1. Functionally different subgroups, each innervating a different part of the peripheral target, were defined within the ciliary population of the avian ciliary ganglion by electrical stimulation of the various ciliary nerve branches. 2. Although neurons innervating defined parts of the peripheral target consistently sent their axons through certain nerves, the technique of retrograde horseradish peroxidase (HRP) transport showed that the ganglion cell bodies were not spatially grouped but distributed throughout the ganglion, both before and after the period of naturally occurring cell death. However, such neurons tended to be clustered into groups of two or greater. 3. Ciliary and choroid populations, however, were found to be for the most spatially separate and recognizable by location and soma size before the period of cell death. Choroid cells did not project out the ciliary nerves even prior to the cell death period, confirming previous observations of selective axon outgrowth in the two populations. 4. Competition for survival was demonstrated within the ciliary population by experimentally removing approximately two-thirds of the neurons by axotomy-induced cell death at stage 32-34 just prior to the normal cell death period. This reduction in the number of competing neurons resulted in rescue of approximately 40% of the neurons that would have died, as assessed both by the number of axon profiles in the remaining intact nerve branch, as well as the number of somata that could be retrogradely labeled from this nerve. 5. It was concluded that many of the neurons that are normally removed during the cell death period are not destined to die, but can be rescued by reducing the number of neurons competing for a limited supply of some aspect of the peripheral target. Further, the postulated interaction with the target was shown to occur relatively late, just prior to the onset of cell death. 6. At the time of the peripheral interaction, the target was found to consist primarily of myoepithelial cells, which had migrated into the target region following the arrival of the ciliary axons. The target per se, therefore, cannot be involved in the selective growth of ciliary axons to the appropriate region. Well-defined synapses were rare, although many axonal endings were observed in close contact with both myoepithelial cells and the sparser differentiated muscle fibers, which increased to account for 60% of the target by the end of the cell death period. 7. Competition was also found to retard the rate of neuronal maturation because intact axons in the partially axotomized ganglion developed more rapidly than control axons, as assessed by axon diameter, conduction velocity, and degree of glial ensheathment. 8. Finally, at least some of the neurons in the partially axotomized ganglion expanded to innervate the peripheral territory of the axotomized branches, suggesting that competition between neurons is involved in the establishment of the observed peripheral innervation pattern.