1. Mech et al. (1987) documented cumulative, negative effects of previous winters’ snow on rates of population increase in moose (Alces alces) and white-tailed deer (Odocoileus virginianus), but noted no effect of predation by wolves (Canis lupus). Those results were contested by Messier (1991), who analysed smoothed versions of the original abundance data and reported no effect of snow accumulation on population dynamics of either species, but strong effects of wolf predation and food competition.
2. McRoberts, Mech & Peterson (1995) contended that the conclusions reached by Messier (1991) were an artefact of the use of smoothed data. In a subsequent re-analysis of the smoothed data, Messier (1995) argued that the lack of an effect of snow after one year precluded the potential for a cumulative effect beyond one year.
3. We re-analysed original and smoothed data on dynamics of moose and white-tailed deer densities using the same methods as Mech et al. (1987) and Messier (1991), but we used a measure of global climatic fluctuation, the North Atlantic Oscillation (NAO) index. The NAO is the atmospheric process determining most interannual variation in snowfall and winter temperatures in northern latitudes, and its phases drive decadal trends in wintertime precipitation.
4. We observed that rates of increase of moose and white-tailed deer in both the original and smoothed data were influenced by global climatic fluctuation at 2- and 3-year lags, as well as by delayed density-dependent feedback and wolf predation.