Some of the myelinated axons in rat peripheral nerves possess marked axoplasmic carbonic anhydrase activity [Riley, Ellis and Bain (1982) J. Histochem. Cytochem. 30, 1275-1288; Riley and Lang (1984) J. Hand Surg. 9A, 112-120]. A mixture of reactive and nonreactive neurons was a general observation in cervical, thoracic and lumbar ganglia. Nonmyelinated axons in lumbar dorsal roots were nonreactive; this was consistent with the lack of carbonic anhydrase in small sensory neurons. The carbonic anhydrase cytochemical method marked the larger afferent or sensory neurons and distinguished them from the smaller sensory neurons which were devoid of carbonic anhydrase activity. Nonmyelinated axons in the lumbar ventral roots were also nonreactive. Examination of muscle spindle innervation revealed staining of the primary sensory and gamma motor endings. This was strongly suggestive that some of the reactive sensory neurons were primary afferents and a portion of the reactive ventral root axons were gamma motor. The reactive central processes of spinal neurons sent collaterals into the grey matter of the spinal cord, entered the dorsal funiculi, and terminated in synaptic glomeruli in the cuneate and gracilis nuclei. Oligodendroglial cells appeared to be the only intrinsic cellular elements of the brain stem and spinal cord that exhibited high carbonic anhydrase activity. Both oligodendroglial and Schwann cells exhibited intense carbonic anhydrase activity in thin pockets of cytoplasm internal to compact myelin. The subcellular distribution of reaction product within sensory neurons and oligodendroglial cells agreed with biochemical reports of cytosol and membrane-bound forms of carbonic anhydrase. A general staining of the cytoplasm was suggestive of soluble carbonic anhydrase fixed in situ by the glutaraldehyde. Clumps of reaction product on the cytoplasmic surface of the endoplasmic reticulum possibly represented membrane-bound enzyme. Most of the membrane-bound carbonic anhydrase was associated with the internal membranes rather than the axolemma or limiting plasma membrane of the axon. In contrast to biochemical reports, a small fraction of neuronal mitochondria exhibited staining in the intracristal spaces. We suggest that the association of carbonic anhydrase with endoplasmic reticulum and mitochondria implicates the enzyme in regulating intracellular calcium because both organelles are known to sequester calcium.