210 Prof. J. Dewar on the Liquefaction of Oxygen 



the order of the probability which attaches to an hypothesis. 

 Accordingly the principle affords a foundation for Donkin's* 

 elaborate calculation of the probability of a physical connexion 

 between binary stars. However, it may be admitted that his 

 assumption concerning the genesis of double stars, though 

 doubtless the simplest that could be made, yet does not 

 exceed rival assumptions in respect of simplicity and conve- 

 nience to the same extent as the assumption of an arithmetic 

 mean in the cases just referred to excels the geometric and 

 all other means. 



To sum up, there has been expressed a doubt [I. (4)] how 

 far the area of applied Inverse Probabilities extends. But, as 

 far as it does extend, so far, in general and with the trifling ex- 

 ception of games of chance [I. (l)-(3)], a foundation of unsta- 

 tistical a priori probabilities is required. The only perfectly 

 solid foundation of this character which exists is, as it were, 

 a shapeless mass of rough experience [IT. (1)] adequate to 

 support an unnumerical a posteriori probability. To afford a 

 basis for the more regular structures of the science that rude 

 foundation must be levelled by the addition [II. (2)] of a 

 somewhat less solid, but still empirical, material. And after 

 all [II. (8)] there remain fissures to be filled up by legiti- 

 mate conjecture. How much such a foundation will support, 

 to what height it is expedient to carry an arithmetical calcu- 

 lation founded thereon, is a question to be determined by that 

 unwritten philosophy and undefinable good sense which, in the 

 order of scientific method, precedes the application of Calculus 

 and is prior to a priori probabilities. 



XXVIII. On the Liquefaction of Oxygen and the Critical 

 Volumes of Fluids. By James Dewar, M.A., F.R.S., 



Professor of Chemistoy at the Royal Institution, fyc. f 



HAYING had occasion to illustrate, in my lectures at the 

 Royal Institution, the liquefaction of oxygen, a short 

 description of the apparatus I have found most convenient for 

 such demonstration may interest the readers of the Philoso- 

 phical Magazine. The two Russian chemists MM. Wrob- 

 lewski and Olzewski, who have recently made such a 

 splendid success in the production and maintenance of low 

 temperature, have used in their researches an enlarged form of 

 the well-known Cailletet apparatus ; but for the purposes of 

 lecture demonstration, which necessarily involves the projection 

 on a screen of the actions taking place, the apparatus repre- 



* Phil. Mag. 4th series, vol. i. pp. 463-466. 

 t Communicated by the Author. 



