Functional (dietary) and numerical responses by lynx (Lynx canadensis) to changing snowshoe hare (Lepus americanus) densities were studied by trailing lynx for 900 km during winter 1964-68 and 1971-75 at Rochester, Alberta. Lynx populations on a 130-km 2 study area responded numerically to changes in hare densities; the low of 3 in 1966-67 and apparent peak of 13 in 1971-72 followed by 1 year a respective low and subsequent peak among hares. There were no kittens present on the study area during five winters of greatest hare scarcity. Early postpartum mortality of kittens, probably starvation related, was at least partly responsible for the decline of lynx populations. Reduced conception rates among yearlings (again probably starvation-related) may have been additional factors in the decline, but their relative importance is still unknown. Lynx also responded functionally to changes in hare densities; frequency of hares in the lynx diet dropped to 5 of 13 food items during 1 winter of hare scarcity and rose to 16 of 17 items when hares were abundant. Carrion was the main alternative food source during the hare low in 1965-67. Ruffed grouse (Bonasa umbellus) was the only alternative prey species of importance to the lynx during hare lows. The mean daily consumption rate by adult lynx was 37 percent greater during winters of hare abundance than during winters of scarcity. Minimum estimates of home-range sizes (averaging 28.0 km 2 ) did not appear related to either hare or lynx densities. Overlap of home ranges occurred in all years. Avoidance behavior appears to separate lynx in both time and space, but probably does not act locally as a density-limiting mechanism.