1904.] 



Circulation of the Atmosphere. 21 



puted by Teisserenc de Bort, with which the observed motions of 

 the upper clouds are in agreement. 



It is the second suggestion, the general circulation in middle latitudes 

 round the pole, with which the present paper deals. 



I propose to refer to certain representations of the average distribu- 

 tion of pressure and the corresponding average winds for January. It 

 may be conceded at once that the motion of air represented by any 

 synoptic chart of a considerable part of the earth's surface in middle or 

 higher latitudes, including ocean areas, for any day in January could 

 not by any stretch of the imagination be regarded as a "steady" 

 motion round the polar axis. The conspicuous features are large cyclonic 

 systems travelling irregularly, and the actual motion is exceedingly 

 complicated. The chart of mean isobars for the month, fig. 1, displays 

 a. large area of low pressure over the North Atlantic, south-east of 

 Greenland, the axis of which lies along the path frequently followed 

 by centres of depressions, and indicates prevailing westerly winds on 

 the southern side of the low pressure and prevailing easterly winds on 

 the northern side. 



But it is also well known that the cyclical distribution of isobars 

 round local centres and corresponding rotatory winds are specially 

 characteristic of the surface as distinguished from the upper air. For 

 a section taken at successively higher levels the pressure diminishes 

 more rapidly over cold areas than over warm areas, because the air, to 

 the weight of which the pressure at the base of the stratum is due, is 

 denser in cold regions than in warm ones. The turbulent character of 

 the motion is moderated as higher levels are reached, and its rotatory 

 character may be lost at a sufficient height.* 



In general accordance with the obliteration of cyclical motion at 

 great height the system of isobars computed by Teisserenc de Bort for 

 the 4000 metre level, fig. 2, shows no isolated low pressure areas, and 

 the lines suggest a circum-polar circulation instead of the system of 

 local cyclonic circulations. The isobars are deflected in certain parts 

 from the lines of latitude, but they form complete circum-polar rings. 

 It is not unreasonable therefore to regard the actual motion at that 

 level on any day as the result of disturbances of "steady " motion, the 

 steady motion, about which the actual motion fluctuates, being repre- 

 sented by flow along the mean monthly isobars. 



Since two consecutive isobars are not strictly speaking parallel to 

 each other throughout their course, but diverge in some regions and 

 converge in others the velocity is different at different points of an 

 isobar. Upon the principles of hydrodynamics the change of velocity, 

 in the absence of any impressed force to produce such a change, implies 

 an alteration of pressure ; hence, where the isobars are widening, the 



* See Hann, ' Meteorologie,' p. 538, and Teisserenc de Bort, 'Brit. Assoc. 

 Keport,' 1903. 



