The application of DNA barcoding to distinguish between two or more closely related taxa has been used more frequently in recent years. The typical approach has been to isolate, amplify and sequence cytochrome c oxidase subunit 1 (COI), a mitochondrial DNA (mtDNA) encoded subunit of respiratory Complex IV. COI has proven useful as a marker for identification purposes because the frequency of mutations for this gene is relatively high. Thus, conserved mutations and variability in the COI sequence can be used to determine relatedness of individuals. Many valid criticisms about the practice have arisen, not the least of which is that intraspecific COI variability has not been examined and compared across the entire range of a given species. Rather, most studies employing DNA barcoding focus on relatively few individuals and even fewer sites. Together, this may underestimate intraspecific variance, confounding efforts to distinguish between intraspecific and interspecific differences. Therefore, criteria for delimiting species may need revision. To test whether intraspecific differences at COI are influenced by geographical scale, a COI-barcode library was constructed for three species of North American mayflies (Ephemeroptera) across their natural distribution. This order is important for water quality monitoring of streams and rivers and hence species level identifications have the potential for great application. These three species were chosen because they had relatively wide distributions (i.e., throughout eastern North America) yet were presently considered single species based on morphological characters. Sampling sites included in the study were widespread and represented a range of geographical diversity. Two of three species examined (i.e., Eurylophella funeralis and Leptophlebia cupida) exhibited genetic differences between individuals that frequently exceeded 2% base pair deviation at the COI locus. There were three or more distinct barcode clusters within each of these two species. Our data suggests that these two species may represent species complexes that are morphologically cryptic. In contrast, genetic differences between individuals for species (Siphloplecton basale) did not greatly exceed the 2% base pair deviation at the COI locus. The presence of morphologically cryptic species within Eurylophella funeralis and Leptophlebia cupida illustrates the need for a robust library of barcodes with morphological vouchers for North American mayflies to resolve the phylogenetics of this group so their contributions in water quality assessments can be maximized.