The Everglades, a wetland ecosystem unique to southern Florida has been degraded by the loss of nearly half of its area to urban and agricultural development as well as by alterations to Florida’s hydrology. Modifications to the flow of water to the Everglades have altered the remaining portion of the Everglades. Most prominent among these changes in the remaining Everglades is the loss of tree islands with a disproportionately greater loss of larger tree islands. Despite their significance as “keystone habitats”, our understanding of how changes to Everglades tree islands will affect fauna is poor. In the work presented, 16 study tree islands of Rock Reef Pass, Everglades National Park, Homestead, FL, were classified into three size classes (large, medium and small) and used animal capture histories collected between February 1994 and December 2005 to investigate the relationship between tree island size classes and indirect indicators of adult female hispid cotton rat (Sigmodon hispidus) and marsh rice rat (Oryzomys palustris) fitness and reproduction. Adult female S. hispidus and O. palustris body weights were used as an indicator of adult female fitness and compared among tree island size classes. The proportions of reproductively active females, the proportions of captures of individuals in the juvenile age class and the number of juveniles per female were used as indirect indicators of female reproduction and compared among island size classes. Animal capture histories were also used to perform multi-state mark-recapture analyses, where sates were defined as three island size classes, to draw inferences about state-specific rates of survival and state transition probabilities for Oryzomys palustris and Sigmodon hispidus as a function of the most salient features of the Everglades ecosystem, namely, its tree islands and its seasonal wet and dry periods. Specifically, Akaike’s, information criterion (AIC) was used to make inferences about factors influencing survival and transition probabilities by comparing the relative fit among models where survival and transition rates are described as functions of tree island size classes and as a function of environmental variables that distinguish seasonal periods (monthly water levels, total monthly rainfall, mean air temperature, days dry, days inundated). The probabilities of individuals remaining within the same island size class were interpreted as an indirect indicator of size-class specific tree island use. Results indicated that greater proportions of S. hispidus adult females were reproductively active and had higher fitness on larger islands than females on smaller islands. The body weights of S. hispidus adult females, the proportions of individuals in the juvenile age class and the number of juveniles per female, were positively correlated with tree island area while the proportions of reproductively active S. hispidus females were not. When compared among tree island size classes, S. hispidus females recaptured on the same large island weighed more than females recaptured on the same and medium or small island suggesting that females with greater fitness are found on larger tree islands. The proportions of reproductively active S. hispidus females and of individuals in the juvenile age class were greater on larger tree islands than on smaller islands, although the number of juveniles per females did not differ among tree islands. The results also indicated that differences in the fitness and reproductive condition of O. palustris females are not as distinct among females on different size tree islands. The body weights O. palustris adult females, the proportions of juveniles and the number of juveniles per female were not correlated with island area, while the proportions of reproductively active females were negatively correlated with tree island area. The body weights of O. palustris adult females recaptured on the same large tree island were higher than those of females on smaller islands while the proportions of females that were reproductively active and the proportions of captures of juveniles were higher on small islands than on large islands. Mark-recapture analysis provided evidence that overall, Sigmodon hispidus survival rates were higher on larger islands than on smaller islands. Evidence was not found that Oryzomys palustris survival rates differed among size classes, perhaps due to the limited spatial scale of this study relative to the spatial sale of O. palustris habitat use. Both species’ survival probabilities differed between seasons with Oryzomys palustris having higher survivorship during wet seasons and Sigmodon hispidus during the dry season. Both Oryzomys palustris and Sigmodon hispidus were more likely to remain on large island than on smaller islands.

The marsh rice rat (Oryzomys palustris) is a semi-aquatic rodent classified as state-threatened in Illinois due to historical wetland loss coupled with being on the northern periphery of its range. The most recent survey for this species in Illinois occurred in 1988, and this study reassessed its status and distribution. From 2007 to 2009, I used live traps to sample for rice rats at previously occupied and random wetland-dominated sites in 5 southern Illinois watersheds. I captured 132 individual rice rats 192 times in 13,248 trap nights. Capture success was 15.10 captures/1,000 trap nights. I detected rice rats within 3 of 5 watersheds, 16 of 48 sites, and 5 new locations. Most rice rats were captured in permanent or semi-permanent emergent wetlands and rice rat occurrence was strongly related with coal mine-associated wetlands. I collected microhabitat measurements at 10% of trap locations and recorded landcover types within and around sites. I analyzed data using t-tests, logistic regression, and occupancy modeling. Percent herbaceous cover and percent visual obstruction (0.0-0.5m) positively influenced rice rat occurrence and were among the most important microhabitat models. The best model in the candidate set of landcover variables included proportion of upland grass in areas surrounding wetlands as a predictor of rice rat occurrence. I believe insufficient evidence exists to warrant removal of the threatened status of the marsh rice rat. However, current management practices, such as wetland restoration, grassland restoration, and mowing of roadside ditches, likely benefit rice rat populations. Slight modification of these management activities may facilitate the eventual delisting of marsh rice rats.