2. When such thresholds and the intensities of the background lights are expressed in terms of the number of quanta absorbed per unit time, it is found that three factors can be correlated with the thresholds measured in various states of light-and dark-adaptation: (i) the intensity of a steady background, (ii) the rate of regeneration of rhodopsin, and (iii) the amount of metarhodopsin II present in the rods.3. The threshold is found to be proportional both to the intensity of a background and to the rate of regeneration, whereas there is a linear relationship between the logarithm of the threshold and the amount of metarhodopsin II. 4. The presence of metarhodopsin elevates all thresholds, the absolute threshold, increment thresholds and the thresholds elevated by regenerating rhodopsin in the same way.5. The saturation of the rods at high background intensities is found to be correlated with the accumulation of significant amounts of metarhodopsin in the rods, caused by the bleaching effect of the background.6. The effect of metarhodopsin on the threshold is independent of the amount of rhodopsin present in the rods.7. The combined effect of all three factors can be expressed in a general formula, given as eqn. (7) on p. 74.8. A background not only reduces the signals from the rods illuminated, but also those from neighbouring unilluminated rods. This effect is rapidly decreased with increasing distance from rods covered by the background. This kind of lateral spread in the retina probably occurs also when the K. 0. DONNER AND TOM REUTER rate of regeneration affects the threshold. The effect of metarhodopsin, on the other hand, appears restricted to those receptors that contain this substance.