THIRD SERIES— REFRANGIBILITY OF HEAT. 207 



tion of the small effect due to the difference of refrangibility. Now, havmg proved 

 in the Thii-d Section that this difference of mean refrangibility is from most sources 

 very small, which yet differ widely in their polarizability, we infer that that ex- 

 planation is probably inadequate, and that we must look for a mechanical theory 

 of heat differing in some particulars from that of light. 



85. (2.) This latter conclusion is farther confuTned by the results of the Se- 

 cond Section, in which is deduced, from the singularly accordant results of wholly 

 distinct series of experiments with heat from those distinct sources, that the phe- 

 nomena of depolarization differ sm-prisingly, numerically speaking, ft-om those of 

 light, whilst in their general character they are entirely similar. The results at 

 which we have arrived oblige us to admit, either that the length of a wave of heat 

 is several times greater than that of a wave of light, or that the velocities of the 

 ordinary and extraordinary ray in doubly refi-acting crystals are totally different 

 from those of light ; or else a combination of these hypotheses. Now, of the two 

 fii-st alternatives we ai-e bound at present, I think, to prefer the latter, since we 

 know nothing of the phenomena of double refraction but from this experiment ; 

 whUst the subsequent experiments on the refractive index, would, according to 

 the prevalent theory of dispersion, seem to shew that the mean length of a wave 

 of heat cannot differ very materially from one of light. This amounts to admit- 

 ting that the doubly refractive energy is more feeble for heat than for hght ; in 

 other words, that a greater thickness of a crystal is required to produce a given 

 effect. The Second and Thii-d Sections also confii-m one another in this respect, 

 that the unifoi-mity of the results of depolarization with heat from different sour- 

 ces, and also of the refrangibUity, would both be highly improbable did the length 

 of a wave materially differ in those instances. 



86. (3.) Of the results of the Third Section, I have already spoken at sufli- 

 cient length (art. 81). The mean index of refraction for aU kinds of heat tried 

 is less than for light ; — it ranges within narrow limits ; — when the heat from dif- 

 ferent sources is unmodified by transmission through diathermant bodies, these 

 limits are very narrow indeed ; — the measure of dispersion is considerable but un- 

 ascertained, and opens a fair field for experiment ; — dispei'sion is probably least 

 for sources of low temperature. 



87. Such are the chief data for speculation afforded by the experimental re- 

 sults contained in this paper : — ^too imperfect perhaps in themselves to form the 

 basis of a mechanical theory of heat, yet such I hope as may be considered to be fit 

 contributions towards its construction at a future period. 



Edinbukgh, \Gth April 1838. 



