Electroretinograms (ERGs) were measured in 15 patients with central retinal vein occlusion (CRVO). Seven of the patients had neovascularization of the iris (NVI) at the time of testing, and two developed NVI within one month of testing; six did not have NVI or any other form of neovascularization (NV) and were no longer considered to be at risk for NV from their present occlusion. The ERGs were recorded as a function of the stimulus intensity and to a 30-Hz flickering stimulus. A Naka-Rushton-type function was fit to b-wave amplitudes, measured as a function of stimulus intensity, to evaluate changes in ERG amplitude and sensitivity. Compared with eyes in the no-NV group, eyes developing NVI had significantly reduced ERG sensitivity and amplitudes. The distributions of sensitivity values in these two groups did not overlap. All of the eyes with NV showed large a- and b-wave and 30-Hz implicit time delays, but only one eye had a b/a-wave amplitude ratio close to or less than 1. A comparison of ERG sensitivity, amplitude data, and flicker timing data with retinal fluorescein angiography in a two-alternative forced-choice analysis showed that ERG sensitivity and amplitude loss were better than retinal fluorescein angiography at discriminating eyes with CRVO and NVI from eyes with CRVO without NVI.