The 80/20 (w/w) azeotropic styrene/acrylonitrile mixture was copolymerized at 35, 50, and 70°C in crosslinked polybutadiene seed latexes of varying particle size (range 113–242 nm) using the t‐butyl or cumene hydroperoxide/sodium formaldehyde sulfoxylate dihydrate/EDTA‐chelated Fe+2 redox initiator. Other initiators used for comparison were: (1) ammonium persulfate at 70°C; (2) ammonium persulfate/sodium bisulfite redox initiator at 35 and 50°C; (3) oil‐soluble, 2,2'‐azobis(2,4‐dimethylvaleronitrile) at 50°C. The following polymerization parameters were varied systematically: (1) polybutadiene seed latex particle size; (2) monomer/polymer swelling ratio; (3) polymerization temperature; (4) type and concentration of initiator; and (5) mode (and duration) of initiator addition. The following parameters were determined: (1) polymer yield (final conversion); (2) variation of conversion with time; (3) morphology of the latex particles; (4) amount of coagulum; and (5) fraction of styrene/acrylonitrile monomer mixture grafted to the polybutadiene seed latex (defined here as grafting efficiency). For polymerizations using the t‐butyl or cumene hydroperoxide/sodium formaldehyde sulfoxylate dihydrate/EDTA‐chelated Fe+2 redox initiator, the rates of polymerization were faster, the grafting efficiencies were greater, and the amounts of coagulum were greater at 70°C than at 50 or 35°C. The rates of polymerization were decreased by decreasing the concentration of redox initiator or adding it incrementally. A 1.00/1.00/0.08 (w/w) t‐butyl or cumene hydroperoxide/sodium formaldehyde sulfoxylate dihydrate/EDTA‐chelated Fe+2 ratio gave reasonable polymerization rates at 35 and 50°C, and a 0.48/1.00/0.07 ratio, at 70°C. The latexes were stabilized during polymerization with 3% (based on seed solids) nonionic Igepal CO‐990 at 35°C, and 6% at 50 and 70°C. The grafting efficiency increased with decreasing particle size of the polybutadiene seed latex and decreasing monomer/polymer ratio. The rate of polymerization increased with decreasing monomer/polymer ratio. The monomer/polymer ratio also affected the morphology of the particles; higher values gave inclusions of poly(styrene‐co‐acrylonitrile) in the polybutadiene particles. There were no significant differences in polymerization kinetics or grafting efficiency between the t‐butyl and cumene hydroperoxides. Copyright © 1990 John Wiley & Sons, Inc.