COMPARISON OF TIME AND DEPTH OF LAST TILLAGE ON THE GROWTH OF CANADA THISTLE (CIRSIUM-ARVENSE) IN SUMMERFALLOW AND ITS RESPONSE TO GLYPHOSATE

Field studies were conducted in summerfallow in 1990 and 1991 at Bezan son, Alberta Oat. 55 degrees 14'N, long. 118 degrees 22'W), and Melfor t, Saskatchewan Oat. 52 degrees 52'N, long. 104 degrees 36'W), to comp are the development of Canada thistle [Cirsium arvense (L.) Scop.] sho ots following the discontinuance of normal tillage at three times (19 July, 30-31 July and 9-12 Aug.), at three depths (0, 10 and 20 cm), an d its response to glyphosate applied at 0.9 kg ha(-1) 40 d after the l ast tillage (DALT). The tillage treatments had no consistent effect on shoot density when measured at 40 DALT at either location. At Bezanso n, small increases (<5%) in shoot density occurred between 20 and 40 D ALT where tillage was conducted at the shallowest depth, but increases of approximately 20-90% occurred during the same period following til lage at the 10 or 20-cm depth. The proportion of emerged shoots remain ing as rosettes at 40 DALT increased as the time of discontinuing till age was delayed. At Bezanson, the proportion of emerged shoots remaini ng as rosettes at 40 DALT increased from 0.17 +/- 0.05 to 0.59 +/- 0.0 2 when the time of discontinuing tillage was delayed from mid-July to early August, while at Melfort an increase from 0.47 +/- 0.04 to 0.93 +/- 0.02 occurred for a similar delay in time of discontinuing tillage . increasing the depth of last tillage increased the proportion of eme rged shoots remaining as rosettes at 40 DALT at Bezanson, but at Melfo rt the depth of last tillage had no effect on this proportion. The pro portion of shoots remaining as rosettes declined between 20 and 40 DAL T when measured at Bezanson. Factors causing an increase in the propor tion of shoots remaining as rosettes appeared to be associated with ex posure of emerged shoots to reduced daylengths. At both Bezanson and M elfort, shoot density measured 1 yr after tillage-glyphosate treatment s was similar in all plots. Thus, the proportion of shoots remaining a s rosettes at the time of glyphosate application did not appear to inf luence the response of Canada thistle to the herbicide.