This study compares two well-known tumor cell invasion assays: the human amnion model versus the reconstituted basement membrane (RBM) system in the membrane invasion culture system (MICS). Our purpose was to present the quantification of tumor cell invasion using visual counts and radioactivity assessment in a side-by-side comparison and then to determine reasons for discrepancies in data collection and reporting. Basically, the data showed that: (1) fewer tumor cells invade the amnion membrane compared with the RBM, and substantially more variability exists among the data generated from the amnion assay (probably due to differences in membrane thickness); and (2) the invasive ability of the tumor cells appears to be greater using the radiolabel technique in both the amnion and RBM assay, a portion of which appears to be artifactual. Using the RBM, it was possible to sequentially select several subpopulations of highly invasive tumor cells, which was not possible with the human amnion. The invasive and metastatic potentials of these subpopulations were compared with those of established cell lines (selected in vivo). When analyzed independently, a direct relationship was shown between an increase in invasive ability and an increase in metastatic potential for the sublines selected in vitro and the established lines. However, collectively, it is more difficult to correlate the invasive and metastatic profiles of the sublines versus the established lines, which can probably be attributed to selection factors present during the establishment of the individual cell populations.