The present study examined whether different proteins have different effects on whole-body protein turnover in adult rats. The rats were either starved, given a protein-free but energy-sufficient diet (1 MJ/kg body weight (BW) per d) or a diet containing intact casein, hydrolysed casein, or hydrolysed soya-bean protein at a level of 9.1 g/kg BW per d. The diets, which were isoenergetic with the same carbohydrate: fat ratio, were given as a continuous intragastric infusion for at least 4 d. During the last 19 h 15N-glycine (a primed continuous infusion) was given intragastrically and 15N was recovered from urinary ammonia and urea during isotope steady state for measurement of protein synthesis and protein degradation. Compared with starvation the protein-free diet decreased N excretion by 75%, probably by increasing the rate of reutilization of amino acids from endogenous proteins for protein synthesis. The protein diets produced a positive N balance which was independent of the protein source. Intact and hydrolysed casein increased protein synthesis 2.6- and 2.0-fold respectively, compared with the protein-free diet. Protein degradation increased 1.4- and 1.2-fold respectively. Hydrolysed soya-bean protein did not increase protein synthesis but decreased protein degradation by 35% compared with the protein-free diet. Compared with the hydrolysed soya-bean protein, intact casein resulted in 2.2- and 2.8-fold higher rates of protein synthesis and degradation respectively. These results are not easily explained by known sources of misinterpretation associated with the 15N-glycine method. Hydrolysed casein and hydrolysed soya-bean protein produced similar concentrations of insulin-like growth factor-1, insulin, glucagon, and corticosterone. The difference in amino acid composition between the dietary proteins was reflected in plasma amino acid composition and this is suggested to be responsible for the different effect on protein turnover. Preliminary results from this study have previously been published in abstract form (Nielsen et al. 1991).