48 Professors Perry and Ayrtonofl a neglected Principle 



shape of A EF, prevents AB recording any lateral motions, 

 since the motion of A B parallel to itself is so small as to be im- 

 perceptible. C D turns about the pin C, and is prolonged to G, 

 where it is attached by a pivot to an eye rigidly attached to 

 the ball : G C D therefore records lateral motions in one direc- 

 tion, say north-south ; but is not effected by east-west motions 

 or by vertical motions of the ball, as these latter only cause 

 the pin C to move vertically in the slot. The third arm, not 

 shown in the figure, by a somewhat similar arrangement of 

 levers, only records east-west motions. All the motions are 

 recorded on one plane, on the same band of paper; so that the 

 curves would be somewhat as shown in fig. 6. Drawing any 

 line A B at right angles to the motion of the paper, we see that 

 at that moment of time the ball was moving from south to 

 north, from west to east, and from up to down ; and from the 

 shape of the curves we can determine the position, velocity, 

 and accelerati'on in magnitude and direction of the ball at that 

 or any other instant of time : the complete law of the motion 

 of the ball is therefore recorded. Should the box be slightly 

 tilted and some of the springs elongated or shortened during 

 the disturbance, then the motions will not be strictly north- 

 south, east-west, &c. ; but it is evident that this cannot pro- 

 duce any serious discrepancy in the indications unless the 

 earthquake motions be exceedingly violent, and when this is 

 the case it will not be very difficult to eliminate the errors. 



It is evident the points B and D may be above the ball in- 

 stead of below it as in the figure; and this arrangement would 

 be preferable when we wish to surround the ball with a liquid, 

 as the paper could then be kept quite clear of the liquid. 



Section H. 



Mr. Mallet is of opinion that there is no turning action of the 

 ground during earthquakes ; but it is possible this conclusion 

 may be perhaps a little premature, since any explanation that 

 has been given of the observed twisting of columns, based on 

 considerations of the attachment at the base, might also apply 

 to the twisting of rock in its natural position. To test whether 

 any such turning action really exists, a simple apparatus, such 

 as is shown in fig. 7, might be employed. H J is an iron fly- 

 wheel rigidly attached by a stretched thick wire K L to a rigid 

 iron framework K M N, or by a thinner wire if lateral motion 

 of the flywheel is prevented by a guide. An arm J P carries 

 a pencil at P touching a band of paper (the one, for example, 

 employed in the previous seismometer) moved by clockwork 

 parallel to M X. If the periodic time of the torsional vibration 



