In the present field study, the capability of Canada thistle to develop shoots from intact roots and root fragments at different soil depths was studied. The experiments were performed on four sites with high-density Canada thistle, with three or four replications per treatment. At each site, the soil in the plots was removed layer by layer (to 30 or 40 cm, depending on the site), within a 1 by 1-m quadrat, and spread out on a plastic sheet. All roots and other plant parts were removed, and the soil was either replaced without any root material (two sites), or the roots of the thistles were cut into 10-cm-long fragments and replaced into the source holes (two sites). The measured variables were shoot number and biomass. The number of shoots of Canada thistle decreased with increasing depth (P < 0.001) and increased with time. Additionally, the two factors interacted (P < 0.001) such that shoot development was slower from greater depths. Roots from ≤ 20 cm depth produced higher biomasses than did roots from below 20 cm depth. Replacement of root fragments did not affect the amount of biomass produced. It was concluded that the intact root system contributed considerably more to the total biomass produced by Canada thistle than did the root fragments in the upper soil layers.