A Q-switched ruby laser was used to produce retinal lesions in rabbits and monkeys. The data were obtained with special attention given to eliminating inhomogeneities from the portion of the laser beam entering the eye; correcting each eye for refractive error to produce minimal retinal impact area; estimating the retinal area impacted; monitoring of the laser output in space, time, and amplitude; and to detailed clinical and pathological observations. Eight microjoules in pigmented rabbits and 22μj in rhesus monkeys for 8-nano-second pulses gave a 50% probability of producing a clinically observable lesion. Histologically, lesions were seen at approximately half this level. Energy density or power density at the retina is the determining factor for threshold damage. Histologically, glial reaction was conspicuously deficient. However, two instances of local pigmentcell hyperplasia were observed.