The genetic structure of populations of Cronartium ribicola was studied by sampling nine populations from five provinces in eastern Canada and generating DNA profiles using nine random amplified polymorphic DNA markers. Most of the total gene diversity (H(t) = 0.386) was present within populations (H(w) = 0.370), resulting in a low level of genetic differentiation among populations in northeastern North America (F(st) = 0.062). A hierarchical analysis of genetic structure using an analysis of molecular variance (AMOVA) revealed no statistically significant genetic differentiation among provinces or among regions. Yet, genetic differentiation among populations within regions or provinces was small (AMOVA phi(st) = 0.078) but statistically significant (P < 0.001) and was several orders of magnitude larger than differentiation among provinces. This is consistent with a scenario of subpopulations within a metapopulation, in which random drift following migration and new colonization are major evolutionary forces. A phenetic analysis using genetic distances revealed no apparent correlation between genetic distance and the province of origin of the populations. The hypothesis of isolation-by-distance in the eastern populations of C. ribicola was rejected by computing Mantel correlation coefficients between genetic and geographic distance matrices (P > 0.05). These results show that eastern Canadian provinces are part of the same white pine blister rust epidemiological unit. Nursery distribution systems are controlled provincially, with virtually no seedling movement among provinces; therefore, infected nursery material may not play an important role in the dissemination of this disease. Long-distance spore dispersal across provincial boundaries appears to be an epidemiologically important factor for this pathogen.

The fine-level genetic structure of the white pine blister rust agent, Cronartium ribicola, was investigated by sampling multiple monokaryotic spermogonia directly on cankers in four eastern Canadian white pine (Pinus strobus) plantations and assessing genetic variability, using random amplified polymorphic DNA (RAPD) markers. Ninety-eight percent of the cankers surveyed contained a single DNA haplotype, suggesting spermogonia within cankers are the result of clonal reproduction. A single canker contained two haplotypes that were divided between the upper and lower parts of the canker, suggesting it represented two confluent cankers. In contrast, genotypic diversity was high among cankers. Thirty-seven haplotypes were found among forty-three cankers sampled, and an analysis of molecular variance indicated that 93% (P < 0.001) of the total genetic diversity was attributable to sampling of different cankers, strongly suggesting that multiple infections do not take place in the white pine blister rust pathosystem, i.e., a canker is the result of infection by a single genotype. This result is in contrast with the high level of genetic diversity previously reported among dikaryotic aecidia within cankers and is consistent with the hypothesis that variability in the aecidial stage is the result of outcrossing between resident spermogonia and alien spermatia. The genetic structure of the spermogonial stage, which is the vegetative extension of infection by basidiospores and, therefore, the indirect result of meiosis, was consistent with random mating; the observed genotypic diversity was not significantly different (P > 0.05) from the genotypic diversity expected under the assumption of panmixis. The results indicate that monokaryotic cankers can be genotyped by sampling a single unopened spermogonia per canker.