DISTRIBUTION IN THE ATMUSPIIKKE OVKR ENGLAND. 277 



air, in its attempt to get back from A to C, will produce and maintain tin- distri- 

 bution of temperature which we know to exist when C lies about 9 km. high in an 

 area of low pressure and A at the same height in an area of high pressure. The 

 force must exist in tin- neiglilxmring levels to that of AC, but may decrease rapidly 

 in intensity as that level is left either above or l)elow. Furthermore, A and C are not 

 necessarily points, but may represent areas of large extent and of any form, regular 

 or irregular. The only difficulty is to find the source of the necessary horizontal 

 force. We can think of nothing save wind to supply this, and it could IKS supplied by 

 a wind of sufficient velocity and of sufficient lateral and vertical extent. If we guess 

 at the curvature, for after all it can only be a guess, we can calculate the required 

 velocity. Near the earth's surface winds of 50 miles per hour, and extending 

 laterally over 400 or 500 miles, produce differences of pressure sufficient to show on a 

 weather chart the isobaric system of a deep cyclonic storm. The slope of the isobaric 

 surfaces at 9 km. is greater, in the ratio of 3:2, and hence the winds must be 

 stronger, but the velocities need not exceed 75 miles per hour. But the momentum 

 of such a wind system is very great, and it is hard to see how it can be originated 

 and how destroyed. 



The Isothermal Region. 



The explanation that has l>een given of the temperature distribution carries with it 

 an explanation of the fact that the value of H e is greater over the high-pressure area 

 than over the low. For if we take a line starting from A and rising vertically, then 

 turning and descending vertically to C, the displacement of the air along such a 

 line from A to C will carry with it the displacement of any peculiarity of the 

 temperature gradient. For the same vertical displacement simply alters the tempe- 

 rature of each element by the. same amount, but leaves unaltered the difference of 

 temperature between any two elements of the vertical parts. Thus the isothermal 

 region, which begins where the abrupt change of the temperature gradient sets in, is 

 shifted bodily upwards over A and downwards over C, and the invariable phenomenon 

 of a high value for H e with the anticyclone and the low one with the cyclone is 

 produced. The slow warming by radiation over A does not alter the level, for 

 presumably the neighbouring strata are about equally warmed, but the readjustment 

 required to meet this warming carries the isothermal upward again, and thus a 

 considerable increase of its height may be brought about.* 



Velocity of the Wind at Various Levels. 



If this theory of the local circulation be correct, it follows that the winds must 

 continue upwards to a considerable height. In fact, it seems inevitable that the 



* May 30. Since this was written it has been found that at least for the English ascents there is a very 

 close relationship between the value of He and the pressure at 10 km. 



