We examined the feasibility of performing non-radioactive in situ hybridization (ISH) in flow cytometrically sorted (tumor) cells with a chromosome #1 specific centromere probe. The study was performed in a model system of HL60 cells mixed with different quantities of HeLa cells. These latter cells were sorted directly onto poly-l-lysine coated glass slides on the basis of their keratin content, a cytoskeletal component not present in HL60 cells. Overall morphology of the separated HeLa cells was excellent and, after the ISH procedure, the appropriate number of ISH spots was observed in more than 85% of the sorted cells. This percentage did not differ significantly in cell mixtures with different percentages of HeLa cells (down to 1%). Sorting of HeLa cells in different phases of the cell cycle, and subsequent ISH, revealed the same spot number for chromosome #1 in all cell cycle stages, including mitosis. In the latter phase of the cell cycle we did not find a duplication of the chromosome #1 centromere, not even after sorting of the mitotic cells on the basis of specific labeling with an antibody to mitotin. The early G2 mitotin negative fraction, however, showed a significant percentage of cells with a duplicate spot number, most likely representing a tetraploid cell fraction in this HeLa cell culture. The protocol that evolved from these model studies was applied to cell suspensions of malignant body cavity effusions as well as solid bladder carcinomas. In several of these cases numerical chromosome aberrations could be detected by ISH more evidently after sorting on the basis of keratin labeling.