1913-14.] 
105 
Principia Atmospherica. 
In this table the gradual conversion of a southerly component into a 
northerly component associated with higher temperature to the westward 
is very noticeable. 
It will be seen that the isobars above 4 kilometres are, roughly speak- 
ing, at right angles to those in the lowest kilometre, which is, of course, in 
accordance with the wind observations ; but that the isotherms, with some 
fluctuations, particularly in the second kilometre, are similarly arranged at 
the top and at the bottom. That is to say, the upper winds are flowing 
from the north-west with the higher temperature on the south-west side, 
while the lower winds are moving transversely from the south-west with a 
distribution of temperature parallel to that of the upper air, but in this case 
the isotherms are across the wind. 
These results are represented in fig. 4, which was originally drawn to 
the same horizontal scale as the larger chart of the Daily Weather Report, 
and it is clear that in the lowest stage the columns of warmer air brought 
in by the south-westerly current are being carried underneath the parallel 
columns of the upper current. Up to 4 k., where the wind has become 
westerly, we have a distribution which produces the same effect. The 
wind is always carrying warmer air under colder air, and as, by Proposi- 
tion 1, a southerly current tends to thicken and a northerly current to give 
way, the pushing under of the warmer air becomes more effective, until 
instability occurs and rainfall sets in. The irregularities which are shown 
in the distribution of temperature are probably due to previous convectioii a 
We have here, therefore, the assurance of rainfall conditions as the 
south-westerly wind pursues its course under the north-westerly in front 
of the approaching depression. The rainy condition of that part of a 
depression is thus directly accounted for. 
The characteristic rainfall of a cyclonic depression is generally associated 
with a general convergence of the surface isobars, but this hypothesis is 
difficult to follow into details, because the convergence is general over the 
area, while the rainfall is local. The analysis of the conditions of the upper 
air here set out shows that there is good reason for rainfall in the upper 
layers, to which the doctrine of general convergence cannot safely be held 
to apply. 
To the examples which are taken from Mr Cave’s work, I may add one 
for October 16, 1913, which was reported to me by Mr J. S. Dines in con- 
nection with his work for the branch Meteorological Office at South 
Farnborough. 
On that day, at Pyrton Hill, where the sounding was made, there was 
