We visually censused Warner sucker cohorts in a low gradient stream reach and determined microhabitat use from random availability data. For above-water visual censuses (AVC), we counted aggregations while slowly wading through Honey Creek three times both years. Underwater visual censuses (UVC) were done in 1993 by snorkeling. Mean focal point velocity (FPV) during AVC significantly differed by year but not by census date, and mean depth did not differ for either. In 1993 we found suckers used microhabitats with FPV between >3 to 6 cm/s (χ₁² = 3.93, P < 0.05) and depth between >20 to 40 cm (χ₁² = 4.5, P < 0.05). Suckers avoided areas where FPV exceeded 15 cm/s (χ₁² = 18.7, P < 0.001) and depth between >60 to 80 cm (χ₁² = 8.1, P < 0.005). Aggregate abundance was significantly related to both distance to (r₆₄ = −0.45, P = 0.0002) and percentage (r₆₄ = 0.44, P = 0.0003) submerged vegetation. We found 97% of suckers (N/m³) in riffle/run habitat during UVC. When mean flow was used as a continuous variable, we found suckers occupied habitats >3 cm³/s and avoided habitats <2 cm³/s in a disproportionate manner relative to availability. These results suggest young suckers select vegetated areas with moderate flow and relatively shallow depths during first few months of life. This information improves our understanding of ecological habits of early life stages of a rare western catostomid during stream residence, which could be useful for conservation management in low gradient stream reaches of Warner Basin.