Results of selection over 135 generations for high scutellar bristle number in three further lines derived from Oregon-RC complement those on the four lines reported previously (Sheldon and Milton 1972). All lines differed widely in pattern of scutellar response, correlated response in abdominal and posterior central scutellar bristles, sex differences, and behaviour on relaxing selection, though the selection limits reached in the present three lines were lower than in the previous lines.The early scutellar response supports the conclusion in the previous paper that early selection was mainly for poor regulation of the scute locus at the canalization level of control, since the probit span of the four-bristle class decreased as the mean scutellar number increased in the selection lines, and increased in the relaxed lines as the mean scutellar score decreased.Introduction of the set allele into the selection lines confirmed the earlier findings that the selected background increased scutellar bristles in set to a much lesser extent than in sc+ flies, and that considerable interaction of set and sc+ with selected background occurred. The degree of dominance of sc+ over set was increased in two lines out of five and decreased in one line. The latter indication of breakdown in the dominance level of control was not supported by results of the effect of an extra dose of sc+ in males, where all lines were similar to the unselected level.The correlated response in abdominal bristles was positive, similar in size in the three lines, and smaller than in the previous lines. The changes in abdominal score on relaxation of selection for scutellar bristles do not correlate well with the variable reductions in scutellar score across 10 relaxed lines. Positive, zero or negative genetic correlations between the two characters seem to be possible depending on the line and stage of response. In relation to abdominal scores, changes in dominance of sc+ over set occurred but were not correlated with the dominance changes for scutellar bristles. The effect of an extra dose of sc+ on abdominal bristles was similar in all six lines observed, adding 2·6-4·1 bristles, much smaller than the scutellar effect, where sc+/sc+· Y males exceeded sc+ /sc+ females. _1.1."