The paper reports a detailed statistical study of breakout azimuths in 48 oil wells widely distributed in the Alberta sedimentary basin, bringing the number of oil wells contributing azimuths to 94 for the western Canadian basin. The azimuths show significant regional variation between the northern, central, and southern parts of the basin. Twenty-one wells, in which breakouts cover depth ranges greater than 600 m, were used to investigate regression of breakout azimuths on depth. Ten wells give positive regression coefficients, 11 are negative, and no coefficient is significant at 95% confidence level. If the dominant northwest–southeast orientation of the long axes of breakouts gives the orientation of the lesser horizontal principal stress Sh, following the hypothesis of Bell and Gough, the inferred stress orientations indicate that throughout the whole basin the direction of the maximum horizontal stress SH is northeast–southwest. Stress measurements, by strain-relief techniques in a mine and from hydraulic fracture in wells, support the stress orientation given by the breakouts. The insignificant regression of breakout azimuths on depth supports the view that the orientation data represent stress in the lithosphere rather than in the sediments only. Directions of the lesser horizontal compression Sh, from Zoback and Zoback in the United States and from breakout studies in western Canada, are combined to suggest that the Mid-Continent stress province of North America may include the western Canadian basin and the Canadian Shield as well as the central United States. Coherent stress with the observed orientation, over the continent east of the Rocky Mountains, would result from northeastward basal drag on the North American plate, as Zoback and Zoback have pointed out. Basal drag to the northeast could arise either if the plate were sliding southwestward over a passive asthenosphere, as suggested by Zoback and Zoback, or if northeastward mantle flow were driving the asthenosphere and the plate northeastward.