This paper provides an overview of results from a long-term (38 generations) selection experiment. Lines were developed from individual phenotypic selection for high or low body weight at 8 wk of age. Included are data for the selected lines, sublines in which selection was relaxed, crosses of the selected lines, and sublines in which the sex-linked dw gene was introduced. Periodically (and in some cases every generation) data were obtained for unselected traits. These unselected traits included feed consumption and intake behavior, reproduction, allomorphic relationships, and metabolic, immunological, endocrine, and molecular factors. These responses have been integrated into a resource allocation paradigm.

Weights of internal organs and levels of digestive enzymes were obtained through the first 15 days posthatch for cockerels from three lines of chickens known to differ greatly in body weight. On Day 15 body weights from the fastest growing line were eight times greater than those from the slowest growing line. Differences among lines were found for weights at hatching and for growth patterns (both absolute and relative to body weight) of the vitelline residue, heart, lungs, liver, pancreas, crop, proventriculus, gizzard, and segments of the small intestine. Line differences were also evident for levels of trypsin, chymotrypsin, and amylase in the pancreas and contents of the small intestine. Ranking of lines for these traits varied with age. In all lines weights of the small intestine, liver, and pancreas increased relatively more than did total body weight during the 1st wk posthatch, after which the relationship reversed.

The selection of chickens for high (HA) and for low (LA) antibody titers to sheep erythrocytes has produced differences in the selected trait and in the correlated responses in body weight, egg production, and erythrocyte antigens. The response to selection continued through 14 generations. There was considerable divergence between lines for erythrocyte alloantigen systems, including the major histocompatibility complex. Females from Line LA were heavier as juveniles and lighter as adults, matured at a younger age, and had higher egg production than those from Line HA. There were no differences between lines for the incidence of defective eggs laid-except for the percentage of eggs with double-yolks, which was greater for Line LA than HA. The phenotypic correlations of antibody response with growth and with reproductive traits were very low; the genetic correlations were moderate to high for most of these traits.

A long-term selection experiment for high (HWS) and low (LWS) BW at 8 wk of age (BW8) was conducted in White Plymouth Rock chickens. Over 54 generations of selection, responses to bidirectional selection were profound. Increase in BW8 in line HWS was linear, and there was a significant quadratic response in line LWS for BW at both 4 and 8 wk of age. Although there is no indication that line HWS has come close to approaching a selection limit in more than 50 generations, selection limits occurred in line LWS chickens at generation 48 for females and generation 50 for males. Evidence also exists that one or more beneficial mutations have occurred in line HWS, aiding in progressive increases in BW8 over generations. Analyses of ratios of BW at 4 wk of age with those at 8 wk of age (ratio 4/8) revealed that LWS females grew proportionately faster through 4 wk of age than LWS males or HWS chickens. Comparisons of the selected lines with contemporary lines in which selection had been relaxed (discontinued) indicated that, in line HWS, the relaxed lines generally regressed toward original (preselection) values, suggesting that the linear response to single-trait selection was at least partially due to continued genetic variance. In LWS chickens, a series of plateaus in selection response occurred, but relaxed contemporary lines still regressed toward preselection values for BW8. In spite of the length of this selection experiment (54 generations), genetic variance and beneficial mutations have allowed continued, linear response to selection for increased BW8. Response to selection for decreased BW8 has been tempered by physiological barriers that have decreased survival of young chicks or the ability of females to reproduce. These findings are discussed in a historical perspective.