MGR586 DNA fingerprinting has been widely used to characterize population diversity of the rice blast pathogen, Pyricularia grisea. However, the frequency and distribution of particular haplotypes (individuals) within MGR-delimited lineages has not been examined in the United States. MGR586 DNA fingerprinting, mitochondrial DNA (mtDNA) restriction fragment length polymorphism (RFLPs), and virulence phenotyping were used to examine genetic diversity of P. grisea in Arkansas. A total of 470 monoconidial isolates were recovered from eight rice cultivars in 18 commercial fields in nine counties in Arkansas. All isolates were examined for nuclear DNA RFLPs with the MGR586 DNA fingerprint probe, and both the MGR lineage (isolates with >80% similarity) and the haplotype frequencies were determined. Four distinct MGR586 DNA fingerprint lineages (designated A, B, C, and D) were identified among the 470 field isolates. All four lineages were found in 9 of the 18 locations. Three lineages were found in four locations, two lineages in three locations, and only a single lineage was found at two locations. In all, 10, 19, 16, and 13 haplotypes (isolates which had MGR586 DNA fingerprints which differed by 1 to 20%) were identified within lineages A, B, C, and D, respectively, among the 470 isolates examined. Within each lineage, a single haplotype (clone) predominated, representing 51 to 71% of the isolates collected for each of the four lineages. Overall, 60% of the 470 isolates belonged to one of only four haplotypes (A1, B1, C1, and D1) and these four predominant haplotypes were recovered from between 7 and 14 of the 18 locations sampled, indicating a widespread distribution of these four clones. These data indicate an exceptionally low level of genetic diversity in the regional rice blast pathogen population in Arkansas relative to several other populations of P. grisea examined from tropical environments. In addition, no mtDNA RFLPs were detected among representative haplotypes within each of the lineages, indicating a single mtDNA haplotype was present in the population. Examination of virulence indicated that two races predominated in the regional collection. All 30 isolates in lineages A and C tested had an IB-49 virulence phenotype. Out of 30 isolates in lineages B and D, 29 had an IC-17 virulence phenotype. One isolate in lineage B, isolated from a highly susceptible cultivar (L201), had an IG-1 virulence phenotype. The frequencies of the four lineages varied among the locations sampled and may have been due, in part, to the cultivar from which isolates were recovered. A single lineage was recovered from two cultivars, Mars and Millie. Although only a single field of each of these cultivars was sampled, the data indicate that certain cultivars grown in Arkansas may serve as a "bottleneck", selecting out specific lineages in the regional population. To test this hypothesis, an additional 283 isolates were recovered from replicated plots of cvs. M204 and Mars located within commercial rice fields at two locations during two seasons. All four MGR586 lineages were recovered from each location. However, there was a strong bias for lineage B on cv. M204 (79% of all isolates) and a strong bias for lineage A on cv. Mars (95% of all isolates), indicating some cultivars were effective in excluding certain lineages.