It has recently been shown that the ‘timing’ of recurrent epidemics in humans (when in the year they occur) carries important epidemiological information. Here, we use data on cowpox virus in field voles to explore whether related patterns are observable in wildlife (variable abundance) populations. We also combine these data with a model to explore how seasonal infection patterns may differ between humans and wildlife. The focal patterns found in humans, relating the timing of epidemic peaks to their size, and to the size of succeeding peaks, were not repeated. However, an association found in humans between timing and the supply of susceptible hosts was found, though the pattern was not the same. Indeed, timing in our system was associated with both the number and the rate of recruitment of susceptible hosts. A plentiful supply of susceptibles throughout the summer gave rise to a steady rise in infecteds and a late peak. A meagre supply was often insufficient to sustain an increase in infecteds, leading to an early peak followed by a decline. Model results emphasise the importance of the interplay between seasonal infection and recruitment and suggest that our empirical patterns have a relevance extending beyond our own system.