The reproducibility of survival time in the L 5222 leukemia in rats was tested by transfer of varying numbers of untreated and X-irradiated cells. A linear relationship between log cell dose and survival time was established for the range of unirradiated cell doses between 10-8 to "10-0", resulting in survival times from 6 to 16 days, with very little variation between individual animals of each group. This narrow deviation of survival time makes it possible to use mean survival time as a measuring parameter instead of the cell dose required to kill 50 percent of recipient animals. From the longer survival times observed after transfer of a given number of X-irradiated cells, the number of viable cells transferred in the inoculum could be calculated and thus the degree of cell death due to X-radiation. Agains a correlation between log cell dose and survival time was found but host survival was prolonged when small numbers (similar to 10-2) of viable irradiated cells were transferred. It is suggested that this is not due to a change in proliferation kinetics but rather to host factors, such as an immunological reaction. Thus the L 5222 leukaemia seems to be a good model in its reproducibility of survival time, and may have some similarity to human acute leukaemia when low numbers of treated cells are concerned.