A novel model of tumor progression in metastatic cancer is described which grew out of attempts to derive stable non-metastatic variants from a highly metastatic mouse tumor called MDAY-D2. The variants were obtained by selection of so-called lectin-resistant (LecR) membrane mutants using toxic concentrations of wheat germ agglutinin (WGA) as the selective agent, after mutagenesis. Cloned WGAR variants almost all appeared to be highly tumorigenic and metastatic, but displayed altered growth properties which were highly suggestive of major cellular phenotypic alterations occurring prior to metastasis. This were confirmed with the discovery that spontaneous visceral metastases always consisted of WGA-sensitive (WGAS) 'revertant' tumor cells. Such revertants also arose at the site of the subcutaneous inoculation and, with time, comprised an increasing proportion of the tumor cells at that location. The WGAS/high metastatic phenotype was stable in vitro or in vivo, implying the WGAR leads to WGAS shift had an underlying genetic basis. Thus, it appeared that the WGAR tumor cells could not metastasize, because of either an intrinsic cellular defect or a host imposed barrier, but that this block could be circumvented through a genetic change in the WGAR tumor cells which was accompanied by reversion of the WGAR phenotype. Non-tumorigenic (tum-) WGAR variants were also obtained, but in these cases the mutagenesis treatment itself appeared responsible for development of the tum- phenotype. The reduced tumorigenicity had an underlying immunological basis, a finding which could be exploited to immunotherapeutically treat established visceral metastases of poorly immunogenic tumors. Throughout these studies, emphasis was placed on the considerable potential of using tumor cell populations having various stable drug-resistant genetic markers to monitor aspects of tumorigenesis, tumor progression, and metastasis.