The effects of protein source (PS), soybean meal (SBM) vs. canola meal (CM) in corn silage-based diets, and of forage source (FS), corn silage vs. grass silage in CM-based diets, were examined in 9 rumen-cannulated cows (103 ± 28 DIM, 704 ± 65 kg BW), in 3 3 × 3 Latin squares (LS; 21-d periods). The 3 isoenergetic (1.60 Mcal/kg) and iso-N (16.0% CP) diets were: SC = SBM (17.7% DM)+corn silage; CC = CM (27.4% DM)+corn silage; and CG = CM (20.6% DM)+grass silage. Milk yield and composition, and DMI were monitored on d 19–21; on d 21, 2 samples of rumen fluid were collected. On d 12 of period 2, the PS of that period was incubated for 16 h in the rumen of each cow to prepare 16-h residues. After the completion of the LS, cows continue to receive their period-3 ration for 1 wk. To determine CP intestinal digestibility, small nylon bags (2 × 2 cm) containing 80 mg of the 16-h residues of the PS of that period were inserted into the abomasum of each cow and recovered in the feces (10 bags recovered per cow). Data of the LS were analyzed using the Mixed procedure of SAS, using contrasts to test the effect of PS and FS. Results are presented in the Table, ns = 0.20. Intestinal digestibility of CP was higher (P < 0.01) for SBM vs. CM, averaging 94.3 vs. 74.4 ± 1.2%. Substituting SBM by CM in a corn-based diet tended to increase milk yield whereas a positive effect of CM inclusion was larger when the cows were fed a grass-based compared with a corn-based diet. Table 1.

Item	Diet			SEM	P-value
	SC	CC	CG		SC vs. CC	CC vs. CG
DMI, kg/d	26.0	27.5	25.9	0.5	0.05	0.04
N intake, g/d	672	696	681	18	NS	NS
Milk yield, kg/d	43.7	46.0	48.9	0.8	0.08	0.03
True protein yield, g/d	1,301	1,330	1,389	36	NS	0.20
Fat yield, g/d	1,540	1,639	1,753	33	0.06	0.03
Lactose yield, g/d	1,942	2,003	2,187	46	NS	0.02
BW change, kg/21d	−0.4	−1.4	12.0	5.2	NS	0.09
Rumen pH	5.90	5.87	6.23	0.12	NS	0.06
Rumen ammonia, mM	10.5	7.9	8.0	0.7	0.01	NS
Rumen total VFA, mM	109.4	99.2	95.5	3.2	0.03	NS
Rumen acetate:propionate	2.1	2.2	2.9	0.1	NS	0.01