6 THE PRESIDENTIAL ADDRESS. 
surface which most nearly satisfies this requirement. Of more recent 
interest are the irregularities in the intensity of gravity, which have been 
found to exist over wide areas, by the highly trained Survey of India, by 
the Coast and Geodetic Survey of the United States, and by various 
observers on the continent of Europe. Briefly, the general result is this, 
that in mountainous regions the observed value of gravity is abnormally 
low, whilst on oceanic islands, and so far as can be ascertained on the sea, 
it is abnormally large, when all allowance has been made for altitude and the 
normal variation with latitude. The fact that this has been found to be the 
case in so many different places, shows that we have here to deal with no 
casual phenomenon. The accepted explanation, originated by Archdeacon 
Pratt, of Calcutta, in 1859, and since developed especially by Hayford and 
Bowie, of the U.S. Survey, is that if we imagine a level surface to be drawn 
at a depth of about 100 kilometres, the stratum of matter above this, 
though varying in density from point to point, is approximately uniform, 
in the sense that equal areas of the surface in question bear equal weights. 
The altitude of the mountains is held to be compensated by the inferior 
density of the underlying matter, whilst the oceanic hollows are made up 
for by increased density beneath. Leaving aside the technical evidence on 
which this hypothesis is based, there are one or two points to be noticed. 
In the first place it suggests, as is highly plausible on other grounds, that 
the matter in the interior of the earth, below the stratum referred to, is in 
a state of purely hydrostatic stress, 7.e. of pressure uniform in all directions. 
So far as this stratum is concerned, it might be floating on an internal 
globe of liquid, although no assertion is really made, or is necessary, to 
this effect. But in the stratum itself, shearing forces must be present, and 
it is necessary to consider whether the actual material is strong enough to 
withstand the weight of continents and mountains, and the lack of lateral 
support due to the oceanic depressions. The researches of Professor Love 
and others show that this question can fairly be answered in the affirmative. 
The accurate determination of the acceleration of gravity at any place 
is, of course, a matter of great delicacy. Not to mention other points, 
in the pendulum method the yielding of the support due to the reaction of 
the pendulum as it swings to and fro affects the time of oscillation. It 
may be recalled that so far back as 1818 Kater, in his absolute determina- 
tion of the length of the seconds pendulum in London, was on his guard 
against this effect, and devised a test to make sure that it was in his case 
negligible. In a portable apparatus, such as is used for comparative 
determinations, it is difficult to give sufficient rigidity to the support, and 
