When tissues of limited extent are irradiated with penetrating ionizing radiation, the absorbed dose, i.e., the energy absorbed in a volume divided by the mass of that volume, is sensibly constant throughout the medium, provided the ratio is taken over a sufficiently large volume. Thus if a mouse has received 100 rads of hard X-rays or high-energy fast neutrons, each gram of the animal absorbs very nearly 10,000 ergs. The energy density in masses that are not very much smaller is also 104 ergs/gm. However, when one considers masses of tissue that are of a size comparable to the one of cell nuclei or subnuclear structures, the energy absorbed per unit mass may differ from this value because of the random nature of energy deposition. Whereas masses of the order of 1 gm are traversed by a vast number of charged particles, subcellular structures are traversed by but a few, and, when the volume under consideration becomes very small, these variations become exceedingly large. In fact, ultimately the ratio of E/m may frequently be zero, since it may be found that no charged particle traverses a volume having dimensions of the order of a fraction of a micron. Thus material that has been subjected to uniform irradiation and has received a well-defined absorbed dose exhibits strong variations of energy density on a microscopic scale. A knowledge of these local variations is essential for an understanding of the biological action of radiation, since it must be assumed that the energy instrumental in the inactivation of the cell originates in a volume that is certainly no larger and probably a good deal smaller than the cell. The fact that the same absorbed dose delivered by radiations of different LET can give rise to different degrees of biological effect is perhaps the most convincing evidence that the efficacy of an absorbed dose must depend on the manner in which the energy is delivered on a microscopic scale.