ELASTIC WAVES IN AMHERST SANDSTONE 17 
than the effective density of the bar. This agreement is illustrated by 
Figure 6. The circled points on the curve are calculated from the ob- 
served velocity and measured density of the bar. ‘The points indicated 
by crosses are those derived from the static measurement. In plotting 
the latter no effort was made to evaluate accurately the constant K 
because the cross section of the bar was somewhat irregular. The 
constant was chosen to make one point fall on the dynamic curve. 
The others then fell on the curve within the limits of experimental 
error, which was approximately 2 per cent. 
DISCUSSION 
The conclusion to be drawn from the work here described is that 
the observed changes in the velocity of sound in the sample used are 
primarily due to the observed changes in Young’s modulus. In order 
to formulate an explanation for the phenomena it is therefore neces- 
sary to account for the change in Young’s modulus as the water con- 
tent of the sample is varied. Sufficient data are not available at the 
present time to permit the formulation of a definite explanation of the 
observed phenomena, but the following discussion is of interest in this 
connection. 
Since the mineral constituents of the rock are highly insoluble in 
water, it seems safe to conclude that the elastic properties of the in- 
dividual rock particles themselves are not affected by moisture. The 
effect may then be localized in the bond holding the particles together. 
If we picture the rock as made up of particles held together by a 
bonding material which combines with water as does glue or gelatine, 
the rock should act somewhat as was observed. A comparatively small 
amount of a bonding material evenly distributed throughout the 
sample could, by softening or dissolving, release a large number of 
bonds and thus cause a large change in the velocity. This explanation 
fits the observed facts, since the velocity would tend to approach a 
minimum value determined by the amount and distribution of the 
bond present. Furthermore, the removal of water would allow the 
bond to recement the particles and thus restore the rock approximate- 
ly to its original condition. 
An immediate objection to this picture as applied to the sandstone 
under investigation is that the bond in the sandstone seems to be a 
siliceous silt which is only slightly soluble. Furthermore the percentage 
of water-soluble constituents in the bar was found to be extremely 
small. However, since the absolute amount of bonding material pres- 
ent is minute compared with the total mass of the bar it is not neces- 
sary that a large absolute amount of material be dissolved or softened 
677 
