In order to disentangle the contribution of host and parasite biology to host specificity, we compared the structure and population dynamics of the Gyrodactylus (von Nordmann, 1832) flatworm community living on sympatric three-spined Gasterosteus aculeatus L. and nine-spined Pungitius pungitius (L.) stickleback. Between April 2002 and March 2003, a small lowland creek was sampled monthly. Species identity of about 75% of the worms per host was determined with a genetic nuclear marker (ITS1). Each stickleback species hosted a characteristic gill- and fin-parasitic Gyrodactylus: G. arcuatus Bychowsky, 1933 and G. gasterostei Gläser, 1974 respectively infecting the three-spined stickleback, with G. rarus Wegener, 1910 and G. pungitii Malmberg, 1964 infecting the nine-spined stickleback. Host size and seasonal dynamics were strong determinants of parasite abundance. A strong interaction between host and parasite species determined infection levels and affected three levels of parasite organisation: community structure, population structure and topographical specialisation. Community and population structure were shaped by asymmetric cross-infections, resulting in a net transmission of the Gyrodactylus species typical of the nine-spined stickleback towards the three-spined stickleback. Host density was not a major determinant of parasite exchange. Aggregation and topographical specialisation of the Gyrodactylus species of the three-spined stickleback were more pronounced than that of the nine-spined stickleback.
Parasites with high host specificity maximally depend on their hosts, which should increase the likelihood of coevolution. However, coevolution requires reciprocal selection exerted by the host and the parasite, and thus a considerable level of parasite virulence. In species of the monogenean ectoparasite genus Gyrodactylus consecutive generations are confronted with a single host, which may constrain the evolution of virulence. Transmission, which is often important in the ecology of Gyrodactylus species, may have the opposite effect, but may also lead to the avoidance of coevolutionary arms races. We investigated the potential outcome of coevolution between Gyrodactylus gasterostei Gläser, 1974 and its host, the three-spined stickleback (Gasterosteus aculeatus L.) by determining the strength of genotype by genotype (G×G) interactions on two levels: within and between sympatric and allopatric host populations. To do so, we compared the parasite's infection dynamics on laboratory-reared sympatric (Belgian) and allopatric (German) hosts. We found that a parasite line successfully infected a range of sympatric host genotypes (represented by families), while it failed to establish on allopatric hosts. Phylogeographic studies suggest that neutral genetic divergence between the host populations cannot explain this dramatic difference. Provided that this result can be generalised towards other parasite lines, we conclude that coevolution in this host-parasite system is more likely to lead to local adaptation on the population level than to G×G interactions within populations.