Southwestern Naturalist 44:57 63. MARION, K. R., AND O. J. SEXTON. 1972. The reproductive cycle of the lizard Sceloporus malachiticus in Costa Rica. Copeia 1972:517 526. MÉNDEZ-DE LA CRUZ, F. R., L. J. GUILLETTE JR., M. VILLAGRÁN-SANTA CRUZ, AND G. CASAS-ANDREU. 1988. Reproductive and fat body cycle of the viviparous lizard Sceloporus mucronatus (Sauria, Iguanidae). Journal of Herpetology 22:1 12. MÉNDEZ-DE LA CRUZ, F. R., M. F. ORTIZ, AND O. CUELLAR. 1992. Geographic variation of reproductive traits in a Mexican viviparous lizard, Sceloporus torquatus. Biogeographica 68:149 156. MÉNDEZ-DE LA CRUZ, F. R., L. J. GUILLETTE JR., AND M. VILLAGRÁN-SANTA CRUZ. 1993. Differential atresia of ovarian follicles and its effect on the clutch size of two populations of the viviparous lizard Sceloporus mucronatus. Functional Ecology 7:535 540. PARKER, W. S. 1994. Demography of the Fence Lizard, Sceloporus undulatus, in northern Mississippi. Copeia 1994:136 152. RODRÍGUEZ-ROMERO, F. 1996. Estudio comparativo de los parámetros asociados al tamaño de camada o nidada en lacertilios emparentados. Unpubl. bachelor thesis, Universidad Nacional Autónoma de México, México, D. F., México.. 1999. Estudio comparativo de algunos aspectos de la inversion parental en lacertilios de ambientes tropical y templado. Unpubl. master s thesis, Universidad Nacional Autónoma de México, México, D. F., México. ROFF, D. A. 1992. The Evolution of Life Histories: Theory and Analysis. Chapman and Hall, New York. RZEDOWSKI, J. 1981. Vegetación de México. Editorial Limusa, México. SMITH, G. R., R. E. BALLINGER, AND B. R. ROSE. 1995. Reproduction in Sceloporus virgatus from the Chiricahua Mountains of southeastern Arizona with emphasis on annual variation. Herpetologica 51:342 349. TINKLE, D. W. 1972. The dynamics of a Utah population of Sceloporus undulatus. Herpetologica 28:351 359.. 1973. A population of the Sagebrush Lizard Sceloporus graciosus, in southern Utah. Copeia 1973:284 295.. 1976. Comparative data on the population ecology of the Desert Spiny Lizard, Sceloporus magister. Herpetologica 32:1 6. TINKLE, D. W., AND R. E. BALLINGER. 1972. Sceloporus undulatus: a study of the intraespecific comparative demography of a lizard. Ecology 53:570 584. TINKLE, D. W., A. E. DUNHAM, AND J. D. CONGDON. 1993. Life history and demographic variation in the lizard Sceloporus graciosus: a long-term study. Ecology 74:2413 2429. VINEGAR, M. B. 1975a. Demography of the Striped Plateau Lizard, Sceloporus virgatus. Ecology 56:172 182.. 1975b. Life history phenomena in two populations of the lizard Sceloporus undulautus in southwestern New Mexico. American Midland Naturalist 93:388 402. VITT, L. J., AND J. D. CONGDON. 1978. Body shape, reproductive effort and relative clutch mass in

SHORTER COMMUNICATIONS 443 lizards: resolution of a paradox. American Naturalist 112:595 608. R. SWAIN. 2001. Geographic and annual variation in life-history traits in a temperate zone Australian skink. Journal of Herpetology 35:194 203. WAPSTRA, E., AND WIENS, J. J., AND T. W. REEDER. 1997. Phylogeny of the spiny lizards (Sceloporus) based on molecular and morphological evidence. Herpetological Monographs 11:1 101. Accepted: 16 May 2004. Journal of Herpetology, Vol. 38, No. 3, pp. 443 447, 2004 Copyright 2004 Society for the Study of Amphibians and Reptiles Characteristics of Burrows Used by Juvenile and Neonate Desert Tortoises (Gopherus agassizii) during Hibernation LISA C. HAZARD 1,2 AND DAVID J. MORAFKA 3 1 Department of Organismic Biology, Ecology and Evolution, University of California, Los Angeles, P.O. Box 951606, Los Angeles, California 90095-1606, USA 3 Research Associate, Department of Herpetology, California Academy of Sciences, Golden Gate Park, San Francisco, California 94118, USA ABSTRACT. Behavior of young tortoises released from seminatural nurseries could be affected by the length of time spent within the nursery before release. We tested whether neonate (under two months) and juvenile (8 9 years) Desert Tortoises selected hibernation burrows with differing characteristics after release from their natal pen. Burrow habitat (canopy cover and landscape slope) did not differ between age classes. Juvenile tortoises were larger than neonates and, therefore, used larger burrows than neonates, but their burrows were a closer fit to tortoise size than were the neonate burrows. Juvenile burrow orientation differed significantly from a uniform distribution, with a mean direction of 1628 (SSE); the burrows of neonates were not oriented in any particular direction. Selectivity of juveniles compared to neonates may have contributed to higher levels of movement by juveniles between release and hibernation. These age-related differences in behavior should be incorporated into nursery-based management plans. Desert Tortoises (Gopherus agassizii) face serious population declines from a variety of causes, including disease, habitat loss, and predation. Hatcheries have been proposed as one mechanism for managing Desert Tortoise populations. Eggs laid in hatcheries would be protected from predation, and young tortoises could be similarly protected (Morafka et al., 1997). Older juveniles are larger and have more ossified shells and, therefore, may be more resistant to predation than neonates. However, behavior of young chelonians released from seminatural hatcheries could be affected by the length of time spent within the hatchery before release. Previously, we examined differences in dispersal behavior between juvenile (8 9 years) and neonate (two months) Desert Tortoises (Hazard and Morafka, 2002). Recently released juvenile tortoises were more active than neonates and took longer to settle in a hibernation burrow. We hypothesized that juveniles were more active in part because they were more 2 Corresponding Author. Present address: Department of Ecology and Evolutionary Biology, Earth and Marine Sciences Building, University of California, Santa Cruz, California 95064, USA; E-mail: hazard@ biology.ucsc.edu selective about their hibernation burrows and, therefore, had to move around more to find or excavate suitable burrows. Here, we examine differences in the characteristics of the hibernation burrows chosen by these tortoises, to evaluate potential selectivity by the two age classes. MATERIALS AND METHODS We studied Desert Tortoises at the juvenile tortoise nursery at the U.S. Army National Training Center at Fort Irwin, California (Morafka et al., 1997). In October 1999, 12 neonates (hatched in the nursery within the previous two months) and 12 nursery-raised juvenile tortoises (8 9 years) were fitted with radiotransmitters (Holohil model BB2G, weighing 1.8 g), released near the pen, and periodically tracked (detailed methods in Hazard and Morafka, 2002). Tortoises ceased moving to new locations by day 34 (20 November 1999), and it was assumed that they were hibernating for the winter, though juvenile tortoises are facultative hibernators and may become active in winter if thermal conditions permit (Wilson et al., 1999a). We located winter burrows for 22 tortoises (12 juvenile, 10 neonate) and marked them with pin flags in November 1999. We evaluated all burrows in February 2000, after the animals had emerged and

444 SHORTER COMMUNICATIONS Table 1. Characteristics of neonate and juvenile Desert Tortoises, their hibernation burrows, and two unoccupied burrows nearest to each occupied burrow. Means 6 SD or (for burrow facing direction) angular deviation. Statistics for canopy cover are Chi-square values; all others are F-values. Effects Occupied burrows Unoccupied burrows Age class Burrow status Age class 3 burrow status Variable Neonate Juvenile Neonate Juvenile Statistic P Statistic P Statistic P Habitat Distance to nearest unoccupied burrow (cm) Distance to 2nd nearest unoccupied burrow (cm) Number of burrows with 0 50%/51 100% canopy cover Distance to base of nearest shrub (cm) 114 6 94 165 6 140 0.942 0.343 176 6 129 225 6 172 0.558 0.464 5/5 9/3 15/3 15/9.0.0001 1.00 0.734 0.392 3.395 0.065 52 6 48 82 6 78 67 6 43 69 6 61 1.143 0.289 0.006 0.938 0.939 0.336 Landscape slope (8) 2.6 6 1.2 2.5 6 0.5 4.3 6 4.6 2.5 6 0.4 2.056 0.157 1.687 0.199 1.687 0.199 Burrow slope (8) 8.3 6 4.5 6.3 6 3.8 8.6 6 5.0 4.9 6 2.5 7.263 0.009 0.240 0.626 0.600 0.442 Burrow size Tortoise width (mm) 38.6 6 1.9 69.9 6 6.1 209.6,0.0001 Tortoise height (mm) 22.8 6 1.1 38.6 6 3.4 182.6,0.0001 Burrow width (mm) 77.9 6 17.2 114.9 6 23.4 15.9 0.0008 Burrow height (mm) 42.0 6 11.5 58.7 6 8.8 14.2 0.0013 Burrow width: tortoise 1.83 6 0.78 1.68 6 0.40 3.43 0.0797 width Burrow height: tortoise 1.84 6 0.46 1.53 6 0.30 3.39 0.0813 height Burrow width 3 height: tortoise width 3 3.67 6 1.3 2.60 6 0.95 6.12 0.023 height Burrow orientation Facing direction of burrow 72.5 6 73.8 162.3 6 52.6 78.1 6 67.1 111.7 6 76.3 (8) Rayleigh s test z-statistic and P-value z1,10 5 0.288 P. 0.50 z1,12 5 4.011 P 5 0.015 z1,20 5 1.963 P. 0.10 z1,24 5 0.306 P. 0.50

SHORTER COMMUNICATIONS 445 Fig. 1. Relationships between burrow size and Desert Tortoise size for neonate (squares) and juvenile (circles) Desert Tortoises. Burrow width (closed symbols) and height (open symbols) were significantly correlated with tortoise width and height when age classes were pooled. Width: y 5 1.09x þ 38.1; R 2 5 0.403; P 5 0.002. Height: y 5 0.925x þ 22.1; R 2 5 0.368; P 5 0.0035. moved to new locations, to avoid disturbing them. We measured characteristics of the burrow chosen by the tortoise and of the two nearest unoccupied burrows that we judged (based on size) to be potentially usable by that animal. We recorded canopy species (if present), percentage canopy cover over the burrow, distance to the base of the nearest shrub, direction the mouth of the burrow faced, height and width of the burrow, slope of the ground immediately outside the burrow (burrow slope) and slope of the surrounding area (landscape slope). We also recorded distance from the two nearest unoccupied burrows to the tortoise s burrow. Compass bearings were corrected to true north. Data are presented as mean 6 SD. Circular statistics were calculated according to Zar (1984); all other statistics were calculated using JMP (SAS Institute, Inc.). A P-value of 0.05 or less was considered significant. RESULTS The dominant shrub species at the site were creosote (Larrea tridentata), box thorn (Lycium pallidum), and bur sage (Ambrosia dumosa); shrub species were pooled for statistical analysis. Percent canopy cover over burrows was bimodally distributed; therefore, burrows were categorized for analysis as having, 50% or. 50% cover. We found no significant differences between ages classes or between unoccupied and occupied burrows in canopy cover use, distance from burrow to the base of the nearest shrub, or overall landscape slope (Table 1). Burrow slope was typically steeper than landscape slope. Burrows associated with neonates (occupied or unoccupied) had a significantly steeper slope than those associated with juveniles; within age class, there were no differences between occupied or unoccupied burrows (Table 1). Appropriately sized alternative burrows (primarily rodent burrows) were typically found within 2 m of the hibernation burrow. There were no differences between age classes in distance to the nearest or second nearest unoccupied burrow (Table 1). Juvenile tortoises were larger than neonate tortoises, and burrows used by juveniles were significantly larger Fig. 2. Facing direction of hibernation burrows of (A) neonate (N 5 10) and (B) juvenile (N 5 12) Desert Tortoises. Compass directions corrected to true north (08). Neonate mean vector direction was 72.58 but was not significantly different from random orientation (Rayleigh Test P 5 0.975). Juvenile mean vector direction was 162.3 and was nonrandomly oriented (Rayleigh Test P 5 0.015). than neonate burrows (Table 1). Burrow height and width were both significantly correlated with tortoise size when age classes were pooled (Fig. 1). To evaluate fit of burrow size to tortoise size, we examined the ratios of burrow width to tortoise width, burrow height to tortoise height, and burrow area to tortoise area (width 3 height). Width ratio and height ratio did not differ between ages; however, relative area of the burrow mouth (burrow width 3 height/tortoise width 3 height) was significantly higher for neonates (Table 1). Unoccupied burrows were selected by us in part

446 SHORTER COMMUNICATIONS based on their size (roughly appropriate size for the individual tortoise); therefore, size of nearest neighbor burrows was not analyzed statistically. Juvenile tortoises selected burrows that faced, on average, south-southeast (mean direction 1628), with a range of 798 to 2338; mean direction differed significantly from a uniform distribution (P 5 0.015; Fig. 2B). Orientation of burrows used by neonate tortoises did not differ from a uniform distribution (P 5 0.975; Fig. 2A). Available burrows did not appear to have a bias in their orientation; facing direction of the nearest unoccupied burrows for both neonates and juveniles did not differ from uniform distributions (Table 1). DISCUSSION Juvenile tortoises do not appear to search farther for appropriately sized burrows than neonates do, because there was no difference between ages in distance to suitable unoccupied burrows. No differences in canopy cover, shrub species preference, or landscape slope were found, so juveniles do not appear to have selected burrows differently from neonates based on those criteria. Occupied and unoccupied neonate burrows had steeper slopes than did juvenile burrows, but this may be because the two age classes moved into slightly different habitats (Hazard and Morafka, 2002). Both juveniles and neonates made similar use of canopy cover: 80% of juveniles and 75% of neonates in this study had burrows within the canopy margin of a shrub. Juveniles confined in the pens were comparable; 80% of juvenile burrows within the natal pen were underneath the canopy (Wilson et al., 1999b). The sizes of burrows used by juveniles were more similar to the sizes of the tortoises than were burrows used by neonates. Because juveniles often excavated preformed rodent burrows that may have been initially slightly smaller than the tortoise s cross-sectional dimensions, the burrows height and width became similar to those of the occupying tortoise. In contrast, neonate DesertTortoisesfrequentlyusedexistingrodentburrows that may have been substantially taller or wider than the tortoise and, thus, required little or no excavation. Juvenile tortoises selected hibernation burrows that were nonrandomly oriented and faced, on average, south-southeast. The range was relatively narrow (798 2338; Fig. 2). Juveniles kept within the pens used burrows with an average facing direction of 718 (east northeasterly) but a range that spanned the full compass (Wilson et al., 1999b). Burrows of juvenile tortoises at sites throughout the Mojave Desert tended to face westerly to southeasterly (Berry and Turner, 1986). Hibernacula of adult desert tortoises in the San Pedro River Valley in Arizona (Bailey et al., 1995) and the Whitewater Hills in California (Lovich and Daniels, 2000) were found primarily on south-facing slopes; burrow orientation itself was not measured in these studies. Juvenile tortoises may have preferred burrows that faced the morning sun, allowing them to thermoregulate near the mouth of the burrow early in the day. Another possibility is that the burrows were selected because of their orientation relative to the slope of the landscape, as appears to be the case with G. polyphemus (McCoy et al., 1993). Direction of the local slope of each burrow was not measured in this study, but the overall landscape in the area used by the juveniles sloped downhill to the east. Regardless of the cause, the older tortoises exhibited a directional bias not seen in the neonates. If this bias caused juvenile tortoises to search longer for suitable burrows or burrow locations, it could explain the higher postrelease activity level of juvenile tortoises compared to neonates (Hazard and Morafka, 2002). The increased activity level seen in juveniles could result in higher exposure to predation risk, possibly negating any benefits of larger size. Although no mortality was observed in either group during the 34 days between release and hibernation (Hazard and Morafka, 2002), sample size was relatively small. Differences in burrow selectivity may reflect ontogenetic changes in dispersal behavior. Neonates may be predisposed to disperse as quickly as possible to a safe location in which to hibernate and wait out the dry autumn, emerging in the spring to forage and find a more permanent home. Not only is fall forage absent from the western Mojave Desert where summer monsoons are rare to nonexistent, but neonates function in the fall as postnatal lecithotrophs, surviving on the energetic and hydric reserves provided by or derived from residual yolk mass (Lance and Morafka, 2001). In contrast, eight- to nine-year-old juveniles who have been in the same location for years may not be prepared to disperse and when released in the fall may have been more focused on finding a suitable permanent burrow, not just a hibernaculum. Rather than dispersing when released, many of the juvenile tortoises in this study initially returned to the perimeter of the home pen (Hazard and Morafka, 2002), and some actually attempted to excavate under the predatorresistant hardware cloth barrier in an apparent attempt to return to their home burrows. These age-related differences in behavior need to be incorporated into future management plans involving long-term use of nurseries for conservation of desert tortoises. Acknowledgments. We thank S. Hillard and M. Marolda for assistance with mounting transmitters on tortoises; L. Bell, L. Cunningham, K. Emmerich, A. Johnson, M. Mendoza, B. Parker, and C. Todd for assistance with radio-tracking; and W. Alley for assistance with statistical analysis. Comments from two anonymous reviewers greatly improved the manuscript. Funding was awarded to the California State University, Dominguez Hills Foundation by the U.S. Army National Training Center, Fort Irwin California Directorate of Public Works, Department of Cultural and Natural Resources. Special thanks are extended to Department Manager M. Quillman for funding and support. This research was conducted under USFWS recovery permit CSUDH-5 issued to D. J. Morafka and a memorandum of understanding from the California Department of Fish and Game and was approved by the Chancellor s Animal Research Committee of the University of California, Los Angeles Office for the Protection of Research Subjects. LITERATURE CITED BAILEY, S. J., C. R. SCHWALBE, AND C. H. LOWE. 1995. Hibernaculum use by a population of Desert Tortoises (Gopherus agassizii) in the Sonoran Desert. Journal of Herpetology 29:361 369.

SHORTER COMMUNICATIONS 447 BERRY, K. H., AND F. B. TURNER. 1986. Spring activities and habits of juvenile Desert Tortoises, Gopherus agassizii, in California. Copeia 1986:1010 1012. HAZARD, L. C., AND D. J. MORAFKA. 2002. Comparative dispersion of juvenile and neonate Desert Tortoises (Gopherus agassizii): a preliminary assessment of age effects. Chelonian Conservation and Biology 4: 406 409. LANCE, V. A., AND D. J. MORAFKA. 2001. Post natal lecithotroph: a new age class in the ontogeny of reptiles. Herpetological Monographs 15:124 134. LOVICH, J. E., AND R. DANIELS. 2000. Environmental characteristics of Desert Tortoise (Gopherus agassizii) burrow locations in an altered industrial landscape. Chelonian Conservation and Biology 3: 714 721. MCCOY, E. D., H. R. MUSHINSKY, AND D. S. WILSON. 1993. Pattern in the compass orientation of Gopher Tortoise burrows at different spatial scales. Global Ecological and Biogeographical Letters 3:33 40. MORAFKA, D. J., K. H. BERRY, AND E. K. SPANGENBERG. 1997. Predator-proof field enclosures for enhancing hatching success and survivorship of juvenile tortoises: a critical evaluation. In J. Van Abbema (ed.), Proceedings: Conservation, Restoration, and Management of Tortoises and Turtles An International Conference, pp. 147 165. New York Turtle and Tortoise Society, New York. WILSON, D. S., D. J. MORAFKA, C.R.TRACY, AND K. A. NAGY. 1999a. Winter activity of juvenile Desert Tortoises (Gopherus agassizii) in the Mojave Desert. Journal of Herpetology 33:496 501. WILSON, D. S., C. R. TRACY, K. A. NAGY, AND D. J. MORAFKA. 1999b. Physical and microhabitat characteristics of burrows used by juvenile Desert Tortoises (Gopherus agassizii). Chelonian Conservation and Biology 3:448 453. ZAR, J. H. 1984. Biostatistical Analysis. Prentice-Hall, Inc., Englewood Cliffs, NJ. Accepted: 24 May 2004. Journal of Herpetology, Vol. 38, No. 3, pp. 447 451, 2004 Copyright 2004 Society for the Study of Amphibians and Reptiles Field Body Temperatures of Pregnant and Nonpregnant Females of Three Species of Viviparous Skinks (Mabuya) from Southeastern Brazil DAVOR VRCIBRADIC 1,2 AND CARLOS F. D. ROCHA 1,3 1 Setor de Ecologia, Departamento de Biologia Animal e Vegetal, Instituto de Biologia, Universidade do Estado do Rio de Janeiro, Rua São Francisco Xavier 524, Maracanã, 20550-011, Rio de Janeiro, RJ, Brazil 2 Programa de Pós-Graduação em Ecologia, Instituto de Biologia, Universidade Estadual de Campinas, CP 6109, 13081-970, Campinas, São Paulo, Brazil ABSTRACT. Many lizards are known to alter their thermal ecology during pregnancy, although body temperatures of pregnant/gravid females may either increase or decrease, depending on the species. Most of the data available on this phenomenon come from temperate taxa. In the present study, we compared field body temperatures (T b ) of pregnant females with those of nonpregnant females and males of three species of viviparous skinks (Mabuya agilis, Mabuya macrorhyncha, and Mabuya frenata) from southeastern Brazil. We found that pregnant females did not differ in T b from nonpregnant animals (including males). Thus, reproductive condition did not influence body temperatures regulation by these skinks to a significant degree. Thermal ecology of lizards is often related to reproductive condition, especially in pregnant or gravid females, since the most appropriate temperatures for the developing embryos may be different from the body temperatures normally attained by active lizards (e.g., Beuchat, 1986, 1988; Daut and Andrews, 1993; Andrews et al., 1997). In some lizard species and/or populations, pregnant females tend to regulate lower body temperatures (T b ) during activity than nonpregnant females (Garrick, 1974; Patterson and Davies, 1978; Beuchat, 1986; Heulin, 1987; Andrews and Rose, 1994; Tosini and Avery, 1996; Andrews 3 Corresponding Author. E-mail: cfdrocha@uerj.br et al., 1997), whereas in other cases pregnant females regulate higher T b s than nonpregnant ones (Werner and Whitaker, 1978; Stewart, 1984; Hailey et al., 1987; Daut and Andrews, 1993; Rock et al., 2000), and there are also cases in which T b s of pregnant and nonpregnant females do not differ significantly (Mayhew, 1963; Schall, 1977; Schwarzkopf and Shine, 1991; Andrews et al., 1999). All of the aforementioned studies involve taxa from temperate regions or high elevations, where ambient temperatures can vary widely; relationships between thermal ecology and female reproductive condition remain largely unknown for lowland tropical lizards. In tropical regions, the seasonal variation in environmental temperature tends to be relatively mild, however; there are regions (the so-called seasonal

448 SHORTER COMMUNICATIONS tropics) in which there is a clear, albeit not extreme, difference in ambient temperature between the warmest and the coolest periods of the year. Such seasonal variation in thermal environment may influence thermal ecology of some lizards inhabiting such areas, so that their body temperatures during activity may also vary throughout the year (e.g., Huey et al., 1977; Rocha, 1995; Teixeira-Filho et al., 1995, 1996; Menezes et al., 2000). In other cases, such as our studies on three species of Brazilian skinks of the genus Mabuya (Rocha and Vrcibradic, 1996; Vrcibradic and Rocha, 1998a), no significant seasonal variation in lizard T b was detected, even though environmental temperatures varied significantly throughout the year. In those studies, however, T b s of males and females were pooled together for between-season comparisons, which could have obscured any significant effect of sex or reproductive condition on thermal ecology. In the present study, we compared field body temperatures (T b ) during activity of pregnant females with those of nonpregnant females and males of three species of viviparous skinks of the genus Mabuya from southeastern Brazil (Mabuya agilis, Mabuya macrorhyncha, and Mabuya frenata), to verify whether pregnant females tend to have different (either higher or lower) T b s than nonpregnant animals. Comparisons were made within the same season (wet or dry), to control for possible effects of seasonality on lizard body temperatures. MATERIALS AND METHODS Mabuya agilis and M. macrorhyncha were collected in various restinga habitats (restingas are open coastal sandy habitats with xeric-adapted vegetation; see Eiten, 1992) along the coasts of the states of Espírito Santo and Rio de Janeiro, between latitudes 188419S and 238039S (altitudes near sea-level); an additional sample of M. macrorhyncha (N 5 18) was collected at the island of Queimada Grande (248309S; 468419W), off the southern coast of São Paulo state (for a description of the area see Duarte et al., 1995). Mabuya frenata was collected at a disturbed habitat within the cerrado-atlantic forest transition zone, in Valinhos (228569S; 468559W; altitude approximately 700 m), São Paulo state (for a description of the area see Vrcibradic and Rocha, 1998a,b). The climate is seasonally variable in all areas, with a wetter, warmer season from October to April in the coastal areas and from October to March in Valinhos, and a drier, colder season from April to September in the coastal areas and from March to September in Valinhos. We measured the body temperatures of a total of 62 M. agilis (body mass 4 15 g), 127 M. macrorhyncha (body mass 4 11 g), and 129 M. frenata (body mass 4 14 g). Most of the animals were collected between 1995 and 2000, during several different months (mostly in summer and spring), except for the M. frenata sample of Valinhos, collected monthly from December 1993 to December 1994 (see Vrcibradic and Rocha, 1998a), and for the M. agilis and M. macrorhyncha samples of Barra de Maricá (which comprised a substantial portion of the total sample for each species), collected in various months between 1989 and 1996. Very few data (N 5 6) for pregnant M. agilis was collected during dry season months, so those data were not used in the present study. Also, we could not obtain data for pregnant M. frenata during wet season months. All animals used in this study were collected during daylight hours, mostly between 0800 and 1700 h. The body temperatures (T b ) of the animals were measured immediately after their capture, with a fast-reading Schultheis thermometer (to the nearest 0.28C). Only body temperatures of active individuals were considered. Also, we did not consider individuals that were badly damaged during collection procedures, nor those captured after more than 30 sec had elapsed since the first capture attempt (animals not killed instantly during capture were euthanased with ether). Following Rocha and Vrcibradic (1999), we considered as pregnant females those in reproductive stages 5 (moderately to well-developed embryos) and 6 (near-term or term foetuses) and as nonpregnant those in stages 1 (no yolking follicles or oviductal ova), 2 (yolking follicles but no oviductal ova) and 3 (oviductal ova or small embryo sacs present). Mean T b s were compared between pregnant and nonpregnant females (and between males and females within seasons) for each species of Mabuya using one-way ANOVA. RESULTS AND DISCUSSION Values of T b for females and males of the three Mabuya species, as well as the statistics for the comparisons between pregnant and nonpregnant animals (females and/or males) within seasons, are given in Table 1. T b s of pregnant females were significantly higher than those of males for M. macrorhyncha in the dry season (F 1,36 5 10.47; P, 0.005). None of the remaining comparisons were statistically significant. Mean T b s for pregnant females of the three Mabuya species were around 32 338C, a temperature considered most adequate for the development of embryos of other viviparous lizards (e.g., Sceloporus jarrovi and Sceloporus grammicus [Phrynosomatidae]: Beuchat, 1986, 1988; Andrews et al., 1997; Mathies and Andrews, 1997; Eulamprus tympanum [Scincidae]: Schwarzkopf and Shine, 1991). Therefore, it is possible that females of those species do not alter their T b s when pregnant simply because the optimal temperatures for their embryos/foetuses may be coincident with the T b s normally regulated by them in their respective habitats, independently of reproductive condition. Nevertheless, other studies indicate that for some viviparous lizard species the preferred body temperatures for pregnant females may be either higher (e.g., Sceloporus bicanthalis (Phrynosomatidae): Andrews et al., 1999) or lower (e.g., Lacerta vivipara (Lacertidae): Van Damme et al., 1987; GvozdíkandCastilla,2001) than328c, andthisillustrates the variability in thermal requirements for embryonic development within the Lacertilia. Unfortunately, there are no data on preferred temperatures of pregnant females for any species of Mabuya; their ideal temperature for embryonic development remains unknown. There was virtually no difference in T b between pregnant and nonpregnant animals collected in wet season months, which may reflect in part the homogeneously warm environmental temperatures prevalent during spring summer periods in tropical latitudes. During the dry season (autumn winter), when daily and monthly ambient temperatures tend to be lower and more variable, there seemed to be a tendency for pregnant females to be active at higher T b s than males (although the difference was not statistically different