Reanastomosis of the injured peripheral nerve, in clinical reality, can never bring the severed ends together in a way to coapt fasciculi, let alone axons. However, some good recovery is occasionally met even in unsatisfactory anastomosis. In this regard, attempts were conducted to study the three dimensional structures of myelinated axons of the normal peripheral nerve with special reference to the possible existence of axonal anastomosis. Seven-hundred-and-fifty thin sections of the epoxy resin embedded sciatic nerve of Sprague-Dawley rat were prepared in 0.7 um thickness. Three dimensional reconstruction displayed patterns of branching directing proximally and floating segmentally myelinated fibers. In separate study, the histomorphologic changes of the distal segment of the injured nerve and regenerating patterns were severed under operating microscope. Thirty nerves were left unrepaired for control group, and remaining 30 nerves were anastomosed with 10-0 nylon for repair group. Each 10 specimens were prepared for light and electron microscopy at 1, 2 and 3 months after anastomosis. In control groups at 1, 2, and 3 months, myelin tubules disappeared and degeneration went on. In repair groups, some loose scaffolds formed by undigested basal laminae, fibroblasts, collagen fibrils were observed, encircling regenerating fasciculus. Some regener tion with myelination was observed within scaffolds of basal laminae, fibroblast foot processes, collagen fibrils. These findings suggest that loose scaffolds in the distal segment of the injured nerve might serve as a guide to the nerve ending in the random nerve regeneration directing distally. The number and size of regenerating axon vary in 1, 2, and 3 months. In second months, the number of regenerating axons increase more than that of first month group and normal. In the third months, it decrease and it looks like normal. This means that some regenerating axon continue mature and others are atropied. So we suggest two possible mechanisms of the target specific regeneration. First, there may be a communicating axon which connect to the nearest axon through the communicating branch. This communicating axon can transport the signal to the specific distal axon. Second, injured proximal axons may sprout multiply and be connected multiply to the severed distal axons, and correctively connected axon may mature and aberrantly connected axon may atropied.