There has been a question as to whether the optical plankton counter (OPC) accurately measures zooplankton density and biomass in large lakes, oceans, and estuaries. Results from our Lake Michigan surveys in 1998--2000 often revealed poor agreement between the OPC-2T and 153-¿m mesh plankton net samples. The most serious common problem was overestimation of zooplankton biomass by the OPC relative to net tows. Such overestimates have been attributed to the orientation of zooplankton assuming that their maximal silhouette area is detected by the OPC, coincidence, and the presence of nonzooplankton particles. Experiments using an OPC-1L with polymer microspheres, nylon rods, or live zooplankton confirmed that the OPC accurately counts and sizes zooplankton at low zooplankton concentrations typically found in Lake Michigan and that a shape factor correction often used need not be applied. Most overestimation of biomass was associated with the presence of nonzooplankton particles. Worst agreement was seen in shallow nearshore zones during periods of high total suspended matter (TSM), which consisted of fine sediments and large resuspended particles such as sediment-phytoplankton aggregates, and benthic plant and animal debris. Best agreement was found under low TSM conditions associated with offshore waters during the stratified period.