Sanderson

 
"The majority of tumours are pigmented, and the first evidence of their presence is the development of an area of dark pigmentation in normal skin, or of a darker area in a melanocytic naevus. This change is often associated with a sensation, which may amount to itching or merely be an increased awareness of the presence of the lesion.
 
On previously normal skin the tumour may develop in one of two ways. It may become raised, erode the epidermis, or ulcerate and invade downward to produce a macular or slightly raised plaque . . ."
 
"Malignant melanoma of the tumid variety begins as a small, dark area on normal skin, or as a localized, but usually rapidly spreading, increase in pigmentation in a melanocytic naevus. The area may itch and bleed, as in the superficial plaque, and early ulceration is not uncommon. As the tumour enlarges it often appears to be erupting in mushroom fashion through the epidermis, and further growth leads to a fungating, or even pedunculated, lesion. A spreading and irregular-shaped halo of pigmentation is not infrequently found. A curious lightening of the central tumour sometimes occurs when the halo begins to spread. The pigmented halo may occasionally be surrounded by a further leucodermic halo, as in Sutton's naevus. Downward extension leads to deep induration and fixation to the underlying structures. Lymphatic dissemination locally can produce satellite lesions or radially extending streaks of pigmentation and, more distantly, enlargement of regional lymph nodes."
 
Sanderson KV. Tumours of the Skin. In: Rook A, Wilkinson DS, Ebling FJG, eds. Textbook of Dermatology. 2nd Edition. Great Britain: Blackwell Scientific Publications, 1972:1982–3.
 

Brief critique

 
It is true that the first evidence of melanoma usually is a locus of pigmentation, but the same is true for the appearance of such disparate lesions as freckles, solar lentigines, and Clark's nevi. Moreover, at first the pigmentation in melanoma is not "dark," and hardly ever is melanoma associated with any sensation, including itching. As a rule, by the time that the lesion is eroded or ulcerated, it no longer is curable; those signs are often a reflection of the thickness of the neoplasm and its compromising effects on the integrity of the epidermis. The notion that melanoma, at the outset, is "localized, but usually rapidly spreading" is incorrect; when still a macule or a patch, melanoma evolves very slowly. As a consequence, there is no such thing as "early ulceration" of melanoma; by the time that ulceration has occurred, the neoplasm is many years old and usually quite thick. The same is true of melanomas that appear "to be erupting in mushroom fashion," "fungating," or "pedunculated." What is described by Sanderson as a "spreading and irregular-shaped halo pigmentation" is not truly a "halo" like that seen in a halo nevus and in rare examples of melanoma, but a sign of regression of melanoma. Sanderson made that clear when he wrote that a "curious lightening of the central tumor sometimes occurs when the halo begins to spread." Last, "lymphatic dissemination" is an indication of metastasis of melanoma and death nearly inevitably from the effects of it.