We tested the importance of mesozooplankton feeding and defecation for the downward flux of biogenic carbon (C) in the Gulf of St. Lawrence, a marine coastal environment characterized by high zooplankton abundance. Five stations were sampled over nine cruises between July 1992 and June 1994. The rates of chlorophyll (Chl) ingestion and C defecation were determined for the major copepod species. Free-drifting, short-term (24 h) sediment traps were deployed at the sampled sites at 50 and 150 m, and the trap contents were analyzed for fecal pellets (microscopy), Chl and pheopigments (chromatography). The flux at 50 m of C associated with fecal pellets ranged between 5 and 138 mg C m −2 d −1, while total POC flux varied between 78 and 302 mg C m −2 d −1. Of the total POC found in sediment traps at 50 m, close to 50%, on average, was composed of zooplankton fecal pellets. In contrast, direct algal (Chl) sinking was <10%. The fecal pellet contribution varied seasonally, with a maximum (>73%) in June 1994 at all stations. This maximum did not necessarily coincide with the highest rate of production of fecal pellets in the surface waters, indicating that other factors influenced the sinking of fecal pellets. A shallow mixed-layer depth seemed to favor fast removal from the surface. The average fecal pellet contribution to the total downward POC flux was slightly higher for stations 4 and 5 (52% compared with 39% for the other stations). These stations were characterized by zooplankton communities that differed from those of the other stations, smaller copepods such as Temora longicornis frequently dominated at stations 4 and 5, Calanus spp. dominated at stations 1, 2 and 6. High fluxes of fecal pellets frequently occurred when the index of herbivory (Chl ingestion : total C ingestion) was low, transferring downwards carbon of heterotrophic origin. The contribution of C-transformed pheopigments to total POC flux was low (6% on average). It co-varied with the degree of herbivory in the overlying waters. The production of fecal pellets in the top 50 m co-varied with Chl ingestion, but not with the index of herbivory, consistent with an algal food intake representing <50% of total C ingested. Phytoplankton ingestion was highest when primary production (not algal biomass) was maximum. In consequence, mesozooplankton feeding did not exert a strong control over the phytoplankton biomass of the Gulf of St. Lawrence (at most 10%), but it was more closely tied to primary production. Phytoplankton biomass is more likely controlled by physical factors in this region. Mesozooplankton fecal pellets contributed significantly to the downward flux of biogenic C all through the year in this environment, transferring alternatively C of autotrophic and heterotrophic origin.