Kishchuk, B. E., Quideau, S., Wang, Y. and Prescott, C. 2014. Long-term soil response to variable-retention harvesting in the EMEND (Ecosystem Management Emulating Natural Disturbance) experiment, northwestern Alberta. Can. J. Soil Sci. 94: 263–279. We report on soil responses to variable-retention (VR) harvesting from a large-scale experiment (>1000 ha). Samples were taken prior to treatment, and 1 and 6 yr after treatment under cover types representing the successional trajectory for boreal mixedwood forests in northwestern Alberta, Canada. Variable-retention harvesting at six overstory-retention levels [100 (control), 75, 50, 20, 10, and 0% (clearcut)] were applied to 10-ha experimental units under four cover types: deciduous-dominated (80–95% trembling aspen); deciduous-dominated with coniferous understory (80–95% trembling aspen with white spruce understory at 60–80% of full stocking); mixed coniferous-deciduous (35–65% of each); and coniferous-dominated (80–95% white spruce). Only a few clear differences in soil properties attributable to VR harvesting were evident 6 yr post-harvest: (1) both extractable NH4-N in forest floor and 0–7 cm mineral-soil, and forest floor exchangeable K decreased with increasing canopy removal, and (2) forest floor exchangeable Ca increased with canopy removal. There was a decreasing trend in forest floor and mineral soil C (kg ha−1) in undisturbed stands between 1998 and 2005. Differences in soil properties among cover types included higher pH and N concentration in forest floors, and higher cation exchange capacity and exchangeable Ca and Mg in mineral soils in deciduous-dominated stands. Deciduous-dominated stands appear to have distinct soil properties that change under stand development.