Prof. Kelland in the Phil. Mag. for May 1842. 487 



viz. " "die shall not stop to discuss this formula as the subject 

 is too uncertain to allo'w us to pursue it into detail." In the 

 second place, I would ask Professor Kelland, is it possible that 

 he thinks this formula capable of accounting for dispersion 

 independently of the hypothesis of finite intervals ? Is it not 

 very evident, except that hypothesis be true, that kAx is ex- 

 tremely small, and the formula becomes 



u^=~V^ + 2(Q-M)2(f^'j, 



which gives a value of u quite independent of the length of 



the wave ( the constant k = — J. Why then has Professor 



Kelland produced this expression as equivalent to mine ? 



I need say nothing more on this point, except that Professor 

 Kelland has no where else even attempted to account for di- 

 spersion independently of the hypothesis of finite intervals, and 

 therefore I may fairly lay claim to oi'iginality on this head at 

 least. 



Professor Kelland asserts, on what grounds I know not, 

 that the law of molecular force must be such as to make C zero. 

 I have shown in the paper before alluded to, which was read 

 before the Camb. Phil, Soc. last April, that if such be the 

 case the whole universe is in a state of neutral equilibrium. 

 I refer Professor Kelland to a paper in the Camb. Phil. Trans., 

 vol. vii. p. 97, by Mr. Earnshaw, where I think quite enough 

 is proved to show that the Newtonian law cannot be the law 

 of molecular force. 



In answer to the remark made by Professor Kelland at the 

 foot of page 376, on my needless generality, a few words will 

 suffice ; the axis of x is not arbitrary, it was assumed by me to 

 be an axis of symmetry, and every axis is not an axis of sym- 

 metry. To have arrived at my result, therefore, in the way re- 

 commended by Professor Kelland, I should have had first to 

 prove that any line whatever drawn through the medium may 

 be regarded as an axis of symmetry * ; the simple method I 

 have pursued avoids all this. 



[The author promises the continuation of his former paper 

 very shortly. — Edit.] 



• Or in other words, that any three axes satisfy the condition 

 except n I) <j be all even, whicii certainly should not be assumed. 



