The very rapid development in the last few years of techniques based on use of the polymerase chain reaction (PCR) for characterizing molecular lesions in leukaemia and lymphoma now offers the opportunity for monitoring residual disease at a sensitivity of one malignant cell in 10(5) or 10(6) normal cells. Maximal specificity is presumably achieved when the DNA sequences amplified are truly leukaemia-specific, such as BCR/ABL in chronic myelogenous leukemia, RARA PML/RARA in t(15;17) acute myelogenous leukemia, DEK/CAN in t(6;9) AML, PBX1/E2A in t(1;19) acute lymphoblastic leukemia (ALL), or TAL-1 deletions in other T-ALLs. Comparable sensitivity may be achieved by using immunoglobulin heavy chain (IGH) and T-cell receptor (TCR) gene rearrangements if a clonospecific probe can be generated. However, the presence of similar sequences in IgH genes from normal B lymphocytes may decrease the specificity. For clinical purposes the crucial issues are the following. Can PCR techniques be used for confirmation of diagnosis and evaluation of extent of disease? Can PCR data obtained in remission provide information about the probability of cure or of relapse? Can techniques be developed to quantitate the PCR product and thereby increase its predictive value? These and other issues were addressed at the 4th Workshop of the Molecular Biology/BMT Study Group that took place in Bristol UK on 9-10 May 1992.