From George Gabriel Stokes   30th June 1864

Malahide Ireland | 30^th June 1864

My dear Tyndall,

I have read carefully over the few last pages of your paper¹ which you sent me, and must own that I am not and never was satisfied that the connexion between good radiation and bad conduction & v.v.² is established on theoretical grounds, or even rendered very probable. If a small portion of a heated mass were obliged to get rid of a given portion of heat in a given time, whatever the nature of the mass, and must do so either by conduction or radiation, then no doubt to show that the molecules could readily impart their motion to the ether would suffice to show the substance to be a bad conductor if by conduction be meant transference of motion (motion of the molecular kind which constitutes heat, not motion of the particles (or small parts) of the body as a whole, such as belongs to sound-waves) by pure molecular action, i.e. without going through the process of setting either in motion (internal radiation) and being set in motion by ether (internal absorption). Conduction however as usually understood includes the effects of molecular radiation and absorption (as for instance in the case of the metals which even in so small a thickness as the 1/10000th part of an inch would I imagine be sensibly opaque to radiant heat as they are to light). Indeed Poisson³ goes so far as to make conduction nothing but internal radiation and absorption. I think with you that it is more probable that a portion of the transfer of motion takes place simply by the action of molecular forces.

But suppose we could separate completely conduction from the effects of molecular radiation and absorption, so as to isolate the portion of the transfer due to purely molecular forces. Would your reasoning hold good then? I think not. I will use your own illustration a little modified

Let the 0 and ●

represent

molecules of the

same mass but

different size

the 0 being

more bulky,

and let the | and

|| represent

elastic connecting rods, the first the more slender. I should expect the big molecules to lose the more motion by imparting motion to the surrounding ether and the molecules connected by the stouter rods to lose more motion by communication to their neighbours. Then fig 1 would picture a body which was a good radiator and bad conductor, fig 2 a bad radiator and bad conductor, fig 3 a good radiator and good conductor, fig 4 a bad radiator and good conductor. Cases 2 and 3 seem to me a priori as likely to occur as 1 and 4. On the whole I <herafter LT Typescript only> must confess that what you have written seems to me hardly to show a finger post pointing to the true road.

Yours very truly | G. G. Stokes

I return the MS to Mr White.⁴

RI MS JT/1/S/236

RI MS JT/1/TYP/4/1394–5