side of reality. Since this umrealistic efficiency is paired with the value m = ¥.4, it is consequently likely that this value of m is also too high. The procedure described for constructing both the TNT and the nuclear curves shown on Fig. 4.3 can be performed equally well using values of m otner than the most probable value of 3.4. Other appropriate values of m as indicated on Fig. 4.2 are 3.0, representing both conventional cube root scaling and the lower limit of slope on the basis of the 10 per cent uncertainty in experimental values postulated earlier, and 4,1 representing the upper limit. together on Fig. 4.4. Both curves have been plotted Since, for military purposes, it is believed that the data for extrapolation should be available in the simplest possible form for quick use without computation, the nuclear curves shown on Figs. 4.3 and 4.4 have been re-plotted in the form of radius in feet against charge depth in feet, with yleld as a parameter. This has been done on Fig. 4.5, in which for each yield shown poth the most probable value (m = 3.4) and the limiting values m = 3.9 and 4.1 are shown. The estimates for this soil for tne most probable value of m (m = 3.4) are re-plotted on Fig. 4.6. Range of uncertainty (m = 3.0 and m = 4.1) are indicated by short horizontal bars attached to each of the parametric yleld curves. , The same kind of analysis has been carried through for dry clay, dry sand, wet clay, and sandstone and the results of these analyses are included in Figs. 4.7 through 4.10. In the case of these other soils no nuclear data are available and hence the efficiencies found in the Nevada soil have been used in the following fashion. For the most probable value of the scaling exponent m in each of these other soils, the variation of efficiency with depth at Nevada for m = 3.4 has been used. Similarly, for the lowest value of m for each of these other soils the same variation of efficiency with depth nas been used as was found at Nevada for the lowest value of m there, namely, 3.0. The corresponding analysis has been made for the upper limiting value of mn. The most probable and limiting values of m for all the soils re- ported here are listed in the table below. In each case, the available data have been plotted in the same form as was shown on Figs. 4.1 and 4,2, the best straight line was drawn fcr those points and then values of radius 10 per cent above and below the curve were marked at the up- per and lower limits of the charge sizes considered.* By this procedure, the limiting values of m have the greatest range for those soils in which no large TNT charges have been fired, and this is appropriate, since in fact the extrapolation is less certain in such cases. In the case of wet clay, Fig. 4.€, so little TNT data are available that crater radius has been predicted only for the most probable value of the scaling exponent nm. * It was decided not to review TNT cata from charges less than 200 lb.