November 30, 191 1] 



NATURE 



141 



Der Fanamakanal. By Max D. Fiegel. Pp. viii + 

 183. (Berlin : Dietrich Reimer— Ernst Vohsen, 

 191 1.) Price 4 marks. 

 Mr. Fiegel describes clearly the course of the canal 

 now in process of construction by the United States 

 Government ; also the ' engineering- works and 

 machinery, and the commercial and political aspects of 

 the enterprise. His book will provide German readers 

 with an informative account of the position about a 

 rear ago of what in three years' time promises to be 

 one of the most remarkable human schemes ever 

 brought to a successful conclusion. 



LETTERS TO THE EDITOR. 



{The Editor does not hold himself responsible for opinions 

 expressed by his correspondents. Neither can he undertake 

 to return, or to correspond with the writers of, rejected 

 manuscripts intended for this or any other part of Nature. 

 No notice is taken of anonymous communications.] 



The Weather of igii. 



Sir Edward Fry's letter in Nature of November i6, and 

 Commander Hepvvorth's reply in the issue of November 23, 

 prompt me to give expression to some ideas which I have 

 been discussing with various meteorological friends during 

 the last few months. What I have to say will not attempt 

 in explanation of the phenomena of the remarkable summer 

 nd autumn of this year, but it will give an indication of 

 I he direction in which, for my part, I hope to look for an 

 'xplanation of those phenomena. 



I agree with Sir Edward Fry that the prevalence of anti- 

 vclones in a particular direction, or the continuance of 

 ;iven winds, are only part of the phenomena to be ex- 

 ilained ; I think, too, that the statement may be rightly 

 inderstood in a sense slightly different from that which 

 Sir Edward intended. I mean that this summer furnished 

 a good deal of evidence for a proposition which includes the 

 inference that the distribution of pressure is not in itself 

 a complete explanation of the weather. I will state my 

 proposition now, and afterwards explain why I make it. 

 It is that the main outlines of the distribution of pressure 

 are imposed upon the surface layers of the atmosphere by 

 transmission from a region g or 10 kilometres high — a 

 region which is above what may be called the physical 

 laboratory, where rain and thick clouds are made — and that 

 the phenomena of weather are due, not to the mere exist- 

 ence of the air currents which correspond with the distribu- 

 tion of surface pressure, but to their heterogeneity. 

 Weather as represented primarily by rainfall is dependent 

 on the convection of moist air, while pressure distribution 

 is governed by changes which take place above or nearly 

 at the top of the convective region of the atmosphere. In 

 other words, the dynamics of the atmosphere is controlled 

 in the upner air, while the phvsics of the atmosphere is a 

 matter which concerns the lower layers. 



The remarkable summer has provided evidence in support 

 of this proposition by furnishing a number of examples of 

 pressure distributions which might well have been rainy, 

 and were not. The Coronation festivities were somehow 

 preserved from the copious rainfall which, according to the 

 pressure distribution, was their due. If pressure distribu- 

 tion is the cause of rainfall, 100,000 children at the Crystal 

 Palace on June 30 ought to have got wet through, but they 

 did not. There are many other instances of the same kind 

 which I need not quote. Let us look at the matter from 

 the other side. 



.Some time during the summer Mr. W. H. Dines sent to 

 me the results of some work which he has done upon the 

 correlation of various data for the upper air. .'Xmong them 

 was the correlation coefficient between the variations of 

 pressure at the surface and at the level of 9 kilometres. 

 For certain groups of ascents it was so large as to show 

 a close approximation to pronortionality. That is, of 

 course, not surprising, because the variations at the surface 

 and at kilometres are certainly not independent ; but 

 what was surprising to me was that the standard deviations 

 of pressure at the two levels in the rn«'-'s und^r invf>stign- 



No. 2iq6. vol. SSI 



tion were very nearly equal. Thence it follows that the 

 pressure variations at 9 kilometres level (with two-thirds 

 of the atmosphere below it) are not merely proportional to 

 the variations at the surface, but the same in magnitude ; 

 and as they are certainly transmitted to the surface, it 

 follows, further, that the variations at the surface are 

 practically accounted for by the variations that occur ar 

 9 kilometres. We are accustomed to urge the importance 

 of the study of the upper air for increasing our knowledge 

 of meteorology ; but, so far as I know, we have not recog- 

 nised that it was so directly responsible for steering our 

 surface air currents, 



Mr. Dines 's variations were those shown between the 

 individual balloon ascents and their average. Looking into 

 a series of charts for the upper air recently published by 

 Prof. Rotch, I came upon another step in the proof. In 

 the charts Prof. Rotch gives the average wind velocity at 

 30,000 feet (9 kilometres) and the atmospheric density there. 

 The product of these two is about the same as for a point 

 near the surface, whence it follows that for Blue Hill the 

 pressure gradient at 9 kilometres, which is proportional to 

 the product, is the same as for the surface ; in other words, 

 not only are the chronological changes transmitted to the 

 surface from 9 kilometres, but the average pressure dis- 

 tributions are similarly transmitted. In a discussion at the 

 Meteorological Office on October 23 I was reminded that 

 these conclusions are not new. The inverse proportionality 

 of velocity and density of air is known as Egnell's law ; 

 and the approximate constancy of pressure gradients Up to 

 9 kilometres was pointed out to the British Association by 

 Gold and Harwood in 1909. But the idea of looking to 

 the level of 9 kilometres for the outlines in full scale of 

 our surface distribution of pressure is new ; and it seems 

 to me to be possibly the beginning of a new era in the 

 endeavour to explain such phenomena as those of the past 

 summer. 



I cannot, at this stage, give particulars as to the details 

 of the application of so general a proposition to special 

 cases, such as circular revolving storms, the northern sides 

 of which may be confined to the lower strata, nor can I 

 say whether the application of the proposition is limited 

 to certain parts of the world. I think it must be. It 

 will be remembered that M. Teissercnc de Bort computed 

 mean isobars at 4 kilometres for January and July that 

 showed the average circulation of the upper air in each 

 hemisphere as a great cyclonic depression, with centres at 

 the poles. It is not likely that there is any great changr 

 of distribution between 4 kilometres and 9 kilometres. At 

 Blue Hill the winds at 3 kilometres vary between W.S.W. 

 and N.N.W., and these, again, should agree with the 

 pressure distribution. We know from the study of the 

 points where sounding balloons land that the westerly 

 circulation is not always to be found aloft. We know, also, 

 that at 9 kilometres the variations of temperature from day 

 to day are as large as, or larger than, those at the surface. 

 Hence we may conclude that the pressure distribution af 

 9 kilometres corresponds with a cyclonic circulation of 

 westerly winds round the pole, periodically, but perhaps 

 not regularly, invaded by winds from some northerly 

 quarter, with marked changes of temperature. This will 

 be recognised as merely a rough description of a series of 

 V-shaped depressions, which also, on the average of the 

 month, would give a westerly circulation. Possibly, in 

 realitv, the V-shaped depressions at 9 kilometres are made 

 up of comparatively warm westerly winds with repeated 

 incursions of cold air from the north or north-west, giving 

 phenomena similar to those which have been described in 

 recent discussions of line squalls. In the results of Mr. 

 Cave's observations of pilot balloons there is evidence that 

 V-shaped depressions at the surface sometimes originate 

 with northerly winds at high levels. 



If this is so, our failure to explain the outlines of the 

 distribution of pressure by means of the surface conditions 

 is itself explained. Mr. Dines's recent paper before the 

 Royal Society shows how futile is the endeavour to explain 

 them by reference to temperatures in the lavers below 

 9 kilometres. Thev come from above, and their shape at 

 any time is governed bv causes in the ronsider.ntion of 

 which we must treat the globe as a whole. Our first step 

 in explaining, for example, the recent succession of_ gales 

 w.nild hr to note whether the westerly circulation in thf. 



