The first and second neural projections of the insect eye

Abstract

1. The patterns of projection of some of the perpendicular neurones between the retina and medulla of the optic lobes of various insects have been studied. Axon paths have been studied from consecutive semi-thin plastic sections cut transversely and stained with toluidine blue. The termination positions and the paths of axons are both highly ordered and predictable. 2. In all insects with fused-rhabdomere eyes the axons of one ommatidium project to one cartridge of the lamina and the array of cartridges duplicates the array of ommatidia. In insects with open-rhabdomere eyes visual information is distributed amongst a number of lamina cartridges so that each cartridge receives information originating from one visual axis. 3. In both open- and fused-rhabdomere types the cartridge, array of the lamina is exactly duplicated in the medulla but by the intervention of the chiasma is reversed about a dorso-ventral plane. The axons of one lamina cartridge contribute to one medulla cartridge. Thus, in all insects studied, the visual field is projected exactly from retina to medulla. 4. Most of the retinula axons from one ommatidium terminate in the lamina but usually one pair passes directly to the medulla. These are from the central retinula cells (open-rhabdomere eyes) or from the small retinula cells (apposition type fused-rhabdomere eyes). Retinal responses are known mainly only for the short retinula axons so that visual information delivered to the medulla cartridge is still largely unresolved. 5. The lamina neuropile probably contains the elements responsible for the lateral correlation between parallel receptor inputs which is necessary for movement perception, but units with long lasting responses which could act as the delay circuit of movement perception are unknown. 6. The occurrence of errors in termination of the first projection of the optic lobe of the fly, which are reported for the first time in this work, provide no direct clues to the developmental processes by which such a morphologically complex system arises. Nevertheless errors may arise within a sequence of growth processes which are fundamentally quite simple and not obvious from knowledge of the generalized perfect pattern of connections.