Colonization of cottonseed by M. phaseolina during the early stages of infection was affected by inoculum density, temperature, incubation period (time of seed exposure to inoculum), and pathogenicity of isolates. The higher the inoculum density, the earlier the isolation of M. phaseolina from seeds and the higher its isolation frequency. The three main effects of ANOVA (temperature, incubation period, and isolate) were all very highly significant sources of variation in isolation frequency of M. phaseolina from seeds regardless of the tested cultivar (Giza 75, Giza 80, or Giza 85). Temperature x incubation period interaction was a very highly significant source of variation in isolation frequency from Giza 75 (P = 0.0000) and Giza 80 (P = 0.0001), while it was a nonsignificant source of variation (P = 0.2925) in isolation frequency from Giza 85. All the other source of variation were nonsignificant (P > 0.05) regardless of the tested cultivar.