: 



1904] Circulation of the Atmosphere. 25 



increase with height in an atmosphere in which the distribution of 

 isobaric lines is similar at all heights. 



The velocities entered upon the diagram lie between 20 and 66 miles 

 per hour, and these are not inconsistent with observed cloud velocities, 

 so that a steady motion with the velocities indicated is a reasonable 

 representation of the average conditions so far as they are known. 

 As regards the direction of the slope of pressure, that direction is 

 determined by the relation of the direction of motion of the air to the 

 direction of rotation of the earth. Whatever be the direction of motion 

 of the air, the horizontal acceleration arising from the earth's rotation 

 along the normal to the path, and to the left, in the northern hemi- 

 sphere. On that side, therefore, the low pressure must lie. For 

 motion along a parallel of latitude the slope of pressure will be 

 " downward " towards the pole for motion from west to east, and for 

 otion from east to west " downward" towards the equator. 

 Hence we may conclude that the distribution of pressure at the 

 000 metre level is favourable for a steady circulation of air round 

 he polar axis, with an average velocity of about 50 miles per hour, 

 'he direction of the motion is from west to east, with divergence in the 

 orthern hemisphere towards the north over the Pacific and Atlantic 

 ceans, and back again to the south over the Continental land areas. 

 n the southern hemisphere the motion follows the lines of latitude 

 ore closely and the deflections are less marked, though the influences 

 f the land projections are similar. 



Below the 4000 metre level is a region of cyclonic depressions with 

 tatory motion, lying between the tropical high-pressure belt and 

 ther anticyclonic regions to be found in the far north. The motion 

 near the surface level corresponds with the surface distribution of 

 pressure, which is made up of the distribution at the 4000 metre 

 level, transmitted to the surface, and the pressure clue to the weight 

 of the stratum below the 4000 metre level. I now exhibit (fig. 3) the 

 distribution of this remainder of the surface pressure when the distri- 

 bution at the 4000 metre level has been deducted. It has been 

 obtained by plotting the difference between the surface distribution of 

 .fig. 1, and Teisserenc de Bort's distribution at the 4000 metre level, 

 fig. 2. 



The result is very remarkable. A comparison of figs. 2 and 3 shows 

 at once that the general trend of the lines is very similar to that for 

 the upper air but the direction of the gradient is reversed. This is in 

 -accord with the theoretical calculation of the direction of the gradient, 

 because the direction of motion for such a pressure distribution is the 

 reverse of that represented in fig. 2. The result of the pressure distri- 

 bution due to the lower stratum alone would be a circulation round the polar 

 axis from east to west along lines almost identical with the lines of .flow for 

 the upper air but in the reverse direction. 



