The recovery in urine of an intrajugular infusion of physiological amounts of allantoin was measured in four sheep nourished by an intragastric infusion of volatile fatty acids and casein (to eliminate rumen fermentation). The recovery was 72% (S.E.M. 7) and the remainder was presumed to have been lost by diffusion into the gut and degradation by gut microflora. Measured in two sheep, allantoin was removed from the blood at a fractional rate of 0.30 h-1, and excreted in urine at 0.23 h-1. Calculation based on creatinine excretion showed glomerular filtration rate and tubular reabsorption of allantoin to be unchanged by the intravenous infusion. Maximal tubular reabsorption at 1.28 mmol day-1 was saturated by the load of endogenous allantoin alone. In a second experiment with seven normally fed sheep (28-50 kg live weight, all given 1 kg feed), urinary excretion and plasma concentration of allantoin were linearly related. However, the errors were such that plasma allantoin concentration would be of little value as a predictor of urinary excretion. There was a nearly twofold range in allantoin excretion (the larger animals excreting less), which implied that the supply of microbial biomass to the host animal per unit of feed ingested could be profoundly affected by feeding level.