The fecal pellet-plot method has been used extensively for snowshoe hare (Lepus americanus) population studies across the species' range, but potential biases associated with the technique have not been addressed adequately. We studied hare pellet-plots in northern Idaho to quantify pellet decomposition rates across environmental gradients, and conducted feeding trials on captive hares to assess the role of diet on pellet production rates. We found that across our study area pellet numbers tended to be higher on plots with high vegetative cover, which likely was a reflection of hare habitat choice rather than lesser pellet decomposition in such habitat. A pellet decomposition experiment indicated that pellet persistence was negatively related to moisture level, and that pellets produced by hares during summer decomposed more quickly than those from winter. We found that only 19% of fecal pellets collected from plots located across northern Idaho were produced by hares during winter. There was a correlation between pellet numbers from plots that were pre-cleared 1 year earlier and estimated numbers of hares on 6 study areas. A similar correlation was lacking for pellet counts from uncleared plots, implying that hare population estimation via pellet-plot counts should involve plot pre-clearing. In captive studies, juvenile hares produced slightly fewer pellets per day per gram of food ingested than adults, but pellet production was similar across diets comprised of 10 different browse species. We conclude that for our study area the fecal pellet-plot method may be subject to notable pellet decomposition bias, and therefore recommend that use of the method elsewhere across the species' range be preceded by assessment of both the pellet-hare density relationship and pellet decomposition rates across habitats.