i 7 6 



NA TURE 



[ [une 19, 1 < 



together, of normal and transverse waves, with a nearly 

 horizontal direction of propagation ; but in that case we 

 should expect to find normal waves occurring alone at 

 the beginning of the earthquake with much greater am- 

 plitude than they actually have. Other still less probable 

 solutions might be referred to ; but it is safe to say that 

 the evidence furnished by these observations goes far to 

 prove that the earthquakes of the Plain of Yedo consist 

 chiefly of distortions, not compressions, of the ground, and 

 emerge at Tokio in a direction not very far from vertical. 

 In the older seismology it was generally assumed not 



only that an earthquake consists mainly of one impulse, 

 but that the motion of the ground has a definite direc- 

 tion, and that that is the same as the direction of propa- 

 gation of the wave. All three assumptions were false. 

 An old piece of seismic apparatus, based on these ideas, 

 was a group of columns of various heights standing on 

 a plane horizontal base. These were intended to show the 

 direction and " intensity of the shock" by falling over. 

 It is clear enough, however, that no appliance of this kind 

 can give intelligible results from earthquakes of such 

 complexity as those described above. The very word 



" shock,'' accurately as it describes the feeling produced 

 by an earthquake, is a singularly inappropriate name for 

 what an apathetic seismograph records. 



As evidence of the accuracy of the apparatus by which 

 the foregoing results were obtained, it should be men- 

 tioned that the records given at the same place by 

 different instruments during the same earthquake were 

 found to agree remarkably well. Further, the instruments 

 were tested experimentally by placing them on a shaky 

 table, and obtaining, side by side, two records of table- 



quakes, one from the so-called " steady-point " of the 

 instrument, and the other from a point in a fixed bracket 

 projecting from a neighbouring wall, and known to be 

 truly steady. When the table was shaken in such a way 

 as to give records resembling those of actual earthquakes, 

 the agreement of the two showed conclusively that the 

 steady-point of the instrument did remain very nearly 

 undisturbed, and that the records were in all important 

 particulars substantially correct. 



We have then the means of accurately observing the 



