ABSTRACT
Tungstic acid precipitable N and nonprecipitable N were measured in feed, rumen and feces of cows fed complete diets, and changes in ratios of these components to chromic oxide were used to estimate net changes in true protein from feed to rumen to feces. Apparent digestible true protein leaving the rumen was assumed to be metabolizable protein. Rumen contents were hand mixed before sampling from rumen-fistulated cows so that samples would be as representative of total rumen contents as possible, and these samples were assumed to represent contents being passed to small intestine. Metabolizable protein percentages of three low protein basal diets containing 1.90, .95 and 0% urea and three similar diets with added soybean meal fit responses of milk yield obtained in a previous experiment much better than metabolizable protein estimated from literature values for various feedstuffs. Net true protein digestion estimates confirmed negative performance from 1.90% urea basal diets compared with performance from other basal diets. An experiment in which rumen samples were taken by stomach tube from intact cows showed rumen samples were not representative of total rumen contents, at least not for chromic oxide percentage, and, consequently, net true protein digestion could not be estimated a well as with fistulated cows.