On the Duties of Ether for Electricity and Magnetism. 305 



exceeding 0'25 per cent, and generally not exceeding O'l per 

 cent. 



5. If, however, the critical density be calculated from the 

 mean densities at low temperatures (say below the boiling- 

 point) only, the error may be considerable ; in the case of 

 normal decane it is probably between 5 and 6 per cent. 



6. As has been pointed out by M. Guye [Archives des 

 Sciences Phys. et Nat. 1894, series 3, vol. xxxi.) the law does 

 not, as a rule, hold good at all for substances the molecules 

 of which differ in complexity in the gaseous and liquid states. 



Appendix. 



Since this paper was written Mr. K. Tsuruta has kindly 

 sent me a copy of his ' Thermodynamic Notes,' No. 10 (Phys. 

 Rev. 1900, x. p. 116) in which he discusses the " law of straight 

 diameter " as applied to benzene. The conclusion he arrives 

 at is that whilst the mean densities obey the law " in quite a 

 satisfactory manner " not only from 80° to the critical-point, 

 but also from 80° down to the triple point, yet a more minute 

 examination showed the so-called diameter " to consist of two 

 nearly straight parts meeting somewhere near the temperature 

 180° 0., or, more strictly, to be a curve whose curvature is so 

 slight that it can be so described." 



This conclusion is practically in agreement with that stated 

 in this paper. 



XXVIII. On the Duties of Ether for Electricity and 



Magnetism. By Lord Kelvin*. 



[Continued from p. 198.] 



^ § 19. TN my paper published in the last number of the 

 JL Philosophical Magazine, of which this is a con- 

 tinuation, I limited myself to a problem of mathematical 

 dynamics ; and merely suggested the possibility of finding 

 in it an explanation of the fundamental difficulty in the 

 Undulatory Theory of Light referred to in the first and 

 last paragraphs (§§ 1, 18). The following communication is 

 the substance of a supplementary statement relating to that 

 paper given orally to the Oongres International de Physique 

 at a meeting held in Paris last Wednesday (August 8). 



§ 20. i.now cannot resist the temptation to speak of 

 efforts which occupy me to find proper assumptions for 

 including something of the allied subjects mentioned in the 

 foot-note on § 1. 



* Communicated by the Author. 



