Two 4 x 4 Latin square trials (4-wk periods; 16 wk total) were conducted to see whether supplementing rumen-protected Met (RPM; fed as Mepron) would allow feeding less crude protein (CP), thereby reducing urinary N excretion, but without losing production. In trial 1, 24 Holsteins were fed 4 diets as total mixed rations containing [dry matter (DM) basis]: 18.6% CP and 0 g of RPM/d; 17.3% CP and 5 g of RPM/d; 16.1% CP and 10 g of RPM/d; or 14.8% CP and 15 g of RPM/d. Dietary CP was reduced by replacing soybean meal with high-moisture shelled corn. All diets contained 21% alfalfa silage, 28% corn silage, 4.5% roasted soybeans, 5.8% soyhulls, 0.6% sodium bicarbonate, 0.5% vitamins and minerals, and 27% neutral detergent fiber. There was no effect of diet on intake, weight gain, or yields of protein, lactose, and solids-not-fat. However, production was greater at 17.3% CP plus RPM and 16.1% CP plus RPM than on the other 2 diets. Apparent N efficiency (milk N: N intake) was greatest on the lowest CP diet containing the most RPM. Linear reductions in milk urea N and urinary N excretion were observed with lower dietary CP. In trial 2, 32 Holsteins were fed 4 diets as total mixed rations, formulated from ingredients used in trial 1 and containing 16.1 or 17.3% CP with 0 or 10 g of RPM/d. On average, cows were calculated to be in negative N balance on all diets because of lower than expected DM intake. There was no effect of RPM supplementation on any production trait. However, higher CP gave small increases in yields of milk, protein, and solids-not-fat and tended to increase DM intake and lactose yield. Apparent N efficiency was greater, and milk urea nitrogen was lower, on 16.1% CP. In trial 1, feeding lower CP diets supplemented with RPM resulted in improved N efficiency and reduced urinary N excretion. However, in trial 2, reducing dietary CP from 17.3 to 16.1% reduced milk secretion, an effect that was not reversed by RPM supplementation at low DM intakes when cows were apparently mobilizing body protein.