EFFECTS OF EXPOSURE GEOMETRY 567 quantitative comparison with bilateral exposures. Evidence from small animal data has been presented indicating that the erzt tissue dose might be the best single parameter in terms of which to express unilateral exposure (29, 30). The small amount of data available are not sufficient to evaluate this relationship in large animals. Although with bilateral and the more complicated exposure geometries use of the midline tissue dose appears to normalize the LD values, no clear-cut single parameter appears to exist for unilateral exposure. It has been suggested that integral dose or gram roentgens might be a suitable measure of dose for comparing “total-body” exposures. Grahn and Sacher (/8) investigated this possibility using rabbits and mice exposed to different energy _ radiations and found that integral dose did not normalize the Dg values. These results, together with data from partial-body irradiation (3/), indicate that the concept of gram roentgens or integral dose is not useful in acute irradiation LD50 studies. “Unilateral”? exposure to the atomic bomb immediate y-radiation appeared to be more effective than unilateral, and perhaps more effective than bilateral irradiation in the laboratory (bomb y-rays were found to be equally effective as laboratory X-radiation in mice; see reference 32). The explanation for this may lie in unevaluated geometry or energy factors, or it may lie in biological factors. A neutron contribution cannot be excluded definitely. The EDs. values were ob- tained in a single exposure with relatively few animals. The swine used were much smaller than those used in the laboratory (24). From the dog and swine data in Tables I and II, anintrinsic energy dependence in going from high-energy X-radiation to undegraded Coy-radiation appears to exist. The higher LD,» for swine exposed to Co® y-radiation at Oak Ridge could be largely explained on the basis of the low dose rate employed (33, 34); however, no such explanation applies to the Co®™ data obtained on dogs at Rochester. There are considerable data indicating a possible low effectiveness of Co® and other y-radiations compared to high-energy X-radiation (35-40). The data are not sufficient to indicate to what degree the apparent difference results from a true energy dependence, and how much is explicable on dosimetric or other grounds. i I01 4 nak At any rate, the difference probably does not exceed 10 or 15%, much less than can result from exposure geometrydifferences. The discrepancies among air-dose LDs9 values is considerably larger for swine than for dogs (Tables I and IT), and the differences in LDso values for unilateral radiation, however expressed, appear to be greater for swine than for dogs. This would be expected, since energy and geometry factors become more pronounced as animal size is increased. Thus, even dogs are not sufficiently large to allow direct quantitative comparisons with man, and animals the size of adult human beings should be used for this purpose. Biological data for multiport, rotational, and crossfire exposure are available 7h 0b