The objectives of this study were to measure performance responses and to evaluate indictors of N utilization in late-lactation cows fed diets with incremental reductions in crude protein (CP) concentration. Holstein cows (n = 128; 224 +/- 54 d in milk) were stratified by parity and days pregnant (86 +/- 25 d) and randomly assigned to 1 of 16 pens in a randomized complete block design. For 3 wk, all cows received a covariate diet containing 16.9% CP [dry matter (DM) basis]. For the subsequent 12 wk, pens were randomly assigned to 1 of 4 treatments that contained 16.2, 14.4, 13.1, or 11.8% CP (DM basis). Diets were offered once daily and contained 32.5% corn silage, 32.5% alfalfa silage, 13.5% high-moisture corn, and 21.5% concentrate mix. A reduction in dietary CP was achieved by replacing soybean meal with soy hulls in the concentrate mix (DM basis). Dry matter intake, milk urea N (MUN; mg/dL), and the yield of milk urea N (g/d) decreased linearly with dietary CP. Compared with a 16.2% CP diet, a 14.4% CP diet did not alter milk yield throughout the study, but the 13.1 and 11.8% CP diets reduced milk yield after 4 and 1 wk, respectively. Furthermore, milk protein percentage was reduced for all dietary CP less than 16.2%, but this negative effect was temporary and disappeared after 7 wk for the 14.4% CP diet. In contrast, MUN adjusted to a new steady state within 1 wk for all dietary treatments. Modeling quadratic responses with a plateau led to predictions of no reduction in fat- and protein-corrected milk (32.6 kg/d) and yields of fat (1.31 kg/d), lactose (1.49 kg/d), and true protein (1.12 kg/d) until dietary CP decreased below 15.5, 15.3, 15.9, and 16.2%, respectively. In this study, MUN and the yield of MUN were highly correlated with N intake, milk protein yield, and fat- and protein-corrected milk. Surprisingly, N use efficiency (milk protein N/intake N) was not correlated with any variables related to N utilization and reached an apparent upper limit of approximately 30%. Although this observation may be associated with feeding diets deficient in metabolizable protein, late-lactation cows in this study adjusted to low dietary CP concentration better than anticipated as milk production was 2.6, 3.6, 6.4, and 8.0 kg/d higher than National Research Council (2001)-predicted metabolizable protein-allowable milk for dietary CP of 16.2, 14.4, 13.1, and 11.8%, respectively.
