350 Prof. Forbes on the Polarization of Heat. 



not being symmetrically arranged, more heat may possibly 

 reach the pile by conduction in one case than in the other, with- 

 out any reference to polarization. 



The question is not, however, as has been stated, " purely 

 mathematical." It must be shown that the alleged effect is 

 not merely of the kind required, but that its amount is such as 

 to produce the observed variation ; and since the temperature 

 acquired by conduction is unknown, and can only be assigned 

 by its effect, the question, on the contrary, is purely experi- 

 mental. 



Had the objectors taken the trouble to try the experiment, 

 the difficulty would never have been urged. I have never 

 denied the interference of the effects of conduction; I have con- 

 tented myself with the assurance that they were of an order 

 sufficiently insignificant not to produce the effects I observed, 

 which in almost every case, on the contrary, they tended to 

 diminish, as I have shown in my memoir ; and for that rea- 

 son I gave my numerical results as approximations only. 

 I never thought of entering into the fatiguing detail which 

 would have been necessary for explaining fully my grounds 

 of confidence in the experimental results at which I arrived, 

 which were chiefly such as must immediately offer themselves 

 to those who would attempt to repeat them, which I naturally 

 considered as preliminary to any criticism upon their sound- 

 ness. 



Melloni has shown that in his experiments the effect of the 

 warmth, or conducted heat, of the interposed plate was always 

 or almost always insensible. I might argue that the effect of 

 the still smaller quantity of heat which could reach the plate 

 B, derived from that already so small in A (fig. 1.), must be 

 insignificant ; and therefore the variation of this quantity, 

 owing to the relative change of position in question, must be 

 almost infinitely small even in my experiments, in which the 

 approximation of the source of heat was much greater than in 

 those of Melloni. And this argument would be incontrover- 

 tible to any one who had tried the experiments, and who was 

 capable of weighing quantitative evidence. But I have a much 

 more direct reply. 



The mathematical distinction which Professor Powell and 

 Mr. Murphy established (and justly) between the cases of 

 fig. 1. and fig. 2. must be expressed by the difference of cer- 

 tain integrals depending on the distances of the elements of 

 the plates from one another in the two cases. But what will 

 be said if I destroy entirely this distinction ; or give it an op- 

 posite character, by pushing the plates nearer to or further 

 from one another in one position than in the other ; or if, in- 



