RELYING heavily on the conclusions of Beer1, Walls2 and Brett3 assert that fishes are generally near-sighted. Although Beer found evidence by retinoscopy of far-sightedness in fishes, he assumed that retinoscopic reflexion occurs from the front retinal surface and thus falsely concluded that fishes are near sighted. By examining the eye through a +10 dioptre lens with a dissecting microscope equipped for vertical illumination, Baylor and Shaw4 found that the reflexion comes from the guanine layer behind the retina. On the other hand, far-sightedness in fishes was reported by Rochon-Duvigneaud5 and Verrier6,7. Moreover, Verrier7 found that on replacing the back of the retina with a screen, the image focused on the screen.