440 Mr. Earnshaw's Reply to Prof. Kelland's Defence 



ruption of my investigations ; which reasons will, I trust, prove 

 satisfactory to the Professor for its having been passed over 

 in silence. 



The quotation which the Professor gives at the bottom of 

 p. 265 from my letter (April) I can assure him was not in- 

 tended to have any reference to his writings. The Professor 

 must also have mistaken my views when he states (p. 266) 

 that I "appear to look for a complete explanation of disper- 

 sion to the very quarter at which I aim my objections," for 

 I look to the direct action of matter, against which I have not 

 as yet brought forward any objection. 



In the middle of p. 266 the Professor begins his reply to 

 my remarks on his defence of his numerical calculations. It 

 appears to me that he is hereupon somewhat inconsistent 

 with himself. For (May, p. 378) his words are, " my cal- 

 culations are affected with an error, in that / have neglected 

 to shorten A ; " but here he writes, " the data are not erro- 

 neous." These two statements seem hardly reconcilable. Also, 

 if " the calculations are affected with an error" I do not com- 

 prehend how they can " strengthen theory." What he states 

 (p. 267) about his " formula admitting as many arbitrary con- 

 stants as you please," amounts to a confession that he em- 

 ployed the common principles of interpolations, instead of 

 theory, which is all I have contended for in this part of the 

 subject. 



The latter part of the Professor's letter is employed in con- 

 troverting my remarks on his proof of the transversality of 

 vibrations. The values of v u' v" which the Professor makes use 

 of in establishing this principle are derived from the equations 

 of motion, which in my last letter I have proved to be non- 

 existent. That letter is therefore a sufficient answer to this 

 part of the Professor's reply. I cannot, however, dismiss the 

 subject without remarking, that the non-existence of normal 

 vibrations is not proved when it has been shown that (u) 

 the velocity of their transmission is imaginary. It must be 

 shown that o is zero, or very much greater or very much less 

 than the velocity of transmission of the transversal vibrations. 

 Por, if it turn out that u is imaginary, the proper inference is, 

 as I have before stated, that the equations of motion have been 

 incorrectly integrated, and the whole investigation needs to be 

 revised. As the remarks which I have made in my last letter 

 respecting the evanescence of the quantity n, and, with it, of 

 the equations of motion extend to all that the Professor has 

 written in his Memoirs on Light, and in his Theory of Pleat, as 

 far as they are respectively dependent on Newton's law of mole- 



