Relative benefit of introducing forage species to the Northern Great Plains have been examined with contradictory conclusions. In most cases, studies were either confounded by time of establishment or treatments were not randomized and lacked independence. We examined aboveground net primary production (ANPP) in northern mixed prairie using a randomized complete block design with four treatments: crested wheatgrass (Agropyron cristatum [L.] Gaertn.), Russian wildrye (Psathyrostachys juncea [Fisch.] Nevski), a native control that was not harvested, and a harvested native. The experiment was conducted in a Stipa–Agropyron–Bouteloua site and a Stipa–Bouteloua site over 13 yr and 12 yr, respectively. The data were analyzed by sampling period (Stipa–Agropyron–Bouteloua: 1, 1994 to 1997; 2, 1998 to 2001; 3, 2002 to 2006; and Stipa–Bouteloua: 1, 1995 to 1998; 2, 1999 to 2002; 3, 2003 to 2006). ANPP among treatments was influenced (P < 0.05) by site and its interaction with treatment and sampling period (1 to 3). ANPP from the native-control, harvested-native, crested wheatgrass, and Russian wildrye treatments was 220.9, 183.9, 300.8, and 189.6 g · m^–2 (SEM  =  11.2), respectively, in the Stipa–Agropyron–Bouteloua site and 122.9, 98.2, 216.3, and 115.9 g · m^–2 (SEM  =  12.0), respectively, in the Stipa–Bouteloua site. Mean ANPP (SEM) within each sampling period (1 to 3) was 186.4 (9.1), 135.4 (5.8), and 263.9 (8.8) g · m^–2 in the Stipa–Agropyron–Bouteloua site, respectively, and 124.5 (6.4), 138.6 (6.1), and 151.3 (10.5) g · m^–2 in the Stipa–Bouteloua site, respectively. Russian wildrye in the Stipa–Bouteloua site and crested wheatgrass in both sites was relatively more productive in the first period after establishment than in subsequent years. The study confirms the relative ANPP advantage of crested wheatgrass over native on the Stipa–Bouteloua site but not on the Stipa–Agropyron–Bouteloua site, whereas Russian wildrye exhibited no ANPP advantage over the native on either site.