The effect of intake on urea production, entry into the digestive tract and return of N to the ornithine cycle was studied in four sheep. Each sheep received 0.6, 1.2 and 1.8 x estimated maintenance energy intake quantities of grass pellets for 9 d. After 4 d of adjustment, N balance measurements were conducted between days 5 and 8. From day 7 to day 9 animals were continuously infused, via the jugular vein, with [15N15N]urea and three urine samples were collected at approximately 2 h intervals 48-54 h after the start of infusion. Total urea and enrichments of [15N15N]- and [14N15N]urea in the urine samples were determined. Urea production was calculated from the isotopic dilution of [15N15N]urea and entry into the gastrointestinal tract (GIT) obtained from the difference between this and urinary urea elimination. Urea which enters the GIT undergoes hydrolysis to liberate NH3 which may be reabsorbed and enter the ornithine cycle in which case the product is [14N15N]urea, based on the probabilities of labelled and unlabelled N providing ureagenic precursors. The quantity of urea-N which returns to the ornithine cycle from the GIT can thus be calculated. Existing models based on this approach yield overestimates of the fate of individual urea molecules due to a failure to allow for multiple recycling of [14N15N]urea species through the GIT. Refinements introduced to cover this resulted in a 33-48% reduction in calculated return of label for the current study. The present model also predicted that 95% of the label movements across the GIT could be accommodated by three or fewer entries and returns of urea-N and 99% by recycling for a maximum of six occasions. Urea-N production increased with intake (P < 0.001) and exceeded digestible N values at all intakes. Urea which entered the digestive tract, both in absolute terms (P < 0.001) and as a proportion of production (0.62, 0.69, 0.73; P = 0.027), increased with intake. The proportion of entry into the digestive tract which was returned to the ornithine cycle remained reasonably constant (0.37-0.41) across all intakes but the absolute amount increased (5.6, 9.2 and 15.0 g N/d; P < 0.001) with intake. If allowance is included for losses of 15N in faeces then the approach offers a relatively simple means of estimating anabolic reuse of urea by digestive tract micro-organisms and can complement data obtained from the technically more demanding arterio-venous and multiple-isotope techniques used hitherto.