Metastasis from thin melanomas is rare and unpredictable. In order to assess the prognostic value of the proliferation marker, MIB-1, immunohistochemical staining was evaluated in a retrospective case-control study of 11 thin melanomas with documented metastasis and 11 control tumors that failed to metastasize. Tumors selected were < 1-mm thick and were individually matched for tumor thickness, date of excision, and patient age and sex. Analysis of MIB-1 expression as both a mean and a maximum level for the case and control groups revealed no association with metastasis. Wilcoxon's matched-pairs signed-rank test had p-values of 0.45 for the maximum values and 0.79 for the mean values. For the 11 thin melanomas that metastasized, there was a weak, yet statistically insignificant, correlation between the proportion of cells positive for MIB-1 and the length of the relapse-free period [Spearman's correlation coefficient = 0.20 for the maximum level (p = 0.56) and 0.19 for the mean level (p = 0.58)]. These results suggest that MIB-1 expression may be of limited value as a prognostic marker for increased risk of metastasis in patients with thin melanomas. MIB-1 immunohistochemistry was also performed on 25 benign and 70 malignant paraffin-embedded melanocytic tumors to evaluate the level of MIB-1 expression at different stages of tumor progression. A progressive increase in MIB-1 expression was seen from benign tumors through to primary melanomas, with the highest level seen in metastatic melanomas. Within the group of primary melanomas, the MIB-1 score was shown to correlate significantly with tumor thickness and Clark's level of invasion (Spearman's correlation coefficient = 0.71 for level and 0.77 for thickness).