Date of this Version
Divergent selection in mice was renewed in 3 independent replicates for high (MH) and low (ML) heat loss. An unselected control (MC) was maintained in all replicates. Heat loss was measured for individual male mice for 15 h, overnight in direct calorimeters. After 16 initial generations of selection followed by 26 generations of relaxed selection, divergent selection resumed for 9 generations. The realized selection applied was very close to the maximum possible selection according to the criteria and protocol. Selection differentials were greater for high than for low selection due to greater variation in the MH line. When corrected for SD, standardized selection differentials were similar for MH and ML selection. Unintended selection in MC was negligible. Realized heritability for divergence was 0.14 ± 0.01, which was considerably less than that realized during the initial generations of selection (0.28 ± 0.03). Realized heritabilities for MH selection (0.16 ± 0.05) and for ML selection (0.07 ± 0.06) were less, especially for ML selection, than were observed in the earlier generations. The difference in heat loss between MH and ML males was 55.7% of the MC mean at generation 51, compared with a difference of 53.6% in generation 15; this difference had decreased to 34.4% at the end of the relaxed selection (generation 42). For feed intake between 8 and 11 wk, MH and ML males differed by 34.0% of the MC mean by the end of the selection process. Body weight at 12 wk for MH and ML males was less than for MC males. Litter size response was positively related to the heat loss response. Conception rate was poorer in MH matings than in MC and ML matings.