Mycoplasma gallisepticum infection dynamics in a house finch population:: seasonal variation in survival, encounter and transmission rate

1. We considered the impact of an emerging pathogen (Mycoplasma gallisepticum Edward and Kanarek) on apparent survival, encounter and transition rates in a population of a novel host (the house finch, Carpodacus mexicanus Muller). We used a multistate analysis of mark-encounter data from individually marked birds. Individual birds were categorized to a particular disease 'state'; transition rates among states, conditional on apparent survival, were analogous to rates of new infection and recovery from infection. We hypothesized that M. gallisepticum infection would reduce the apparent survival of infected individuals, and that the magnitude of this reduction would vary as a function of the physiological condition of the host (which was characterized in our analyses by including a demographic and an environmental surrogate as covariates). 2. We found consistent support for the hypothesis that M. gallisepticum infection resulted in lower apparent survival among infected individuals, and that recovery rates (from infected to non-infected) were greater than infection rates in this population. We also found strong evidence indicating that infected individuals were less likely to be encountered than were non-infected individuals. Although we predicted that both sex and temperature (proxies for physiological condition) would explain a significant proportion of the variation in our data, only marginal influences of both factors on apparent survival, encounter and state transition rates were detected. 3. Our analyses identified several factors that may be important to studies of disease in the wild. First, disease state assignment may be uncertain, which can complicate parameter estimation. Secondly, encounter rate for infected individuals in our study was low relative to that for non-infected individuals, reflecting possible behavioural changes in infected individuals. Low encounter rates reduces precision of estimated parameters, especially for multistate models. Finally, our results (and mark-recapture models in general) assume independence among individual birds. However, we are aware that there is a social structuring in house finches (and in general for many bird species). Accounting for such non-independence may be especially important for situations where the state transitions are directly related to the pattern of social contact.