Host and symbiont often conflict over patterns of symbiont transmission. Symbionts favour dispersal out of the host to avoid competition with close relatives. Migration leads to competition among different symbiotic lineages, with potentially virulent side-effects on the host. The hosts are favoured to restrict symbiont migration and reduce the virulent tendencies of the symbionts. Reduced mixing of symbionts would, in many cases, lower symbiont virulence and increase the mean fitness of the host population. But a host modifier allele that reduced symbiont mixing increases only when directly associated with reduced virulence. The association between modifiers and reduced virulence depends on the particular details of symbiont biology. The importance of this direct association between modifier and virulence was first noted by Hoekstra (1987) when studying the evolution of uniparental inheritance of cytoplasmic elements. I apply Hoekstra's insight to a wide range of host-symbiont life histories, expanding the scope beyond cytoplasmic inheritance and genomic conflict. My comparison of differing symbiont life histories leads to a careful analysis of the conditions under which hosts are favoured to control mixing of their symbionts.