ELASTIC CONSTANTS OF ROCKS. 21 
carefully faced and absolutely parallel to one another. Before the actual 
measurements were made, the rock in every case was brought to a "state of 
ease" in the manner already described. 
The pressure was applied in most cases by a loo-ton Wickstead testing 
machine, which was so carefully adjusted that it was sensitive to a load of 
4 pounds. 
The specimen, having been placed in the press and reduced to a state of ease, 
was then after careful adjustment submitted to loads increasing in successive 
stages of 1,000 pounds until the limit of safety had been reached, when 
the load was reduced successively by the same amounts, accurate readings 
being taken at each increment and decrement of load. The maximum 
load employed in the case of most -rocks was 9,000 pounds, equivalent to 
from 9,000 pounds to about 11,500 pounds per square inch, according to 
whether a square or round prism was employed. In the case, however, of 
some of the stronger rocks a load of as much as 15,000 pounds per square 
inch was employed. 
In the determination of the lateral strain, which was made upon the same 
set of columns as those used for measuring the vertical compression, care 
was taken that the theoretical conditions were realized, and that the material 
was free to expand laterally, as otherwise the values obtained for the lateral 
extension would be inaccurate. In all cases, therefore, the measuring appa- 
ratus was set as nearly as possible upon the central section of the test piece, 
and the ends of the specimen, after being ground smooth, were coated with a 
thin film of oil, so that the polished pressure plates of the machine would 
have as little tendency as possible to prevent freedom of lateral expansion. 
In a number of cases accurate measurements were taken during the suc- 
cessive cycles of loading and unloading to which the specimen was subjected 
in order to bring it to a state of rest. These are recorded in the case of 
the Baveno granite and the Stanstead granite and served to show how the 
hysteresis of the rock may be reduced to a minimum by subjecting the test 
piece to this process. The measurements of each cycle usually occupied 
from 10 to 15 minutes. 
It was at first conjectured that in the case of rocks composed of several 
minerals differences of reading might result from the attachment of 
the extensometer to different portions of the rock, the points of the instru : 
ment being fixed in some cases in grains of one mineral and in other cases in 
grains of another. It was found, however, as has already been mentioned, 
that the measurements on two sets of prism faces made in the manner above 
described, or on the four planes intersecting the vertical columns, where- 
these had been provided with eight pairs of holes, showed that in the case 
of the rocks examined the differences between the several measurements 
on the same prism seem to be unaffected by the circumstance above referred 
