The differential effects of exposure to a moderately hypoosmotic hyponatric solution (0.35 isoosmotic, Na+36 mmol/l) on conduction in myelinated (A) and unmyelinated (C) axons were studiedin vitroon compound action potentials of rabbit vagus nerves in which the perineurial sheath had remained undisturbed. Controls were incubated at 37° C in isoosmotic isonatric solution for 5 h (Group 1a, n = 7) or 7 h (Group 3, n = 3). Other controls were incubated in isoosmotic isonatric solution for 2 h followed by 3 h in isoosmotic hyponatric (Na+36 mmol/l) solution (Group 1b, n = 6); experimental nerves were incubated in isoosmotic isonatric solution for 2 h followed by 3 h in hypoosmotic hyponatric solution (Group 2, n = 7) and, to study recovery, a further 2 h in isoosmotic isonatric solution (Group 4, n = 8). In Group 1b, isoosmotic hyponatric exposure approximately doubled the latency of the A-component (A-CAP) and decreased the A-CAP amplitude to 44 ± 8% of control; the amplitude of the C-component decreased to 65 ± 15% of control. Hypoosmotic hyponatric exposure increased the latency of the A-CAP by 82 ± 10% (mean ± SE,P< 0.001) and extinguished A-CAP within 20 min, whereas the latency increase of the C-component (C-CAP) was more than twice as great and extinction slower and often incomplete; neural wet weight increased 34 ± 4% and neural sodium and potassium contents decreased 55 and 42%, respectively. Recovery in isoosmotic isonatric solution (Group 4) was absent or very small in the case of A-CAP, as regards latency and amplitude but was complete for C-CAP amplitude. Neural wet weights and sodium content also recovered fully, but neural potassium content recovered only about 45%. Electron microscopy revealed hypoosmotic hyponatric structural damage to the larger myelin sheaths; the axons themselves were unaffected. It is concluded that it is probably inadvisable to attempt selective conduction block of sensory C-fibers by application of hypoosmotic solutions to peripheral nerves.