A for ?ln(?S) and mean E for a bin. For the data used in Fig. 3, the regression coefficient for ?ln(?S) was 3.81 ± 0.21 (the SE was obtained from normal distribution regression theory). The regression coefficients (b) for mean E for the low and high rates of gene conversion were 0.236 ± 0.039 and 0.0935 ± 0.0208, respectively, indicating a significant difference between the two regression coefficients in each case. Similar results were obtained with the mel data. 3, which assumed that 15% of the total number of mutations had strongly deleterious effects (tl = 0.044), and cases without any major effect mutations.