248 



NATURE 



[November 6, igig 



sequences, and the principle of ■post hoc propter 

 hoc held full sway; the laws of motion and the 

 more recently discovered laws of thermodynamics 

 were in most cases completely ignored, or at least 

 considered as not being- applicable to meteorology. 

 This has been largely changed for the better, and 

 one does not now expect to find a cold area ex- 

 plained as being due to the descent of air in an 

 anticyclone from a higher and colder region. 

 Perhaps the pendulum has swung too far the 

 other way, and mathematical analysis may some- 

 times be used when it is not applicable. On 

 the assumption that air is a perfect fluid, it follows 

 from a strict mathematical analysis that a sphere 

 exposed to a steady current of wind will offer 

 no resistance to that wind — a result obviously 

 inconsistent with the facts. The assumption made 

 cannot be justified, and one cannot help feeling 

 that great caution should be used in making 

 assumptions if the result of a complex mathe- 

 matical investigation into a me^orological ques- 

 tion is to be trustworthy. Mathematics, however, 

 afford a most useful and often indispensable 

 aid to meteorology, and of late years espe- 

 cially, although far from exclusively, by their 

 means many useful deductions have been 

 drawn. 



It is impossible in a brief article to give any 

 full statement of the present position of meteoro- 

 logy, but a short account of the great access of 

 knowledge that has come to us in the last fifteen 

 years or so by means of observations in the upper 

 air may be of interest, the more so because the 

 great central problem of meteorologists who live 

 in temperate latitudes has always been the genesis 

 and motion of cyclones and anticyclones which 

 bring us our various types of weather, and this 

 problem is most intimately interwoven with the 

 upper-air observations. 



A mass of detail remains to be filled in, but 

 the salient facts of the distribution of tem- 

 perature in the upper air are well established, 

 and, at least for Europe, where some 1 500 observa- 

 tions are available, are beyond dispute. We have 

 also observations from Canada, the United States, 

 and Batavia, and a few from Central Africa and 

 the tropical Atlantic. 



It has been found that the atmosphere is divided 

 into two parts : a lower part, the troposphere, in 

 which there is a lapse rate — that is, a fall of tem- 

 perature with height — of about 6° C. per kilometre 

 (17° F. per mile); and an upper part, the strato- 

 sphere, in which there is no appreciable change of 

 temperature with height. The boundary between 

 the two parts is in these latitudes quite sharp and 

 distinct, but is not so well defined in the tropics. 

 Its height varies with the latitude : for the South 

 of England the mean is 106 km. ; for Scotland it 

 is 98 km. ; and for the equatorial regions it 

 reaches 16 km. It has also for temperate lati- 

 tudes an annual variation, rising in the summer, 

 falling in the winter. It should be added that the 

 usual lapse rate is less than 6° per kilometre in 

 the first three or four kilometres, is more than 6° 

 NO. 2610, VOL. 104] 



above that height, and in regions of excessive 

 cold, such as Canada or Siberia in the winter, may 

 be absent or reversed in the lower strata. With 

 regard to temperature, over the equator the 

 stratosphere may be as cold as —80° C. ; over 

 Europe it has about —54° C. for its mean, but 

 may vary from —40° to —70° C. 



Confining, now, our attention to Europe, there is 

 very little or no correlation between the tempera- 

 ture and the barometric pressure of the air at the 

 surface, but a totally different set of conditions is 

 met with as soon as the very lowest stratum — the 

 first 2000 ft., say — is passed. From i km. 

 and upwards there is a very high correlation 

 indeed between temperature and pressure ; 

 between 4 and 8 km. the correlation coeffi- 

 cients are more than 085 ; they then fall off 

 rapidly, so that there is again no correlation at 

 the boundary between the troposphere and strato- 

 sphere. Above this, in the lower part of the 

 stratosphere, the correlation is negative and 

 reaches —0-30, but falls off with increasing height. 

 Also the correlation between the pressure at 

 9 km. and the temperature at any height 

 excepting the surface and the common boundary 

 is very high, being positive for the troposphere 

 and negative above 12 km. Since a low- 

 pressure area at the surface remains so up to 

 nearly 20 km., the correlation defined above 

 leads to the following rules. In a cyclone the 

 troposphere is relatively cold and the stratosphere 

 warm, and, it may. be added, the boundary between 

 the two is much lower than usual. In an 

 anticyclone the troposphere is warm and the 

 stratosphere cold ; also the common boundary 

 is raised. The actual differences of temperature 

 between a well-marked cyclone and anticyclone 

 in the British Isles are about 10° C, the cyclone 

 being 10° cooler from 3 to 8 km., and the anti- 

 cyclone 10° cooler from 12 km. height and up- 

 wards. In the cyclone the common boundary wi 

 3 to 4 km. lower than in the anticyclone. 3^1^ 



The cause of these differences is still more or 

 less a matter of conjecture and controversy. In 

 my opinion the changes of pressure at heights 

 of 8 or 9 km. are in some way brought about by 

 the accumulated momentum of the general circu- 

 lation, and the temperature changes that follow 

 are easily explained by the laws of mechanics 

 and thermodynamics. Thus I think that tem- 

 perature changes in the upper air are the 

 results, and not the causes, of cyclones and anti- 

 cyclones. 



In addition to the results obtained by observa- 

 tions of temperature and humidity by means of 

 registering balloons, much work in the last fifteen 

 years has been done by means of pilot balloons. 

 A large portion of this remains to be worked up. 

 Also a considerable advance has been made from 

 the theoretical side in our knowledge of the 

 motion of the air particles near the centre of a 

 cyclone, and meteorologists have good cause for 

 congratulation in the steady progress that is 

 taking place. 



