Knowing the relative importance of seasonal environmental variables to subsequent spring yields of alfalfa (Medicago sativa L.) in the Northern United States may be useful in current alfalfa modelling strategies because they do not consider overwintering effects on spring yields and possible changes in the yield response to the environment as the crop ages. The purpose of this study was to define for the first three production years the growing and overwintering environmental variables of greatest importance to subsequent spring dry matter production. The variables were determined using multiple regression analysis on 21 years of climate-yield data from the University of Minnesota Experimental Station at Rosemount. Precipitation in the prehardening period of growth (September in Minnesota) had a positive effect on subsequent spring yield, accounting for 40% of the yield variability in the first year after seeding. However, precipitation during the hardening period (October) was negatively related to yield. Spring yields in the second year of production were positively affected by maximum air temperatures in fall and spring and minimum air temperatures in winter. These variables accounted for 50% of the yield variability. Spring yields in the third production year were positively related to solar radiation in fall and late winter and to spring temperature (as related to growing degree days). Forty-three percent of the yield variability was attributed to these variables. The results indicate that the relative importance of environmental variables in alfalfa spring dry matter production were dependent on stand age. Fall and winter period variables, notably precipitation and air temperature, were found to have a large effect on spring yields and thus should be considered in alfalfa growth models.