446 TJniversitij of California Publications in Geology [Vol. 12 



1854 to 1882 was .058 meters per year. Applying this rate to the 

 interval 1882-1906a we find that Tamalpais must have arrived at D, 

 a distance of 3.04 meters from A in 1906a. Then, at the time of the 

 slip on the San Andreas fault, the strain being thereby relieved, 

 Tamalpais sprang back suddenly to the point C, where it was found 

 by survey III after the earthquake, the measure of the rebound being 

 1.97 meters in azimuth 341°. 



The path of the station on its rebound from D was, however, 

 parallel to the San Andreas fault upon which the slip occurred. It 

 therefore moved along the line hh', and could not have arrived at C 

 by that movement. There are two possible explanations of this incon- 

 gruity: (1) The San Andreas fault moved parallel to itself to the 

 southwest an amount equal to the normal distance from C to hh', which 

 is .58 meters, or (2) the position of the station in 1906p was on. the 

 line bb' and not at the point C, as the Geodetic Survey determined. 

 Let us consider these two possibilities in turn. First let us assume 

 that the trace of a fault having a strike of 144° passes through the 

 Tamalpais geodetic station. This fault trace in 1854 would have had 

 the po.sition aa', in 1906a the position bb', and in 1906p the position 

 cc\ That is to say, if there had been such a fault, it would have 

 participated in the slow strain creep between 1854 and 1906a and on 

 April 18, 1906, it would have sprung back suddenly to the position cc' 

 with the rest of the country. Thus the station by rebound along the 

 line bb' parallel to the direction of slip, and a coincident shift of 

 the line bb' itself could have arrived at the point C, where it was found 

 by the Geodetic Survey. Such a shift would have involved the same 

 shift of the San Andreas fault, and this may be imagined to have 

 been effected either by a slip on a deep, east-west favilt, relieving the 

 longitudinal strain, or by a sudden transfer of a transverse strain 

 to the southwest of the fault. Second, let us assume, as an alternative 

 to the foregoing, that the controlling base line Diablo-Mocho was not 

 immobile in the interval between surveys II and III, but that it moved 

 northerly parallel to itself, or without change of azimuth, by operation 

 of the strain creep which I have recognized as a necessity of the 

 elastic rebound theory in the more western parts of the field. A creep 

 of a meter in azimuth 180° between surveys I and II of this controlling 

 base would have the effect of locating C on the line bb' and so removing 

 all incongruity. 



Chaparral. — Let us next consider the record of Chaparral, which 

 also was located by the surveys I, II, and III. In figure 6, A, B, and C 



