Ecological interactions affecting American chestnut restoration and Allegheny woodrat conservation in Indiana

The Allegheny woodrat (Neotoma magister) is a small rodent native to rocky habitats surrounded by deciduous forests in the eastern United States. Woodrats have recently experienced a severe range-wide decline and are a species of conservation concern throughout their range. In Indiana, the persistence of woodrat populations is threatened by a loss of genetic diversity, so captive propagation was initiated to ameliorate the immediate threat of inbreeding depression and improve local genetic diversity. However, achieving the objectives of captive propagation ultimately depends on the ability of captive-reared individuals to survive in the wild until they can contribute reproductively to the population. Therefore, I monitored survival of hard- and soft-released animals and compared their survival to that of wild juvenile individuals. Survivorship of wild-born individuals was consistently higher than captive-reared individuals. Patterns of survival were best explained by origin (i.e., captive-reared or wild-born) and the interaction between origin and time, with disparities in survival for wild and captive animals decreasing over time. Soft releases improved initial survival, although the importance of the effect of release strategy on survival also decreased over time. Captive propagation may not be the most effective recovery tool for this declining prey species in Indiana, as few individuals likely survived to contribute reproductively to local populations. Allegheny woodrats largely depend on cached hard mast for subsistence over winter and so are sensitive to mast crop failures common to oaks, which were the species most successful at compensating for the loss of American chestnut from their habitat. I provisioned woodrat populations with supplemental food as a way to mitigate this sensitivity and simulate the probable effect of potential chestnut restoration on woodrat populations. I monitored the differential response among populations provisioned consistently, populations receiving a pulsed supplementation, and populations remaining as un-supplemented controls by evaluating population growth rates, apparent survival, recruitment, and individual condition. These measures did not differ statistically among populations assigned to each supplementation treatment, nor did supplementation affect woodrat body mass. Overall, woodrats did not appear to be food limited during the study; mast failures did not occur in the years in which supplementation was provided. Moreover, woodrats were unable to respond positively to a surplus of food resources. In addition to supplementation, I evaluated Allegheny woodrat patterns of preference for chestnuts relative to other hard mast food items to determine whether BC3F2 hybrid chestnut seed represents a desirable food source for woodrats in lieu of pure American chestnut. Woodrats consistently preferred acorns of the red oak group over chestnuts, and BC3F2 chestnut seeds were less preferred than pure American chestnuts, suggesting that woodrats might not perceive these seeds as equivalent food resources. Wide-scale restoration of a hybrid chestnut tree (BC3F 3) is planned, but seed predation by small mammals can have profound impacts on tree species composition and may limit recruitment into plant populations. Therefore, the relative preferences of small mammals for BC3F 3 mast may have important implications for either promoting or limiting the expansion of the BC3F3 from reintroduction sites. I evaluated patterns of preference by small mammal consumers (i.e., white-footed mice, eastern chipmunks, southern flying squirrels, fox squirrels, eastern gray squirrels) for both pure American and BC3F2 chestnuts relative to other hard mast and determined the fate of seeds following selection. The primarily larder-hoarding rodents (mice, chipmunks, flying squirrels) preferred American chestnut seeds over some of the other sources of mast (walnut, hickory, chestnut oak), whereas the scatter-hoarding rodents (fox squirrels, gray squirrels) preferred certain other sources of mast (walnut, hickory, red oak) to chestnut. In general, both groups tended to consume a higher proportion of chestnut than other seed types. In some cases, rodents treated BC 3F2 chestnuts differently than pure American chestnuts, indicating that there are perceived differences between these seeds that influence rodent foraging. Gray squirrels and chipmunks strongly avoided BC3F 2 seeds relative to American chestnut, and fox squirrels left fewer BC3F2 seeds in ground caches within 30 m of feeding stations than American chestnut. The selection of seeds by seed predators is often frequency dependent, such that common or rare mast items are consumed at disproportionately high rates in positive or negative frequency-dependent selection, respectively. I tested for frequency-dependent selection by fox squirrels when chestnut was presented in combination with red or white oak. Fox squirrels showed evidence of positive frequency-dependent selection when combinations of white oak and chestnuts were available, removing more common mast items at a higher rate than expected based on availability. Although I examined seed fate, the observed level of caching was insufficient to allow inferences about how squirrel tendencies to consume or cache might change with frequency of seed types. The loss of chestnut seeds to predation may be dampened if they are planted sparsely throughout white oak stands. However, larger-scale studies of community-level interactions that incorporate chestnut seed fate within the context of different seasons and oak masting events are needed to determine if this outcome can be expected and its longer-term consequences for white oak