Eight dual-flow continuous culture fermenters (1320 mL) were used in two 10-d consecutive periods to study the effects of fiber content and particle size on rumen fermentation, nutrient flow, and the profile and flow of amino acids (AA) from microbial origin. Treatments were arranged in a 2 x 2 factorial; the main factors were fiber content (high fiber [HF] = 67% alfalfa hay, 33% concentrate; low fiber [LF] = 39% alfalfa hay, 61% concentrate) and forage stem particle size (ground not to pass a 3-mm sieve = large size [LS]; ground to pass a 1-mm sieve = small size [SS]). On the last day of the experiment, liquid- (LAB) and solid- (SAB) associated bacteria were isolated from each fermenter for chemical analysis. Microbial N and AA flows were estimated using LAB or SAB composition data. Total volatile fatty acid concentration was higher for LF (122 mM) than for HF (102 mM). The proportion of acetate was higher for HF (63.5%) than for LF (58.2%). When SS was fed instead of LS, there was a reduction in the acetate proportion (63.4% vs 58.3%, respectively) and in the acetate to propionate ratio (3.13 vs. 2.41, respectively). Bacterial N flow was higher, and dietary N flow was lower, in SS compared with LS when LAB were used for calculations. Efficiency of microbial protein synthesis (EMPS) was affected by particle size, when LAB were used for calculations, and by fiber content, when SAB were used for calculations. The AA profile of LAB differed from SAB in 4 of 16 AA. The fiber content and particle size had small effects on microbial AA profile. However, the bacterial AA flow was higher in 13 of 16 AA, when SAB instead of LAB were used for calculations. Estimates of bacterial N flow, EMPS, AA profile, and flows of microbial origin were dependent on the type of microbial population used for calculations. The representativeness of the microbial samples is essential for the accurate prediction of the AA flow from microbial origin.