The concept of geographic variation in phenotypes has a long history in evolutionary biology. Differences in phenotypes can stem from both local adaptation, in which differentiation occurs at the genetic level, and phenotypic plasticity, in which differentiation is a direct effect of an animal's environment. Although not mutually exclusive, each of these pathways contributes to organismal fitness. I examined the effects of environmental factors, including development, temperature, latitude, altitude, and habitat structure on the growth, body morphology and performance of two terrestrial ectotherms. Eleutherodactylus coqui is a direct-developing frog whose life history lends itself to studies of development, and Sceloporus occidentalis is a wide-ranging temperate lizard suited for studies of geographic variation. In the laboratory, I manipulated the development of these ectotherms either by introducing a predation threat or by varying incubation temperature. I then examined the effects of these factors on the animals' phenotype at hatching and, in the case of S. occidentalis, for a number of weeks thereafter. This design allowed me to test the persistence of phenotypic plasticity. Finally, I measured the correlations among morphology, performance and habitat use in four wild populations of lizards. I found that hatching early in response to predation can have dramatic effects on the morphology and performance of E. coqui froglets: animals that hatch early to escape predation suffer a decrease in jumping ability, resulting in a trade-off between life history stages. In S. occidentalis, incubation temperature affects morphology and performance over time, but some morphological characteristics that contribute to performance capacities, such as hindlimb length, vary instead across latitudinal and altitudinal gradients. Finally, despite morphological and performance differences among populations of wild-caught lizards, S. occidentalis represents a generalized species that can utilize many different habitat types without large corresponding changes in body form. Here I argue that developmental responses to the environment can be adaptive, but also that traits that have direct effects on fitness tend to be more canalized in these animals. Further studies on geographic variation should include both of these factors when determining the evolutionary history and possible evolutionary trajectory of ectotherms.

Wood turtles (Glyptemys insculpta) are of conservation interest rangewide. Anecdotal accounts demonstrate that some populations have been decimated since 1850, and recent studies demonstrate that declines are still underway. From 2004–2008 I investigated the ecology of wood turtles in Massachusetts and New Hampshire. I obtained between one and five years of annual home range data for 150 turtles, and evaluated population structure at 31 sites in five major watersheds. Seasonal floods displaced 7% of wood turtles annually in one watershed, and accounted for elevated mortality. Twelve wood turtles were displaced < 16.8 km, and two were displaced over a 65-foot dam. Several turtles overwintered at their displacement site and two returned successfully, indicating that floods are a mechanism of population connectivity. Several homing turtles ended up in new areas. Turtles occupied stream segments with gradient < 1%, lower than generally available. Agricultural machinery accounted for most observed mortality, followed by automobiles and mammals. Female turtles exhibit smaller home ranges in agricultural areas. Older turtles move farther from the river than do young turtles, possibly reflecting their familiarity with a former landscape. Population density ranged from 0–40.4 turtles/river-kilometer. The highest densities occur in central New Hampshire and lower densities occur in the Housatonic watershed. Population density is negatively correlated with agriculture at both riparian and watershed scales, and responds unimodally to forest cover. Wood turtle populations in western Massachusetts are declining by 6.6–11.2% annually. I estimated ages of turtles by assessing shell-wear rates from photographs. Wood turtles regularly achieve ages over 80 years, and like related species, do not exhibit clear signs of senescence. Old wood turtles are reproductively dominant, and their survival rates are twice as high as young turtles. Carapace scutes appear to require 80 years to become worn. Population modeling indicates that wood turtle populations are declining in New England due to anthropogenic and natural factors. Conservation efforts must address the effects of agriculture on adult survival. Climate change may negatively affect northeastern wood turtles through increased flooding. Populations in mountainous areas may be likely candidates for conservation because they don't occupy prime agricultural land, but may be more susceptible to floods.