The effects on production of milking grazed crossbred cattle (32, 53 and 74 cow and heifer dams in experiments 1, 2 and 3 in 1986, 1987 and 1988 respectively) with or without a dietary supplement (2kg dairy meal daily for 3 to 4 months post-partum) were studied in a semi-arid area of Kenya. During the second and third trials, half the calves from the milked and also the unmilked groups were creep fed for 4 months commencing approximately 1 month from birth. Herbage dry matter intake was determined four times during the wet and dry seasons between November 1988 and May 1989, and the deterministic model and computer program of the bioenergetic system of lactating and pregnant cattle of Bruce <i>et al</i> (1984) used to examine the performance of the crossbred cows, and to evaluate its potential applicability. The respective mean extracted milk yields over a six month lactation were 670.6, 414.4 and 371.2 kg for the milked and supplemented (MS) groups, and 563.4, 398.4 and 331.1 kg for the unsupplemented (MNS) groups during experiments 1, 2 and 3 (P> 0.05). Mean daily calf growth rates were affected by milk offtake (P< 0.05). They were respectively, 0.52, 0.48, 0.79 and 0.71 kg during experiment 1; 0.64, 0.61, 0.72 and 0.72 kg during experiment 2; and 0.59, 0.57, 0.66 and 0.66 kg during experiment 3 for the milked (MS and MNS) and the control groups (not milked and supplemented, NMS and the not milked and not supplemented, NMNS). Overall, post-partum supplementation improved dam liveweight gain (P< 0.05) but milk extraction tended to reduce it (P> 0.05). Similarly, calving indices were improved by supplementation while the effect of milking was inconsistent. Both creep feeding of the calves and beef genotypes of the dams were not significant (P> 0.05) factors affecting performance. Similarly, body condition scores (1 to 9 score scale) ranged between condition score 4 and 5 for all treatments. The respective mean daily dry matter intakes were 8.5 and 8.2 during the wet, and 8.6 and 8.2 kg during the dry periods; and the corresponding dry matter digestibilities were 0.72 and 0.74 in the wet, and 0.67 and 0.62 during the dry periods. The discrepancies between observed and predicted values of milk (energy), though not for liveweight, were small, consequently, with a few refinements the bioenergetic model may find a useful role in cattle production systems in the semi-arid tropics. It was concluded that the offtake of milk from suitable crossbred cattle in the semi-arid areas may be increased without significantly reducing offtake of slaughter animals.