204 SCIEXCE PROGRESS. 



isopentane of absolutely ciinstant vapour-pressure: the same 

 is true for carbonic acid as prepared and investigated by 

 Amagat. This gas contained less than "0004 of gaseous 

 impurity. The writer of this article at one time prepared a 

 sample of carbonic acid containing something like "0007 of 

 impurity, and which showed an increase of "iS atmosphere 

 at I 5^ C. Comparing this result with those of Amagat and 

 Andrews for the same substance we come to the conclusion 

 that for carbonic acid Andrews' explanation of the anomaly 

 was right, and that it must be due to the presence of im- 

 purities, and if so in the case of carbonic acid, why not also 

 in the other cases ? At any rate to put it as mildly as 

 possible, we are better not to give up the simple notion of 

 a constant vapour-pressure until it has been more con- 

 clusively proved that pure substances display phenomena 

 that one would expect to find in mixtures. 



An explanation for the alleged behaviour of pure sub- 

 stances during condensation was readily found. If, say, 

 ether behaves like a mixture when condensed, why not sup- 

 pose it to contain two different kinds of molecules, vapour 

 molecules and liquid molecules, the latter probably an ag- 

 glomeration of the former ? Both may be present in both 

 phases, but presumably the liquid molecules will be much 

 more numerous in the liquid than in the vapour. Would a 

 mixture of that kind not show an increase of pressure 

 during compression and condensation ? An affirmative 

 answer will be found in a sentence in \ an der Waals' 

 famous treatise on the continuit\- of the two states. But 

 this is evidently a slip. The difference between a mix- 

 ture of liquid and vapour molecules and a mixture of 

 two substances is that in the former the molecules can 

 change into one another. A simple application of ther- 

 modynamics, for instance of one of Gibbs' rules of phases, 

 shows that the process of condensation will take place at 

 constant pressure just the same as if the molecules 

 did not associate. If this had been remembered by the 

 promoters of the theory, it would have led to more caution 

 in assuming the absolute trustworthiness of experiments. 



The same idea of the existence of liquid molecules as 



