We determined the normal limits for various neuropathic tests in healthy subjects. The study, the Rochester Diabetic Neuropathy Study (RDNS), is noteworthy because of its size (more than 400 subjects), random selection of subjects, and selection of at least 15 men and 15 women without neuropathy, neurologic disease, or diseases predisposing to neuropathy from each hemidecade between 18 and 74 years of age from the population of a defined region (Rochester, MN). Subjects were classified into those with (nonhealthy subjects, RDNS-NS) and without (healthy subjects, RDNS-HS) neuropathy, neurologic or psychiatric disease, or diseases known to predispose to neuropathy. The study provides normal limits for tests used in the RDNS but it has broader uses as well. We found that (1) less than 10% of subjects in the third decade, approximately 20% in the fourth decade, and approximately 30% in the fifth or older decades were placed into the RDNS-NS category; (2) healthy subjects (RDNS-HS) retain their ability to walk on toes and heels regardless of age, excessive weight, or lack of physical fitness, but not their ability to arise from a kneeled position--lost in more than 5% of persons 60 years and older; (3) the frequency of decreased or absent ankle reflexes exceeds 5% in healthy subjects older than 50 years--limiting their value as a sign of diabetic polyneuropathy and necessitating a grading change with age in the neuropathy impairment score.(ABSTRACT TRUNCATED AT 250 WORDS)