76 THE SHORTER-TERM BIOLOGICAL HAZARDS QF A FALLOUT FIELD terest has been discussed [1], and can be sum- marized briefly. The available data are conflicting; however, it appears that such an energy dependence mayexist in mice over the range of 250 to 2,000 KVP, and that 1,000 and 2,000 KVP X-rays may beless effective by a factor of 0.8 or 0.9 (in terms of tissue dose). There are several pieces of evidence that Co” gamma may be even less effective—perhaps 0.7. Part of these differences may be dosimetric in origin; however, they appear to be real as doses are measured at present. With large animals, dogs and swine, there appears to be no such dependence of effect over the range of 250 to 2,000 KVP. Undegraded gamma radiation (Co) appears to be less effective in the dog (Table I), as with mice. It would appear that intrinsic energy dependence over the range of energies of interest is at. most of the order of 10 or 15 percent, a factor much smaller than other sources of uncertainty. In considering the effect of distribution of dose as it affects degree of response, the concern is mainly in comparing one type of unilateral exposure to another, and unilateral to bilateral exposures. It is obvious by now that with iden- tical depth-dose patterns, the same degree of effect, within a few percent, will result from the same dose. In comparing one typeof unilateral irradiation to another, it is of course known that the shallower the curve, the /ess the effect for a given entrance or midline tissue dose. This can be easily seen from the data of Potter [27] and Ellinger [28], and that of Tullis in swine (Table TI). Little difference is noted for dogs irradiated unilaterally with 250 and 2,000 KVP X-rays (Table 1); however, the beams were filtered such that the depth-dose patterns were not greatly different [29]. It is thus clear that differences do exist; however, the data are not sufficiently good to allow quantitative treatment, As for a means of predicting effects with a given unilateral pattern, some data obtained with small animals indicate that the erit tissue dose may be a normalizing quantity [27, 28]. The data in large animals are insufficient to evaluate this point. Integral dose or gram GEOMETRICAL, ENERGY FACTORS--BFFECT OF RADIATIONS ON MAN roentgens has been proposed as a normalizing quantity. Grahn and Sacher [18] have shown that with different types of “total body” irradiation, integral dose is of no value in this regard and the concept does not apply in predicting mortality with partial-bodyirradiation (30]. Evenif integral dose were the normalizing factor, the computations involved are so com- Thus inverse square fall off is appreciable, unlike the fallout field. Also with large animals placed in a standing position among tion. With the cooperation of Col. Trum, additional plex and lengthly that this parameter would have no practical usefulness in hazard evalua- Some additional points will be mentioned in regard to the large animal data in Tables I and II. Looking first at the btlateral data for dogs and swine, it is seen that the air dose LDw’s vary considerably among investigators, but that the LDy’s in terms of midline tissue dose are remarkably constant for X-rays with 8 variety of energies and experimental conditions. The discrepancy between air dose and midline dose is muchlarger for swine than for dogs, which would be expected from the larger swine. This indicates that such data, to be quantitative for man, must be obtained on man-sized animals. Data from dogs or monkeys do not apply directly. It is apparent that the usually quoted LD» values for large animals, in terms of air dose, are much too high, and that there is no true energy dependence of effect over the range of 250 to 2,000 KVP. The LD,»for dogs and swine are approximately equal and considerably below the LDfor mice or rats. No biological data are available for large animals exposed to fallout gammaradiations; however, the LD» in terms of midline tissue dose would be expected to equal those in the tables to a few percent. With regard to the Co” gamma data in Tables I and II, the bigher LDy values may reflect in part the apparent intrinsic energy dependence that has been noted for mice. With the swine exposed to Co® in the multisource field at Oak Ridge, however, additional factors enter. It can be easily shown that approximately 65 percent of the radiation received at any point in air at the “center” of any unit of 3 of the total of 19 sources comes from a distance of approximately 1.5 meters. the sources, a large percentage of the radiation traverses the long axis of the animal, rather than a transverse (shorter) diameter as with animals exposed to bilateral X-irradiation or with man upright in the fallout field. Thus the midline dose would be expected to be relatively quite low compared to the air dose. depth-dose curves were obtained in the Co® field, which indicate that the midline dese in a swine phantom is less than half of the entrance dose. The LD» value (Table ID) is corre- spondingly low in terms of midline tissue dose. From Tables I and IT, it can be seen that in the laboratory, more radiation dose (entrance air or tissue dose) is required to produce a given effect with unilateral than with bilateral exposure. With “unilateral” exposure to the immediate bomb gamma radiation in the field, however, the LD values are lower than for unilateral irradiation in the laboratory, and approximately.equal to bilateral irradiation in the laboratory. This could indicate uncertainties in the field data—the LDy values were obtained in a single determination with 10 animals per point, and the swine used were smaller than those used in the laboratory. It could also mean that the relatively flat curve for bomb immediate gamma resembles in effect bilateral, more than unilateral irradiation. The considerations outlined must be taken into account in hazard evaluation. The prob- lem is analogous to the RBE problem, which gave rise to the dose unit “rem” to more closely estimate hazard than is possible with the roentgen or rad. The dose in rem is equal to the dose in r multiplied by an experimentally determined RBE factor. It would appear that another factor should be introduced, a geometry or g factor, which must be experimentally determined for each situation as is the RBE factor. It is seen from the present paper, that under many circumstances the g factor may greatly exceed in magnitude the RBE factor. The problem of accurate hazard evaluation 77 in large animals and man is seen to be particu- larly complex. It is not possible to use a single quantity such as “r” or “rem” alone to predict, hazard under a variety of circumstances— additional factors to describe the situation considered must be introduced. No one would ask for the “hazard” from a given dose of any common toxic agent such as arsenic without describing the situation farther-—-howthe drug is to be given, the chemical form, part of the bodyreceiving it, time over which it was administered, size of individual, ole. Yet it is frequently expected that a “dose” of radiation in “ry”? or “rem’’ will describe the hazard under all situations. And the difficulties cannot be circumvented by changing a name—introducing, as has been suggested, some arbitrary type of “hazard” unit that supposeldy will indicate what effect can be expected in man. No one unit can ever describe the hazard; other quantities are necessary. Substitution of a “hazard�� unit represents a regression to the “skin erythema dose” days, that nullifies the very great advance made with the introduction of the roentgen unit. The roentgen (or rep or rad) is as good as any presently available single quantity to allow a very general estimate of hazard. Hf greater accuracy of prediction is desired, then the situation must be recognized and treated as a complex one. This is done in other disciplines, and personnel are trained to handle the problem. Quantities in addition to the instrumentreading in r or rads (where dose is measured, type of exposure, type of radiation (RBE), type of biological response of interest, dose rate, body region exposed, etc.) must be taken into account. These factors could be incorporated into one, or a series of nomograms; however they cannot be incorporated into a single “hazard” unit or into a single instrument reading. Perhaps most dangerous in attempts to devise a hazard unit is that it will involve combinations of several factors in unknown proportions. Thus one trained and conversant will not he able to sort out the important quantities that would allow accurate evaluation of the hazard. LD for man.—The consideration of the