We determined genetic variance and gene flow across multiple scales (reaches, headwater segments, and catchments) to examine the dispersal ability of the flatworm Polycelis coronata along the Wasatch Mountains of Utah. Multiple models predict patterns of genetic differentiation in stream invertebrates based on dispersal traits and the spatial connectivity of the habitat. The stream hierarchy model predicts genetic differentiation to be low and gene flow to be high between reaches nested in segments, moderate among segments within catchments, and differentiation to be highest and gene flow lowest among catchments, whereas the headwater model predicts the greatest differentiation between headwater segments. Our objective was to determine which model best described genetic patterns observed in P. coronata. Using a nested hierarchical sampling design ensured that if limitations to dispersal had an effect on genetic differentiation, we would be able to identify at what scale these processes operate. We hypothesized genetic variation would be small within headwater segments and reach maximum levels between headwater segments with no increase in differentiation with increasing distance between headwater patches or between drainages. We do not expect high dispersal along the stream network or across the terrestrial environment (actively or passively).We generated DNA sequence data (mitochondrial COI) from 50 sites nested within 24 segments, which were nested in four adjacent catchments. We identified 134 haplotypes from 506 individuals using a 763 bp fragment of mtDNA. Genetic patterns did not conform to the SH model. Evidence from one drainage (Provo River) was consistent with the headwater model. However, high differentiation within sites suggested that the genetic patterns we uncovered may be representative of high ancestral polymorphism among pre-fragmented populations that were historically widespread. Large effective population sizes and no evidence of bottleneck events suggest incomplete lineage cannot be discounted as an explanation of high differentiation at the smallest scales.