The interaction of cultivation and photoperiod on the initiation of Canada thistle rosettes, and the effect of growth stage on control of Canada thistle with glyphosate were determined in a series of field experiments. Under the natural photoperiod occurring in the southern Canadian Prairies, rosettes of Canada thistle were initiated by cultivation to remove the shoot growth during the last week of July. Regrowth remained as rosettes without any stem elongation and formed dense clusters of leaves. Application at the “August rosette stage” improved the effectiveness of the glyphosate treatment and resulted in consistent control of Canada thistle with less than half as much herbicide as recommended for control when applied at the bud-stage. The number of shoots of Canada thistle on plots treated with glyphosate was less than on the intensively summer-fallowed check plot. Check plots received 5 cultivations during the summer-fallow season. One year after application of glyphosate at the rosette stage, the reduction in shoot density was 99% compared to the summer-fallowed check. By year three, without glyphosate applications in years two and three, the benefits of applying glyphosate at the rosette stage rather than the bud-stage were very evident. When applied at the bud-stage the shoot density on plots treated with glyphosate at 2.25 kg ha−1 was 24 m−2 compared to only 10 shoots m−2 when applied at the rosette stage, (35% vs 72% control). Glyphosate at 0.9 kg ha−1 or less than half of the rate recommended for application at the bud-stage, applied to Canada thistle in the rosette stage, resulted in consistently fewer shoots 2 and 3 yr after treatment. Physical removal of shoot top growth, simulating an in-crop herbicide treatment, improved the consistency of control. Two years after application of 0.9 kg ha−1 of glyphosate at the rosette stage, the control of Canada thistle was 98% compared to the summer-fallowed checks.