Movements of the parasitic honey bee mite,Varroa jacobsoni (Oud.) were monitored in several assays as they moved among adult host honey bees,Apis mellifera. We examined the propensity of mites to leave their hosts and to move onto new bee hosts. We also examined their preference for bees of different age and hive function. Mites were standardized by selecting mites from newly emerged worker bees (NEWs). In closed jars, 50% ofVarroa left NEWs irreversibly when no physical path was present for the mites to return to the NEWs; about 90% of mites left newly emerged drones in identical assays. In petri dish arenas, mites were rarely seen off NEW hosts when monitored at 15-min intervals for 4 h; this was the case for single NEWs with one mite (NEWs+) and when a NEW+ and a NEW− (no mites) were placed together in a petri dish. When a NEW+ was held with either a nurse beeor a pollen forager, 25% of the mites moved to the older bees. When both a nurseand a pollen forager were placed in a petri dish with a NEW+, about 50% of the mites transferred to older bees; nurse bees received about 80% of these mites, whereas pollen foragers received significantly fewer mites (about 20%,P 0.05), implying that the factors mediating the mites’ adult-host preference change gradually with bee age. A possible chemical basis for host choice byVarroa is indicated by their greater propensity to move onto freezer-killed nurse bees than onto freezer-killed pollen foragers (P < 0.05) and by their lower movement onto heat-treated bees than onto control bees (P < 0.05). Bee age, hive function, and directional changes in cuticular chemistry are all correlated. Movements of newly emerged mites in relation to these variables may provide insights into their reproductive success inApis mellifera colonies.