A transformed NIH 3T3 fibroblast cell line, Cl-e, normally does not produce spontaneous metastasis from subcutaneous or footpad tumors in nude mice. However, pulmonary tumor nodules are formed when more than 1 x 10(3) cells are injected intravenously into nude mice. Co-injection of 1 x 10(6) heavily irradiated and inactivated cells increases the clonogenic ability of the viable cells in that tumor colonies then occur with as few as 1 x 10(2) viable cells. Utilizing the action of these inactivated cells to enhance the lung colonizing ability of a relatively small number of viable tumor cells, we have developed a novel experimental model of spontaneous metastasis. In this model, a footpad tumor of the nude mouse metastasizes to the lungs following intravenous injection of 1 x 10(6) inactivated cells at a specific time of tumor growth and following tumor foot amputation, whereas no spontaneous metastasis develops without injection of inactivated cells. This model enables us to detect metastatic ability which would otherwise be too low to detect using other assays. In addition, it allows us to evaluate metastatic ability at a specific time point during primary tumor growth, since no metastases can develop during the periods before inactivated cell injection and after tumor amputation. Using this model, we have determined that the metastatic ability of Cl-e tumors in the footpad is constant throughout the exponential and stationary growth phases, even though cells isolated from exponentially growing tumors possess a 3.3-fold greater lung colonizing ability following intravenous injection than those from stationary tumors. This new experimental model may be applicable to other tumor cell lines and to other analyses where metastatic ability during a defined interval of tumor growth is of importance.