r.)l;-i] on The Winds in the Free Air 721 



some cases the wiud at 2 or :-) kilometres is double the srvadient value, 

 or even more. This type is likely to occur when there is a low 

 pressure to the north of the station and when there is a strong tem- 

 perature gradient, such that the low temperatures correspond to the 

 low pressures, and vice versa ; such conditions should theoretically 

 cause an increase in wind velocity in the upper aii', but it is not 

 possible to calculate what the effect should be without knowing the 

 temperatures, not only on the surface, but in the upper air over the 

 region in question. One may, however, calculate what effect surface 

 temptratures would have on the isobars at, say, ? kilometres, assuming 

 that the vertical temperature gradient is the same at every point ; a 

 map constructed to show the isobars, which have been thus calculated, 

 must be looked on as a rough approximation only to the real condi- 

 tions. A map of the isobars at 3 kilometres for May 11, 1907, shows 

 how much steeper was the gradient on this day in the upper air than 

 it was on the surface, a fact which quite accounts for the rapid 

 increase in wind velocity from 2 metres per second at the surface to 

 ly metres per second at 3 kilometres. 



Winds belonging to this class may come from any point of the 

 compass. 



The third class comprises those cases in which the wind, after 

 reaching the gradient velocity in the first h kilometre or so, falls off 

 more or less rapidly in the upper air. This class is almost entirely 

 associated with easterly winds on tlie surface, when there is high 

 pressure to the north and low pressure to the south. An east wind 

 is usually, though not always, a shallow one ; a south-west gale 

 increases in the upper air, but when an easterly gale is blowing, 

 causing such high seas and such dangers to shipping, it is curious to 

 reflect that such a short distance up we should meet with light 

 breezes, or even a complete calm. 



We now come to the class of reversals when the wind in the upper 

 air is very different in direction from that near the surface, and when 

 it often bears no relation to the surface pressure distribution. In a 

 typical case, after an initial increase for a short distance above the 

 surface, we find the wind gradually decreasing as we ascend, till a 

 layer is met with, in which there is a complete calm ; above this we 

 find an entirely different wind, which usually increases as we go 

 higher, as in the case of winds in the second class. It looks at first 

 sight as though there were a discontinuity in the atmosphere, but I 

 hope to show later that this is more apparent than real. A typical 

 example of a reversal occurred on November 6, 190^, when the sur- 

 face wind was easterly with a velocity of 17 metres a second, just 

 below one kilometre ; al)ove this it fell off to a complete calm at 

 3 kilometres ; at -4 kilometres there was a light north-west wind, 

 which increased to a wind of 15 metres per second at 10 kilometres. 

 The weather map for this day is remarkable : over this country there 

 is no sign on the surface of the westerly wind above, but it appears 



