In Exp. 1, four Holstein heifers (112+/-5.5 kg BW) fitted with ruminal cannulas were used in a 4 x 4 Latin square to evaluate the effects of N source on ruminal fermentation and urinary excretion of purine derivatives. A 2 x 2 factorial arrangement of treatments was used; the factors were the type of protein source (soybean meal, SBM, vs a 50:50 mixture of fish meal and corn gluten meal, FMCGM) and the partial substitution of protein source by urea (with vs without). Heifers were allowed to consume concentrate and barley straw on an ad libitum basis. Barley straw:concentrate ratio (12:88) and average ruminal pH (6.25) were not affected (P > 0.05) by treatment. Ruminal NH3 N concentration and urinary excretion of purine derivatives were not affected (P > 0.05) by supplemental N source. In situ CP degradability of supplemented SBM was very low (50%). In Exp. 2, eight dual-flow continuous-culture fermenters were used to study diet effects on microbial fermentation and nutrient flow, using forage:concentrate ratio, solid and liquid passage rates, and pH fluctuation to simulate in vivo conditions. The treatment containing SBM without urea reached the greatest total VFA concentration (P < 0.01), molar percentage of acetate (P < 0.05), and NH3 N concentration (P < 0.05), followed by treatments with partial substitution of protein source by urea, and finally by the treatment containing FMCGM. True OM digestion tended to increase (P = 0.13) in treatments containing SBM. These results suggest that amino N from SBM and NH3 N concentration stimulated nutrient digestion. Microbial protein synthesis was lowest in treatments with FMCGM and without urea, indicating that rapidly available N limited microbial growth. The low CP degradability of SBM observed may have contributed to the limitation in N supply for microbial growth. Efficiency of microbial protein synthesis increased in treatments containing urea (P < 0.05). Protein source affected total (P < 0.05) and essential AA (P < 0.10) flows, which were greater in treatments containing FMCGM. Partial replacement of protein supplements by urea did not affect total and essential AA flows. Because mean dietary protein contribution to total N effluent was 46%, the AA profile of supplemental protein sources had a great impact on total AA flow and its profile.