12 Newcomh and Dutton — Speed of 



would make the corrected result 3*214 ± 0*072 miles, or 5171 

 ±116 metres per second. 



Stopped Clocks. 

 It is natural to suppose that if a clock were stopped by an 

 earthquake and if its error at the time were known it would 

 give the best possible record of the time of advent of the shock. 

 An examiuation of the time reports of this earthquake, how- 

 ever, strongly contradicts this conclusion. A clock may stop 

 at almost any phase of the disturbance. A sensitive one may 

 pass through, an earthquake of considerable violence and not 

 stop at all. A jeweler's clock in Charleston was found going 

 the next morning, and when the telegraph wires were re-opened 

 its error was found to be small, showing that its escapement had 

 missed very few beats, if any. Clocks in Columbia, Savannah, 

 Augusta and Wilmington, N. C, in many cases kept going. 

 Inquiry at Wilmington elicited the reply that no jewelers' 

 clocks had been stopped. Several reports describe clocks whose 

 rates are satisfactorily vouched for but whose times can be ac- 

 counted for only upon the theory that they were stopped by 

 the second powerful shock, which was felt at Charleston about 

 five minutes after the principal one, e. g., Branch ville, S. C, 

 Augusta, Eome, Ga., Cape Canaveral, Camden, Ala., Memphis, 

 Tenn. There are some cities where the time of beginning is 

 well established by independent observation and which also re- 

 port stopped clocks. In every such case the time of the stopped 

 clock is much later. Thus at Nashville the time of beginning 

 was noted by a clock which continued going for 42 seconds 

 and then stopped. Similar means of comparison come from 

 Cincinnati, Covington, Ky., "Pittsburg, Newark, N. J., Brooklyn 

 and New York. And in general wherever stopped clocks can 

 be compared with really good personal observations they invari- 

 ably show a later time and usually a much later one. The dif- 

 ference is plainly due to the fact that it generally takes a con- 

 siderable time and an accumulation of the effects of the vibra- 

 tions of the building upon the pendulum to stop a clock. An 

 attempt has been made to evaluate this difference by taking 

 those cases where a comparison can be made between the read- 

 ings of stopped clocks and independent determinations of the 

 times of the beginning in the same locality. 



Locality. 



Nashville, 



State. 



Tenn. 



Intervals by 



personal obs. 



Seconds. 



144 



Intervals by 



stopped clocks. 



Seconds. 



186 



Eatios. 



1-29 



Weights. 



2 



Covington, 



Ky. 



155 



235 



1*52 



1 



Cincinnati, 



0. 



155 



195 



1*26 



2 



Pittsburg, 



Pa. 



174 ' 



234 



1*34 



1 



Brooklyn, 



N. Y. 



204 



234 



1-15 



1 



New York, 



N. Y. 



204 



249 



1-22 



2 







Mean 



ratio, 



1*28 





