Because of a lack of suitable archival material, it is rarely possible to make retrospective studies of the correlation between the prognosis for a patient with mammary carcinoma and the distribution of nuclear DNA in the cells of the neoplasm. An investigation of the possibility of using sections cut from paraffin-embedded specimens showed that such sections are not suitable for use in retrospective studies of breast carcinoma. Because of such factors as the heterogeneity in size and shape of the nuclei from neoplastic cells and their tendency to mold around each other, determinations of DNA content of cells in sections were extremely difficult; in this particular carcinoma it was found that the distribution of nuclear DNA as obtained from a Feulgen-stained histologic section was not the same as that obtained from a Feulgen-stained imprint smear, and some polyploid tumors were erroneously classified as aneuploid.