This study was conducted to determine if the decreased MP supply predicted by the NRC (2001) when canola meal (CM) substitutes soybean meal (SBM) was supported by direct measurement of net portal absorption of AA or energy-yielding nutrients, plus the impact of the type of forage in CM-based rations. Nine Holstein cows with indwelling catheters in splanchnic blood vessels, 8 also with a ruminal cannula were used to examine the effects of protein source in corn silage-based diets, comparing SBM versus CM, and forage source in CM-based diets, comparing corn versus grass silage. The cows were allocated to a triple 3 × 3 Latin square design with 21-d periods. The 3 experimental diets, formulated to be isoenergetic and isonitrogenous, were based on: 1) SBM and corn silage (SoyCorn); 2) CM and corn silage (CanCorn) and 3) CM and cool-season grass silage (CanGrass). Averages of intake, milk yield and milk composition of the last 3 d of each period were used for statistical analyses. On d 21 of each period, 6 sets of arterial, portal, hepatic and mammary blood samples and 2 ruminal fluid samples were collected. On d 12 of period 2, the protein sources were incubated in nylon bags to determine 16h-ruminal disappearance of DM and N and to obtain 16-h residues. Finally, 5 d after the completion of the Latin square design, the mobile bag technique was used to determine DM and N intestinal disappearance of the 16-h residues of SBM and CM. Pre-planned contrasts were used to compare the effect of the protein source in cows fed corn silage, i.e., SoyCorn versus CanCorn, and the effect of forage in cows fed CM, i.e., CanCorn versus CanGrass. Data of the cow without a rumen canula could not be used because of health problem. In corn silage-based diets, substitution of SBM by CM tended to increase milk (6%) and milk fat (7%) yields. The 8% higher ruminal N disappearance and the 19% decreased MP supply from RUP predicted by NRC (2001) were not supported by the 25% decrease in ruminal ammonia concentration, similar net portal absorption of AA (except 22% higher for Met), and the 14% decrease in urea hepatic removal when CM substituted SBM. Ruminal incubation of CM in nylon bags does not appear suitable for adequate determination of the rumen by-pass of a protein source like CM. Inclusion of grass silage rather than corn silage in CM-based diets tended to increase milk (6%) and increased milk lactose (8%) yields. Neither protein nor forage source resulted in variations of metabolism of energy-yielding nutrients that could explain observed increments in cow performance. The present study indicates no decreased AA availability when CM substitutes SBM. Therefore, substitution of SBM by CM in diets based on corn silage and CM in corn- or grass silage-diets can be used successfully in high producing dairy cows.