Recent studies have suggested that the presence of DNA aneuploidy in stage I cutaneous melanoma carries a poor prognosis. To see if our experience correlated with these reports, we used DNA analysis by flow cytometry of propidium iodide-stained nuclei disaggregated from formalin-fixed paraffin-embedded tissue of biopsy specimens to retrospectively study 55 patients who had cutaneous stage I melanomas. The patients had been treated from 1977 to 1987 with a mean follow-up of 5.4 years. Thirty-nine (71%) of the 55 histograms were diploid, and 16 (29%) of the histograms were aneuploid. DNA content was significantly associated with other conventional prognostic factors, including growth pattern, ulceration, pathologic stage, tumor thickness, and Clark's level. DNA aneuploidy was significantly related to disease-free survival and predicted a poorer prognosis (p less than 0.05), but when stratified for tumor thickness it lost significance. A multivariate discriminant function analysis of 12 factors in melanoma showed six factors to be independently significant in determining prognosis. DNA content (p = 0.034) ranked fifth in importance behind growth pattern (p less than 0.001), ulceration (p less than 0.001), thickness (p = 0.001), and pathologic stage (p less than 0.005). DNA content, although significantly associated with conventional prognostic factors and disease-free survival, is not the best indicator of biologic behavior of melanomas in this study. Further investigation into its usefulness is necessary before DNA content can become a routine diagnostic modality in the work-up of stage I cutaneous melanomas.