1Department of Ecological Botany, Umea University, S-901 87 Umea, Sweden; 2Department of Genetics, Umea University, S-901 87 Umea, Sweden; 3Swedish Biodiversity Centre, Box 7007, S-750 07 Uppsala, Sweden

Background and objectives
The pollinator-transmitted anther-smut fungus Microbotryum violaccum (syn. Ustilago violacea), causes a systemic infection in a wide range of species within the Caryophyllaceae. Since infected plants are permanently sterile, M. violaceum has, potentially, a great effect on the fitness and dynamics of host-plant populations.

In an archipelago in northern Sweden, populations of the perennial, dioecious herb Silene dioica (Caryophyllaceae) are often infected with M. violaceum. This area has undergone a constant rate of land uplift since the last glaciation and the height of an island above sea level is thus correlated with island age and the time since it was first colonized by plants. S. dioica populations established from a small number of founders [2] when islands are 70-150 years old. The populations increase to large sizes and high densities, after which they decrease and eventually become extinct on older islands. In general, populations on young and old islands have low incidences of disease, whereas all populations on intermediate-aged islands are diseased, with incidences varying between 5 and 50%. The large variation in incidence of disease observed among populations partly can be attributed to changes in host population sizes and densities along the successional age axis. However, even if host populations are more likely to be diseased when large, some large and dense populations have escaped epidemics, suggesting that these populations may contain high frequencies of resistant plants. The aim of our studies was therefore to evaluate the importance of resistance in explaining patterns of disease across the archipelago.

Materials and methods
Several approaches were used to determine whether genotypes and populations of S. dioica differ in resistance. First, replicated genotypes were artificially inoculated with M. violaceum and scored for disease. Second, disease levels were monitored on plants exposed to naturally spread disease in field transplantation experiments.

Results and conclusions

Several of the results from our studies suggest that variation in host resistance influences patterns of disease in populations of S. dioica in the archipelago studied. First, there are clear differences in susceptibility among host plant genotypes inoculated manually with fungal isolates. Second, in a transplantation experiments, disease levels showed large variations among transplants from different populations [1].

There are several processes that could affect the resistance structure among S. dioica populations in the study area. First, our results indicate that disease results in selection for resistance within populations [1]. The frequency of resistant plants is therefore expected to increase and disease incidences to decline over time - a pattern observed in the archipelago. Second, electrophoretic studies have shown that the genetic structure of S. dioica populations is influenced by founder effects and restricted gene flow [2]. If resistance has a genetic basis, these processes may also have strong effects on the resistance structure among host populations in the archipelago. For example, populations established by a few fully susceptible plants may remain relatively susceptible and experience high disease levels later on, while populations founded by resistant plants may never reach high disease levels. This may explain why healthy plants from some highly diseased populations still show high levels of susceptibility after exposure to their own pathogen populations under experimental conditions [1]. If newly founded populations are composed mostly of susceptible or resistant genotypes, and if most of the individuals comprising intermediate aged populations have decended from these founders, the process of selection at the level of the archipelago may be slow.

It is important to note that knowledge about the resistance structure alone cannot explain the observed patterns of disease in the archipelago. Plants from some of the older disease-free populations have been found to be highly susceptible to the pathogens present on nearby islands [1]. In these populations, the average low density of flowering plants and the high degree of patchiness may be factors that prevent inoculum from increasing above a threshold needed for successful infection and effective transmission.

1. Carlsson-Graner, U 1997.