200 ROYAL SOCIETY OF CANADA 



studying various phenomena collective^ in an investigation for corre- 

 lation, considerable restraint must be exercised not to draw conclusions 

 as to cause and effect from a limited number of coincidences. For a 

 conclusion once drawn is apt to become an obsession to the investigator, 

 and he is more or less blinded to facts that do not fit his theory. 



In examining the record of microseisms the first que~tion that 

 presents itself is whether the recorded motion is that of the ground or 

 of the pendulum, in the first case the pendulum acts as a steady mass 

 or point, while in the latter case it is set oscillating either by impulses 

 from the ground or by an undulatory movement of the ground. Let us 

 consider the case of microseismic records of the " Sawtooth " type, where 

 we see regular and almost wholly uniform oscillations kept up for hours 

 and longer. If in this case the "ffendulum actual^ oscillates it will do 

 so with the period inherent to it. After recjeiving the first impulse or 

 impact let us suppose it to oscillate, if no further impact were received 

 the oscillations would soon die out and the amplitudes would decrease 

 in the given ratio of the damping coefficient. When a second impulse 

 is given the pendulum will continue its uniform swings provided the 

 time interval from the preceding impulse is that of the period of the 

 pendulum or a multiple thereof. If this is not the case, then we shall 

 have interference and this Avill be shown on the record. But such 

 interference is not present in our supposed diagram, whence we must 

 conclude that even if we admit that the diagram is a record of the os- 

 cillations of the pendulum, in reality it is only a counterpart of the actual 

 movements of the ground, that is, of horizontal to-and-from motions of 

 the earth particles. If the pendulum is kept swinging uniformly it can. 

 only be don,e so by some force acting at intervals of the period of the 

 pendulum. As indicated, this may be done by the periodic oscillating 

 movements of the eartli particles; or the same eil'ect may be produced by 

 rhythmic undulatory movements of the ground. Now, the period of 

 microseisms recorded here lies mostly between 5 and 6 seconds, which it 

 may be rettnarked is also approximately the period at times of the two 

 pendulums, and the period of the undulatory movements manifested in 

 the " principal portion " of tectonic earthquakes is 20 seconds or more, 

 so that for microseisms we find the period only about one quarter of the 

 preceding, provided we admit that the record of the microseisms is due 

 to ^undulatory motion and not to horizontal movements. 'The shorter 

 period might perhaps be assigned to a far thinner part of the crust of 

 the earth being affected in the microseisms than is involved in the un- 

 dulatory motion connected with microseisms. As an analogy we may 

 give the short period of the ripples in water from a breeze, or the much 



