Partial girdles were applied in August, 1957, to one stem of two double-stemmed, 20-year-old Douglas fir (Pseudotsuga menziesii (Mirb.) Franco). The second stem served as control. A third double-stemmed tree was treated in May, 1958. Cone production responses were obtained on all three girdled stems, averaging 7.4 times that of control stems in 1959, and 1.6 and 2.3 times that of control stems in 1961 and 1962. Cone production responses to treatment, and cone crop variation over several years were correlated with reduced bud failure during the period of new shoot elongation. Total number of buds per shoot was initially similar for paired stems. These relationships indicated a predetermined potential for annual cone production, and that cone crop periodicity is determined by later conditions favorable or unfavorable to continued early bud development. Treatment increased both sugars and starch in shoots sampled 40 days after August girdling, but only starch remained elevated the next spring and throughout the decisive May–June period of reproductive bud development. Other factors indicated food reserves to be related only weakly to reproductive bud survival. Cone production reduced carbohydrate concentration in shoots of all ages, growth and number of new shoots, and number of developed buds per shoot. These factors explain the absence of consecutive cone crops in Douglas fir, and suggest that cone inducing treatments should not be applied in good flowering years. Cone production responses on single-stemmed trees girdled at weekly intervals showed an optimum timing coincident with the onset of flowering, a more variable response up to the time of vegetative bud break, then an adverse effect on cone production when girdled later than 1 week after vegetative bud break.