The radiosensitivity of the target cells for myelopoiesis: Comments on the cell kinetics model of Morris et al. (Radiat. Res. 135, 320-331, 1993)

Roberts and Hendry differ with us in matters of opinion in their first point, and are factually incorrect in their second and third points. Further, the spirit of point 1 would appear to be critical of our presentation of a model whose parameters are not uniquely estimated from a large data set, while the gist of point 3 seems to be that we should have introduced more parameters-one effect of which would certainly be to make all parameters even less identifiable! Where is the correct balance between modeling detail (scientific completeness and statistical accuracy) and modeling economy (scientifically useful summarization and statistical precision)? There is no single correct answer, and obviously-as evidenced by this exchange- considerable difference of opinion among those with some experience. Our approach is based on selection of models based on scientific knowledge, and 'cautious' analysis of high-quality data using the model as the framework. Despite Roberts and Hendry's opinions to the contrary, the exercise reported in our paper offers new, if indirect, evidence for the possibly critical role of relatively radioresistant cells in marrow recovery.