POLYMORPHISM AT HIGH PRESSURES. 117 



points for Av (II-VI) at 180° and 200° lie higher than the others would 

 demand, and suggest the same possibility. Furthermore, the direc- 

 tion of curvature would be explained by the existence of a new modi- 

 fication. I was not able to settle the point, however, and have drawn 

 only one curve in the diagram. This is the curve which best fits the 

 equilibrium points. It is very probable, however, that at the triple 

 point II-V-VI the curve II-VI should be steeper than drawn, and the 

 curve II-V less steep. This is indicated by the latent heat relations. 



On the II-III line only two points were determined, but the line 

 has nevertheless been indicated in Table VII as ha\ing a curvature 

 in the normal direction. The existence of curvature in this direction 

 is demanded by the latent heat relations of II-III-IV-V at the approxi- 

 mately quadruple point, and the actual amount of curvature can be 

 very approximately calculated from the data for the other curves. 

 But to detect this curvature experimentally would have been difficult, 

 because the breadth of the band of indift'erence is sufficient to conceal 

 the effect on a ciu-ve so short. 



On the I-II line a considerable element of uncertainty is introduced 

 because the change of volume is so small that equilibrium points can- 

 not be obtained, but the transition runs entirely to completion when 

 it has once started. The transition is fairly rapid, however, and runs 

 without much trespassing of the equilibrium pressure. The points 

 shown in the diagram are the means of the pressures at which the 

 transition ran spontaneously from above and below; these two pres- 

 sures differed by from 50 to 100 kgm. The values of Av cannot be 

 determined with much percentage accuracy because of their smallness; 

 it was not possible to tell from the data whether Av increases or de- 

 creases with rising temperature, and in the computations A» has been 

 assumed to be constant. The points on the I-II line were not deter- 

 mined at the same time as the other points, but nearly one year later. 

 The same specimen of camphor was used. 



In view of the uncertainty of the Av values on the II-V, IV-V, and 

 V-VI curves, I have not attempted in Table VII to give even the most 

 probable values for the change of volume or latent heat. The indi- 

 vidual determinations of these changes of volume are chaotic; except 

 for being of the same order of magnitude they offer no justification for 

 the choice of any one set of values. I have, therefore, not included 

 the experimental points in Figure 12. One is in a position to say this 

 much, however; on the lower ends of the curves II-V and IV-V the 

 two changes of volume should add up to the total change II-IV which 

 is indicated in Figure 12, and which was determined with fair consis- 



