Virulence is thought to be a driving force in host-pathogen coevolution. Theoretical models suggest that virulence is an unavoidable consequence of pathogens evolving towards a high rate of intrahost reproduction. These models predict a positive correlation between the reproductive fitness of a pathogen and its level of virulence. Theoretical models also suggest that the demography and genetic structure of a host population can influence the evolution of virulence. If evolution occurs faster in pathogen populations than in host populations, the predicted result is local adaptation of the pathogen population. In our studies, we used a combination of molecular and physiological markers to test these hypotheses in an agricultural system. We isolated five strains of the fungal pathogen Mycosphaerella graminicola from each of two wheat cultivars that differed in their level of resistance to this pathogen. Each of the 10 fungal strains had distinct genotypes as indicated by different DNA fingerprints. These fungal strains were re-inoculated onto the same two host cultivars in a field experiment and their genotype frequencies were monitored over several generations of asexual reproduction. We also measured the virulence of these 10 fungal strains and correlated it to the reproductive fitness of each fungal strain. We found that host genotypes had a strong impact on the dynamics of the pathogen populations. The pathogen population collected from the moderately resistant cultivar Madsen showed greater stability, higher genotype diversity, and smaller selection coefficients than the pathogen populations collected from the susceptible cultivar Stephens or a mixture of the two host cultivars. The pathogen collection from the mixed host population was midway between the two pure lines for most parameters measured. Our results also revealed that the measures of reproductive fitness and virulence of a pathogen strain were not always correlated. The pathogen strains varied in their patterns of local adaptation, ranging from locally adapted to locally maladapted.

Three cultivars of winter bread wheat (Gene, Madsen and Stephens) were each inoculated as seedlings in the greenhouse with seven or eight individual isolates of Mycosphaerella graminicola collected in 1997 from each of the same cultivars in the field. Isolates collected from Gene were virulent to all three cultivars, while isolates obtained from Madsen and Stephens were virulent to those two cultivars and, in all but one case, avirulent to Gene. At its release in 1992, Gene was resistant to M. graminicola, as indicated by both field observations and greenhouse tests, but by 1995 its resistance had substantially deteriorated. This indicated that its resistance was vertical (sensu Vanderplank) or race-specific, and that commercial cultivation of Gene rapidly selected for strains in the local M. graminicola population that were specifically adapted to overcome its resistance.