Because skin lesions are visible and easily accessible, skin cancers provide us with an excellent in vivo model to study the development of cancers. Cutaneous malignant melanoma and squamous cell carcinoma (SCC) both arise from the epidermis and have an initial progression stage in which proliferation of the neoplastic cells is confined to the epidermis. This stage is called melanoma in situ or SCC in situ. Molecular analyses of melanoma in situ and of solar keratosis, a prototype of early SCC in situ, show that loss of p16(INK4a)/p14(ARF) and dysfunction of p53 play a critical role, respectively. Furthermore, there seems to be potential precursor cells to these in situ lesions, which are not discernible with conventional hematoxylin and eosin-stained sections. The precursor cells have minimal but critical genetic alterations, such as cyclin D1 amplification and p53 mutation, and can be identified using fluorescent in situ hybridization and immunostaining with p53 antibodies, respectively. These precursor cells may be defective in repair response to DNA damage, and would have proliferative or survival advantages over their normal neighboring counterparts in the presence of growth factor stimulation or genotoxic events, such as ultraviolet irradiation. Such precursor clones may be induced at a rather young age, and their number and size increase with accumulating carcinogenic stimuli. If these lesions acquire additional mutations, they could progress to clinically visible lesions of in situ carcinoma. Precise molecular analyses of early stages of skin cancers may have a strong impact on our understanding of in vivo development of cancers in other human organs.