IT is generally assumed that the b-wave of the human retinal action potential is due to the stimulation of the rods. With an intensity of 2,000 lux, the b-waves fuse at a frequency of about 20 flashes per second. At lower intensities the fusion frequency in our experiments decreased. A linear relation was found between fusion frequency and the logarithm of stimulus intensity. However, the subjective fusion frequency, determined by cone activity, is known to reach values around 50–60 flashes per second. These high values are never obtained with the b-waves of the electroretinogram as index. This is but one of several reasons for assuming the b-waves to express rod activity.