The aim of this research was to examine to what extent variation in the relative proportions of solid- (SAB) and liquid-associated rumen bacteria (LAB) in duodenal bacteria have an impact on the estimation of duodenal flow of bacterial N. For this, four dairy cows were fed diets varying in forage: concentrate ratio (80:20, 65:35, 50:50 and 35:65). SAB and LAB were separated from rumen contents four h after the morning feeding. Adenine, cytosine and odd and branched-chain fatty acids were determined both in SAB and LAB and used to estimate bacterial N flow. Bacterial N flows were also calculated using a SAB:LAB ratio in duodenal bacteria, as estimated from the odd and branched-chain fatty acid pattern. Compared with calculations based on the estimated SAB:LAB ratio, estimations based on SAB or LAB only as a bacterial reference on average over- and underestimated bacterial N flow by 37 and 55 g N/d, respectively (P<0.05) when cytosine or adenine were used as bacterial marker. In contrast, due to the small differences in the OBCFA:N ratio between SAB and LAB, these differences were less than 15 g/d when OBCFA were used as bacterial marker. The results suggest that, depending on the marker used, changes in the proportions of SAB and LAB can have a substantial impact on estimated duodenal flow of bacterial N.