98 Lord Rayleigh on Balfour Stewart's Theory of the 



screw-turns, x the pitch of the screw, z the distance from A x 

 to A between the centres of the tubes (the latter being sup- 

 posed circular), y the distance from D to S, and 7? is expressed 

 as a column of water. The effect of temperature on the 

 calculated values of the measured pressures is thus no more 

 than it would be for a gauge containing water only. 



As is probably inevitable, a slow drift occurs in the level 

 of A due to temperature-changes in the room. It is regular, 

 and therefore easily eliminated by timing the observations 

 and alternating zero- with pressure-readings. It has the 

 advantage, moreover, that it prevents all the measurements 

 being made near one position of the screw. 



When the surface of A. is clean and illuminated with con- 

 vergent light, it is just possible to estimate a change of 

 pressure of two ten-thousandths of a millimetre of water with 

 our gauge ; but the surface must be very clean for this and 

 the light faint. 



V. On Balfour Stewart's Theory of the Connexion between 

 Radiation and Absorption. By Lord Rayleigh, FM.S* 



ON a recent occasion f I remarked that Stewart's work 

 appeared to me to be insufficiently recognized upon the 

 Continent. One reason for this is probably the comparative 

 inaccessibility of the Edinburgh Transactions in which his 

 first paper appeared J. Another may be found in the fact that 

 the paper itself is not well arranged, and that the principal 

 conclusion is put forward in the first instance as if it were 

 the result of Stewart's special experiments. The experiments 

 were indeed of great value ; but this course gave an opening 

 to KirchhofPs objection that " this proof [of the law that 

 the absorption of a plate equals its radiation and that for every 

 description of heat§~\ cannot be a strict one, because experi- 

 ments which have only taught us concerning more and less, 

 cannot strictly teach us concerning equality " ||. I am inclined 

 to think that Stewart would have received more recognition 

 if he had never experimented at all ! 



While yielding to no one in admiration for KirchhofT, I 

 can hardly regard him as in this matter an impartial critic. 



* Communicated by the Author. 

 t Phil. Mag. S. 5, vol. 1. p. 539 (1900). 

 | Edin. Trans, vol. xxii. p. 1, March 1858. 

 § The italics are Stewart's. 



1 1 Kirchholf, On the History of Spectrum Analysis, &c, Phil. Mag. 

 voi. xxv. p. 258 (1863). 



