To determine the optimal combination of sample volume and replicationrequired to assess hyporheic taxa richness and abundance, 10 consecutive one-L Bou-Rouch pump samples were collected from each of 9 wells inserted in a 3x3m grid at two sites on the Rhône River, France. Variable and non-linear relationships between sample volume and numbers of taxa and individuals demonstrate that hyporheic densities cannot simply be expressed per L for comparison among studies that have collected different sample volumes. Our results imply that an optimal sampling design for rapid assessment of hyporheic biodiversity and abundance would need at least five samples of 3–5 L to provide reasonable levels of precision (20%). At both sites, 6–7 replicates of 5L sufficed to be able to detect a 50% change in mean taxa richness with a power of 0.90 and α = 0.10. Although this sampling exercise should be repeated whenever accurate population estimates are needed, it appears that the spatial variance of the hyporheos at a fine scale (< 9m²) is great enough to obscure detection of potentially large differences when replication or sample volume is small.