By combining both immunocytochemical and functional investigations, a hypothetical framework will be developed for the molecular mechanisms underlying neuron-glia interactions during development and regeneration of peripheral nerves. In particular, the immunoglobulin-like molecules L1, N-CAM, MAG and P0, the extracellular matrix molecules laminin and tenascin, and the carbohydrates PSA and L2/HNK-1 will be considered. During early stages of limb bud innervation in embryos, L1 and N-CAM are expressed on axons and Schwann cells and are involved in axonal fasciculation, whereas tenascin is thought to be involved in forming a scaffold around the nerve possibly preventing axons and/or Schwann cells from leaving the nerve. PSA has been shown to be involved in pathway selection at initial stages of limb bud innervation. Later on, when motor axons enter muscles, the carbohydrates determine the branching pattern of the nerves. During myelination, L1 appears to play a pivotal role during the formation of the first Schwann cell loops around the prospective myelin-containing axons. MAG and P0 appear also to be functionally involved at initial stages of myelin formation. Additionally, MAG may contribute to the formation and maintenance of non-compacted myelin and axon-Schwann cell apposition whereas P0 is involved in myelin compaction. Under regenerative conditions, L1, N-CAM, laminin, and tenascin are strongly up-regulated by denervated Schwann cells. In vitro observations strongly suggest that these molecules might foster axonal regeneration. The carbohydrate PSA is confined to regrowing axons and is also a candidate to support axonal regrowth. L2/HNK-1, which is found on motor axon-associated Schwann cells, may provide regenerating motor axons with a selective advantage over others resulting in appropriate reinnervation of motor pathways. Since many of the functional studies this review refers to have been performed in vitro, some of the conclusions drawn need reexamination in vivo. Gene manipulations, such as the generation of null mutants followed by a thorough morphological and immunocytochemical investigation may be a powerful tool to resolve this problem.