Ninety patients who fulfilled the ARA criteria for systemic sclerosis were divided into 6 groups on the basis of cutaneous areas of sclerosis. Anticentromere antibody (ACA) was found in 14 of 28 patients of the first 2 groups, the first group being pure sclerodactyly, and the second sclerodactyly plus minimal sclerotic lesions (eyelids, neck, armpits). In the other 4 groups (62 patients) in whom skin sclerosis was more diffuse, ACA was found in one case. Fluorescent antinuclear antibodies (FANA) with a diffusely grainy pattern, those with a nucleolar pattern and the anti-Scl-70 antibody were present in all 6 groups, but were significantly more frequent in the last 4 groups than in Groups 1 and 2. The cumulative survival rate as well as disease duration were found to be significantly longer in the first 2 groups designated "limited cutaneous systemic sclerosis," than in the 5th and 6th groups, i.e., patients in whom the trunk also was involved: "diffuse cutaneous systemic sclerosis." The cumulative survival rate and disease duration of patients with skin sclerosis of the face and limbs, but not of trunk, were not significantly different either from those of patients with limited cutaneous systemic sclerosis or from those of patients with diffuse cutaneous systemic sclerosis. Thus these patients show the same antibody pattern as diffuse cutaneous systemic sclerosis, but the prognosis not significantly different from the other 2 subsets. We suggest they constitute a distinct subset of systemic sclerosis for which we propose the name "intermediate cutaneous systemic sclerosis."