Pairwise invadability analysis is used to examine the evolutionary dynamics of host resistance to microparasitic infection. A continuum of strains of the host differs in susceptibility to infection, with less susceptible strains paying a cost resulting in a lower intrinsic growth rate. With a combination of analytical and graphical pairwise invadability analysis, we show that the evolutionary outcome depends crucially on the shape of the constraint function between resistance and its assumed cost in intrinsic growth rates. When resistance is increasingly costly, a single evolutionarily stable strategy is predicted. Alternatively, with decreasingly costly resistance, we find that the hosts lend to be maximally resistant or not at all resistant. There are conditions under which dimorphism of both these types exists but intermediate resistances do not occur. Independently of the trade-off function used, we are always more likely to get resistant strains of the host when the carrying capacity of the host is high. The pathogenicity of the parasite is also important in determining the likelihood and degree of resistance.