616 Sir Napier Shaw [March 10, 



a hundred soundings on the Continent, to 0.88 for soundings in 

 England grouped for the winter season. Moreover, the standard 

 deviations are of the same order of magnitude at both levels, that 

 is to say, both levels are subject to similar changes. At the same 

 time, the correlation co-efficient between the pressure at the surface 

 and the mean temperature of the 9 -kilometre column is small ; in 

 other words, the temperature of the lower strata of the atmosphere 

 has on the whole little to do with the general distribution of surface 

 pressure in this country. Its effects are local. 



We must, therefore, regard the general flow of air, except in so 

 far as it is disturbed by convection, as governed not by what happens 

 at the surface but by what is imposed upon it from the stratosphere 

 above. It is from there that the general control of the distribution of 

 our pressure comes. It is only modified by what happens below. The 

 upper air, the stratosphere, is the operator, and the lower air the 

 subject operated on. After fifty years of strenuous endeavours to 

 regard the surface as the operator and the upper air as the subject, 

 the exchange of role is very disturbing, but it has its compensations. 

 There are many things which can easily be explained by operations 

 from above, but only with the greatest difficulty by operations from 

 below. Let us indulge in some speculations which follow from 

 supposing that the stratosphere operates upon the troposphere. It 

 makes the troposphere as tuneful as an organ under the alternating 

 rarefaction and compression caused by the changes in the stratosphere. 

 Every cloud is an expression of its action. One can imagine them 

 being developed, like photographic plates, showing first the region of 

 greatest humidity, and then developing further into loci of instability, 

 and remembering each several cloud means the disturbance of the 

 normal circulation, the condensation will alter locally the horizontal 

 distribution of temperature and therefore that of the pressure and 

 wind. On the table are two autochrome photographs of the Western 

 sky at Ditcham Park with a quarter of an hour's interval, taken on a 

 September evening in 1911, with gradually reddening clouds that 

 gradually vanished as they approached from the West. Nothing 

 could be more attractive than to speculate upon such changes in 

 relation to the changes of pressure in the stratosphere. 



The Regime of the Stratosphere. 



Bat our new point of view only shows our problem removed one 

 step further ; we have now to begin again and imagine for ourselves 

 what is the regime of pressure and winds in the stratosphere until 

 the enterprise of meteorologists completes our knowledge of what it 

 actually is. The problem is at any rate much simplified because 

 convection is avoided ; we deal with an atmosphere which, being 

 nearly isothermal, is inherently stable, density goes directly with 



