Genetic markers from horizontal starch gel electrophoresis show clearly that boreal-arctic C. stygium, a bisexual polyploid with n  =  14, is closely related to C. arcticum and C. latifolium, unisexual haploids with n  =  7. One locus of C. stygium shows fixed heterozygosity, suggesting an allopolyploid origin with these two haploids as progenitors. Two other loci, which are heterozygous in >90% of all plants screened, also support this view. It is also possible, however, that C. stygium is an autopolyploid, most likely of C. arcticum, in which meiosis has become regularized. Despite strong morphological differences between them, we detected only one fixed allelic difference between C. arcticum and C. latifolium. For C. subrotundum, another bisexual polyploid with n  =  14, the evidence for allopolyploidy is very strong, with seven fixed heterozygous loci out of 17 scored. One of the lineages involved is similar to that of C. arcticum and C. latifolium, but the other is highly divergent from any known extant species of Cinclidium. It is quite possible, however, that additional, highly divergent species exist or, at least, once existed, as our screening of populations from Alaska turned up a new, highly genetically distinct species that experts have previously mistakenly referred to C. latifolium.