Search

A strain of mice (S1) was successfully selected for large litter size for 31 generations, increasing the mean by 4.2 young per litter. After generation 31, there was no further progress and it was concluded that a selection plateau had been reached. Realized heritability decreased during the course of the experiment from 0.16 ± 0.06 for the first 15 generations to 0.00 ± 0.03 for generations 30 through 45. — In order to explore the nature of the selection plateau, the following groups were derived from line S1 at generation 34 or 35: Upward selection with inbreeding (SF), random (relaxed) selection (SO), and downward selection (SR). Selections were carried out for 10 or 11 generations. The means of SO and SF were similar to those of S1, ruling out any major effect of natural selection or overdominance. SR decreased, the mean averaging 2.3 young per litter below that of S1 during the last three generations. The fact that SR responded to selection indicates that genetic variance was still present in the plateaued population. The SF sublines were crossed when the inbreeding was 95% and a new line, SX, was formed. SX was maintained for three generations and the difference of +0.7 young per litter above the contemporary generations of S1 was significant. The results from this experiment suggest that the selection plateau in line S1 was caused by reduction of additive genetic variance to a very low level. Some nonadditive genetic variance remained, however, and was attributed to recessive alleles at low frequency. In agreement with results reported by Falconer (1971), inbreeding with selection followed by crossing of the inbred sublines proved to be effective in overcoming a selection plateau in litter size.

An increase in mean prolificacy is a goal for many sheep production systems, but it is also desirable that variation be minimized around the optimum value. Lambs born in litters of 3 or 4 can be reared with intensive management, but at higher input costs than for twins. Thus for a mean litter size (LS) goal of 2.0, producers desire the maximum possible proportion of twins with as few singles and litters of 3 or more as possible. Coefficients of variation for ovulation rate (OR) range from about 22 % for Romanov to about 40 % for Booroola Merinos; breed CV’s for LS vary less. The most uniform LS’s are achieved by populations with a uniform OR and high prenatal survival. Variability in Booroola populations results in part from segregation of a gene with large effect on OR, but variation is high within genotypes (FF or F+). Among breeds where high prolificacy appears to be inherited quantitatively, there are differences in variability, with Romanov quite uniform, Finnsheep intermediate, and D’Man more variable. The possibility of a gene with large effect in the D’Man breed, in addition to many favorable quantitative genes for prolificacy, has not been ruled out. In general, these three breeds (and some others) transmit their prolificacy additively, making it possible to choose a wider range of mean prolificacy values by crossing and backcrossing with such breeds than by use of a major gene such as the Booroola. Data on variability of first and later generation crossbreds between non-prolific and different prolific breeds are now available.