A plaque assay system for the virus of epizootic hemorrhagic disease of deer in the WI-2 subline of BHK-21 cells is described. The system uses infection of the cells in suspension before growth as a monolayer under supplemented Leibovitz L-15 medium incorporating gum tragacanth. Crude cell lysates were initially used as a source of virus for the assay system. Since 90% of the infectious virus titer was lost on removal of cell debris by centrifugation, a statistical analysis of the validity of using uncentrifuged lysates for plaquing was carried out. Data showed that, except at low virus concentrations, the plaque counts were consistent with a negative binomial rather than a Poisson distribution. When the cell lysates were clarified by low-speed centrifugation, plaque count variation followed a Poisson distribution even at high mean plaque counts. There was no appreciable overlap error for plaque counts up to at least 100 per 60-mm dish. Using both clarified and unclarified cell lysates, a good linear relationship was found to exist between plaque numbers and relative virus concentration.