1. Four mature Friesian cows were used. For 2 weeks after calving they were fed 3 kg hay (147 g crude protein (nitrogen × 6·25; CP)/kg dry matter (DM)) plus 9 kg concentrates (177 g CP/kg DM) per d and thereafter 4 kg hay plus 12 kg concentrates/d.
2. At approximately the 2nd, 4th and 9th weeks post-partum each cow was given a single intravenous dose of [14C]urea (1 mCi) and [6-3H]glucose (1·5 mCi). Following this, jugular venous blood samples were withdrawn up to 26 h post-injection.
3. The log (specific activity) v time curves were not linear for either metabolite. Glucose irreversible loss rates (IRL) were calculated by a stochastic procedure. The linear part of the urea log (specific activity) v. time curve gave the same estimate for urea IRL rate as a stochastic method of calculation, but urea pool sizes were overestimated so that stochastic analysis of results was preferred.
4. Mean milk yields at the time of the three measurements were 27·1, 30·8 and 27·9 kg milk/d.
5. Urea IRL was significantly lower (P < 0·05) in the first 4 weeks of lactation than in the 9th, and blood urea concentration was lower (P < 0·05) in the 4th than in the 2nd and 9th week, but there was no close correlation between urea IRL and concentration.
6. Glucose IRL rose (P < 0·05) between the 2nd and 9th weeks post-partum, perhaps due to the increase in feed intake. The results were generally consistent with the relationship between milk yield and glucose IRL established by Paterson & Linzell (1974).
7. Urea IRL (y) and milk protein-N (x) production values were used to calculate the efficiency of protein utilization (EPU) as x ÷ (x + 0·35 y). This assumes that 0·35 of urea IRL was due to protein catabolism. EPU was found to be 0·59–0·80 (mean 0·69).
8. The possible contribution of catabolized amino acids to glucose IRL was calculated by assuming that 0·35 of urea IRL was due to protein catabolism and 0·2 of catabolized protein-C passed through the glucose pool. On this basis the contribution of protein-C to glucose-C was < 25 g/kg glucose IRL.