1. Several anterior segments of the lumbosacral spinal cord were deleted in a series of St. 15-16 chick embryos, which is prior to the birthdate of the motoneurones and to limb bud formation. Deleted segments did not regenerate and therefore some muscles were deprived of their normal source of innervation. 2. The projection pattern of motoneurones in remaining lumbosacral segments was assessed electrophsiologically (St. 30-36) and by orthograde and retrograde labelling of motoneurones with horseradish peroxidase (St. 24-36) from the time that axons first enter the limb until after the normal motoneurone cell death period. 3. The projection pattern of remaining segments was not altered by the deletion. At no time did inappropriate segments innervate muscles or regions of the primary muscle mass which had had their enter innervation source removed by the deletion, although they often innervated immediately adjacent regions. 4. The pathways taken by remaining motoneurones both in the plexus and muscle nerves were not different from their normal control pattern. From the time that axons first reached the base of the limb, there was no evidence that they compensated for missing spinal nerves by projecting down inappropriate pathways. 5. With respect to pathway selection, we can conclude that axons are normally not excluded from certain pathways by interactions with other axons, nor do they interact amongst each other in distributing themselves to all available pathways. 6. We can also conclude that once axons enter a muscle or muscle mass, they are not kept within appropriate regions by competition with axons that normally occupy adjacent regions. In summary we can exclude both the process of competition and a timed outgrowth mechanism as playing a major role in the development of specific motor connexions in the chick hind limb. 7. Muscles totally deprived of their innervation by the deletion underwent cleavage, became contractile and appeared to differentiate normally until St. 20. However by St. 35 they had become extremely atrophic. 8. In some cases following the deletion of part of a motoneurone pool, the number of motoneurones per unit of cord in the remaining pool at St. 35-36 was significantly greater than control values. Since these motoneurones were located in the position (both rostro-caudal and transverse) of the control motoneurone pool, we conclude that we were able to rescue some of the neurones that would normally have died during the motoneurone cell death period (St. 30-35).