classification of Clark et al, 1 this case suggests that we should keep verrucous-keratotic malignant melanoma in mind as an important differential diagnosis for verrucous, hyperkeratotic lesions. In this case, the subtle perilesional clinical appearance of dark infiltration at the ridges might have been a clue to the diagnosis of melanoma.The intercellular adhesion between melanoma cells is much weaker than that between normal keratinocytes, because melanocytes or melanoma cells do not have desmosomes. In addition, membranous expression of E-cadherin, which is responsible for melanocyte-keratinocyte adhesion, is sometimes decreased in melanoma progression. 2 Thinning of the epidermis with attenuation of the basal and suprabasal layers, described as ''consumption of the epidermis,'' is frequently seen in areas of direct contact between the epidermis and melanoma cell nests. 3 Consumption of the epidermis occasionally results in the formation of a cleft separating the thin epidermis and melanoma cell nests. 3 Braun-Falco et al 4 reported that the clefts are a reliable diagnostic criterion for malignant melanoma. It has been suggested that the greater the Breslow depth, the greater the consumption of the epidermis in a melanoma lesion. 5 However, the Breslow depth of the current lesion did not appear to be very great, although we were unable to measure it accurately because of the large size of the cleft. The presence of the verrucous architecture may be associated with the presence of epidermal consumption. In the current case, melanoma cells completely replaced the basal layer keratinocytes, forming only a single-layer lining on the basement membrane. This characteristic infiltration pattern of melanoma cells with fragile cell adhesion might have lead to the large cleft with features mimicking the tombstone appearance of pemphigus vulgaris, although in the current lesion, the tombstones were melanoma cells instead of the basal keratinocyte tombstones of pemphigus vulgaris lesions.