The known benefits of fish cleaning in reef communities are increasing. An ongoing experiment, involving the removal of cleaner wrasse Labroides dimidiatus from seven reefs and left undisturbed on nine reefs (64-285 m²) for 15 years, has shown 1) decreases in the abundance and diversity of large visiting and small resident fishes, and abundance of recruited residents, juvenile visitors, and juvenile conspecifics; 2) a decrease in growth and increase in parasitic copepods in individuals of a resident species; and 3) a skew in the size frequency distribution of individuals towards more small fish in two resident species. However, the mechanisms involved in these community-wide and individual effects remain unclear. Cleaner wrasse eat parasitic gnathiid isopods: ~1200 day^-1 individual^-1. An increase in these harmful blood-feeders could explain the observed effects on fish. Therefore, we tested whether their infection rate differed with cleaner presence. As clients, we used a resident (ambon damselfish, Pomacentrus amboinensis) and a visitor (thick-lipped wrasse, Hemigymnus melapterus). On the experimental reefs, fish were placed in traps (~12h, day/night) which gnathiids could enter but not escape. Few gnathiids (0-3) infected damselfish, with more at night; and they were not affected by cleaner presence. In contrast, more gnathiids infected wrasse; with more on cleaner-absent (0-179) than on cleaner-present (0-14) reefs in the day; at night there was no difference between reefs without (0-25) and with (0-24) cleaners. This is the first demonstration that cleaner wrasse affect the infection rate of gnathiids. We suggest gnathiids are a causative agent for the observed changes in fishes, although this could be restricted to visiting fish which are known to be especially prone to gnathiids.