The development of the adult visual system of Drosophila requires the establishment of precise retinotopic connections between retinal photoreceptor cell axons and their synaptic partners in the optic lobe of the brain. To assess the role of axon-axon interactions in retinal axon guidance, we used genetic methods to disrupt the normal spatiotemporal order of retinal axon ingrowth. We examined retinal axon projections to the developing first optic ganglion, the lamina, in two mutants in which reduced numbers of ommatidia develop in the eye imaginal disk. We find that in the developing lamina of these mutants, sine oculis and Ellipse, retinal axons project to proper dorsoventral positions despite the absence of the usual array of neighboring retinal axons. In a second approach, we examined animals that were somatic mosaics for the mutation, glass. In glass- animals, retinal axons project aberrantly and the larval optic nerve is absent. We find that in the developing lamina of glass mosaic animals, wild-type retinal axons project to proper dorsoventral positions despite the misrouted projections of neighboring glass- retinal axons. In addition, wild-type retinal axons project normally in the absence of the larval optic nerve, indicating that the latter is not an essential pioneer for retinal axon navigation. Our observations support the proposal that axon fascicles can make at least some pathfinding decisions independently of other retinal axon fascicles. We suggest that positional guidance cues that might label axon pathways and target destinations contribute to retinotopic pattern formation in the Drosophila visual system.