Mr. C. K. Akin on the Compressibility of Gases. 299 



perimental data of corresponding p's and a's which have been 

 supplied by M. Regnault, immediately and in a legitimate man- 

 ner, for the purpose of determining the constants of any empi- 

 rical formula, to be substituted for the general <f>, by which we 

 may wish to supplement the Boylean law. With this object 

 should be expressed as a function of p, as it would be improper 

 to make it a function of a, or the mean density, and the density 

 strictly corresponding to p is not ascertained*. It would be 

 superfluous to insist here on what would be the best way to avail 

 oneself of the whole body of observations tabulated by M. Reg- 

 nault, as it will become obvious of itself to everyone on inspection 

 of the Tables. Before concluding, however, we may be allowed 

 yet to make this simple observation, that for formula (a) may 

 be advantageously substituted the simpler form 



, p a ID, , 



^fa-j-k— 1 

 (k having the same meaning as in art. 1), which, on account 



of the small value of yt> is sufficiently exact for practical pur- 

 poses ; at least if we are satisfied with ascertaining the nume- 

 rical values of <j> to about ~ , beyond which order the con- 

 cordance of the experimental data to be employed does not pro- 

 mise reliable results. 



P.S. Since the above paper was written, a second volume of 

 M, Regnault' s Relation, &c.f has come to hand, in which some 

 fresh experiments on gaseous compressibility are mentioned 

 which call for a short supplemental remark. It should be stated, 

 in the first place, that these new measurements are not intended 

 to supersede the older, but were only made to meet an incidental 

 want in some other research ; for which purpose a much nar- 

 rower sweep of pressures and a wider range of substances was in 

 ' respect of the former sufficient, and in respect of the latter a 

 necessity. The investigation in this case was carried on after 

 another method, and with rather different apparatus, than in 

 the former; as a consequence of which first departure from 

 precedent (or rather return to the precedent as set by D along), 



* It would be easy, however, to evaluate a formula <fi=f(p) by means of 

 data to be derived from the first established form of <p = F(p) and of for- 

 mula (b). 



f Recently published as vol. xxvi. of the Memoires de VAcademie des 

 Sciences. In an Errata appended to this volume the errors in pp. 418, 419 

 of vol. xxi., mentioned in art. 3, have been corrected. 



