Decades of sediment trap studies have revealed that zooplankton fecal pellets constitute a much smaller fraction of the sedimentary flux than expected from the abundance of copepods and their anticipated production rates of fecal pellets. The explanation for this is thought to be coprophagy (ingestion) of fecal pellets by copepods. We examined fecal pellet clearance rate of Acartia tonsa and Temora longicornis and feeding behaviour of A. tonsa. Pellet clearance rates in A. tonsa and T. longicornis females were similar but low on their own pellets (11 to 22 ml female -1 d -1 ). Our own data together with observations compiled from the literature revealed that copepod fecal pellet clearance rates decrease with increasing relative pellet size and that all species can be described by a common relationship. In A. tonsa, the presence of alternative phytoplankton food increased pellet clearance rate. Direct observations revealed that this was accomplished through the modulating effect of phytoplankton on the feeding behaviour. In the absence of phytoplankton, A. tonsa is nonmotile but perceives sinking fecal pellets at distance. However, only a small fraction of the pellets that came within detection distance elicited an attack. In the presence of phytoplankton food, A. tonsa switched to suspension feeding, and all fecal pellets entrained in the feeding current were encountered. Independent of the feeding mode, fecal pellets were mainly degraded by fragmentation during rejection and only a small fraction was actually ingested (5%). The main impact of A. tonsa on the vertical flux of fecal pellets is therefore through coprorhexy (fragmentation) turning fecal pellets into smaller, slower-sinking particles.