RESUMO
Data were collected from 10 experiments with duodenally cannulated Holstein dairy cows (271 combinations of cow and period) to evaluate the relationship between milk purine metabolites and microbial nitrogen flow. Experiments evaluated the effects of dietary factors on microbial N production and included: 1) supplemental sources of ruminally protected amino acids; 2) grass silage treated with fibrolytic enzymes; 3) bacterial inoculation of corn silage; 4) a comparison of corn and barley grain; 5) ruminal starch availability as affected by corn silages of varying maturity; and, 6) ruminal starch availability as affected by corn silages harvested at varying chop length and with or without mechanical processing. The coefficient of determination for individual experiments for the relationship between microbial nitrogen flow and allantoin or uric acid excretion in milk ranged from -0.01 to 0.77 and -0.06 to 0.22 respectively. Across all experiments, the coefficients of determination between microbial nitrogen flow and allantoin or uric acid excretion in milk were r2 = 0.09 and 0.01 respectively. Milk allantoin output was used to develop a prediction equation estimating microbial nitrogen flow to the duodenum. The greatest predictive value (r2 = 0.25) relationship was observed across non-bST experiments and included the factors of allantoin excretion, milk yield, dry matter intake, and days in milk. When milk yield was included in the model, the predictive value improved (r2 = 0.44). Based on these data, milk uric acid excretion alone cannot be used to predict microbial N production accurately [corrected].