The data from two lines of swine, descended from a common stock, that were maintained as closed populations for six generations were used in this study. One line (H) was full-fed from weaning to 150 lb., hand-fed from 150 lb. to parturition at levels which permitted gains of 1.5–2.0 lb. daily and again full-fed during lactation. The other line (L) received 70% as much feed as the H group throughout. All selections were made as each pig reached 150 lb. using the following index: I=B+2W+35G, where B is the number farrowed alive and W is the number weaned in the litter in which the individual pig was born, and G is the average rate of daily gain of the individual pig from weaning to 150 lb. There was no significant difference between H and L in average number of pigs farrowed alive per litter (8.85 vs. 9.26). However, there was a highly significant difference between H and L in average number of pigs weaned per litter (5.65 vs. 6.63), average weaning weight of the individual pig (36.5 lb. vs. 28.5 lb.) and average rate of daily gain from weaning to 150 lb. (1.31 lb. vs. 0.89 lb.). Selection apparently was ineffective in increasing litter size in either line. However, selection for rate of daily gain appeared to be effective in both lines. As determined by the method of path coefficients, approximately equal emphasis was placed on prolificacy (B+2W) as on average daily gain in H, while in L more emphasis was placed on prolificacy than on average daily gain. Using weighted selection differentials expressed in terms of standard deviations, somewhat less selection was practiced for prolificacy than for average daily gain in H, while in L there was considerably more selection for prolificacy than for average daily gain.