18 SECTIONAL ADDRESSES. 
place and from time to time, but one remarkable result has come out of 
the investigation, and that is that the average decrease is practically the 
same in all parts of the world. Near the ground the conditions are com- 
plicated ; here the rate of decrease is largely affected by such factors as 
the kind of surface, whether land or water, the time of day and the 
time of year. If we omit for this reason the two lower kilometres of the 
atmosphere, we are able to state that the rate of decrease of temperature 
with height, to which I shall refer as the ‘lapse rate,’ is the same in all 
parts of the world, from the equator to the poles. The lapse rate is not 
the same at all heights, but increases regularly as one ascends. Between 
two and four kilometres above sea-level the rate of decrease is 5°6° C. for 
each kilometre of ascent; the rate is greater at greater heights, until 
towards the top of the troposphere, say between six and eight kilometres, 
the rate is 7°1° C. per kilometre. 
The importance of these results lies in the bearing they have on the 
possibility of vertical motion in the atmosphere. Whether air will rise 
or fall as the result of differences of temperature depends not only on an 
initial difference of temperature but also on the lapse rate in the sur- 
rounding atmosphere. When dry air rises its temperature falls on 
account of adiabatic expansion 10° C. for each kilometre of ascent. From 
the observed values of the lapse rate given above it will be seen that if a 
mass of air is as much as 10° C. warmer than its surroundings it cannot 
rise much more than two kilometres before it has no buoyancy left. The 
question of ascending and descending air is, however, very complicated 
on account of the condensation of the water vapour carried with it. The 
vertical motion of the atmosphere cannot be determined simply from 
consideration of the lapse rate of temperature in the atmosphere. We 
have also to take into account the pressure and vapour content of the 
moving air. This can best be done by considerations of entropy. 
Sir Napier Shaw has prepared diagrams showing the entropy through- 
out the normal atmosphere. These show surfaces of constant entropy 
which are nearly horizontal, but they slope upwards from the equator, to 
the poles, especially in the lower layers (Fig. 1).’ If these surfaces could be 
made visible, we should see a series of layers lying one above the other 
like the strata in a geological specimen of stratified rock. 
The advantage of this method of representing the thermal structure 
of the atmosphere lies in the fact that entropy in the atmosphere bears a 
close analogy to density in an incompressible fluid. Just as a fluid in 
equilibrium sets itself with the layers of equal density horizontal, so the 
atmosphere is in equilibrium if the isentropic layers are horizontal. Also 
when a portion of fluid is displaced it will return to its appropriate density 
layer, so a mass of air displaced will also return to its appropriate entropy 
layer. 
Any mass of air retains its initial entropy no matter what its position 
in the atmosphere, unless heat is added to it or extracted. In the former 
case the entropy is increased while in the latter it is decreased. Adding 
heat to air is, therefore, analogous to changing the density of an incom- 
pressible fluid. It must, however, be remembered that increasing entropy 
is equivalent to decreasing density, so that for equilibrium the numerical 
value of the entropy layers must merease upwards. 
1 Facing p. 24. 
