534 SCIENCE PROGRESS 



air of the surface is capable of rising to the heights which are 

 actually characterised by convection in the equatorial regions, 

 and that, if a hollow column could be filled with it and pro- 

 tected by a rigid wall from its environment, it would give rise 

 to a difference of pressure at the surface of the same order 

 as those found between the centres and margins of tropical 

 cyclones, and rather larger than is generally recorded." 



The writer then considers the manner in which convection 

 might be expected to achieve the results just described. He 

 points out that in the atmosphere convection apparently 

 proceeds either by " threads " or " bubbles." The " thread " 

 process is evidently operative on an ordinary sunny day in 

 dry regions and results in the heating of a considerable thickness 

 of air near the surface in such a manner as to produce a vertical 

 gradient of temperature appropriate for dry air which is in 

 convective equilibrium. With the " bubble " type of con- 

 vection, on the other hand, a large mass of warm air is pushed 

 up by the convergence of the air beneath it. It must be 

 supposed that as it moves upwards a certain amount of the 

 surrounding air is dragged with it, at the expense of some of 

 the ascensional force. By the passage of a succession of bubbles 

 in this manner the air originally over the area will have been 

 removed ; the external air will have converged towards a 

 central axis, and the beginning of revolving fluid will have 

 been set up. 



" Continued further, the same process will continue to 

 remove the internal portion of the revolving column until the 

 rotation has become sufficiently developed to resist further 

 convergence towards the centre. By that time, with the aid 

 of the original vorticity of the earth's rotation, we shall have 

 reached at all levels the condition of a simple vortex with a 

 ring of maximum velocity, within which the pressure is kept 

 low through the continual removal of air by what may be 

 called the scouring action of the ascending bubbles. The axis 

 then becomes practically unapproachable because the air 

 that aims towards it is always deviated from its course. It 

 takes part in the circulation and misses the convergence. So 

 we get a dynamical system of great stability which admits 

 air to the region of the axis only along the immediate surface, 

 where the motion cannot reach the limit of protection, because 

 it is retarded by friction. 



" So far we have a warm core with an environment the 

 temperature of which, except at the very bottom, is governed 

 by the dynamical cooling due to the convergence towards 

 the axis. If the air of the environment contains sufficient 

 moisture, cloud will form ; and with the formation of cloud 

 instability is probable, which will cause further condensation 



