Several methods have been used to diagnose diabetic polyneuropathy and to quantitate the degree of affection of peripheral nerves. Using a newly developed scoring system we compared bedside clinical examination with neurophysiological examination in a group of 78 diabetic patients. Individual scores for clinical examination were significantly correlated with scores for neurophysiological examination (r = 0.7, p < 0.0005). All 78 patients had at least one clinical symptom or sign of polyneuropathy. Clinical examination indicated polyneuropathy in three patients with neuropathic complaints, while neurophysiological examination in these patients showed no abnormalities. In 12 out of 14 patients with normal neurophysiological sensory nerve function, clinical examination showed at least one abnormal sensory modality. Comparing the four different sensory modalities, light touch sense and pinprick sense indicated polyneuropathy better than vibration or position senses. An abnormal Hoffmann reflex of the soleus muscle was always associated with a decreased or absent ankle jerk. The scoring system for the clinical examination proved useful for diagnosing and quantitating the severity of diabetic polyneuropathy. Clinical sensory deficits could not be inferred from the results of neurophysiological testing of sensory nerve function. Pinprick sense, light touch sense, and ankle jerks were the most important parameters in the clinical diagnosis of diabetic polyneuropathy.