Transgenic mice bearing an oncogene targetted for expression in a specific tissue can reveal how that oncogene influences differentiation and help to delineate the pathways to malignancy. To explore lymphoid neoplasia, we have made strains of transgenic mice bearing different oncogenes driven by the immunoglobulin heavy chain enhancer (E mu), which promotes expression within lymphocytes and certain myeloid cells. The prototype E mu-myc mice succumb to pre-B and B cell lymphomas, following a preneoplastic phase in which cycling pre-B cells are overproduced. The similar fate of E mu-N-myc mice suggests that N-myc and myc have overlapping functions. Surprisingly, E mu-N-ras mice develop T lymphomas and macrophage tumours but no B lineage tumours; thus the ability of ras to initiate tumorigenesis may be lineage specific. Similarly, the high predisposition of E mu-v-abl mice to develop plasmacytomas may indicate that v-abl is oncogenic only at certain stages of B cell maturation. The bcl-2 gene promotes cell survival rather than proliferation, and E mu-bcl-2 mice produce copious resting B lymphocytes. The random onset and monoclonality of tumours in the transgenic strains argues for spontaneous genetic alterations that cooperate with the trans-oncogene. Indeed, most plasmacytomas of E mu-v-abl mice bear spontaneous myc rearrangements. Moreover, a minority of E mu-myc B lymphomas exhibit ras mutation, and the tumorigenesis can be reconstructed by crossing E mu-myc and E mu-ras mice, or by retroviral delivery of v-ras or v-raf, either in vitro or in vivo. To access novel cooperating oncogenes, we are using a retrovirus lacking an oncogene as an insertional mutagen. This approach should be applicable to any trans-oncogenic strain and help to delineate the genetic events that trigger malignant clones.