The author earlier discovered (1932) that in plasma cultures, liquefaction and the breaking down of the fibrin network supporting the cells results (1) for malignant cells[long dash]in detachment from the fibrin, degeneration, and death; (2) for non-malignant connective-tissue cells[long dash]remaining adherent to the cover glass and, without renewal of the culture medium, survival (in many cases) after the malignant cells have died; this suggests that the membrane of the malignant cell may be less adhesive than that of normal cells. To test this, healthy areas of transplantable tumors, cut into small fragments and washed in Ringer''s, were explanted into mouse or rat serum or plasma with or without mouse embryo extract. The cultures were stained vitally with 0.5% trypan blue, and daily photographic records were kept of their mode of growth. The non-malignant cells present consisted of cells of the stroma and cells of the monocyte-macrophage series, which vary in number according to the extent of the resistance which the animal from which the tumor was removed had opposed to the malignant growth. The malignant cells can be distinguished from the non-malignant cells (1) by the failure of the former to accumulate trypan blue to the cultures; (2) by vital staining with a basic dye, such as neutral red; and by (3) their general cytological characters, such as size of nuclei and nucleoli, and cytoplasmic vacuolation. In plasma cultures both malignant and non-malignant cells migrate from the explants; in serum cultures only non-malignant cells wander out. In serum cultures the explants of carcinomata soon become rounded. The carcinoma cells inside the explants divide mitotically. In serum cultures the explants of sarcomata tend quickly to disintegrate[long dash]the sarcoma cells rounding off and later floating freely in the medium. Mitosis occurs in rounded sarcoma cells. The non-malignant cells which wander out in serum cultures exhibit a wide variety of forms, but they all stain vitally with trypan blue. Plasma cultures made at the same time do not exhibit good growth of malignant cells. The different behavior of malignant cells in plasma, and in serum, may be the result of an alteration in their plasma membrane making them unable to adhere to glass, though able to use the fibrin network of a plasma clot as a support for their movement. This may be brought about by a greater amount of fats or lipins in the plasma membrane. To such an alteration may be attributed also the failure of malignant cells to segregate the water soluble acid dyes, since these may be unable to penetrate such a membrane. The results of this research render invalid conclusions drawn concerning the specific destructive action of immune sera on cancer cells in serum cultures. Tests for any further claims of this kind are formulated.