The energy deposition in the nucleus of cells exposed to the 10B(n, alpha)7Li neutron capture reaction has been calculated and compared to the measured biological effect of this reaction. It was found that a considerable distribution of hit sizes to the nucleus occurs. The comparison of hit size frequency with the observed survival indicates that not every hit, independent of its size, can lead to cell death. This implies the existence of a hit size effectiveness function. The analysis shows that the location of boron relative to the radiation-sensitive volume of the cell is of great importance and that average dose values alone are of limited use for predicting the biological effect of this reaction. Boron accumulating in the cell nucleus is much more efficient in cell killing than the same amount of boron uniformly distributed; its presence in one cell, however, has little effect on its neighboring cells in a tissue. When boron is present on the cell surface of a tissue (as presumably delivered by antibodies), its cell-killing effect is greatly reduced compared to that in uniform distribution. However, in this case much of the dose to one cell comes from neutron capture reactions occurring on the surface of its neighbor cells. These data have implications for the choice of boron carries in neutron capture therapy. The mathematical analysis carried out here is similar to that proposed recently for low-level exposure effects of radiation, taking mutation and/or carcinogenesis as biological effects. The results here show that high-level exposure to high-LET particles (resulting in cell killing) should be treated in an analogous manner.