In three separate experiments, groups of four, three and six mature, rumen-cannulated Simmental cows were fed diets designed to supply different quantities of energy (E) and nitrogen (N) to the rumen microbes. Experiment 1 (straw plus concentrate plus urea) had balanced E and N supplies; Expt 2 (hay and four levels of soyabean meal) had different sub-optimum N levels at the same intake as Expt 1; and Expt 3 (hay plus maize plus two levels of urea) reproduced the N shortage in Expt 2 at a higher level of intake.
The amount of total N excreted in the urine (TUN) was 574 mg/kg LW0·75 in Expt 1 and 420 mg/kg LW0·75 of this N was excreted as total purine nitrogen (TPN). In Expt 2, TUN increased significantly (P < 0·05) with increasing soyabean levels; TPN also increased, reaching the level observed in Expt 1 when soyabean meal supplementation was highest. In Expt 3, TUN and TPN increased with increasing dietary urea concentrations; TUN and TPN were always higher than in Expts 1 and 2.
Estimated microbial nitrogen supply (EMNS, based on TPN) was always considerably lower than estimates based on fermentable non-protein OM and crude protein (EMNR-EN) or an assumed yield of microbial N per kg OM apparently digested in the rumen (EMNR-OM). Regression analyses of EMNS on EMNR-EN, EMNR-OM or OM intake had moderate r2 values (0·76, 0·65 and 0·62 respectively) but the constant terms were significantly different from zero. The regressions of EMNR-EN and EMNR-OM on TPN gave angular coefficients of 15·81 and 11·47 respectively.
The correlation between rumen liquid parameters (total count, bacterial DM, ATP and nitrogen) and OM intake, EMNS or EMNR increased with sampling time (from 09.00 to 16.00 h). The EMNR-OM produced correlation coefficients similar to those obtained with OM intake; these correlations were numerically higher than those obtained with EMNS or EMNR-EN.