The Diurnal Variation of Level at Kimberley. 
305 
consequence of the surface tension of water, will be from the warmer 
lower levels to the colder upper levels.''' In the heat of the day, on the 
other hand, the same surface tension will tend to direct the flow of 
moisture downwards, that is to say, so much moisture as is not 
evaporated. It will readily be allowed therefore that the movements of 
moisture in a damp earth are probably in some measure competent to 
set up some such diurnal oscillation of the pendulum as is actually 
observed. At the same time it is not so obvious that the same explanation 
will apply to Kimberley. The soil of Kimberley is not often very wet : 
mostly it is very dry, and it is a curious fact that when it is dryest, 
namely, at the end of the winter, the range of the pendulum's excursion 
from west to east is twice as great as it is in summer. 
Not so much connection as might have been expected can be traced 
between the variations of weather and the movements of the pendulum at 
Kimberley. Cloud and variations of barometric pressure are perhaps the 
most potent disturbers in a small way of the regular diurnal march of 
the pendulum. Whether the cloud disturbance arises because under a 
cloudy sky there is usually more moisture in the air, and so less 
evaporation from the soil, or whether it arises because the clouds 
mitigate the direct heat of the sun is a question. As to the baro- 
metric pressure, one would expect the elastic crust of the earth to 
fall under a high pressure and to rise under a low pressure. Hence 
we should expect the crust to sink under the increased load of air at the 
time of the morning maximum of pressure at, say, 10 a.m., and to rise 
under the lightened load at the time of the afternoon minimum of 
pressure, say, at 4 p.m. That is, we should expect the pendulum to swing 
away from the place at which the local time is 4 p.m., and to swing 
towards the place where it is 10 a.m. There are no very obvious indica- 
tions that such is the case. Nor does the pendulum respond always in a 
very marked fashion to the passage of a barometric depression. This is 
probably due to the fact that the usual type of South African storm passes 
across the country in a direction S.W.-N.E., and as often as not exag- 
gerates rather than obstructs the normal diurnal sequence of meteorological 
phenomena. Did our storms travel from W. to E., as certain amateur 
meteorologists assert, or from E. to W. as others believe, the effect upon 
the pendulum would of course be much more marked. 
* Dr. W. J. Humphreys in pointing this out (" Note on the Movement of Water in 
Soils," Bulletin of the Mount Weather Ohservatorxf, vol. i., part 2) remarks that the 
temperature effect upon evaporation, condensation, and surface tension tends to conserve 
whatever moisture is in the earth. Surely Nature is not quite so beneficent ! Would it 
not be more correct to argue that the temperature effect promotes loss ? For surface 
tension brings the underground water to the surface at night just where it may be the 
more readily evaporated by day. 
