Solanum microdontum is a diploid potato species with features that make it a good model for research into management and use of germplasm in the genebank. Its taxonomic status is unambiguous and it is in the taxonomic series of wild species closest to cultivated forms. It is represented by about 100 populations in the genebank-not too many for comprehensive evaluation, yet not too few to make prioritization of the most valuable populations worthwhile. This species is also particularly rich in desirable traits, often exhibiting very broad segregation. We here report use of DNA markers for selection of a core set of populations, and assessing whether that core captures populations with the most desirable evaluation results for economic traits. DNA was extracted from bulks of 27 plants from each of 94 populations to generate AFLPs. A total of 1,741 informative loci were detected. AFLP loci were treated as though they were traits, with the banded condition considered to be the desired state to include in a core set. At least one band unique to a population was present in 45 populations, and these 45 populations together captured 98 % of all bands. Adding another 14 populations for a core of 59 populations captured 100 % of bands. This core set was assessed for whether it encompassed those populations known to have useful traits, including nutritional and quality components; and disease, stress and pest resistances. As with AFLP bands, 25 of 26 of the most desirable phenotypic traits were also found in populations in the core set of 59 populations. The most desirable status of 3 traits is lost by selecting a core of 45 populations. We conclude that these core sets would be a rational starting point when prospecting for new useful traits in microdontum. © 2013 Potato Association of America.