ContextIntercropping is used to improve crop productivity and sustainability. However, total and partial yields of intercrops vary widely due to both intercropping system and environment. Understanding the factors controlling this variability would be of value for crop producers deciding whether and how to adopt intercropping. ObjectiveOur objective was to determine the dominant factors controlling total and partial yields of pulse-oilseed intercrops in a semiarid region. MethodWe conducted a field study for three years in southern Alberta under both dryland and irrigated conditions with lentil (Lens culinaris) and pea (Pisum sativum) intercropped with canola (Brassica napus) or mustard (Sinapis alba). Oilseed plant density and N fertility was varied and competition for soil and fertilizer N was measured using 15N. ResultsCrop productivity expressed using the Land Equivalent Ratio (LER) ranged from 0.98 to 1.75, similar to the range reported in the literature. High LER occurred when monocrop oilseed yields were strongly limited by N and increased when the oilseed crop was more competitive for fertilizer and soil N. In 2018, canola was about 4- to 6-fold more competitive for fertilizer N than pea or lentil and LER was high when N was limiting. In 2019 and 2020, mustard was about 3-fold more competitive than lentil but no more competitive than pea and LER was only higher when mustard was intercropped with lentil and N was limiting. Modifying oilseed plant density primarily impacted partial yields, not LER. Application of N fertilizer at 50 kg ha-1 reduced LER by 12%, whether applied at seeding or at five weeks. Over-yielding in this study was largely controlled by complementary N use, with variation in total and partial yields largely accounted for by the deficit in N supply for the oilseed crop, relative crop competiveness for soil N, and plant density. ImplicationsThe primary benefit of intercropping pulse and oilseed crops is sustained crop productivity with reduced N fertilizer inputs or under variable N sufficiency. Reliable prediction of total and partial yields should be feasible and allow crop producers to evaluate whether and how to incorporate intercrops into crop rotations. Studies on intercrops of legumes with non-legumes should include an N-sufficient non-legume control and benefit from the use of 15N to measure competition for soil N.