ot 
mencement, have supplied fairly reliable data for the determination 
of the earth’s figure, for they were already assisted by the highly 
developed theories of geodesy. 
Thus it comes about that even now, at the close of the same cen- 
tury, we can still use, for the study of gravity on the earth, the 
material obtained at the beginning of this century; indeed, it is 
almost all that we can use, for new material only exists to a small 
extent, and this does not always exceed the old data in quality and 
trustworthiness. 
In the attempts to make gravity determinations useful to geodesy, 
two periods are to be distinguished ; one at the beginning, and one 
at the end of our century. These are separated from each other by 
a long interval, during which nothing, or very little of any use, 
was accomplished. 
In the first period are included those numerous and excellent de- 
terminations of gravity to which we owe, for the most part, our 
present knowledge of the figure of the earth as indicated by grav- 
ity determinations, and which also afford us information as to the 
distribution of gravity over the earth generally. 
The names of those men who have furnished these important and 
valuable materials to science are well known to every one. 
This fertile period comes, we may say, to a sudden termination 
with the fundamental investigations of Bessel. 
The second period of these inquiries, which brings us down to 
the present time, was developed, first, by the measurement of de- 
grees of the meridian in Europe, which has now become the inter- 
national measurement of the earth, and has included gravity 
determinations in its programme. 
During the second period of these inquiries, the determinations of 
gravity in Europe have yielded but few good results, inferior to those 
of the first period as regards accuracy, because it was supposed that 
the relative determinations which were previously employed, and 
which are so trustworthy, might be replaced by absolute measure- 
ments. But however accurately adso/ute measurements are carried 
out, they are always affected by numerous, and for the most part 
also by greater, errors than the relative ones; they are therefore but 
little, if at all, suited for the investigation of details; for the un- 
avoidable errors of the absolute determinations are mostly larger 
than the very small differences which they are intended to ascertain. 
Moreover, there exist in the apparatus employed defects by which 
