VELOCITY OF ELASTIC WAVES IN GRANITE 69 
percent of pore space, show an abnormally high compressibility at very low 
pressures; the value at a pressure of only a few hundred megabars being 
nearly double that at 1000 megabars.” 
After suggesting two more possible explanations and citing examples both 
contradictory and confirmatory, they proceed to a fourth: “Very often the 
abnormally high compressibility at low pressures is associated with ccarse- 
ness of grain—for example, the granites and marbles, which are compara- 
tively coarse-grained, show a high initial compressibility. The Sudbury 
diabase, however, is moderately coarse-grained and does not show much of 
this effect, while the serpentine, which is very fine-grained, shows a con- 
siderable decrease in compressibility and this tendency to decrease persists 
even at high pressures”. 
In this same connection, it is interesting to note that Adams and Coker! 
had investigated the effects of grain size, including a granite in the trial runs, 
and concluded: “It will thus be seen that there is no correspondence between 
the coarseness of grain and the magnitude of the variations in the readings 
obtained.” 
This situation, which clearly proved rather difficult, and hinges essentially 
on the high values for the compressibility of granite obtained by Adams and 
Coker, is left by Adams and Williamson as follows: “About all that can be 
said concerning the way in which the compressibility of recks changes with 
pressure is: For pressures above 2000 megabars the compressibility does net 
change very much; at pressures of only a few hundred megabars the com- 
pressibility is likely to be notably higher; and that the change of compressi- 
bility at low pressures, while connected with the admixture of minerals of 
different compressibility and with a looseness of structure existing in some 
coarse-grained rocks, cannot with certainty be predicted in advance.” 
The compressibility of Quincy granite, 2.29X10-"% cm?/dynes, reported 
earlier in this paper, has an important bearing on the question under dis- 
cussion. The issue is somewhat confused by the fact that some writers con- 
tend that high-stress, statically determined compressibilities may be con- 
siderably different from those determined dynamically, which involve low 
stresses. If, for the moment, we assume the equivalence of statically and 
dynamically determined values, it is at once apparent that the compressi- 
bility of Quincy granite as plotted in Fig. 6 agrees with the high-pressure 
values of Adams and Williamson and eliminates the anomalous change at 
low pressures advocated by them. If, on the other hand, we deny this equiva- 
lance, it seems that it is almost inescapable to assume a variation in dynami- 
cally determined compressibility similar to that shown by Adams and Wil- 
liamson. The dotted curve in Fig. 6 represents this assumed variation. Ac- 
cording to this curve, the seismically effective compressibility at 2000 mega- 
bars would be roughly one half that observed at Quincy for pressures of a 
few megabars. Barring compensating changes in density and rigidity, such 
a decrease in compressibility would give a velocity of approximately 7.5 
km/sec. for longitudinal waves in granite at 2000 megabars pressure. This is 
far in excess of the figure 5.6 km/sec. observed by Jeffreys® and others as a 
213 
