The fate of the leukaemic clone after bone marrow transplantation (BMT) for patients with chronic myeloid leukaemia (CML) may determine the long-term prognosis. Several groups have reported the use of the polymerase chain reaction (PCR), a highly sensitive technique for detecting residual leukaemia, to study blood and bone marrow of patients in apparent cytogenetic and clinical remission at various intervals after BMT. The limitations of the technique relate to the need in CML to use mRNA rather than DNA as test material, the undefined sensitivity of the assay, the lack of knowledge of cell lineage when the test is positive and the high risk with present techniques of false positive and false negative findings. These points notwithstanding, the collected data suggest that most longterm survivors have no evidence of leukaemia detectable by PCR; in contrast many of the patients transplanted within the last 5 years do have evidence of residual leukaemia. This suggests that in a significant number of patients the leukaemic clone may survive for several years after BMT before it is eradicated (or falls below the threshold for detection by PCR). If this is confirmed the finding of residual leukaemia by PCR in the first few years post-BMT would have limited prognostic significance.