Developing new plant-derived medicines for treating malignant tumors involves the selection of plants with antitumor activity. Only a few plants show such activity, and testing them on animals with transplanted tumors is laborious and restricts the scope for routine analysis. Therefore, for many years efforts have been devoted to selecting preparations from their effects on tumor cell suspensions in vitro. Here I consider methods of selecting preparations based on the optical differentiat ion of living and dead cells in aseitic fluid f rom mouse Ehrlich carcinoma with Schrek straining [1]. The preparation is mixed in a 1:1 ratio with fresh ascitic fluid and the mixture is kept in a thermostat at 37~ for 3 h. During that time, it acts on the cells, some of which die. Then an eosin solution is added to the mixture. The dead cells are stained, while the live ones remain colorless. In existing methods, the ratio of the number of dead cells to the total number is determined by counting under the microscope. This has low accuracy and restricted use in routine analysis because of visual fatigue on prolonged microscope use. My method defines the proportion of cells killed by the extract without the use of a microscope by means of a colorimeter. After the mixture of ascitic fluid (1 ml) and plant extract (1 ml) has been kept in the thermostat for 3 h and the dead cells have been stained with eosin (2 ml of 0.05% aqueous solution), the material is passed through a centrifuge, and the deposited cells are washed three times with distilled water. To the washed material, one adds 3 ml of 0.2 M potassium hydroxide solution and heats the mixture on a water bath for 15 min. The cells dissolve completely and the eosin f rom the stained ones becomes dark green. The color strength under otherwise identical conditions is directly related to the number of stained (i.e., dead) ceils. The colorimeter is used with a graph to determine the proportion of cells killed by the preparation. The calibration curve is constructed from two standard specimens: one where all the cells have been killed (for example, by heating or with mercury chloride) and the other where the cells have been exposed only to physiological saline. The standard specimens are prepared in the same way as the others and assayed. The extinction is plotted as a function of the proportion of dead cells with those two points, through which one draws a straight line, which is used in interpreting the colorimeter readings. When the samples and standards are prepared, all the solutions must be handled under identical conditions: in each series, one should use carefully mixed ascitic fluid, with each centrifugate washed with the same amount of water and so on. If the color provided by a standard specimen differs so much from that for the samples that the colorimeter scale does not allow one to compare them, one can dilute the solutions with 0.2 M potassium hydroxide, e.g., by a factor of two, and then take the proportion of dead cells as 50%, which corresponds to the dilution. I also examined how color in the specimens affects the result. Even highly colored aqueous extracts f rom dried plants do not influence the result substantially. Plant juices sometimes may be adsorbed by the ceils in ascitic fluid and thus strengthen the color in the solutions prepared for colorimetry. When tests are made on highly colored plant juices, one should use two solutions for each sample for colorimetry, not one. One of them is prepared as described above, and the other with the difference that after incubation in the thermostat one adds not eosin but 2 ml of water. The color in the first solution is due to the eosin and the plant juice residues. The color in the second is due only to the residues, whose concentration is as in the first. One takes the difference in the extinctions, and then determines the proportion of cells killed by the extract or juice.