Dimpling of the skin of the thighs and buttocks is commonly referred to as cellulite, and it afflicts women much more frequently than men. Whereas many therapies that presume cellulite is caused by an abnormality of adipose tissue have gained recent popularity, the basic pathophysiology of cellulite has not been clearly identified. Theoretically, cellulite could reflect differences in adipose tissue biochemistry or connective tissue structure of affected versus unaffected individuals and/or of affected versus unaffected regions within an individual. We report here on direct experimental examination of these possibilities. Seven healthy adult subjects (five women, two men; four affected, three unaffected) underwent sonography of the thigh, measurement of regional in vivo subcutaneous adipose tissue metabolism (catecholaminergic responsiveness and blood flow) by microdialysis probe studies of the abdomen and the thigh, and full-thickness wedge biopsy of the thigh under local anesthesia. The presence of cellulite was defined as evidence of dimpling of the skin of the posterolateral thigh when the subject stood with the affected leg flexed to 90 degrees at the hip and knee. Any continuous area of skin at least 3 cm in diameter in which no dimpling was evident was designated as "unaffected." In all affected individuals, studies were performed to include both affected and unaffected areas of the thigh. In vitro pathologic examination of wedge biopsies and in vivo sonographic examination of the thigh both showed a diffuse pattern of extrusion of underlying adipose tissue into the reticular dermis in affected, but not unaffected, individuals. In vitro and in vivo studies also demonstrated that women had a diffuse pattern of irregular and discontinuous connective tissue immediately below the dermis, but this same layer of connective tissue was smooth and continuous in men. This connective tissue layer was more irregular and discontinuous in affected versus unaffected individuals. No significant differences were noted in subcutaneous adipose tissue morphology, lipolytic responsiveness, or regional blood flow between affected and unaffected sites within individuals. There is a sexual dimorphism in the structural characteristics of subdermal connective tissue that predisposes women to develop the irregular extrusion of adipose tissue into the dermis, which characterizes cellulite. These gender-related differences are diffuse and not localized only to affected areas. There is no evidence of any primary role for adipose tissue physiology, blood flow, or biochemistry in the etiology of cellulite, although the connective tissue of the female thigh is structured to accentuate differences in small subdermal adipose tissue depots.