Many studies have noted an association between amplification of single oncogenes and poor prognosis in breast cancer. We are investigating whether measurement of amplification in a larger number of proto-oncogenes increases the reliability of the prognostic information provided. As the first stage of this investigation, amplification (of c-erbB-2, cycD1, int-2, c-myc and MDM2), aneuploidy and altered expression of p53, which all indicate genetic instability, were studied in 117 primary breast adenocarcinomas. Amplification was correlated with aneuploidy (p=0.002) but not with altered expression of p53 even though the tumours with p53 overexpression were all aneuploid. Our results suggest that measurement of amplification is a potentially valuable prognostic factor.