Most wild potato germplasm in genebanks is collected, preserved, and evaluated as botanical seed populations that may be highly heterozygous and heterogeneous. However, some species are selfers so potentially very homozygous, perhaps also homogeneous. Intrapopulation heterogeneity increases sampling error that can undermine consistency in seed regeneration in the genebank, screening results, germplasm collecting, and estimates of taxonomic relationships. Thus, knowledge of genetic heterogeneity (GH) can predict the need to commit more resources for larger sample sizes or replication when populations of a given species are being regenerated, evaluated, collected, and classified. This study investigated within-population GH in 32 potato populations comprising four different breeding systems observed in Solanum species: 5. fendleri (2n=4x=48, disomic selfer), 5. jamesii (2n=2x=24, outcrosser), S. sucrense (2n=4x=48, tetrasomic outcrosser), and S. verrucosum (2n=2x=24, selfer). RAPD markers were used to estimate heterogeneity among 24 individuals per population. Populations of S. verrucosum. were quite homogeneous with an average GH of 6.0%. Similarly low heterogeneity was detected among the eight populations of S. fendleri (average GH=7.1%). In contrast, S. jamesii and S. sucrense had a much higher GH of 29.4% and 44.1%, respectively. These results demonstrate and quantify the great difference in intrapopulation heterogeneity among wild potato species. Calculations based on intrapopulation heterogeneity indicate that samples should be composed of 25 to 30 random plants for low sample variation that is uniform for all species.