Sf9 insect cells infected with a recombinant baculovirus expressing beta-galactosidase and suspended in fresh medium (TNM-FH supplemented with 10% fetal bovine serum) at the time of infection were cultured in shake flasks at various combinations of initial cell density and multiplicity of infection (MOI). The effects of cell density and MOI on beta-galactosidase production were quantitatively analyzed by plotting the beta-galactosidase yield against the time integral of the viable cell density from the time of infection to the time when the beta-galactosidase production reached a plateau. The beta-galactosidase yield had a maximum value at a viable cell density time integral of approximately 8 x 10(6) cells.d/cm(3) for each MOI used in a range from 0.01 to 10 plaque-forming units per cell (pfu/cell). Since glucose and fructose were exhausted when the culture reached 8 x 106 cells.d/cm(3), it was concluded that protein production in a high-cell-density culture was limited by nutrient depletion in the culture medium, and hence the nutritional capacity of the medium was able to be determined as the viable cell density time integral at which the maximum product yield was attained. In cultures infected at a low MOI (less than or equal to 1 pfu/cell), the specific productivity, and thereby the yield, of beta-galactosidase declined with decreasing MOI due to the reduction in the proportion of initially infected cells. These results indicate that production of a recombinant protein in a culture with medium replacement at the time of infection can be optimized if the cells are infected at a high MOI (greater than or equal to 1 pfu/cell) and at a cell density such that the viable cell density time integral reaches the nutritional capacity just as the protein production is completed.