262 



SCIENCE PROGRESS. 



closed in a small glass tube) which was applied by the 

 writer of this article to obtain definite results at the critical 

 point for so-called pure substances, as described in part i., 

 was really devised for experiments on mixtures. This rod 

 may be moved up and down inside the narrow tubes by an 

 electromagnet which is outside the tube and the heating 

 jacket surrounding it : it accelerates the process of mixing 

 even at very large volumes, and enables one to obtain almost 

 identical readings with diminishing and with increasing 

 volume. But it is especially at smaller volumes and in the 

 critical region that stirring is very effective. The critical 

 phenomena obtained are practically free from retardation 

 due to slow diffusion and differ materially from the phe- 

 nomena as described by Andrews and others. 



%^ 



Fig. 2 {a b c d) gives the aspect of a tube containing our 

 mixture of 5/6 CO2 and 1/6 air at 18" C. at gradually 

 diminishing: volume without stirring; as described above. 

 In a the condensation has just begun, in h the liquid has 

 increased and the liquid surface is flatter, in c the quantity 

 of liquid is again more but the liquid surface is just in the 

 act of disappearing (critical phenomenon), in d the mixture 

 is homogeneous. Fig. 2 {e. f g /i) shows the behaviour of 

 the same mixture at the same temperature (18 C.) but with 

 stirring ; e and f are very much the same as a and d, the 

 surface, however, being better defined. In g the quantity 

 of liquid has passed through a niaxwium and is decreas- 

 ing with well-defined surface ; in // the liquid is nearly 

 all evaporated. No critical phenomenon takes place but 

 a kind of condensation which is called " retrograde 

 condensation " instead. Evidently the same would 

 be found without stirring- if one could afford to wait 



