Heterokaryosis was recently reported in the chestnut blight fungus, Cryphonectria parasitica, in which individuals contain nuclei that are isogenic except at the mating-type locus (MAT). MAT heterokaryons were found in several natural populations, including a putatively clonal population in West Salem, Wisconsin, providing an opportunity to address the question of how heterokaryons arise. We represented relationships among RFLP fingerprint haplotypes as networks in which loop formation is considered evidence of recombination. From 1990 to 1995, this population was clonal, as indicated by a simple haplotype network without loops, and the correlation of vegetative compatibility (vc) types and mating types with haplotype lineages. By 1999, we observed loops in the haplotype network involving isolates of two vc types (WS-2 and WS-3). Isolates with haplotypes in the loops were either MAT heterokaryons, carried the opposite mating type from other isolates of the same vc type, and/or had two alleles at two or more codominant SCAR (sequence-characterized amplified region) loci. Segregation of markers and recombination were evident among single-spore isolates from one heterokaryon; these single-spore isolates had novel fingerprint haplotypes, also within the loops. In contrast, vc type WS-1, which comprises 85% of the population, was represented by a simple network with no loops, indicating a clonal lineage varying only by mutation. Almost all isolates of WS-1 had the same mating type; the exceptions were five isolates that were MAT heterokaryons. These results are consistent with the hypothesis that heterokaryons formed between vegetatively incompatible individuals, and recombination occurred by a parasexual process.