Developing genetic markers for the identification of recently diverged groups, such as ecotypes or species complexes, remains difficult due to challenges with incomplete lineage sorting, hybridization and introgression. Genome-wide scans of single nucleotide polymorphisms (SNPs) have proven useful for inferring patterns of genetic differentiation at the population level. In combination with a new analytical technique, the discriminant analysis of principal components (DAPC), and within the framework of iterative taxonomy, it may also be possible to extract a combination of SNPs as markers for the delimitation of closely related groups. In addition, since DAPC identifies the loci contributing the most to group clustering, it may be possible to link putative biological function to differences that define group boundaries. We tested this technique on two ecotypes of the hemlock looper (Lambdina fiscellaria), which differ in terms of number of larval stadia, developmental rate and fecundity. It was possible to separately cluster the two ecotypes with 95% correct assignment using 27 SNPs. We also determined that a storage hexamerin carried eight of these SNPs, including the two highest contributing loci, of which the top contributor was nonsynonymous. Other studies have found this protein to be highly expressed just before metamorphosis, pointing to a possible connection between its role in clustering ecotypes and its biological function. These SNP markers can now be further developed for high throughput delimitation of individuals of unknown ecotype identity.