ON THE FACTS OF EARTHQUAKE PHENOMENA. 287 



found however scarcely needful. Both the chronographs were screened 

 from wind by being used within the fir-wood cases in which they were 

 transported, the front sides of the cases being alone removed, or occasionally 

 by a canvas screen ; the weights of course descended in the closed maho- 

 gany clock cases of the instruments. 



From what has been stated as to the effect of changing the moving 

 weights of these instruments, it will be obvious that with any given weight 

 it becomes necessary to rate the instrument, or to ascertain to what interval 

 of actual time each revolution of the dials corresponds. This was done 

 by winding up each instrument a given number of revolutions of the weight 

 barrel, and having placed the instrument firmly before a good seconds' clock, 

 noting repeatedly in how many seconds and parts of seconds it ran down 

 the same number of revolutions ; the clock used had a loud beat and a gra- 

 duated arc to the pendulum, so as to enable a second to be divided by the 

 eye into four parts at least. 



Experimental rating of Wheatstone's (the Firing) Chronograph, 19th of 



October 1849. 



Chronograph wound up in every case 8 revolutions. 



Exp. 



No. of experiments=8)319-82 



39*98 average 

 for time of eight revolutions of large hand; .-. value of one revolution of 



large hand is = ^ =4"-9975, hence value of one division of large dial 



8 



=^-^5^=0"-41646=l revolution of small dial, and 1 division of small dial 

 12 



=2::ii^=o"-oi388. 



30 

 To check this rating more completely, a method different from Wheat- 

 stone's, and I believe new, was adopted. A given number of seconds or beats 

 of the clock was arbitrarily taken, the chronograph released at the first beat, 

 and stopped dead at the last beat ; 20 beats were taken, and in 20 seconds the 

 same chronograph ran down by its own registry. 



Exp. 1. in 20" 4 revolutions — 7° small dial, 



z. ,, 4 „ 5 ,, 



3. „ 4 „ +7 5» 



4. „ 4 5, A „ 



Nos. 1 and 3 rejected as bad experiments, then the average is =4 rev. 

 — 3°-5 small dial. 



Taking then the value of one division of small dial from the first or pre- 

 ceding set of experiments =0"-01388, then 4 rev. — 3°-5 X-01388"=4 rev. 

 — 0"-0486; therefore 20"— 0"-0486=19"-9514=the true time of 4 revolu- 

 tions of large dial, or =4"-988 time of one revolution of large dial 



