The historical and contemporary population genetic structure of the chickpea Ascochyta blight pathogen, Ascochyta rabiei (teleomorph: Didymella rabiei), was determined in the US Pacific Northwest (PNW) using 17 putative AFLP loci, four genetically characterized, sequence-tagged microsatellite loci (STMS) and the mating type locus (MAT). A single multilocus genotype of A. rabiei (MAT1-1) was detected in 1983, which represented the first recorded appearance of Ascochyta blight of chickpea in the PNW. During the following year many additional alleles, including the other mating type allele (MAT1-2), were detected. By 1987, all alleles currently found in the PNW had been introduced. Highly significant genetic differentiation was detected among contemporary subpopulations from different hosts and geographical locations indicating restricted gene flow and/or genetic drift occurring within and among subpopulations and possible selection by host cultivar. Two distinct populations were inferred with high posterior probability which correlated to host of origin and date of sample using Bayesian model-based population structure analyses of multilocus genotypes. Allele frequencies, genotype distributions and population assignment probabilities were significantly different between the historical and contemporary samples of isolates and between isolates sampled from a resistance screening nursery and those sampled from commercial chickpea fields. A random mating model could not be rejected in any subpopulation, indicating the importance of the sexual stage of the fungus both as a source of primary inoculum for Ascochyta blight epidemics and potentially adaptive genotypic diversity.