Patterns of disease and host resistance in spatially structured systems

We use data from species of the anther-smut fungi and the host plants Lychnis alpina and Silene dioica to show that spatial structuring at different scales can influence patterns of disease and host resistance. Patterns of disease and host resistance were surveyed in an archipelago subject to land-uplift where populations of S. dioica constitute an age-structured metapopulation, and in three contrasting areas within the mainland range of L. alpina, where population distributions range from continuous, through patchy but spatially connected to highly isolated demes. In S. dioica, disease levels depend on the age, size and density of local patches and populations. Disease is most predictably found in larger dense host patches and populations of intermediate age, and more frequently goes extinct in small old populations. The rate of local disease spread is affected by the level of host resistance; S. dioica populations showing an increase in disease over time are more susceptible than populations where the disease has remained at low levels. Among-population variation in resistance is driven by founding events and populations remain differentiated due to limited gene flow between islands. As observed in the L. alpina system, when populations are more connected, a greater fraction of populations have disease present. Results from a simulation model argue that, while increased dispersal in connected systems can increase disease spread, it may also favour selection of host resistance which ultimately reduces disease levels within populations. This could explain the observed lower disease prevalence in L. alpina in regions where populations are more continuous.