Vapour-Densities of some Carbon Compounds. 515 



The values of pv/mT were therefore transferred to curves, 

 which are reproduced in PI. XX. : and the curves represent- 

 ing compressibilities, having also been mapped, were shifted 

 in position, so as to become continuous with those in which 

 the true value, and not merely the variation of pv/mT with 

 pressure, was known. From these two sets of data the 

 complete curves were drawn. The lower limit of pressure 

 lies approximately at from 40 to 60 millimetres. 



These cm-ves were extrapolated until they cut the axis of 

 zero pressure. The justice of thus extrapolating the curves 

 will be considered later. 



From the weight of a litre of oxygen at 0" C, namelv 

 l-42i:'61, and from the found coefficient of expansion with 

 temperature, 0-0036694, the molecular weights of the sub 

 stances employed were calculated, as shown in Table IV. 



Table IV. 

 Molecular Weights at 129" 6^. 



K=Talue of P'l- T for one gram-molecule of oxygen at 129-6" at zero pre-=ure 



=62423. 

 K' rvalue of Pi- T for one gram of substance at 1296^, and at zero pressure. 

 K K' = found molecular weight =M'. 

 M=calculated molecular weight on the bases 0=32. C = 12. H=l-(Xi7. 



Substance. 



K'. 



^L' = KK'. 



.^r. 



Difference ' 

 1 par: in 



Methvl Alcohol 



Ethyl Ether 



Hexane 



194-73 

 84-345 

 72-475 

 .54-62 

 54-44 

 79-81 

 67-44 



32056 

 74-01 

 86-13 

 114-29 

 114-66 

 78-21 

 92-56 



32-03 

 74-07 

 86-10 

 114-13 

 114-13 

 78-04 

 92-06 



nil 



1250 

 2.500 

 714 

 213 

 4.54 

 182 



Di-isobutvl 



formal Octane 



Benzene 



Toluene . . 





The absolute value of the molecular weight M' in each 

 case depends on the coefficient of expansion accepted for 

 oxygen ; and this, as has been shown by researches (unpub- 

 lished) made in the laboratory- of University College, varies 

 with the pressure. Were the number 0*003675 to be ac- 

 cepted instead of 0*36694, the value of M^ for methyl-alcohol 

 would be 32*105, and that for ether 74*1 2. But such an 

 assumption would have no influence on the main question, 

 viz. the variation in the discrepancies between the values of 

 M and M'. From the last column these are seen to var\- 

 from 1 part in 1250 (minus) to 1 part in 182 (plus). The 

 discrepancy is made more apparent by accepting for ether the 

 " theoretical '' molecular weight ; the discrepancy is then ; — 

 Methyl ale. Ether. Hexane. Di-isobutyl. ^orm. Octane. Benzene. Toluene. 

 1 part in 588... x 833 455 ' 182 333 159 



2 L 2 



