Four densities of a snail (Juga silicula) and a caddisfly (Dicosmoecus gilvipes) were introduced into separate laboratory streams, and their effects on algal biomass and community structure were monitored for 32 d. Tiles in an ungrazed control stream were covered by thick algal mats by day 32, and were composed primarily of Scenedesmus spp., Characium, and a variety of diatoms. Biomass and community structure of algal assemblages in the stream with the lowest density of snails were very similar to those in the control stream. In the other streams with snails, an inverse relationship developed between algal biomass and snail density after day 16. By day 32, the algal assemblages in the streams with high snail densities were dominated by adnate diatoms (e.g., Achnanthes lanceolata), and basal cells and short filaments of Stigeoclorium tenue. In contrast to the streams with snails, algal biomass was relatively low in all streams with caddisflies. The differences in algal biomass and structure between the streams with the lowest and highest densities of caddisflies were much smaller than those between streams with the lowest and highest densities of snails. On day 32, the taxonomic and physiognomic structure of the algal assemblages in all the streams with caddisflies resembled that in the streams with higher densities of snails. Scanning electron micrographs showed that even at the highest densities, neither snails nor caddisflies could completely remove the algal assemblage. It is concluded that grazing can substantially influence algal growth form and assemblage physiognomy in lotic ecosystems.