This study has examined the structural features and distribution of 'doubly myelinated' axons in normal adult and aged mice. Investigation focused on the superior cervical ganglion (SCG) and paravertebral sympathetic ganglia, which were extensively serial-sectioned for light and electron microscopy. In the SCG, the principal features of doubly myelinated regions were that an apparently normal myelinated axon was enclosed for part of its length by an additional (outer) myelinating Schwann cell. The separate nature of the inner and outer Schwann cells was emphasized by the consistent presence of individual nuclei in each, and by the presence of endoneurial space, often containing collagen fibrils, between the inner and outer cells. In some cases more than a single outer Schwann cell was present, arranged serially along the inner myelinated fibre. While double myelination forms through a mechanism involving displacement of an original myelinating Schwann cell by an interposed Schwann cell (see companion paper), we here provide evidence that in some instances the outer Schwann cell fails to retain any direct axonal contact, either with the axon centrally enclosed within the configuration or with any neighbouring axon. In contrast to the rat, delicate cytoplasmic processes often extended from the lateral extremes of outer Schwann cells. However, again no evidence for axonal contact was found, and similar processes also extended from the paranodal region of some singly myelinated non-displaced Schwann cells. Without exception the outer myelin sheath remained structurally intact, and characteristically underwent a series of conformational changes (progressive infolding of the paranodes and new areas of myelin compaction) which infer a continuing capacity of the outer Schwann cell to translocate myelin-specific components in a co-ordinated manner. A basal lamina was always present on the 'abaxonal' plasma membrane of the outer cell, but not on the 'adaxonal' surface except in areas involved in infolding, thus retaining the polarity which existed at the time of displacement from the axon. At single cross-sectional levels through the SCG, up to approximately 4% of myelinated axons were involved in double myelination. Double myelination was not detected in the sciatic nerve or in the paravertebral ganglia, thus indicating a predilection for the SCG as a site of development of these configurations. Though not challenging the role of the axon in initiating the formation of myelin, these data indicate that in this tissue myelin maintenance does not require direct contact between axonal and Schwann cell plasma membranes.