In order to use the ERG to track the effects of potential photoreceptor rescue treatments, we have compared retinal histology to the ERG in light damage. Male albino CD rats (40) were purchased at 7 weeks of age and reared in 50lx cyclic light until 8 weeks old. They were exposed to a range of light intensities using white fluorescent light (1000, 1500, 2000, 2500 or 3000lx) for 24 or 48hr (n=5 per group). Controls remained in dim cyclic light. Seven days after exposure, dark and light adapted ERGs were recorded from threshold up to 200cdm−2using 50ms Ganzfeld white light stimuli. The STR, and scotopic and photopic b -wave thresholds and amplitudes were measured. After recording the ERG, the eyes were removed from the animals in each of the five 48hr light exposed groups and control group for histological measurements. These included: (1) outer nuclear layer width in rod photoreceptor cell number (cell count) and micrometers, and (2) outer+inner segment layer width along the vertical meridian in the inferior retina. The product of cell count and outer+inner segment length was calculated. All histological measures showed a statistically significant linear relationship to light exposure intensity (P<0.0001):r2=0.94 (cell count), 0.90 (outer nuclear layer width), 0.77 (outer+inner segment length). The log of the scotopic b -wave threshold and log amplitude showed a significant linear correlation to all histological parameters (P<0.0001) and there was no significant difference between b -wave threshold and amplitude for any one of the histology measures used. However, overall, log b -wave threshold was significantly better correlated to histology P<0.02. Only log b -wave amplitude showed a significant increase in variability in light damaged retinas (P<0.02).The b -wave threshold intensity increased 0.33logcdm−2and the maximum amplitude decreased 0.23logμ V with each 10% decrease in cell number in the outer nuclear layer. The sensitivity of the scotopic threshold response, which originates from third order neurons, changed much more slowly with cell loss, than did the b -wave (P<0.0005) and was well fit by a linear relationship to cell loss. The increase in photopicb -wave threshold was not significant for a cell loss of less than 70–80%. Neither the photopic or scotopicb -wave could be reliably recorded with more than 80% cell loss, but the scotopic threshold response remained. Both the scotopic and photopic ERG showed similar waveform changes near the threshold, including loss of the positive going b -wave and the predominance of a negative going response. Outer nuclear layer cell counts in this study showed the same relationship to log b -wave threshold elevation, as has been previously shown for whole retinal rhodopsin content in light damage, indicating that regional histology measurements can be good indicators of overall cell survival. Both the b -wave threshold and amplitude can be reliably used to track photoreceptor cell loss due to the damaging effects of constant light, but the scotopic threshold response may be more useful in severe damage.