Selection for reproduction rate in Australian Merino sheep: direct responses

Two groups of Australian Merinos have been under selection for low (O) and high (T) reproduction rate since 1954, and a third (B) for high rate since 1959. The O and T groups were founded on Peppin ewes from one flock which had borne singles (or twins) in two successive years, mated to single- (or twin-) born rams from the same flock. The B group was founded on two quintuplet rams and ewes born in multiple births (13 in one intake, 91 in a second), from a medium-wool non-Peppin flock belonging to the Seears Brothers, of 'Booroola', Cooma, which had been under selection for multiple births for 10–15 years. In 1959, the lambing percentage of the flock (lambs born to ewes joined) was stated to be 170-180. Selection for both sexes in the early years of the experiment was based on the presence (or absence) of multiple births in the record of the dam's first three lambings, or in that of the ewe herself, again during her first three lambings. In later years this was combined with a dam's ranking coefficient based on all available records, the coefficient being based on deviations from average performance at the ages of record, and the heritability of repeated records. The sex of her litter mates was found to have no effect on a ewe's lambing performance. The litter size in which she was born had no effect on her lambing performance when she belonged to the O or B group, but twin-born ewes in the T group were sometimes at a disadvantage compared with singles-for their lambings at 2–4 years if their dams were adult, and for those at 5–7 years if their dams were 2-year-olds. Selection response was analysed by examining T – O and B – O differences in performance for ewes born in successive calendar years, the two criteria of performance being the percentage of ewes with at least one multiple birth in their first three lambings, and ranking coefficients based on the same lambings (at 2–4 years). A regression of each difference on time is an estimate of the annual rate of genetic change due to selection; no attempt has been made to relate such changes to what might be predicted from parameters. Adjustments were made for the handicap suffered by twin-born T group ewes by adding 10.0 to the percentage showing at least one multiple birth for those whose dams were adult, 3.4 to the 2-4-year-old ranking coefficient of those ex adults, and 2.6 to the 5-7-year-old ranking coefficient of those ex 2-year-old dams. No unbiased correction for age of dam was possible; instead, T – O and B – O differences were calculated separately for progeny of adults and progeny of 2-year-olds, and a weighted mean difference computed. The groups ran at Deniliquin until 1964, and at Armidale from 1965. In spite of the change in environment, and a bad drought in 1965, regression lines could be fitted to all differences except the 2-4-year-old ranking coefficient in group T. Initial responses in percentage of ewes with at least one multiple birth in their fist three lambings were 35.2 and 40.1 for T – O and B – O respectively, and the linear regression coefficients, measuring continuing response, were 2.39 ± 0.58 and 5.52 ± 1.33. The initial responses account respectively for 68 and 62% of the average superiority of the high lines during the experiment. Initial responses for ranking coefficients at 2-4 years of age were 8.2 and 21.7 for T – O and B – O respectively, corresponding approximately to 0.26 and 0 + 72 lambs born/ewe/lambing. Continuing responses fluctuated more than for percentage of ewes with multiple births; regression coefficients of 0.67 ± 0.17 were fitted for T – O from 1954–56 to 1964, and 3.6 ± 1.2 for B – O from 1961 to 1968. The former is equivalent to 0.02, and the latter to 0.11 lambs born per ewe per year. Initial responses contributed 57–76% of the average superiority of the high line (T or B) over O, depending on the regression coefficient used. For the 1968 drop, the difference in ranking coefficient at 2–4 years was equivalent to 0.38 lambs born/ewe/lambing for T – O, and 1.04 lambs born/ewe/lambing for B – O. Ranking coefficients based on lambings at 5–7 years are the result of selection on the ewes themselves, in addition to genetic gains. The T – O and B – O differences in these coefficients showed approximately the same rate of annual change as those at 2–4 years. Corrections were made to some data from T ewes in estimating response, but in practice twin ewes might give a lower performance at their early lambings. However, the actual performance of the flocks shows that any such disadvantage was not serious. In 1972, ewes aged 2–7 years in the three groups bore respectively 1.11, 1.36 and 2.10 lambs per ewe joined. Clearly, the reproduction rate has responded to selection, the rate of response being far higher in the B group than the T. Clearly, also, initial selection of the base animals made a major contribution to the superiority of the high lines, though there has also been a continuing response.