Differential persistence of individual deer mice in a population: The role of disease mediated by environmental stressors

Background/Question/Methods Endemic pathogens with long co-evolutionary histories with their hosts are predicted to have evolved to have a negligible impact on host fitness. We tested this hypothesis by examining the effects of Sin Nombre virus (SNV) infection on the survival of deer mice (Peromyscus maniculatus). Because the negative effects of pathogens may be intensified with environmental stressors, we also investigated the role of human disturbance, environmental conditions and high population density. From May 2002 to September 2007, we conducted seasonal (May and September) live-trapping in sagebrush steppe habitat in 19 sites in central Utah that differed in degree of disturbance by off-road vehicles. Captured deer mice were weighed, sexed, uniquely marked with eartags and a blood sample collected for subsequent testing for SNV antibodies. Animals were released at the site of capture. Sampling periods were classified as “dry” or “wet” based on precipitation. Generalized linear mixed models were used to estimate the probability of surviving to a later season as a function of infection status and its interaction with disturbance, precipitation, density, and the individual’s sex and weight; site ID was used as a random factor.

Results/Conclusions

A total of 3768 unique deer mice were captured over the six years of the study, of which only 264 (7%) were captured at a latter season. Males were twice more likely to persist on the site between sampling periods than females. Deer mice survival probability was negatively affected by density, particularly in highly disturbed sites. Survival on low disturbance sites exhibited strong seasonality with lower persistence over winter than over summer. In contrast, deer mice on high disturbance sites had low survival regardless of the season. Survival was affected by the interaction between disturbance, precipitation, and the infection status. Survival of uninfected individuals during dry conditions was greatly reduced at high disturbance sites compared to low disturbance sites; whereas infected individuals had lower survival probabilities in all conditions. Furthermore, survival probability was predicted by the interaction between infection status and the initial body mass. Large, uninfected deer mice had greater survival probabilities compared to large infected ones. The positive effect of weight on survival probabilities was reduced for infected individuals. Our results indicate that SNV infection in deer mice affects survival and that the effect is complex and differentially mediated by a variety of environmental factors.