310 Proceedings of Poyal Society of Edinburgh. [sess. 
eight years’ observations already published by Piazzi Smyth. These 
I have used as they were given, without any regard to the probable 
corrections. As my object was, however, to get an approximate 
estimate of the amount of heat stored in the rock at different 
times, and not to discuss the conductivity of the material, it was 
not necessary to pay much attention to comparatively small errors 
of observation. The probable heterogeneity of the different layers 
and the surface irregularities of the rock itself will give rise to 
disturbances as important as any that might arise from neglect 
of slight and (as Mr Heath has pointed out) not very certain 
corrections. 
It would be of great interest to apply similar calculations to 
underground temperatures in other parts of the globe, especially in 
parts which are blessed with fairly steady sunshine. 
In regard to the general form of the curves of underground 
temperature, there is one feature which I do not remember to have 
seen commented upon. The feature is apparent in all, but most 
evident in the curve for the thermometer nearest the surface. It 
is the sharpness of the crest as compared with the trough. The 
reason of this is at once recognised when we observe that exactly 
the same feature is distinctly characteristic of the lower solar radia- 
tion curve, but not so of the higher curve. In other words, in the 
higher latitude the low altitude of the sun and the shortness of 
the day combine during the winter months to produce a marked 
effect upon the law of absorption of solar energy. In lower 
latitudes this effect is hardly appreciable, and at the equator a per- 
fectly symmetrical semi-annual variation of comparatively small 
amplitude is to he expected. It is instructive to compare the annual 
variations of solar radiation already given for two different latitudes 
with the corresponding variation at a place on the equator. The 
results, obtained in exactly the same way, are as follows : 
