Pulmonary embolism secondary to deep vein thrombosis is a frequent cause of death in stroke patients. In a multicentre study of deep vein thrombosis prophylaxis, 112 patients with cerebral infarction and leg paresis were given aspirin 300 mg three times a day (t.d.s.) alone or with dipyridamole 100 mg t.d.s. To screen for deep vein thrombosis liquid crystal thermography of the legs was performed daily for 15 days on all patients. Those patients with positive thermography underwent immediate X-ray venography of the appropriate limb as the definitive investigation for venous thrombosis. Twenty-two patients had positive thermograms, of whom 16 had confirmed deep vein thrombosis as demonstrated by X-ray venography. Only 8 of the 22 had clinical signs of deep vein thrombosis and 2 of those had a negative venogram. Of the 14 patients with positive thermography but negative clinical signs 10 had positive venograms. Difference in the incidence of deep vein thrombosis in the two treatment groups was not demonstrated. It is concluded that occult deep venous thrombosis is common after ischaemic stroke and it can occur without clinical signs. Liquid crystal thermography is a simple, rapid and cheap screening test that will allow the detection of clinically unrecognized thrombosis.