Abstract
To the Editor: Judzewitsch et al. (Jan. 20 issue)1 documented small changes in nerve conduction velocity during sorbinil therapy of diabetic patients, but the techniques they used were antiquated and relatively inaccurate; variations of several meters per second (at least 5 per cent) of the mean nerve conduction velocity are found when the same normal subjects are studied repeatedly.2 Most errors are spatial, arising from difficulty in knowing where on the nerve action potentials occur (information only crudely determined by the position of stimulating electrodes on the skin), measuring the true conduction distance, and reproducibly positioning recording electrodes relative to.