Eight newborn lambs were given milk-substitutes based on casein (n 4) or a soya-bean-protein isolate (n 4). On the morning before a 10 d collection period, the milk-substitutes were supplemented with 59Fe as ferric chloride and 65Zn as zinc chloride. Faeces and urine were collected daily and determinations of 59Fe and 65Zn were made on these separately. During a second 10 d period the dosing procedure and collection were repeated but the diet of each lamb was reversed. Mean retention of 59Fe in lambs given the casein diet was 0.50 (SE 0.12) and in those given the soya-bean-protein diet 0.06 (SE 0.05). There were large variations among animals but the treatment effect was significant at P less than 0.05. Retention of 65Zn was significantly greater than that of 59Fe for all lambs but the effect of dietary treatment was similar to that for Fe. 65Zn retention values for casein and soya-bean-protein diets were 0.84 (SE 0.06) and 0.52 (SE 0.03) respectively, and the difference was significant at P less than 0.01. There were only very small amounts of 59Fe and 65Zn in urine, representing 0.002-0.003 of the dose.