Generalisations drawn from various studies on the assemblage composition of the hyporheic zone sampled using a Bou-Rouch pump are hampered by uncertainty about the comparability of different sample volumes and the uneven taxonomic resolution across studies. To asses the effect of sample volume on perceived assemblage composition, we compared results from 10 consecutive one-L samples collected using a Bou-Rouch pump from each of nine wells inserted in a 9m² area in the down-welling hyporheic zone of the Rhône River, France, and in a weakly-upwelling zone of a slowly-flowing anabranch. At the upwelling site, assemblage composition, numbers of taxa, and density of individuals per L in the first few one-L samples were significantly different from those of subsequent samples. Although numbers of taxa and individuals levelled off after the third to fifth L, perceived assemblage composition continued to change with increasing sample volume. In contrast, consecutive samples from most of the wells at the downwelling site collected progressively fewer individuals. However, assemblage composition was also strongly associate with sample volume, largely due to differences in relative abundance. Despite these effects of sample volume at both sites, differences in perceived assemblage composition among wells were consistent at volumes of 3-5 L in most cases. There was evidence for a 'filtering effect' with small-bodied taxa such as nematodes, ostracods and cladocerans being preferentially collected in the early samples. However, larger taxa (e.g., isopods) tended to occur equally often in earlier and later samples. Taxonomic resolution of specimens to order instead of species level did not significantly change the general ordination patterns of assemblage composition among wells at both sites. Overall, our results imply that a sample volume of 5L and broad-level taxonomic resolution would suffice to reveal spatial variation in assemblage composition between these two sites and, at a finer scale, among wells.