The authors acknowledge that application of "their diagnostic guidelines . . . will result in misdiagnosis in a significant percentage of patients," an admission that leads them to suggest "it is better to err in the direction of over-diagnosis." But morphologic diagnosis should be predicated on criteria that are repeatable and reliable, that is, there should be neither "under-diagnosis," nor "over-diagnosis," but accurate diagnosis.

Dobson and Abele state correctly that the criteria they set forth for "clinical features that suggest a malignant melanoma" do not work and some reasons why are as follows : 1) All melanomas begin much smaller than 1 cm in diameter and the optimal time to diagnose melanoma is when it is small, flat, and curable; 2) Most melanomas do not "contain shades of red, white, gray, blue, brown, and black." 3) "Irregular" as a description of a border is inadequate not only because "regular" in this sense has never been defined, but some Clark's nevi also show "angular indentation and notching." 4) The criticism just made of "irregular" for attributes of a border applies equally to "irregular"characteristics of a surface, and, moreover, many small congenital nevi have an uneven surface. 5) "Increase in size," cannot be determined by morphological observation. 6) "Change in color," like "increase in size" usually is learned from history. 7) Ulceration never is seen in flat and slightly elevated lesions. By the time ulceration has occurred, a melanoma usually is thick and has metastasized, and, for that reason, the authors are right to state that ulceration (and inflammation) are "late changes."