Using the primary cultured cells of cancerous thoracic fluid, cancerous ascites and mammary cancer tissue obtained from cancer patients at operation, these cells were cultured in the presence of antitumor agents such as Mitomycin (MMC), Chromomycin (Chr. A(3)) and Cobalt protoporphyrin (Copp) either singly of in combination, and the proliferation of the primary cultured cells was observed. The results of the study are briefly summarized as follows. 1. Cultures in the presence of a single antitumor agent: On addition of MMC or Chr. A(3) in the concentration of 10(-1)γ/ml and 10(-2)γ/ml respectively shows a marked inhibitory effect or destructive effect on the proliferation of the prinary cultured cells. Copp, on the other hand, does not show any inhibitory effect at the concentration of 10γ/ml, 10(-1)γ/ml, or 10(-2)γ/ml. 2. Cultures in the presence of two antitumor agents: When 10(-1)γ/ml of MMC and 1γ/ml Copp are added, there can be observed no cumulative inhibitory effect. In the presence of 10(-1)γ/ml Chr. A(3) together with 10(-1)γ/ml MMC there can be seen a cumulative effect resulting in a marked inhibition of the proliferation. 3. After treating the primary cultured cells with Chr. A(3) at a low concentration for 48 hours, then cultured for 96hrs: When the primary cultured cells are cultured in the presence of Chr. A(3) at the concentration of 10(-2)γ, or 10(-4)γ/ml for 48 hours, and after removing Chr. A(3) by washing, when the cells are further cultured for 96 hours, there can be observed an adverse effect; namely, the growth of the primary cultured cells is accelerated. 4. Cultures with continuous addition of antitumor agent at a low concentration: In the presence of 10(-3)γ/ml MMC there can be observed a marked adverse effect, but on further addition of 10(-4)γ/ml MMC adverse effect can not longer be seen. On the addition of Chr. A(3) at the concentration of 10(-4)γ/ml, similarly a marked acceleration of the growth can be seen, but at a lower concentration of 10(-4)γ/ml there can be observed only a slight accelerative effect. These results indicate that the addition of antitumor agent at a certain low concentration, acting on the cancer cells themselves may bring about an accelerative effect (adverse effect). Therefore, it is desirable to consider the adverse effect so far thought to be induced by the unbalance between the cumulative effect of the resistance of host and tumor proliferative capacity, from a new angle. Hence, on administering antitumor agent in vivo we should do it with utmost precaution. Acknowledgement: The author wishes to express profound thanks to Prof. Sanae Tanaka and Dr. Kunzo Orita for kind guidance throughout this work.