The relationship between the number of substitutions per silent site (S) and the number of nucleotides separating the two members of a duplicate pair on the same chromosome. The data for Arabidopisis thaliana (At) are derived from the complete sequences of chromosomes 2 and 4; those for Caenorhabditis elegans (Ce) are derived from the complete genomic sequence; and those for Drosophila melanogaster (Dm) are pairs of duplicates contained within the same scaffolds. Least-squares regression involve the natural logarithm of physical distance. For At, slope = 0.111 (0.003), r2 = 0.115, n = 811; for Ce, slope = 0.091 (0.003), r2 = 0.028, n = 1027; for Dm, slope = 0.183 (0.009), r2 = 0.031, n = 250. All regressions are significant at the P = 0.01 level, although the proportion of the variance explained by the model (r2) is very low in all cases. In no case is the intercept significantly different from zero.