The effects of acute, protracted, or fractionated exposures to fission neutrons on survival times of female BALB/c mice were examined and compared. Mice were given single, brief exposures or exposures given in equal fractions at either 1- or 30-day intervals to doses of 0, 2.5, 5, 10, 20, 50, and 200 rad at the Health Physics Research Reactor (HPRR) or protracted exposures at rates ranging from 0.1 to 10 rad/day using a moderated 252Cf source to doses of 0, 2.5, 5, 10, 20, and 40 rad. The 252Cf source was moderated to have a similar spectron to that of the HPRR facility. After single or fractionated exposures the extent of life shortening increased rapidly over the 0-50 rad range and then began the plateau. No simple model adequately described the dose response over this entire dose range. Over the 0-50 rad dose range for exposures at the HPRR and over the 0-40 rad dose range for protracted exposures the dose response could be adequately described by either a linear model or a square root of the dose regression model except when the dose was fractionated using a 30-day interval. In this instance a linear model provided an adequate fit while a square root of the dose model could be rejected. No increase in effectiveness after fractionation or protraction was observed for neutron-induced life shortening at doses below 50 rad, while at 50 and 200 rad an increase in effectiveness was observed in this and in previous studies. These data were interpreted to suggest that in the dose range below 20-40 rad the dose-effect curve for life shortening may be linear and begins to flatten at higher doses rather than continuously bending at low doses.