April 27, 19 16] 



NATURE 



^93 



the idea of the low pressure in the warm ascending 

 air of the cyclone as the driving force, whatever be 

 the area covered by the circulation. The obser\'ations 

 of the upf)er air have made us familiar with certain 

 facts about the height of the atmosphere that make 

 such an idea too improbable. The convective atmo- 

 sphere is only about lo kilometres thick. The region 

 in which conv^^ction can operate is therefore a thin 

 skin represented by a centimetre in the case of a map 

 on the millionth scale, on which looo miles is about 

 6 ft. in length. A cyclone is often regarded as a 

 towering structure which may produce curious effects 

 by tilting its axis, but that is clearly illusory ; the idea 

 that descending air over northern France is operating 

 ■ conjunction with rising air over Iceland to produce 

 flow of air along the line joining them is an un- 

 productive way of representing the facts. 



The idea of the ordinary- cyclones and anticyclones 

 in our latitudes as foci of centripetal and centrifugal 

 motion is an illusion. In all ordinary cases of cyclone 

 the oonv^ergence of the paths of air towards the 

 centre is itself an illusion, because the motion of the 

 cyclone makes it miss its apparent aim, and we 

 get in actual fact paradoxical cases of air which, 

 always seeking a place of lower pressure, yet makes 

 its way to a place of higher pressure, because the 

 pressure has been raised over its path ; and though 

 it always seeks the centre, in reality it goes further 

 away from it. If it wanted to reach it, it was a 

 mistake to aim at it; if it wanted to get near, it 

 should have aimed to get away. There certainly is 

 convergence and convection, but it is local and not 

 general over the cyclone. The idea which is conveyed 

 by convergence in spiral paths to the centre of a 

 moving cyclone is an illusion. It did not even require 

 observation of the upper air to tell us that.- 



Take the time required for the operating forces 

 to produce any such wind velocities as we find in 

 actual experience. In one hour an ordinary pressure- 

 difference would produce a velocity of looo metres per 

 second if it were free to act. The time required to 

 generate a velocity of, say, lo metres per second is 

 infinitesimal compared with the time during which 

 we see the forces in operation ; these last for hours, 

 or even days, while a minute would suffice for the 

 production of all the velocities exhibited; the motion 

 of the air which we register on anemometers is not 

 accelerating motion but uniform motion, except for 

 the effect of turbulence and local conv^ection ; so we 

 must picture to ourselves the air of cyclones as being 

 under the operation of balanced forces, not unbalanced 

 forces. I wish to suggest that the idea of air being 

 I accelerated by the forces we see on the map is another 

 illusion so far as the upper air is concerned. 



The ostensible reason for supposing that the dis- 

 tribution of pressure created by convection is pushing 

 air from high to low is due to the fact that the charted 

 winds show the air at the surface crossing the isobars 

 from high to low; the observations with kites and 

 pilot balloons suggest that the effect is peculiar to the 

 surface. If the driving force from high to low were 

 the operative force which produces the wind of a 

 cyclonic depression, we should expect to find its opera- 

 tion more strongly marked as we get higher up, 

 because the friction of the surface would not interfere 

 with it; but the fact is quite otherwise. The move- 

 ment across isobars becomes less and less marked as 

 we -ascend. It is much less at Pendennis Castle than 

 It is at Falmouth Observatorv. a mile awav. We 

 cannot be sure that it exists at all at 1500 ft., because 

 we cannot draw the isobars at that level with the 

 necessar}' accuracy ; the consensus of our observations 

 goes to show that there is no real evidence of con- 



- f?^, " Life-histor>' of Surface Air-ruirents." Ry W. X. Shaw and 

 ^. K. l.empfert. M.O publication No. 174. 



NO. 2426, VOL. 97j 



vergence at that level. There the centrifugal force of 

 the air travelling over the moving earth, combined 

 with the centrifugal force due to the curvature of the 

 air's path, is sufficient to balance the force due to 

 pressure, and there is no component of motion towards 

 the centre.3 



What happens nearer the surface is that the friction 

 of the surface converts part of the energy of the 

 motion of the wind into eddy motion and the air does 

 not move fast enough on the right path to keep up 

 the balance. Consequently, it drifts inwards as a 

 pendulum does when its motion is retarded, but the 

 lower air cannot hold back the air far above it; the 

 effect of viscosity in that direction was shown by 

 Helmholtz to be negligible. The effect of the eddy 

 motion is very limited in height. 



Observations in the Upper Air in Relation to the 

 Convection Theory. 



But the greatest blow to the illusion that I have 

 portrayed comes directly from the observations of the 

 upper air; the convection theory- requires that the air 

 of the cyclone should be warmer than that of the 

 anticyclone, but, as a matter of fact, the new observa- 

 tions show that the opposite is the case. 



In a paper published by the Royal Society, Mr. 

 W. H. Dines * gave the mean values of the observa- 

 tions of temperature in the upper air of this countr\- 

 arranged according to the pressure at the ground. 

 From his results the following table has been com- 

 piled : — 



Table of Averaf^e Values of the Pressure, Tempera- 

 ture, and Density of Air vi High and Low Pressure. 



High pressure Low pressure 



Height . ■ s , ' . 



Pressure Temp. Densitv Densitv Temp. Pressure 



1 000- ft. k. mb. A g/m^ g m^ A mb. 



32809 10 273 226 421 382 225 247 



29-528 9 317 233 474 444 226 288 



26247 8 366 240 531 514 227 335 



22-966 7 422 247 595 583 232 388 



19-685 6 483 254 662 652 240 449 



16-406 5 552 261 736 724 248 516 



13124 4 628 267 818 807 255 591 



9843 3 713 272 911 893 263 675 



6-562 2 807 277 1 01 2 992 269 767 



3-281 I 913 279 1 137 HOC 275 870 



o o 103 1 282 1270 1226 279 984 



The figures show that a pressure-difference of 

 26mb. exists at the level of lo kilometres where con- 

 vection has ceased to exist. The difference is accen- 

 tuated to the extent of 2imb. as the surface is reached 

 by the existence of the high pressure transmitted from 

 above, in spite of the relative coldness of the air at 

 the lower pressure. The diagram included in Mr. 

 Dines 's paper showed that there is a remarkable 

 change at the top of the troposphere. Above the level 

 for which values are given in the table, the high 

 is colder than the low, reversing the state of things in 

 the troposphere. 



We cannot resist the conclusion that the pressure- 

 differences of cyclone and anticyclone are not local 

 surface effects at all* we must seek their origin in 

 the upper air where there is no convection. They are 

 little affected by the lower stratum of 9 kilometres, 

 which, roughly, marks the range of the effect of 

 heating at the surface. 



The idea of warm air in the lower layers causing 

 the low pressures which are recorded on our baro- 

 meters is therefore an illusion. 



Thus it will be seen that the observations of the 



5 See the four reports on wind structure to the Advisory Co-nmittee for 

 Aeronautics by W. X Shaw and J. S. Dines, also " Barometric Gradiert 

 and Wind Force," by Ernest Gold. M.O. Publication, Xo". 190. 



* See M.O. Publication No. 210b. Gcophy-sical Memoirs No. 2. 



