Three successive experiments, of similar design, were carried out during 1986–88 at Mount Derrimut, Australia. Mature Merino ewes were mated to Poll Dorset rams and then allocated to either a control group (C) or a treatment group (R). Ewes from each treatment were slaughtered 60, 70, 100 or 140 days post-conception and the development of their foetuses was compared in terms of body size and muscle characteristics. In Expts 1 and 2, some ewes were allowed to lamb and the progeny in each group were slaughtered after reaching a body weight of 35 kg, for comparison of growth rates, muscle characteristics and meat quality.

In each experiment, treatment extended from mating to day 70 of pregnancy and, during this period, both groups were housed. Ewes in group C were kept as one group and fed ad libitum and ewes in group R were individually penned and fed a restricted ration of the same diet as that given to group C in order to achieve a steady loss of body weight. In Expt 1, this loss was 8 kg but, in Expts 2 and 3, feed intake was controlled according to condition score and, during this period, group R ewes lost 25–35% of their body weight at mating. After day 70, all ewes were kept grazing and were offered supplementary feed at rates sufficient for a steady increase in ewe body weights.

Foetuses in group R were lower in body weight (P<0·05), crown-rump length (P<0·05) and girth (P<0·01). However, birth weights and mean ages of the 35 kg lambs at slaughter did not differ significantly between treatments.

There were no significant differences between treatments for the semitendinosus (ST) and semimembranosus (SM) muscles in total muscle weight, DNA content, protein content, nor in the ratios of muscle weight[ratio ]DNA and protein[ratio ]DNA. However, there were significant differences between experiments, which indicated that cell size in the ST and SM muscles was lower in Expt 2 than in Expt 1.

In day 70 foetuses, the cross-sectional area of α fibres was greater (P<0·05) in group R than in group C but by day 140 the difference was no longer significant. At day 70, there was also a positive correlation (r=0·65, P<0·01) between the cross-sectional area of β fibres and the number of α fibres surrounding each of them. There were no significant differences between treatments at any age in the percentages of βR, αR and αW fibres.

Meat from group R lambs was more tender than that from group C lambs as indicated by significantly lower means in the SM muscle for adhesion (P<0·01) and Warner–Bratzler Peak Force (WB PF) measurements (P<0·05).

Although some effects of nutritional restriction were found, severe feed shortage in early pregnancy in sheep is unlikely to have significant effects on the production of prime lamb meat provided that adequate nutrition is available during late pregnancy and post-natal growth.