Expansion and Contraction of Building Stone. 139 
tric state of the stone was not recorded, we can take no account of 
it in our deductions. ‘These discrepancies, however, will have but 
little effect upon the general result ; for it will be observed, that there 
is always an increase in the length of stone for an increase of tem- 
perature, when any two experiments are considered which are re- 
moved from each other by several degrees. 
From the facts ascertained concerning the expansion of other sub- 
stances, we may assume that the expansion of stone is uniform, and 
that, within the range of our experiments, each of the stones increas- 
ed in length by a common difference for each degree of the ther- 
mometer. ‘To find an approximate value for this common differ- 
ence, say for the granite, we subtract the first observed length from 
the last, and, if these experiments were accurate, the difference 
-0470, would be ninety six times the common difference ; ninety six 
being the difference in degrees between the extreme temperatures : 
the same operation being performed with the second experiment, 
and that next the last; the difference .0298, (the difference in 
lengths,) should be eighty six times the common difference. By 
thus comparing the extreme experiments of those which remain, we 
obiain the following table. 
Experiments. aseeen Diff. in lengths. 
1 and 31 96 + .0470 
g 2and 30 86 -+ .0298 
3 and 29 82 |° +.0433 
4 and 28 80 + .0428 
5 and 27 78 +.0501 
6 and 26 76 +.0406 
Tand 25 74. + .0466 
8 and 24 72 ++ .0373 
9 and 23 48 -+ .0180 
10 and 22 36 + .0063 
bl and: 3 25 — .0001 
12 and 20 | 20 +..0125 
13 and 19 16 + .00354 
14 and 18 16 — .0034 
15 and 17 11 — .0034 
Lota Care .3708 
We have neglected the sixteenth experiment, because we cannot 
employ it without using some other experiment twice, thus giving 
the latter an undue influence; and because the middle term should 
have the least weight in determining the common difference. 
