(footnote cont.) The corresponding HE data from Nevada taken from Tables A.4 and A.6 giveapvalue for R,/W1/3 of about 0.8. If one uses the suggested form of the: equation and hence assumes that the effect of soil is indepen- dent of: the effect of charge size, then one might say that craters in the Marshalls should be expected to be 1.8 to 2.0 times as large (in radius) as ‘craters from identical charge sizes and depths in Nevada. In aiatmilar manner it is found that the value for R,/wi/3 WE for megaton surface shots in the Marshalis is about 1.0, while that for the kiloton surface.‘shot in Nevada is 0.34, which implies that Marshall craters will dé ‘some three times larger than Nevada craters. Actually, if the amall butfinite value of De/w2/3 1/3 is taken into account, particularly for the JANGLE surface shot, the analysis suggsts that scaled crater radii for nuclear charges in the Marshalls are twice as large as for those in Nevada, Since this is the same figure that was obtained for HE craters, it is tempting and not implausible to say that all scaled crater radii in the Marshalls will be very close to twice those in Nevada. While the precise asta. quoted from the AFSWP memorandum were not at hand during the development of the extrapolation method described in section 4.2, some prior discussion of them was held with Dr. Stephenson by telephone. At that timeit was Dr. Stephenson's feeling that the data themselves were somewhat! unreliable because all the craters were water-washed before measurenient.. In addition it seems improper to assume that the characteristics, «for cratering purposes, of the water-saturated coral sand involved-th the HE tests are identical with the characteristics of the more coherent+water-saturated coral rock involved in the nuclear shots. rmia i ; 59