Samples of the New Zealand cockle Austrovenus stutchburyi (Veneridae) collected at two different intertidal levels were found to be heavily infected by the echinostome trematode Curtuteria australis which use the mud flat whelk Cominella glandiformis and shorebirds as first and definitive host, respectively. Parasite loads were on average more than 2-fold higher at the high site than at the low site, whereas the density of Curtuteria australis metacercariae per square metre showed the opposite trend. This pattern appeared unrelated to the abundance of first intermediate hosts and C. australis prevalence within this host that did not differ between sites. At both sites there was a positive relationship between parasite load and cockle condition, a likely by-product of the relationship between condition-based filtration capacity and the rate at which individual cockles encounter the free-swimming trematode larvae. The results suggest that the between-site variation in parasitism is a consequence of the much higher density of cockles at the low site. There, the dense cockle population is able to clear the water column of parasites and consequently individual cockles virtually ‘compete’ for parasite larvae. This will depress mean infection intensity per cockle but elevate mean parasite density per square metre.