On the specificity of a histochemical method for localizing dihydrofolate dehydrogenase