[ :553 ] ./\ f^J 



:S 



XLIY. On the Practical AttaUiment of the Thermodffnamic 

 Scale of Temperature. — Part 11. B/j J. Rose-Innes, 

 MA., B.Sc.^ 



IX tlio first part of tliis paper, which was puhlished in the 

 Philosophical Magazine for July 1901, 1 remarked that 

 the question ot* finding any temperature on the thermo- 

 dynamic scale resolves itself into two minor problems. These 

 two minor problems are : — (i.) To find the absolute value 

 of the freezing-point of water ; and (ii.) to find the value of 

 the interval between the freezing-point of water and the 

 temperature under discussion {loc. cit. pp. 135-136). It 

 was also remarked that these two problems are fairly dis- 

 tinct though not entirely so ; and I hope in this second 

 part of my paper to trace out the connexion between them 

 in greater detail than was done in the first part. 



One of the most important physical constants is the 

 absolute temperature of the freezing-point, but its value is 

 not known with any great accuracy at the present time. In 

 the fi.rst part of this paper I gave two tables containing 

 estimates of the freezing-point. The first table was based 

 upon M. Chappuis's experiments with the constant-volume 

 thermometer, M. Amagat^s experiments on isothermal com- 

 pressibility, and the well-known Joule-Thomson experiments. 

 The following results were obtained : — 



Hydrogen 273*153 



Xitrogen 273-361 



{loc. cit. ^. 137). The second table was based upon Regnault's 

 measurements of the expansion of gases under constant 

 pressure, and the Joule-Thomson experiments. The numbers 

 obtained were as follows : — 



Hydrogen , 273*00 



Air 273*16 



The difference between the two estimates in either table 

 is surprisingly large when we consider the skill of the 

 experimenters furnishing the numerical data. An interest- 

 ing question arises as to the cause of this difi^erence : and an 

 easy way of disposing of the whole matter is to assume that 

 the difficulty is caused entirely by errors in the observed 

 values of the Joule-Thomson effect. There is certainly a good 

 deal to be said in favour of this assumption, as the Joule- 

 Thomson experiments were so difficult to carry out that they 

 were doubtless subject to a considerable experimental error. 

 But this assumption, though plausible, ought not to be 

 * Communicated by the Author. 



Phil. Mag. S. 6. Vol. 6. Xo. 33. Sept. 1903. 2 A 



