398 NEWELL AND MILLER 1 -.5 2 o (¥V'] * [utV*])cos! agoa G _=~ Teas’ reo ofn ) ap 1 au Z=1- os G30 88 ? The symbol T represents the rate of diabatic heating. The other symbols have their usual meteorological significance. Cross sections of the eddy heat and momentum transports and mean zonal wind and temperature were constructed from the data reported by Buch’® and Peixoto'4 for the region below 100 mb and from data reported by Murakami’ and Peng!* and other, unpublished data for the levels of 100 mb and above. The diabatic heating, T, was obtained from the work of Davis® for the 5- to 10-km region and from the work of Kennedy’ for the 10- to 30-km region. The terms in the equation for V were then evaluated at grid points 2.5 km apart in height and 10 degrees apart in latitude. The final re- sults appear in Fig. 2. It is immediately evident that a coherent threecell structure is present with small equatorward velocities in middle latitudes in both seasons and a larger poleward flow in equatorial regions. The magnitudes in the lower stratosphere are very much lower than those given by Oort.*4 The possible error in the results depends on the accuracy of the cross sections; therefore it is difficult to specify the error quantitatively. Those involving wind velocities are likely to be biased because of the selectivity introduced by the early loss of balloons at high wind speeds (see Ref. 31). In addition, values in the vicinity of the tropics are no doubt dependent on the 26-month oscillation. Latent heat is not included in the diabatic-heating distributions used, but this does not affect the values calculated for the lower stratosphere. In fact, a comparison of the magnitudes of the two terms in the numerator of the equation for V showsthat, although the heat-forcing function predominates or is very important in the determination of V at 5 km, its importance decreases with height. In the 12.5- to 17.5-km calculations, the momentum-forcing function is predominant (in numerator and denominator); therefore thermal effects play a secondary role. This pattern perhaps is to be expected since, as we noted earlier, the tropospheric heat engine is driven by the heating effects whereas the lower stratosphere appears to be dynamically coupled to the troposphere. As far as trace substances are concerned, the preceding findings seem to imply that the mean meridional circulations do not contribute Significantly to the mass budget in the middle-latitude lower strato-