It is assumed that initial dose rates of 10-20 cGy/hour and total doses of 1500-2000 cGy, delivered with effective half-lives of a few days, are reasonable starting points for assessing the radiobiological effects of such low and declining dose rates. Such doses might kill 2 or 3 logs of cells out of the 9 or 10 required for tumor eradication; this is well known. The emphasis in this paper is on the change of effectiveness (per Gy) with change in dose rate. Biological effectiveness, in terms of log cell kill per Gy, will be less than that of higher dose rates because there is more time for repair of sublethal radiation injury. This loss in Relative Effectiveness, RE, is unlikely to exceed 20% in most types of tumor cell, compared with conventional external beam radiation schedules. Therefore, a tumor or metastatic deposit that would require 60-70 Gy to sterilize it using conventional radiotherapy with 2 Gy fractions (or traditional radioactive implants at 50 cGy/hr) would require 70-84 Gy at the lower dose rates available from radioimmunotherapy. This is the major dose rate effect and so far we have ignored proliferation. In the dose-rate range of 10-300 cGy/hr in vitro, highly variable depending on type of cell, division might be prevented but not progression through the cell cycle. Additional cell kill is observed because cells accumulate in the radiosensitive G2 phase; this is the inverse dose-rate effect. However, that range of dose rates comes from in vitro experiments where cells are doubling in number every 0.5 to 1 day. In vivo they double more slowly so it is possible that the unpredictable benefit of G2 accumulation, or partial synchrony, could occur at lower dose rates than 10 cGy/hr in human tumors. The dose rates necessary to kill cells at exactly the rate they are proliferating (which of course would not alone eradicate tumors or metastases) can be calculated, if values are assumed for intrinsic radiosensitivity and doubling rate of cells. In median ranges, dose rates of 2-3 cGy/hr should just counteract proliferation. Higher dose rates, maintained for a few days, could cause 2 or 3 logs of cell kill which would be observed as partial regression, as has been reported clinically and in animal experiments.