In vitro models are required for the study of these cancers, and several cell lines have already been established and characterised (Lafargues & Ozzello, 1958;Soule et al., 1973;Cailleau et al., 1974;Engel et al., 1978;Whitehead et al., 1983;Yamane et al., 1984;Chu et al., 1985;Vandewalle et al., 1987). Nevertheless, owing to the heterogenity and the diversity of mammary cancers, a great number of cell models is necessary to understand the reasons for this diversity and the effect of anticancer drugs on tumour cells.Cytogenetic studies of mammary adenocarcinoma cell lines are essential for comprehension of the pathogenesis of these cancers (Trent, 1985;Gebhart et al., 1986). The implication of chromosomal alterations in these pathologies has opened a new and promising route towards better knowledge of these cancers (Cervenka & Koulischer, 1973). Chromosomal alterations are generally numerous, and markers often demonstrate hyperploidy in these cancers (Sandberg, 1980). Demonstration of the minimum genetic alterations indispensable for cell transformation is difficult, and might be easier on cells with a karyotype closer to normal. Sandberg and Wolman mentioned the existence of such cells, but most of their results concerned karyotype studies without chromosome banding (Sandberg, 1980;Wolman, 1983 [-', hyaluronidase 25 IU for 20 ml, Hepes buffer 4.8 g 1' in distilled water). Cells were then fixed in acetic acid:methanol (1:3) and dropped onto grease-free, cooled slides for chromosome counting and examination. R bands were obtained by heat denaturation of the chromosomes according to the method of Dutrillaux and Lejeune (1971). Xenografted CAL51 cells were plated and studied in vitro in the same manner.