Mast seeding causes strong fluctuations in populations of forest animals. Thus, this phenomenon can be used as a natural experiment to examine how variation in host abundance affects parasite loads. We investigated fleas infesting yellow-necked mice in beech forest after 2 mast and 2 non-mast years. We tested 2 mutually exclusive scenarios: (1) as predicted by classical models of density-dependent transmission, an increase in host density will cause an increase in ectoparasite abundance (defined as the number of parasites per host), versus (2) an increase in host density will cause a decline in flea abundance ("dilution," which is thought to occur when parasite population growth is slower than that of the host). In addition, we assessed whether masting alters the relationship between host traits (sex and body mass) and flea abundance. We found a hump-shaped relationship between host and flea abundance. Thus, the most basic predictions are too simple to describe ectoparasite dynamics in this system. In addition, masting modified seasonal dynamics of flea abundance, but did not affect the relationship between host traits and flea abundance (individuals with the highest body mass hosted the most fleas; after controlling for body mass, parasite abundance did not vary between sexes). Our results demonstrate that pulses of tree reproduction can indirectly, through changes in host densities, drive patterns of ectoparasite infestation.