Stable H isotope ratios (D/H, expressed as dD) hold promise as an additional tool for elucidating food sources for consumers in aquatic ecosystems. We tested the applicability of dD as a food source tracer in streams of New Brunswick, Canada. First, we analyzed dD and d13C in biofilm and terrestrial leaves and compared signal-to-noise ratios (variability within sources vs variability between sources) between the 2 elements. Signal-to-noise ratios were roughly similar, and 23 of the 31 sites had isotopically distinct food sources based on dD compared to 20 of 31 based on d13C. Second, we used mixing models to estimate % aquatic H and % aquatic C for benthic invertebrates at a subset of sites. Of 16 samples, only 1 had d13C that was outside the range of the 2 food sources (yielding % aquatic C . 100%), but 12 of the 16 samples had dD outside the range of the food sources, a result suggesting confounding effects of water and lipids on total body H content. Last, we analyzed dD in laboratory-reared consumers (brook trout and water striders) and in their diet before and after lipid extraction to estimate diet-tissue fractionation. Large differences between consumer and diet were apparent before lipid extraction, but no significant differences were found after lipid extraction. All of these measures indicate that dD could serve as a complementary, but not alternative, isotopic method for estimating food sources for consumers in streams. Further laboratory trials are needed to explore the influence of lipids on dD values.