Recent studies have suggested that differences in the initial low-dose region of the radiation survival curves for human tumor cells might explain the differences in clinical response of tumors to fractionated radiation treatment. In this study, which is described in two companion papers, we investigated this hypothesis directly using animal model systems. In the present paper we determined in vitro radiation survival curves for eight murine tumor cell lines of varying histopathological type and: (a) measured survival at the 2 Gy and 8 Gy dose levels, (b) fitted parameters to the linear quadratic and two component multi-target equation models of cellular survival and (c) calculated mean inactivation doses. We found that the choice of the data fitting procedure affected the absolute value, relative ranking, and power to discriminate between the cell lines of these calculated parameters. A detailed statistical study indicated that the measured surviving fraction at 2 Gy (SF2) was the best discriminant of intrinsic radiosensitivity between the eight tumor cell lines. When these same cell lines were assayed for intracellular glutathione (GSH) levels, no correlation was found between levels of GSH and the SF2 value. Determining the SF2 value may be the method of choice to describe the low-dose region of the radiation survival curve, as it precludes the necessity of choosing a model to fit the survival data, it has excellent discriminatory powers, and it represents the survival in the radiotherapeutically relevant region of the in vitro radiation survival curve. Furthermore, as demonstrated in the companion paper, it correlates with cell survival in the tumors following 10 fractions of 2 Gy given in vivo.