The aims of this study were 1) to determine whether transfer of blood urea to the gastrointestinal tract (GIT) or the efficiency of capture of urea N within the GIT is more limiting for urea N salvage, and 2) to establish the relationship between plasma urea concentration and recycling of urea N to the GIT. We used an i.v. urea infusion model in sheep to elevate the urea entry rate and plasma concentrations, thus avoiding direct manipulation of the rumen environment that otherwise occurs when feeding additional N. Four growing sheep (28.1 +/- 0.6 kg of BW) were fed a low-protein (6.8% CP, DM basis) diet and assigned to 4 rates of i.v. urea infusion (0, 3.8, 7.5, or 11.3 g of urea N/d; 10-d periods) in a balanced 4 x 4 Latin square design. Nitrogen retention (d 6 to 9), urea kinetics([(15)N2]urea infusion over 80 h), and plasma AA were determined. Urea infusion increased apparent total tract digestibility of N (29.9 to 41.3%) and DM (47.5 to 58.9%), and N retention (1.45 to 5.46 g/d). The plasma urea N entry rate increased (5.1 to 21.8 g/d) with urea infusion, as did the amount of urea N entering the GIT (4.1 to 13.2 g/d). Urea N transfer to the GIT increased with plasma urea concentration, but the increases were smaller at greater concentrations of plasma urea. Anabolic use of urea N within the GIT also increased with urea infusion (1.43 to 2.98 g/d; P = 0.003), but anabolic use as a proportion of GIT entry was low and decreased (35 to 22%; P = 0.003) with urea infusions. Consequently, much (44 to 67%) of the urea N transferred to the GIT returned to the liver for resynthesis of urea (1.8 to 9.2 g/d; P < 0.05). The present results suggest that transfer of blood urea to the GIT is 1) highly related to blood urea concentration, and 2) less limiting for N retention than is the efficiency of capture of recycled urea N by microbes within the GIT.