Lambs grazed vegetative wheat crops in south-east Australia with and without Mg or Na supplements. Relative to the requirements for lamb growth, wheat forage was marginal for Mg, adequate for Ca, excessive for K and very low in Na. In Experiment 1, a combined Mg/Ca/Na supplement led to a 54% increase in lamb liveweight gain compared with unsupplemented lambs. Further experiments were then conducted to separate the responses to Mg or Na. In Experiment 2, Mg or Na supplements increased liveweight gain by 24 and 37%, respectively. In Experiment 3, Mg intake by lambs was increased by fertilizing the wheat crop with MgSO 4 . This increased wheat forage Mg from below (0.10% DM) to above (0.17% DM) that required by lambs and resulted in 24% faster liveweight gain at a stocking rate of 18/ha. At higher stocking rates (33, 47 sheep/ha), the response to Mg declined as herbage supply became increasingly limited. In the fi nal experiment, Na-supplemented lambs grew 25% faster than unsupplemented * lambs. The observed responses to Mg and Na may be separate responses, but diets low in Na and high in K have been shown to impair ruminal Mg absorption. The wheat forage in Experiments 1-4 contained 3-4% of DM as K and very low Na concentrations. We present evidence suggesting that the response to Na could partly be due to improved Mg absorption, arising from a reduction in dietary K:Na ratio.