The recognition of dysplastic nevi from photographs can aid in population surveys of nevi and in epidemiological studies of melanoma risk. The reproducibility of techniques for recognizing nevi as dysplastic or for scoring them according to the degree of dysplasia has not been measured. Using photographs of 300 nevi taken in the course of a case-control study of melanoma, we assessed the agreement among six clinicians in independently categorizing nevi as dysplastic and in grading the degree of dysplasia. On average, reviewers agreed with each other 77% of the time in classifying a nevus as dysplastic or normal. Pairwise agreement within one point on a six-point scale occurred 87% of the time on average. These results suggest that criteria for recognizing nevi as clinically dysplastic from photographs can be applied reproducibility.