which yield u = 1.35 x 10° cm/sec = 4900 km per hour. Recalling the earlier flow description and using the above calculated equatorial extreme, a reasonable picture of winds for any time near the equator may be deduced. To avoid arriving at invalid conclusions, emphasis must be placed on the several simplifying assumptions used. Probably the most serious assumption is that of ideal fluid flow, ideal implying an inviscid fluid and a continuum concept. Estimates of mean free path and collision frequency at 660 km altitude are 100 km and 1 in 100 seconds, respectively.” Disregard of electric or magnetic forces is probably another serious shortcoming. Characterizing the diurnal buige with a sine function instead of the observed sharply defined peak (Figure 1) probably would readjust the shape of the resultant velocity field (Figure 3) but would not drastically change its main features. Restricting analysis to the equator eliminates consideration of Coriolis force which becomes 4 times as large as meridional pressure gradient at about 35 degrees latitude. A small error arises also from stating that the pressure field is sym- metric about the equator instead of about the ecliptic which is inclined 23-1/2 degrees to the equatorial plane. In support of the reported density observations and the resulting air motion just computed, there is a qualitative observation from operation Hardtack I which supports the previous flow descriptions.” At 1050hours, 1 August 1958, GMT, a nuclear explosive test was conducted over Johnston Island. Calculation performed from Reed's* observations of the behavior of the resulting debris cloud suggests a south wind (at the point marked "J" in Figure 3) of about 6700 km per hour near 1500 km altitude. This, of course, is some 2-1/2 times higher than the previously considered altitudes, but nevertheless represents the nearest thing to a direct observation yet available. Summary From reported diurnal density changes and a cloud drift observation, it is concluded that very large wind velocities are probable at great altitudes (600 km), and that these winds vary in a regular diurnal pattern. There yet remains the task of obtaining direct observations to verify these conclusions. References 1. Priester, W., H. A. Martin, and K. Kramp, 1960: atmosphere, Nature, 188, 4746, 202-204, 2. Kalimann-Bijl, H. K., 1961: Daytime and nightime atmospheric properties derived from rocket and satellite observations, J. Geophys. Res., 66, 3, 787. 3. Minzner, R. A., K. S. W. Champion, and H. L. Pond, 1959: Force Cambridge Research Center. 4, Reed, J. W., A high-altitude wind observation from nuclear explosion ''Teak''. J. Meteor. 46 Diurnal and seasonal density variations in the upper The ARDC model atmosphere, 1959, Air To be submitted to