192 



NATURE 



[April 27, 19 16 



balloons, and Glaisher had followed them up by a 

 large number of ascents for the British Association, 

 which reached their climax in the famous ascent with 

 Coxwell in 1862. They added a good deal to our 

 knowledge but very little to our ideas. They told us 

 that the atmosphere showed continual decrease of 

 temperature with height, and that surprised nobody; 

 it was a natural incident in the gradual transition 

 from the temperature of the surface of the earth to 

 the absolute zero of space. "The nicely calculated 

 less or more " was not of vital importance. Cyclones 

 and anticyclones obviously belonged to the upper air, 

 the regions where clouds are formed and dissipated, 

 where rain and snow and hail are produced, but bal- 

 loon ascents told us little about them beyond confirm- 

 ing the surmise that there are great ascending currents 

 associated with certain forms of cloud. 



The only real information to be got about the atmo- 

 sphere in upper regions was that contained in observa- 

 tions of pressure at the surface, which is the cumula- 

 tive result of the whole thickness of the atmosphere, 

 and the amount of rain, hail, or snow which falls 

 from above. There were also observations of the 

 forms of cloud and their motion, and, if we please, of 

 their position. The rest is necessarily speculation, so 

 that out of these observations meteorologists were 

 obliged to imagine for themselves what cyclones and 

 anticyclones are, how far up they extend, how they 

 are produced and maintained, what kind of air they 

 are made of, and so on. 



Observations of the Upper Air. 



Speculation can do a great deal with the atmosphere. 

 It goes beyond the reach of our balloons, and tells us 

 of the substitution of hydrogen and the rarer gases for 

 oxygen and nitrogen in the region of the meteor and 

 the solar electron. But from the year 1896 onwards 

 there has been a systematic collection of facts about 

 the upper air by using kites to carry instruments up 

 to heights of 3 kilometres, or occasionally more; 

 ballons-sondes which carry instruments up to heights 

 of 35 kilometres (20 miles or more) ; and pilot balloons 

 which give the direction and velocity of the wind at 

 various levels up to 10 kilometres, sometimes more. 



Comparison of Fact with Speculation. 



This investigation has given us a wealth of informa- 

 tion about the upper air. The principal result is the 

 division of the atmosphere into two layers : a lower 

 layer about 10 kilometres thick, the troposphere, the 

 region of convection ; and an upper layer, the strato- 

 sphere, in which there is no convection. We can use 

 the information to test some of the generally accepted 

 ideas about cyclones and anticyclones bv comparing 

 the results of speculation with the new facts. Many 

 of the pictures which we imagined now appear to 

 have been illusions. Those of us, for example, who 

 thought that because the air was warmed from the 

 bottom, the upper part would be free from sudden 

 changes of temperature such as we get at the surface 

 were rapidly and rudely disappointed. Simplicity is 

 not apparently the characteristic of the upper air. 



The Convection Theory of Cyclones and Anticyclones. 



Before giving you other examples, let me quote the 

 description by which Galton introduced the name 

 "anticyclone," because the mental picture of the 

 structure of cyclones and anticyclones which has 

 guided the thoughts of the majority of meteorologists 

 has been formed by the gradual elaboration of the 

 ideas contained in that description : — 



" Most meteorologists are agreed that a circum- 

 scribed area of barometric depression is usually a locus 

 of light ascending currents, and therefore of an in- 



NO. 2426, VOL. 97] 



draught of surface winds which create a retrograde 

 whirl (in our hemisphere)." 



"Conversely, we ought to admit that a similar 

 area of barometric elevation is usually a locus of dense 

 descending currents, and therefore of a dispersion of 

 a cold, dry atmosphere, plunging from the higher 

 regions upon the surface of the earth, which, flowing 

 away radially on all sides, becomes at length imbued 

 with a lateral motion due to the above-mentioned cause, 

 though acting in a different manner and in opposite 

 directions" (Proc. Roy. Soc, vol. xii., 1862-1863, 



P- 385)- 



Out of that there gradually grew the conception, 

 on the one hand, of the central area of a cyclone on 

 the map as a centre of centripetal motion, a focus 

 of attraction for the surrounding air, and of the 

 general area of the cyclone as a region of ascending 

 warm air producing rain or snow ; round the central 

 region the air moves inward with a counter-clockwise 

 motion in spiral curves. On the other hand, the 

 conception of the central area of an anticyclone is of 

 a centre of centrifugal motion, a region of repulsion ; 

 the general area of an anticyclone is a region of 

 descending cold air moving with a clockwise motion 

 spirally outwards. The fundamental dynamical idea 

 is that of air driven like gas along a pipe from 

 high pressure to low pressure, retarded by the friction 

 of the surface, and diverted from its direct object by 

 the rotation of the earth. 



For future reference, let us separate the three 

 elements of this picture and keep them distinct. First, 

 the circulation, counter-clockwise in a cyclone, clock- 

 wise in an anticyclone. Second, the convergence 

 across the circulation from high to low. Third, the 

 convection, or vertical motion, which apf>ears as 

 ascending air in the cyclone and descending air In 

 the anticyclone. 



According to the conception which developed on the 

 lines of Galton 's description, and found ready accept- 

 ance, the circulation is Incidental to the convergence ; 

 the convergence Is universal, the convection general. 



It is another example of the facilis descensus Averni. 

 The very simple piecing together of the three parts 

 makes it almost obvious that the third element, the 

 convection, is the effective cause of the whole 

 dynamical process ; it Is natural to regard convection 

 as the ascent of warm air in a relatively cold environ- 

 ment, causing low pressure on account of the rela- 

 tively high temperature of the ascending air; and 

 high pressure as the natural corollary of cold descend- 

 ing air. The convergence, or motion across the 

 isobars, is the primary result of the distribution of 

 pressure, and the circulation is merely the deviation 

 from the straight path caused by the rotation of the 

 earth. The theory is quite simple and quite self-con- 

 tained, and it has this great advantage : that the cause 

 which it assigns for the cyclone, namely, the convec- 

 tion of warmed air, has always been regarded as the 

 cause of winds ; it has been accepted as explaining 

 land- and sea-breezes, the trade winds and the mon- 

 soons ; and If it is also accepted as explaining the 

 cyclone and anticyclone, which are the modern 

 meteorological names for the diverse winds of the 

 temperate latitudes, we can see in the Idea a beautiful 

 unity in meteorological theory. The origin of all 

 the winds is thereby assigned directly to what we 

 know must be their ultimate cause, namely, the 

 warming of the lowest layers of the air by the 

 warmed surface of sea or land. If we doubt its 

 efficiency in one case, there seems no good reason for 

 holding to it In the others. 



It seems a pity that an illusion which apparently 

 does such good service should be shattered ; but it 

 cannot face the facts of the upper air. 



You will notice that the whole matter depends upon 



