Abstract Electron microscope studies of peripheral nerve fibres, made after the application of various histochemical tests, demonstrate that the latter probably cause changes in the myelin; however, only in the case of Baker's phospholipid reaction, and as a result of the turpentine extraction of osmicated nerve do these techniques cause the type of change which would be reflected under the light microscope as a radial structure. Enzyme histochemical techniques do this with only a few fibres. It is suggested that in the radial bands described as “neurokeratin network” the protein is largely enzyme protein and that it forms a complex with the phospholipid which resists disruption. It is possible that regional localization of polysaccharide has this effect. This hypothesis reconciles the results of histochemical and electron microscope studies. There is no evidence that lipid solvents remove the lipid fraction of myelin leaving the protein behind as a sort of skeleton nor that the material biochemists have isolated and which they call “neurokeratin” has anything at all to do with what histologists have, for 80 years, called the “neurokeratin network” of peripheral nerves. Lipid solvents cause a complete collapse of the lipo-protein complex in radial segments of the myelin sheath, leaving areas of more or less intact myelin between the collapsed areas.