260 GROVE KARL GILBERT— DAVIS tMEM0IKS [v£™xi; 



fluence, and can not check in an important way the growth of the city. The destiny of San Francisco depends 

 on the capacity and security of its harbor, on the wealth of the country behind it, and on its geographical rela- 

 tions to the commerce of the Pacific. Whatever the earthquake danger may be, it is a thing to be dealt with 

 on the ground by skillful engineering, not avoided by flight; and the proper basis for all protective measures is 

 the fullest possible information as to the extent and character of the danger. 



This wise counsel is repeated at the close of an address on "Earthquake forecasts," which 

 is analyzed below. 



gilbert's conception of an earthquake 



Gilbert was appointed by the Governor of California as a member of a State commission 

 to investigate the effects and causes of the San Francisco disaster. He also took part in a study 

 for the United States Geological Survey; and his chief report was published in its Bulletin 324 

 in 1907. He there presented a brief statement of bis conception of the "motions constituting 

 the earthquake in the region of its destructive intensity," from which the following passages 

 are taken : 



The San Francisco earthquake had its origin, wholly or chiefly, in a new slipping on the plane of an old 

 fault [The San Andreas fault, passing a little to the west of the city]. The trend of the fault is northwest 

 and southeast, and it . . . has a length of 300 miles, possibly more. Nothing is known of its depth. . . . The 

 fact of recurrence on the same plane shows that the rock faces in contact had not become welded, so that the 

 molecular force which there resisted motion was less than the cohesion of solid rock, and may have been little 

 stronger than adhesion. A tract of the crust including the fault plane had come to be affected by a system of 

 slowly increasing shearing strains, and the associated stresses were the forces directly causing the fault. When 

 the stress component coincident with the fault plane at some point became greater than the adhesion (or cohesion) 

 a local slipping took place. This caused a redistribution of strains and stresses, the local relief of strain being 

 followed by an increase of strain and stress in all adjoining tracts of the fault plane, with the result that the 

 adhesion was overcome in those tracts and the area of incipient faulting thereby enlarged. Thus from the 

 initial tract the lesion was propagated as a sort of wave through all the fault plane. 



At the initial tract a small movement sufficed to relieve the local strain, and the motion was then arrested 

 by friction, but the movement was renewed by reaction from other tracts, and it alternately started and stopped 

 till the accumulated stresses had spent themselves. There was a similar rhythmic sequence in other parts of the 

 fault, the frequency of the alternations depending on local conditions; and the total movement of dislocation 

 at each point was accomplished by a series of steps and not by a single leap. The time consumed in these reac- 

 tions was not infinitesimal. The rate of propagation of changes in strain was the same order of magnitude as 

 that of earthquake waves in general, and the rate of propagation of initiation of movement on the old fault 

 plane may have been somewhat slower because of the necessity of accumulating a certain amount of stress 

 increment to overcome the adhesion. It is probable that the completion of the fault required more than one 

 minute, and it may have required more than two minutes. It is even possible that the displacement had been 

 completed at the initial point before it began at the most remote. 



In the succession of slippings and stoppings at any point of the fault plane each separate slip communicated 

 a jar or pulse to the surrounding rock, and this pulse was propagated in all directions. The earthquake at any 

 locality in the neighborhood of the fault consisted of such pulses from different directions. The general distribu- 

 tion of intensity indicates that the pulses weakened in transmission somewhat rapidly, whence it may be inferred 

 that the particular pulses constituting the dominant elements of the local earthquake were those from the nearer 

 parts of the fault. If this conception of the earthquake is correct, the rhythm observed in the region of high 

 intensity was a phenomenon distinct from the rhythm of harmonic waves. It was essentially a frictional rhythm, 

 dependent on the relation of certain rock strains and rock stresses to the resistances afforded by adhesion and 

 sliding friction. It was irregular not only because the intervals of local starting and stopping were unequal, 

 but because it was derived from a considerable area of the fault surface, in which the local rhythms were neither 

 harmonious nor synchronous. 



The compounding of unevenly spaced pulses from different points of the fault plane caused both reenforce- 

 ment and interference, introducing a character analogous to beats in music, but without the regularity of musical 

 beats. It also at times made oscillatory motions swifter in one direction than the other, so that reciprocal 

 accelerations were not always symmetrically arranged. In less technical language, the motion was jerky and 

 included abrupt phases that were almost blows. The compounding also introduced variety in the direction of 

 motion, especially at the end, when for a short time the pulses from remoter parts of the zone of origin ceased 

 to be overpowered by those from the nearer parts. The motion in that closing phase of the violent part of the 

 earthquake has been compared by an observer to the motion of a vessel in a choppy sea; and I conceive that 

 this comparison is the expression of a veritable analogy. 



Few readers of the present memoir are likely to be so well informed as to the nature of 

 earthquakes as not to profit by this analytical summary. 



