1. How light acts in controlling flowering of several varieties of chrysanthemums was studied with greenhouse-grown plants from rooted cuttings and transferred to plant-growth rooms for the duration of the experiments. The radiation treatments, which were of 8 days' duration, were preceded and followed by non-inductive photoperiodic conditions. In most experiments the treatments consisted of brief irradiations with red or far red or both near the middle of the dark period, and the plants were dissected 2 weeks later to observe the flowering conditions. The stage of floral initiation was designated by numbers from 0 to 10; 0 represents a vegetative condition of the terminal meristem, and 10 the most advanced condition, in which perianth primordia were present on all florets. Dissection results were found to agree very closely with macroscopic observations made on other branches of the same plants after the plants had developed for several weeks after the light treatments. 2. Floral initiation was markedly reduced by 1 minute of filtered red light from an eighteen-tube fluorescent light source and in most experiments was completely inhibited by 16 or 27 minutes. Floral initiation was repromoted in many varieties of chrysanthemum by a few minutes of far-red radiant energy from incandescent lamps equipped with appropriate filters. Repeated alternations of red and far-red radiant energies resulted, respectively, in repeated inhibition and repromotion of floral initiation. 3. If the far-red treatment is separated from the red treatment by a dark interval, repromotion of floral initiation becomes less as the duration of darkness increases. Little effect is usually found for 15-minute dark periods, but reversal fails almost completely if the dark period is as long as 90 minutes. 4. Far red alone or following a brief red treatment inhibits flowering if the duration of far-red treatment is about 81 minutes. 5. The practical and theoretical significance of the results are discussed.