The value of DNA single-cell cytometry for the detection of neoplasia in Feulgen-stained cerebrospinal fluid cytological specimens was tested on 34 cases of Non-Hodgkin's lymphoma or leukemia and on 66 cases of viral or bacterial meningitis as a disease control group. The DNA content of 200 randomly chosen nuclei was measured on one pre-existing, cytologically representative slide per case, using a TV-image analysis system TAS-plus (Leitz, Germany). Neoplasia was diagnosed, if at least three nuclei with a DNA content above 5c (5cEE > or = 3) were found. The sensitivity investigating only one slide per case was 79.4% (27/34), the specificity 78.8% (52/66). Three lymphomas and 7 inflammatory cases were classified as suspicious (0 < 5cEE < 3). In 4 lymphoma cases (11.8%) a false-negative diagnosis and in 7 cases (10.6%) of viral meningitis a false-positive diagnosis were made. No false-positive diagnosis occurred in bacterial meningitis. While the false-negative diagnoses may be due to the only slightly increased number of cells in cerebrospinal fluid, no final explanation for increased DNA values after viral infection can be given. Therefore, before using DNA single-cell cytometry to prove the malignant character of lymphocytic pleocytosis in cerebrospinal fluid, viral meningitis has to be clinically excluded.