22 Prof. J. Le Conte on the Discrepancy between the Computed 



The results obtained from Dr. Gregory's experiments cannot 

 be regarded as comparable in accuracy with some of the others ; 

 for the distances were too small, and he was not provided with 

 adequate means of measuring with precision intervals of time to 

 a minute fraction of a second. Nevertheless they are interesting 

 as demonstrating that the violence of sounds does not exercise 

 any appreciable influence on the velocity. The experiments of 

 Drs. Moll and Van Beek, from the fact that the distance was 

 large, that they employed an instrument capable of measuring 

 intervals of one hundredth of a second, and that the signals were 

 strictly simultaneous, unquestionably furnish the most trust- 

 worthy results. Assuming, therefore, the mean result of these 

 experiments as the standard, the following Table will illustrate 

 the accordance of the several theoretical formulae with fact. In 

 applying Laplace's formula, I have taken the value of k to be 

 equal to 1*41, in accordance with the recent experiments of 

 Masson. 



Authority. 



Theory. 



Experiment. 



Difference in 

 metres. 



Velocity in 



metres per 



second. 



Velocity in 



metresper 



second. 



Laplace 



Ivory 



332-4343 

 323-2706 

 332-9307 

 3318780 

 342-8798 

 3451375 



332-4117 



+ 00226 



- 91411 

 + 0-5190 



- 0-5337 

 + 10-4681 

 + 12-7258 



Herapath ... 



Cballis 



Potter 



Earnshaw ... 



With these revised data, it thus appears that Laplace's formula 

 gives a velocity which exceeds that furnished by the most reliable 

 experiments, by about 0*074 of an English foot. It is obvious, 

 therefore, that what Mr. Earnshaw characterizes as the great 

 defect of Laplace's explanation, viz. "that the result it gives 

 does not come up to experiment," fails to be sustained by facts, 

 and fully warrants an emphatic endorsement of Sir John Her- 

 schel's judgment, that " in inquiries of such delicacy, and where 

 the effects of minute errors of experiment become so much 

 magnified, it seems hardly candid to. desire a more perfect coin- 

 cidence"*. 



It will be observed that Mr. Earnshaw's formula (for ordinary 

 sounds) gives a result which surpasses the experimental velocity 

 by 41*75 English feet ; and yet the chief merit claimed for his 

 theory is, that it gives a velocity for sound •* which agrees 

 accurately with the experimental value"! With the results of 



* " Treatise on Sound" in Encyc. Metrop. Article 78. 



