Ipsilateral retinal projections (IRP) in the optic chiasm (OC) vary considerably. Most animal groups possess laterally situated eyes and no or few IRP, but, e.g., cats and primates have frontal eyes and high proportions of IRP. The traditional hypothesis that bifocal vision developed to enable predation or to increase perception in restricted light conditions applies mainly to mammals. The eye-forelimb (EF) hypothesis presented here suggests that the reception of visual feedback of limb movements in the limb steering cerebral hemisphere was the fundamental mechanism behind the OC evolution. In other words, that evolutionary change in the OC was necessary to preserve hemispheric autonomy. In the majority of vertebrates, motor processing, tactile, proprioceptive, and visual information involved in steering the hand (limb, paw, fin) is primarily received only in the contralateral hemisphere, while multisensory information from the ipsilateral limb is minimal. Since the involved motor nuclei, somatosensory areas, and vision neurons are situated in same hemisphere, the neuronal pathways involved will be relatively short, optimizing the size of the brain. That would not have been possible without, evolutionary modifications of IRP. Multiple axon-guidance genes, which determine whether axons will cross the midline or not, have shaped the OC anatomy. Evolutionary change in the OC seems to be key to preserving hemispheric autonomy when the body and eye evolve to fit new ecological niches. The EF hypothesis may explain the low proportion of IRP in birds, reptiles, and most fishes; the relatively high proportions of IRP in limbless vertebrates; high proportions of IRP in arboreal, in contrast to ground-dwelling, marsupials; the lack of IRP in dolphins; abundant IRP in primates and most predatory mammals, and why IRP emanate exclusively from the temporal retina. The EF hypothesis seams applicable to vertebrates in general and hence more parsimonious than traditional hypotheses.