DATA on body weights at birth, 21, 56, 98 and 140 days, average daily gain from 56 days to a live weight of about 79.4 kg and days required to reach 79.4 kg were analyzed to evaluate the correlated responses of these traits to selection for high and low backfat thickness in two lines of Durocs and two lines of Yorkshires. Based on data collected through 13 and 11 generations of selection, respectively, the correlated responses given by both the pre- and postweaning growth measures differed rather markedly between the two Duroc and two Yorkshire lines. In Durocs each of the five weights studied gave evidence of being negatively correlated genetically with backfat thickness whereas in Yorkshires they gave responses indicative of traits which are important components of fitness. For both the pre- and postweaning weights the rates of change in Durocs were generally greater in the low line. In Yorkshires, the preweaning weights declined faster in the high than in the low line) whereas the reverse was the case for the postweaning weights. In both Durocs and Yorkshires, the patterns of the correlated responses shown by days on test agreed reasonably well with those of the pig weights. However, post-weaning daily gain in Durocs gave results suggesting a significant improvement of this trait in both the high and low lines, whereas in Yorkshires daily gain decreased significantly in the low line without changing significantly in the high line. Least squares analyses incorporating terms for various interaction effects showed that in both Durocs and Yorkshires upward selection resulted in a rather pronounced narrowing of the sex difference in backfat thickness between gilts and barrows, while downward selection also produced a narrowing of the difference between boars and gilts in Durocs. The significant sex by line interactions shown by 140-day weight, days on test and daily gain in Yorkshires were attributed to greater heterogeneity in the line differences between both boars and gilts and between boars and barrows than in those between gilts and barrows. In Durocs, genetic correlations obtained between the various growth measures and backfat thickness from offspring-midparent covariances agreed reasonably well, both in sign and magnitude, with those based on correlated selection responses. In contrast, there were only two traits in Yorkshires, i.e., birth weight and 56-day weight, for which the signs of the correlations with backfat thickness agreed by the two methods. The results indicate that selection for lower fatness should be effective in improving growth rate in Durocs, whereas in Yorkshires emphasis on greater leanness would most likely result in slower growing pigs.