From a prospective study of 552 Raynaud's phenomena, the authors have analyzed the relationships between digital topography of the attacks, etiologic diagnosis and type of hand vascularization. Five topographic subsets were differentiated: median (one of the three median fingers), ulnar (the 2 or 3 last fingers), radial (the 2 or 3 first fingers), all fingers excluding thumb, all fingers including thumb. The respective role of radial and ulnar arteries in hand vascularization was assessed by Allen's test. It has been found 33% Raynaud's phenomena with involvement of all fingers excluding thumb, 31% with median topography, 29% with involvement of all fingers including thumb and 7% with ulnar or radial topography. Unilaterally (9% Raynaud's phenomena) preferentially points out a syndrome with neurological origin essentially (ulnar or carpal tunnels and thoracic outlet syndromes). Factorial Analysis of Correspondences has been used to study the topographic repartition of the attacks according to etiology: homogeneous involvement of all fingers including severe involvement of thumb has to evoke the diagnosis of scleroderma, whereas involvement of the median fingers, of all fingers excluding thumb or with a mild involvement of thumb is still consistent with primary Raynaud's disease. Strictly lateral topography (ulnar, radial) suggests the diagnosis of Raynaud's syndrome of neurological cause essentially. It has not been found any correlation between digital topography of Raynaud's phenomenon and type of hand vascularization; this can probably be explained by the clear preponderance of neurological etiologies as compared to vascular ones.(ABSTRACT TRUNCATED AT 250 WORDS)