We have studied a diverse group of lymphoproliferative disorders (acute lymphoblastic leukaemia (ALL), chronic lymphocytic leukaemia (CLL), non-Hodgkin lymphoma (NHL) and myeloma) to determine if V kappa gene use is random or disease-specific and whether somatic mutations (a late event in B-cell differentiation) can provide additional information on the type of B cell involved in the neoplastic clone. In this group of disorders V kappa gene selection is not random and some members of each V kappa family are preferentially rearranged. V kappa genes from the distal portion of the locus are seldom used, possibly because rearrangement of the proximal locus by deletion is more efficient than rearrangement of the distal locus by inversion. Although pseudogenes account for 46% of the V genes in the kappa locus none were ever rearranged, even in non-functional rearrangements of lambda-producing leukemias, suggesting the existence of a mechanism which down-regulates the rearrangement of pseudogenes. N regions were noted at the VJ junction in 20% of alleles (in six CLL, three NHL, two ALL and one myeloma) possibly the result of kappa-chain recombination during the early period of B-cell maturation in which TdT is expressed. Nucleotide addition or imprecise joining at the VJ junction, resulting in a shift in the reading frame, were the commonest causes of non-functional rearrangement. The occurrence of somatic mutation broadly correlated with the stage of B-cell maturation from which the different disorders are thought to arise. However, there was no strict association and somatic mutations were demonstrated in 'typical CLL' while V kappa genes were germline in some follicular lymphomas; these findings suggest either heterogeneity in the stage of B-cell maturation at which these disorders arise or some variability in the process of somatic mutation.