In answer to the last question, I would say that it would require energy of the order 1025-i026 ergs to produce the water vapour observed in the middle stratosphere. Krakatoa is estimated to have released N loz7 ergs so that if one assumes a residence time of the order of centuries (which may not be too unreasonable, having regard to the observed hold-up of W,,, over the equator), this idea may be tenable. I would be happier, though, if Krakatoa had released 1028-1029 ergs.( left uncertain, I should like to suggest that it could be explained by the long track over warm water followed by the air currents before entering the rain-producing depressions. Presumably the patterns of warm and cold water, as affecting evaporation, could produce inequalities in the distribution of rainfall at sea much as relief does on land, although not, perhaps, quite to the same extent.After reading Dr. Tucker's paper I was much perplexed by his A and B stations as bearing on orographic precipitation. Naturally, one would not expect orographic movement in flat islands like Tiree and Benbecula, far from high ground, though some other stations in the A group are quite near higher ground. But why should Mildenhall in the B group be expected to have an orographic effect when it lies in flat country with an average annual rainfall of only 21in., well away from the boulder clay plateau of West Suffolk which is sufficiently high to carry a rainfall of about 26 in.? Apparently, therefore, Mildenhall should lie in rain-shadow of this plateau to the south-west, with a negative orographic factor. It seems particularly desirable to clear up this issue because in Dr. Tucker's Table 4, Mildenhall has the largest error after Blackbushe, which on his own showing is out of keeping with small orographic effect. Lastly, I should like to raise the general question of whether the ocean as a whole, occupying some five-eighths of the surface of the globe, receives more water back from direct rainfall or from the in-flowing rivers. In view of the great intensity and persistence of orographic rainfall and of the many high mountain ranges which border the ocean I would hazard the guess that land drainage into the sea accounts for the larger proportion.Dr. R. C. RAINEY : In view of the very great interest of such estimates of precipitation for other ocean areas, Dr. Tucker might care to comment on a possible limitation of his method (and particularly of its geographical extension) which is not explicitly mentioned in his paper. It is the possibility that the rates of precipitation encountered in some sea areas might differ systematically from those described by the same present-weather code number at the land stations used for the determination of x, y and z. Might one not expect the intensity of ' heavy precipitation,' in particular, to show some systematic geographical variation, in association with, e.g., sea temperature? Warm and cool ocean currents might be envisaged as introducing possible complications, analogous to the orographic effects shown by land st...

Since, unless I have made an error, the equations of motion are not invariant to reflections through the midpoint there is no compelling a priori reson for expecting the solutions to have the symmetry Dr. Scorer suggests. Admitting the qualitative differences of symmetry, I can see reasons both for regarding the experiment as responsive to friction at the bottom (and lacking symmetry for that reason) and for seeing more symmetry of this type in the atmosphere than might a t first glance appear. With respect to the first, some related experiments in a large hemisphere have shown that one can have at very low Rossby numbers a similar interior distribution with positive thermal wind and at times waves either with westerlies (maximum at top) or with easterlies (maximum at bottom) almost everwhere. The difference was due to absence or presence of air torque at the top (the rotation was rather high at 10 r.p.m. or more). This certainly suggests that presence of the bottom can determine such things as the overall mean zonal current. Although the measured mean zonal components at the bottom in the present experiment are all easterly, the variations with radius are such as to look quite like a surface wind profile displaced toward easterlies except in the extreme strength of the polar easterlies. I conclude that in this respect the experimental base is behaving more like the ground than at first appears. With respect to the second matter of heating effects, I think one has to regard the interior in the experiment as corresponding, for any fairly detailed comparisons, to a middle latitude strip where advection effects tend to dominate and short time-changes are quasi-adiabatic. The deep tropics then serve the wall function of transferring heat to considerable heights and the motion north of the sub-tropical ridge is, like the experiment, quasi-adiabatic and to a considerable extent independent of the source arrangement at the ground below.Mr. Sawyer's question concerns a more subtle possibility of frictional influence. In some cases experimental jets give a good deal of the impression of separated boundary layers. I am inclined, however, to think that the jet is not essentially connected with friction although that possibility cannot be ruled out at present. A variety of types of experiment exhibit jets in such a way as to suggest that it is an inviscid phenomenon but all involve boundaries, different though they be, that might be a required part of the mechanism. I think it will be very much worth while to run some tests with as much variation in this boundary condition as possible (especially with as nearly a free boundary at the bottom as is feasible and with lids in contact at the top) in order to see what variations occur in the jet. It might then be possible to draw a more definite conclusion.In answer to Mr. Gordon's question I may say that it is feasible to obtain at least some direct variation of the Coriolis parameter by using a container figured to the equilibrium paraboloid for some rather high rotation, say 60...

The mean meridional circulation derived from wind observations is combined with surface torque and mountain torque measurements to obtain a complete angular momentum budget for the northern Hemisphere between ISON and 7 0 ' N for summer and winter. A previously proposed scheme of vertical flux of angular momentum involving downward flux everywhere except in the lower layers of the tropical easterlies is confirmed. The layer ofzero vertical fluxin the tropics is approximately at the level of maximum easterly wind.

SUMMARYThe dynamics of mean zonal flow over the Equator are described using a momentum budget. First the budget is drawn up for the troposphere, and the estimates of the mechanisms involved are shown to be consistent with the observed zonal wind/height profile. The technique is then applied to the stratosphere up to 30 km where the approximately $6-month oscillation in zonal winds is observed. The necessity for a similar oscillation in the momentum convergence due to large-scale horizontal eddies is deduced, and this oscillation is verified from wind observations. T h e vertical velocity/height profile associated with the zonal wind oscillation is derived and shown to be constant in time. The downward motion above 20 krn and upward motion below is shown to be consistent with observed ozone and water vapour observations. Finally, types of synoptic systems are suggested which could be associated with the observed pattern of zonal flow and momentum convergence.

Zonal winds in the stratosphere for a period of 20 years have been analysed at nine stations between 7% and 66% centred on the longitude of eastern Australia. The amplitude, period and phase of the third ('quasi-biennial') cycle are determined. Although a coherent pattern emerges in the variation of period with latitude and height it is concluded that, at least in the tropics, the representative period is 832 days (27.35 months). The well-known maximum in amplitude of this third cycle in the tropics decreases to a minimum a t about 30"s but increases again polewards, reaching a maximum at about SO'S at the highest levels of observation (6mb). The phase of the cycle 'leads' at high levels and low latitudes and also at the height and latitude of the polar night jet. Below these maxima there is a downward propagation of the cycle but little phase variation with height exists in middle latitudes. For the tropics, the variation with latitude of amplitude and phase is consistent with lateral diffusion of a forced equatorial cycle, with eddy diffusion coefficients of 1 x 109cm2s-1 at 80mb, decreasing to 3 x 10*cm2s-' at 25 mb.