118 BRIDGMAN. 



tency experimentally; with a little less confidence, but still with fair 



accuracy, one may say that near 120° the changes of volume II-V 



and V-VI must add to the experimental values given in the figure for 



II-VI; and with much less accuracy one may demand that at the 



triple point IV-V-VI the changes IV-V and Y-Vl check with the 



value given in the figure for IV-VI. It must be remembered that the 



change IV-VI could not be determined on the equilibrium line, and so 



in all prol)ability is not accurate. 



With regard to computing the latent heat, the difficulty of determin- 



dr 

 ing -7- accurately is greatly exaggerated because some of the transition 



curves run so nearly vertical; a very small change in the angular 

 direction of a curve produces an enormous percentage change in the 

 value of dr/dp, and so in the latent heat. Under the circumstances 

 the only means of getting Av and AH is one of trial and error, demand- 

 ing that the additive relations at the triple points shall hold. It 

 did not seem to me that the accuracy of the rest of the work would 

 justify such an attempt. In the figures and the table I have, there- 

 fore, given the values of Av and AH for only five of the curves. In 

 order of certainty these are: III-IV, II-III, II-VI, I-II, and VI-IV, 

 the last being of a higher order of uncertainty than the others. 



On the III-IV line satisfactory measurements were made of the 

 velocity of transition and the width of the band of indifference. These 

 results have been described in a previous paper. 



On the III-IV curve also rough values of the difference of compressi- 

 bility could be determined. Ill is more compressible than IV over 

 the entire curve; at about 7000 kgm. the difference is of the order of 

 O.O525 cm.^ per gm. per kgm., rising to perhaps twice this value at the 

 lower end, and falling to not less than two thirds of it at the upper 

 end. The uncertainty in these values, and also those of AH and Av 

 is so great that it did not seem worth while to try to get A/3 or ACp. 

 Because of incompleteness of the transition etc. it was not possible to 

 get good values of Aa or any of the other curves. 



In a recent paper Wallerant ^^ has described camphor as having 

 three modifications at atmospheric pressure. It crystallizes from the 

 melt in the cubic system. On cooling, the cubic modification changes 

 at 97° to a feebly doubly refracting rhombohedric form, which at 

 —28° is transformed again to a strongly doubly refracting rhombohed- 

 ric form. These transition points of Wallerant are both about 10° 



23 F. Wallerant, C. R., 158, 597 (1914). 



