THE TIME AND ORIGIN OF THE SHOCK, 15 
San Francisco Bay show that this conclusion is applicable to all that part of the fault; 
but further south the isoseismals, where not greatly affected by soft ground, lie close 
to the fault, indicating a smaller depth. 
Both methods of determining the depth of the fault agree in indicating that it is com- 
paratively shallow; the only considerations opposing this are: its length, its compara- 
tive straightness, its independence of the topography, which it seems to have controlled 
rather than to have followed, and the very considerable geologic time which has elapsed 
since movements were first inaugurated along the rift (vol. 1, pp. 48-52). All these 
facts undoubtedly suggest great depth, but our ignorance of the causes leading to the 
fault movements makes us attach greater weight to the more definite conclusions already 
arrived at, and to regard the movement of April 18, 1906, as comparatively shallow. It 
is the general belief of geologists that fractures of the rock are confined to a crust of small 
thickness; Professor Van Hise estimates that about 12 km. is the greatest depth to which 
they can reach, and he bases this estimate on the consideration that the weight of the 
overlying rock is sufficient at that depth to prevent the formation of cracks or crevices. 
He writes: “In rocks which were bent when so deeply buried that cracks or crevices 
could not form even temporarily, it is probable that the material flowed to its new posi- 
tion quietly, without shock, under the enormous stress to which it was subjected.’’ ? 
But this is not a sufficient criterion; rock can fracture by shearing without the forma- 
tion of crevices just as a block can slide on a second one without separating from it; in 
the case of the California earthquake there is no necessity for believing that the two 
sides of the fault did not always remain in contact while they were slipping past each 
other, and, as is pointed out further on, the movement near the ends of the fault is taken 
up by elastic or plastic distortion. 
The temperature increasing with the depth increases the plasticity of the rock, but 
the increasing pressure increases its rigidity to a greater extent, at least for forces like 
those due to elastic vibrations and the tides of short periods, which do not continue to 
act for a very long time in the same direction; but for long-continued forces in the same 
direction, provided they do not increase too rapidly in intensity, the plasticity probably 
allows slow deformation and prevents the forces from ever reaching the ultimate strength 
of the rock. 
The question which must be answered to determine the depth to which fractures can 
occur is: At what depth does the plasticity of rock become sufficient to enable it to 
yield to the stress-difference, which may exist there, rapidly enough to prevent this 
stress-difference from reaching the ultimate strength of the rock? Unfortunately we 
do not know any of the elements of the problem, neither the plasticity of the rock as 
dependent on pressure and temperature, nor the rate at which stress-differences accu- 
mulate at distances below surface. It is probable that the point at which a fracture first 
occurs is not the lowest point to which it extends; for when the break comes, the forces 
are suddenly transferred to nearby points, and thus the fracture may be carried to depths 
where no fracture would take place otherwise. 
There is very little observational evidence bearing on the question we are discussing. 
The Appalachian Mountains are characterized in Pennsylvania by open folds and few 
faults; as we follow the range to the southwest the folds become closer and the faults 
increase, and in Tennessee, North Carolina, Georgia, and Alabama, the faulting becomes 
excessive. Mr. Bailey Willis has pointed out ” that the thickness of the sediments above 
the Cambro-Silurian limestone was about 23,000 feet (7,000 meters) in Pennsylvania, 
10,000 feet (3,000 meters) in southwestern Virginia, and only 4,000 feet (1,200 meters) 
in Alabama; and he thinks the differences in folding and faulting are due to the differ- 
ences of the loads when the deformations took place. This indicates that faults are 
very shallow. 


1 Principles of Pre-Cambrian Geology, 16th Ann. Rep. U. S. Geol. Surv., 1894-95, pp. 593-595. 
2 Mechanics of Appalachian Structure, 13th Ann. Rep. U.S. Geol. Surv., 1891-92, p. 269. 
