It has been thought that adipocytes lack proliferative ability and do not revert to precursor cells. However, numerous findings that challenge this notion have also been reported. The idea that adipocytes dedifferentiate to fibroblast-like cells with increasing cell number was reported in 1975. This possibility has been ignored despite knowledge gained in the 1990s regarding adipocyte differentiation. Several studies on proliferation and dedifferentiation of adipocytes have been published, most of which were conducted from the perspective of regenerative medicine. However, the concept of proliferation of adipocytes remains unclear. In this study, we postulate a new population of adipocytes, which consist of small sized cells (less than 20 μm in diameter) expressing adipocyte markers, such as adiponectin and peroxisome proliferator-activated receptor γ (PPARγ), but not possessing large lipid droplets. These cells show marked ability to incorporate 5-bromo-2'-deoxyuridine (BrdU), for which reason we termed them "small proliferative adipocytes (SPA)". In addition, SPA are observed in the stromal vascular fraction. Since SPA are morphologically different from mature adipocytes, we regarded them as committed progenitor cells. When proliferation of adipocytes in vivo is assessed by measuring BrdU incorporation and expression levels of proliferating cell nuclear antigen (PCNA) in isolated fractions of adipocytes from adipose tissues, subcutaneous SPA proliferate less actively than visceral SPA. Treatment with pioglitazone increases the number of proliferating SPA in subcutaneous, but not visceral, fat, suggesting that SPA may be important in regulating systemic insulin sensitivity and glucose metabolism.