614 Sir Napier Shaw [March lo, 



round its sun is taken as a type for the air instead of the motion of a 

 falling stone. 



While we are considering illusions, let me add another example 

 depending upon what was at one time, and possibly is still, a com- 

 monplace of physical teaching in regard to the relation of barometric 

 changes to weather. 



It is this : moist air is lighter, bulk for bulk, than dry air, and 

 consequently pressure is low where the air is moist. That is why a 

 low barometer is indicative of rain ; the moist air causes the low 

 pressure. This is not true to fact. Mr. Dines has recently examined 

 the correlation between the humidity of the troposphere and the 

 pressure at the surface — the coefficient is quite insignificant ; there 

 is no relation between moist air and low pressure on the map. 



Structure of the Atmosphere according to the Ohservations 

 of the UiJiier Air. 



But if the ideas which were common in meteorological practice 

 fifty years as^o are now to be regarded as illusory, let us consider what 

 we have in their place. We go back to the three elements, the circula- 

 tion, the convergence and the convection. As to the circulation, we 

 now think of it as it is exhibited in the upper air, and instead of 

 regarding it as an incidental disturbance of the motion from high 

 to low, we regard it as the foundation of atmospheric structure ; as 

 the motion of air which is persistent because the pressure-gradient 

 is balanced by the centrifugal action of the earth's rotation, which we 

 may call the geostrophic component, and of the curvature of the path 

 over the earth's surface, which we may call the cyclostrophic com- 

 ponent. If the balance between velocity and pressure is not perfect 

 the difference from perfection can be only infinitesimal, because in 

 the free atmosphere the air must always begin to adjust itself to 

 the strophic balance from the moment that any infinitesimal change 

 becomes operative and the power of adjustment arising from the 

 extreme mobility of the air prevents any finite perturbation being set 

 up, except temporarily in those regions where violent convection is 

 operative. It is only through the mobile air that perturbation can 

 be transmitted. AYe no longer picture to ourselves the air as being 

 somehow held firm without moving until a pressure distribution is 

 set up and then let go ; the first symptom of pressure-difference will be 

 the occasion of motion, the distribution and velocity grow together ; 

 they adjust themselves automatically. The whole history of the 

 general motion of the atmosphere is the story of the constant pursuit 

 of the strophic balance, the adjustment of velocity to pressure, 

 constantly disturbed by infinitesimal changes. 



Near the surface things are much more complicated, because there 

 is turbulence due to the interference of the surface, and the obstacle 



