The present experiment was designed to study the effects of factors likely to alter microbial purine yield from the rumen on urinary excretion of allantoin-nitrogen (UAN). Sixteen mature Clun Forest-Welsh crossbred wethers were used in a 2 x 2 x 2 x 2 factorial design to investigate the effects of (1) level of intake, (2) wheat :nutritionally improved straw (NIS) ratio, (3) sodium bicarbonate inclusion and (4) monensin inclusion on diet digestibilities, fractional outflow rates of solids and liquids from the rumen and urinary allantoin excretion. Each treatment occurred in each of two experimental periods. The treatments were designed to influence microbial purine yield via changes in rumen outflow rate and microbial maintenance coefficient. Increasing the proportion of NIS and increasing feeding level decreased digestibility and increased the fractional outflow rate of solids. Increasing the level of intake increased the fractional outflow rate of liquids. Urinary allantoin excretion (/kg live weight07s per d) was significantly increased by an increased proportion of wheat in the diet and increased level of intake, and significantly reduced by NaHCO,. There was a significant interaction effect such that increasing level of intake did not increase UAN with the high-NIS diet, despite an increased fractional outflow of solids from the rumen, in contrast to the increase observed with the high-wheat diet. Taken together with other observations it is suggested that high sodium concentrations in the diet reduce the efficiency of microbial synthesis, probably by increasing the energy cost of maintaining osmolarity. Monensin had no overall effect on UAN but there were significant interactions between monensin and dietary Na; the inhibitory effect of monensin on UAN was eliminated or reversed in the presence of added NaHCO,. This is consistent with theories that monensin increases the net influx of hydrogen ions into microbial cells and that this influx can be reduced by increased extracellular sodium ions.Urinary allantoin : Rumen : Microbial maintenance coefficient : Sheep Nucleic acids have been used widely to indicate the microbial content of digesta taken from the abomasum or duodenum (McAllan & Smith, 1969). This approach was improved by the measurement of their component purine bases (Schelling et al. 1982), which are more robust and readily measured (Zinn & Owens, 1986). To estimate the yield of microbial material from the rumen there was still a need for invasive techniques to assess the quantity of digesta flowing through the lower gut. Rys et al. (1975) proposed the use of urinary purine derivatives as indices of microbial yield from the rumen. Chen et ai. (1990), using animals sustained by intragastric infusion, demonstrated a direct relationship between the supply of exogenous purine to the lower gut of sheep and their recovery as purine derivatives (hypoxanthine, xanthine, allantoin and uric acid) in urine. When purine derivative excretion exceeded 0.6 mmol/kg live weight (W)" 75 per d there was no net