Stewart, R. D., Ratnayake, R. K. and Jennings, K. Microdosimetric Model for the Induction of Cell Killing through Medium-Borne Signals. Radiat. Res. 165, 460–469 (2006). Microbeam, medium-transfer and low-dose experiments have demonstrated that intercellular signals can initiate many of the same biological events and processes as direct exposure to ionizing radiation. These phenomena cast doubt on cell-autonomous modes of action and the linear, no-threshold carcinogenesis paradigm. To account for the effects of intercellular signals, new approaches are needed to relate dosimetric quantities to the emission and processing of signals by irradiated and unirradiated cells. In this paper, microdosimetric principles are used to develop a stochastic model to relate absorbed dose to the emission and processing of cell death signals by unirradiated cells. Our analyses of published results of medium transfer experiments performed using HPV-G human keratinocytes suggest that the emission of death signals is a bi-exponential function of dose with a distinct plateau in the 5- to 100-mGy range. However, the emission of death signals by HPV-G cells may not become fully saturated until the absorbed dose becomes larger than 0.6 Gy. Similar saturation effects have been observed in microbeam and medium-transfer experiments with other mammalian cell lines. The model predicts that the cell-killing effect of medium-borne death signals decreases exponentially as the absorbed dose becomes small compared to the frequency-mean specific energy per radiation event.