spectra. Thus, energy distribution of the light source would have to be defined and equated to absorption spectra of the photore ceptors. Laplacian transforms would be extremely complex and would need considerable simplification before being solvable. We have assumed that the stimulus was uniformly distributed over the retina and was stationary; for nonuniform spatial organiza tions and moving stimuli, mathematical analysis would be ex tremely complex, if not impossible, and we still have gotten past the photoreceptors. If other features such as stray light, electrical coupling of cones, and HCs and possibly ACs, as well as negative feedback are added, problems are further compounded, which leads one to the conclusion that analog modeling, even though it requires considerable development of hardware, nevertheless, in the long run, affords an accessible method for modeling a realistic retina. ACKNOWLEDGMENT