Same as Fig. 1, but with ﹩\nu =600﹩ Hz and ﹩i=90^{\circ }﹩, and varying q and R for the numerical calculations. Setting ﹩R=20﹩ km (long dashed line) has the effect of decreasing the amplitude at high energies; compare this to the solid line. Decreasing q from 200 to 10 (dotted line) affects the energy dependence very little, showing that it is difficult to constrain the surface pattern created by the mode.