A total of 88 samples of seedlings infected with postemergence damping-off or rotted roots of adult plants were obtained from different cotton-producing areas in Egypt. Isolation from these samples revealed the following results: (1) M. phaseolina was isolated from 37.5% of the samples examined. Its isolation frequency (8.3%) was the fourth in rank after Fusarium spp. (35.2%), Rhizoctonia solani (24.4%), and Aspergillus spp. (9.6%). (2) The percentage of samples from which M. phaseolina was isolated and percentage of isolation frequency of this fungus were highly correlated (r = 0.943, P < 0.001). (3) Significant positive and negative correlations were observed between the frequency of root colonies of M. phaseolina and the frequency of root colonies of other fungi. (4) Variation in isolation frequency of M. phaseolina over locations did not follow a definite geographical pattern. Thus, the isolation frequency of M. phaseolina of the samples obtained from Minufiya or Gharbiya (Middle Delta) was not significantly different from that of the samples obtained from Sharqiya (East Delta). On the other hand, within the east Delta region, a significant difference in isolation frequency was observed between Sharqiya and Daqahilya. (5) M. phaseolina was completely absent from the samples obtained in April, while it maintained almost the same level of isolation frequency over June, July and August. Isolation frequency of M. phaseolina from Giza 75 was not significantly different from that of Giza 85. M. phaseolina is a seedborne pathogen. However, frequency of M. phaseolina recovery from seeds was affected by cultivar, isolation method, and degree of soil infestation with the fungus. A total of 189 isolates of M. phaseolina from different hosts were tested for pathogenicity on seedlings of cotton cultivar Giza 75 under greenhouse conditions in 1994, 1995, 1996 and 1997. These isolates were obtained from 16 hosts; however, cotton isolates (123 isolates) were the predominant group representing 65.08% of the tested isolates. Isolates of the other hosts ranged from 1 to 16. Analysis of the data obtained from the four pathogenicity tests revealed the following results: (1) Of the 123 isolates recovered from cotton, only 79 (64.22%) were pathogenic on cotton cultivar Giza 75. On the other hand, of the 66 isolates obtained from the other hosts, 56 (84.85%) were pathogenic on Giza 75. (2) Most of the pathogenic isolates recovered from cotton were concentrated in the Lower Egypt governorates. Thus, of the 62 cotton isolates recovered from the Lower Egypt Governorates, 50 (80.65%) were pathogenic on Giza 75. On the contrary, of the 61 cotton isolates recovered from the Upper Egypt Governorates (Giza, Minya, Assiut, and Sohag) only 29 (47.54%) were pathogenic on this cultivar. (3) In each pathogenicity test, the percentage of isolates which significantly affected postemergence damping-off was much greater than that of the isolates which significantly affected preemergence damping-off, plant height, or dry weight. (4) Significant negative correlations were observed between the variables used in assessing pathogenicity of the isolates (pre- and postemergence damping-off and the variables used in assessing vigor of seedlings (plant height and dry weight). A computerized program was used to develop two regression models that predicted pathogenicity of M. phaseolina isolates as a functionof seed germination and length of radical after the exposure of seeds to the effects of toxins secreted by the tested isolates. The first model was simple and linear. It accounted for 0.70 of the variation in pathogenicity of the isolates. The second model was multiple and nonlinear. It accounted for 0.92 of the variation in pathogenicity of the isolates. The superiority of the second model in prediction is clear; however, its application took 12 days while the application of the first model took only four days.