A g AN INVESTIGATION INTO THE 
the set. The greatest differences obtained are those found in the finest 
grained rock in the series, viz, the Sudbury diabase. 
It is evident, therefore, that the different moduli of elasticity of the constit- 
uent grains of a rock do not introduce any perceptible error in measurements 
made by this method, when a column an inch in diameter is employed, and 
when the rocks are not coarser in grain than the Baveno granite. In fact, 
while surrounded on all sides by other grains, no individual grain can expand 
freely, as it would if subjected to compression unhampered by any surrounding 
medium, and thus the anisotropic character of the individual grains produces 
but little effect on the elasticity of the rock as a whole. 
These experiments also show that in the case of rocks composed of several 
minerals it makes no perceptible difference whether the points of attach- 
ment of the instrument are embedded in the grains of one mineral or of 
another. 
The chief source of error and the one to which the variations observed are 
for the most part to be attributed seems to be a mechanical one, viz, the 
difficulty of getting an ideal contact between these points of attachment and 
the specimen to be measured, especially in view of the extremely small 
dimensions of the movement to be measured. 
The question of the influence of temperature on the elasticity and compressi- 
bility of rocks is of course one which has an important bearing on certain 
problems of geophysics. The only investigation of this subject, so far as can 
be ascertained, consists of a few preliminary experiments by Nagaoka and 
Kasakabe.* In these the torsion method was employed, and the experi- 
ments were carried out on a single rock, viz, sandstone. This rock, as has 
already been mentioned, being porou? and stratified in character, is a material 
whose elastic properties are far from ideal. The results are summed up by 
the authors in the following words : 
Preliminary experiments with sandstone show that the modulus of elasticity is much 
affected by the variation of temperature, i. e., about 0.5 percent per degree. It does not, 
however, necessarily diminish with the increase of temperature where the temperature is 
low, i. e., it is maximum about 9 C. 
As has been shown however, the values for the elastic constants obtained 
by this torsion and bending method have yielded results which can not in all 
cases be correct and which differ very considerably from those obtained by 
the more direct and simple method which has been employed in the present 
paper. These results bearing on the variation of elasticity induced by changes 
of temperature, especially in view of the fact that they are stated by the 
*Modulus of Elasticity of Rocks and Velocities of Seismic Waves. Publications of the 
Earthquake Investigation Committee, No. 17. Tokyo, 1904, p. 43. 
