eS a 
i a oe 
PRESENT STATE OF OUR KNOWLEDGE OF THE UPPER ATMOSPHERE. 101 
1908, indicated that even to heights greater than 10 km. the wind had a 
component directed outwards from the region of high pressure, or was 
parallel to the general direction of the surface isobars and in the sense of 
the gradient wind at the surface. On January 3, 1908, on the other hand, 
the direction of the wind over Munich changed after 3-4 km., and the flow 
above this height up to 8 km. was directed inward towards the region of 
the surface high pressure. The English ascents indicate that the relative 
coldness of cyclones ceases at a lower altitude there than over the continent, 
and this tends to support the idea that the energy of the cyclonic motion 
is used up in extending the cyclone to greater heights and is gradually 
converted into the potential’energy of the anticyclone. Finality can be 
reached only by an examination of individual cases in which the observa- 
tions are extensive enough to furnish a good representation of the distribu- 
tion of pressure and wind at great heights. 
The results so far obtained show the need that exists for a series of 
ascents in the middle of the great Atlantic low-pressure system simul- 
taneously with ascents in Europe and America. The general drift of 
registering balloons is from high to low pressures, although there are excep- 
tions which are possibly due to the balloon entering at high altitudes a 
westerly current, which is caused by the general temperature and pressure 
distribution over the earth and may at times remain unaffected by shallow 
disturbances near the surface. The greater relative humidity over cyclones 
would tend to diminish the intensity in the upper air, but it is quite in- 
sufficient to bring about a reversal of the gradient between high and low 
pressure areas. 
For the surface layers Gold ' showed (1) that near the centre of cyclones 
the gradient of temperature up to 2 km. coincided very nearly with the 
adiabatic gradient for saturated air; (2) that in winter the gradient in 
the central region of anticyclones up to 3 km. was quite irregular, tem- 
or increasing and decreasing in different layers in different ascents, 
ut, on the whole, varying little from the surface value ; (3) that in summer 
the gradient in the central regions of anticyclones was regular in the first 
kilometre and nearly equal to the adiabatic gradient for dry air, but that 
above this level the fall of temperature was frequently arrested, showing 
that the vertical circulation was purely a surface phenomenon and was not 
connected in any way with a general descending current of air. This 
shows that the air up to a height of 3 km. in anticyclones is practically 
an inert mass taking little part in the general circulation. The result 
may be compared with the deduction arrived at by Shaw and Lempfert ” 
from a consideration of the air currents at the surface. They say, ‘ We 
have failed to identify the central areas of well marked deihoytones as 
regions of origin of surface air currents. . . . These latter are for the most 
part inert and comparatively isolated masses of air, taking little part in 
the circulation which goes on around them.’ ... ‘ The areas of descending 
air seem to be (a) the shoulders or protuberances of anticyclones, in par- 
ticular the regions of comparatively high pressure between two consecutive 
cyclonic depressions, and therefore also between two anticyclones or (0) the 
extension of an anticyclone between a depression and its secondary.’ If 
there is descending air in the upper atmosphere over an anticyclone (as 
indeed there must be if it maintains or increases its intensity) this air will 
not be considerably affected by radiation between 5 and 10 km., and the 
! Barometric Gradient and Wind Force, M.O. No. 190. 
2 Life History of Surface Air Currents, M.O. 174, p. 24. 
