754 REVIEWS 



subjects to which the records of this "land of earthquakes" make excep- 

 tional contributions. Under the time distribution, the annual, diurnal, 

 lunar, and other periodicities are considered. The most impressive feature 

 of the geographical distribution lies in the fact that the frequency of the 

 Japanese earthquakes increases toward the sea border, especially toward 

 ihat portion of it which overlooks the great "deeps," and not toward the 

 mountainous axis of the islands. 



The instruments used in Japanese investigations are described, illus- 

 trated and discvissed, as well as the nature of the oscillations which they 

 detect and record.. 



In the discussion of the velocities of propagation of the seismic vibra- 

 tions to distant points, preference is given to the view that the paths are 

 essentially parallel to the surface, and not chords, as held by Milne, Knott, 

 Button, and others (see previous review). The grounds for this are not 

 made quite clear; at least they are not quite clear to the reviewer. The view 

 involves the supposition that the ratio of elasticity to density in the different 

 horizons near the surface of the earth varies so greatly that, while the main 

 vibrations move at about 3.3 kilometers per second, the advance tremors 

 move at about 14 kilometers per second, or more than four times as rapidly. 

 Professor Nagaoka's investigations of the elastic constants of rocks (see 

 table in Button's work, pp. 230, 231) give computed velocities for the 

 fastest or normal wave ranging from 1.19 to 7.05 kilometers per second, 

 the average of sixty-seven determinations made upon various surface 

 rocks being 3.85 kilometers per second. When some little allowance is 

 made for the fractured condition of the crust, this tallies well with the 

 observed superficial rate of propagation, 3.3 kilometers per second. Select- 

 ing the Archean and eruptive rocks as being more nearly representative 

 of the material of the sub-crust, eighteen determinations give an average 

 velocity of only 3.75 kilometers per second. The compact Paleozoic rocks 

 prove to be as elastic as the eruptives. The average of the highest eight 

 of the whole list is only 4.95, or a little more than one-third the maximum 

 velocity of the preliminary tremor. If the velocities of these select examples 

 were computed for an unlimited rock m.ediura, instead of a bar, they would 

 average 5.79 kilometers per second, which is still less than half the requisite 

 velocity. As no single rock was found to have much more than half the 

 requisite velocity computed on the assumptioii that the preliminary tremors 

 went around the spheroid parallel to the surface, and as the speed of the 

 fastest form of wave in an unlimited medium of steel under surface pressure 

 is theoretically only aboiit 6.2 kilometers per second, it seems a rather 

 arbitrary assumption that any substratum in the outer part of the earth 



