1. Using a phosphoroscope, delayed luminescence and prompt chlorophyll fluorescence from isolated chloroplasts have been compared during the induction period. 2. Two distinct decay components of delayed luminescence were measured a "fast" component (from approximately 1 ms to approximately 6 ms) and a "slow" component (at approximately 6 ms). 3. The fast luminescence component often did not correlate with the fluorescence changes while the slow component significantly changed its intensity during the induction period in a manner which could usually be linearly correlated with variable portion of the fluorescence yield change. 4. This correlation was evident after preillumination with far-red light or after allowing a considerable time for dark relaxation. 5. The close relationship between the slow luminescence component and variable fluorescence yield was observed with a large range of light intensities and also in the presence of 3-(3,4-dichlorophenyl)-1,1-dimethylurea which considerably changes the fluorescence induction kinetics. 6. Valinomycin and other antibiotics reduced the amplitude of the 6 ms (slow) luminescence without affecting its relation with the fluorescence induction suggesting possibly that a constant electrical gradient exist in the dark or formed very rapidly in the light, which effects the emission intensity. 7. Changes in salt levels of suspending media equally affected the amplitude of both delayed luminescence and variable fluorescence under conditions when the reduction of Q is maximal and constant. 8. The results are discussed in terms of several models. It is concluded that the model of independent Photosystem II units together with photosynthetic back reaction concept is incompatible with the data. Other alternative models (the "lake" model and photosynthetic back reaction; recombination of charges in the antenna chlorophyll; the "w" hypothesis) were in closer agreement with the results.