Four Holstein cows with ruminal and duodenal cannulas were assigned to a 4 x 4 Latin square at each of four stages of lactation (peak, wk 4; early, wk 8 to 12; mid, wk 17 to 21; and late, wk 27 to 31). Treatments were duodenal infusions of 1) Met, 2) Lys, 3) Met plus Lys, and 4) casein; periods were 7 to 10 d. Quantities of DL-Met, L-Lys, and casein infused at the four stages of lactation were 12, 30, and 400; 12, 30, and 400; 10, 25, and 333; and 8, 20, and 266 g/d, respectively. Rations were composed of corn and grass-legume silages, corn meal, wheat middlings, soybean meal, and distillers dried grains with solubles. Intake of CP (percentage of NRC) and percentage of total CP from corn sources were (peak) 87, 56; (early) 90, 71; (mid) 98, 73; and (late) 114, 77. Using content and yield of milk protein as primary response criteria, Lys appeared to be first-limiting and Met second-limiting at peak lactation, their infusion together resulted in the same production of milk (40 kg/d) and milk protein (1135 g/d) as did casein. Lysine was first-limiting in early lactation, but whether Met was second-limiting was questionable. The two AA were colimiting in midlactation. There appeared to be no AA deficiencies in late lactation. Amounts of Lys and Met (percentage of total essential AA) in duodenal digesta during peak (12.4, 3.7), early (12.6, 3.5), and midlactation (14.9, 3.9) were not adequate for optimal AA utilization.