The development of problematic weed populations is a concern in western Canadian fields where canola and pea are grown in a 4-yr sequence with spring cereal grains. Weed densities were examined at a site near Melfort, Saskatchewan, Canada, from 1994 to 1997 in seven zero-till managed crop rotations. Four rotations that included canola, pea, or flax in at least 3 of 4 yr (HBF: high broadleaf–crop frequency) were compared with three rotations that included broadleaf crops grown in 2 of 4 yr (LBF: low broadleaf–crop frequency). Spring wheat and barley were the cereal crops in rotation. Residual (postherbicide application) weed density for each weed species in a given year was summed across all phases for each rotation to reflect the overall weed infestation. Four annual broadleaf weed species were most abundant in 1996 and a second group of three species, having a variety of reproductive strategies, became progressively less abundant as the study progressed. The difference between the HBF and LBF rotations for the density of these species varied and was most prominent in years when environmental conditions were conducive for their growth. More frequent applications of ethafluralin, with its residual weed control, best explained why wild oat and catchweed bedstraw generally were less abundant in the HBF rotations. Of particular interest was the 8 plants m−-2 greater density of dandelion and perennial sowthistle in the HBF vs. LBF rotations in the last year of the study. It is thought that the limited herbicide options for the control of these species could present a future problem if they continued to develop in the HBF rotations. Differences in herbicide use between the HBF and LBF rotations were considered the primary factor controlling the rotation effects on weed density.