The phenotype of a host-parasite association is under the joint control of both protagonists. We have seen that these genotype by genotype interactions can lead to profound differences in the quantitative traits expressed during infection when different combinations are considered. More importantly, these interactions are still present at the local scale, between parasite strains from the same population. We argue that if new combinations of genotypes are frequent, then host and parasite may be unable to achieve any global adaptation of one to the other. These interactions should thus be considered as another level of heterogeneity, similar to environmental heterogeneity, that affects host-parasite coevolution.
The shared control of phenotypic traits of the infection, including host and parasite fitness, may also influence the relationship between these traits. Indeed, even if, in theory, host and parasite fitness are traded off against each other, as for any two functions that use a common resource, variation in the compatibility of particular parasites to particular hosts may render the relationship between host and parasite fitness unpredictable. Increased parasite transmission does not necessarily come at an increased cost to all hosts. As a consequence, virulence evolution may not always be manageable by modifying transmission success except in systems with little variation in basic compatibility or little host genetic diversity. Virulence management of human diseases, which has aroused much interest but remains controversial, might not be globally applicable.