We evaluated the marrow toxicity of 200 and 300 cGy total-body irradiation (TBI) delivered at 10 and 60 cGy/min, respectively, in dogs not rescued by marrow transplant. Additionally, we compared toxicities after 300 cGy fractionated TBI (100 cGy fractions) to that after single-dose TBI at 10 and 60 cGy/min. Marrow toxicities were assessed on the basis of peripheral blood cell count changes and mortality from radiation-induced pancytopenia. TBI doses studied were just below the dose at which all dogs die despite optimal support. Specifically, 18 dogs were given single doses of 200 cGy TBI, delivered at either 10 (n=13) or 60 (n=5) cGy/min. Thirty-one dogs received 300 cGy TBI at 10 cGy/min, delivered as either single doses (n=21) or three fractions of 100 cGy each (n=10). Seventeen dogs were given 300 cGy TBI at 60 cGy/min, administered either as single doses (n=5) or three fractions of 100 cGy each (n=10). Within the limitations of the experimental design, three conclusions were drawn: 1) with 200 and 300 cGy single-dose TBI, an increase of dose rate from 10 to 60 cGy/min, respectively, caused significant increases in marrow toxicity; 2) at 60 cGy/min, dose fractionation resulted in a significant decrease in marrow toxicities, whereas such a protective effect was not seen at 10 cGy/min; and 3) with fractionated TBI, no significant differences in marrow toxicity were seen between dogs irradiated at 60 and 10 cGy/min. The reduced effectiveness of TBI when a dose of 300 cGy was divided into three fractions of 100 cGy or when dose rate was reduced from 60 cGy/min to 10 cGy/min was consistent with models of radiation toxicity that allow for repair of sublethal injury in DNA.