106 BRIDGMAN. 



is this evident inaccuracy in the method which leaves me unconvinced 

 as to the cogency of his proof that the melting curves of the stable 

 and unstable forms are approximately parallel. 



Carbamide. — This was obtained from Eimer and Amend and was 

 used directly without further purification. The purity was fairly 

 high, as was shown by the melting point. At atmospheric pressure I 

 found for my specimen 131.7°, which agrees within the limits of error 

 of the previous work with the only published value that I have found, 

 132° by Liubarvin.^^ At least six different runs were made with this 

 substance. For all except the least important it was hammered dry 

 and cold into the rec^uisite forms. It may be easily melted into the 

 forms, but not without some slight decomposition, so that it seemed 

 best to avoid this possibility. Under pressure decomposition takes 

 place with increasing rapidity above 150° so that accurate measure- 

 ments could not be made above this. The carbamide was placed in 

 different sorts of containers for the different runs. At first it was 

 hammered dry into the nickel steel shell and used with a mercury seal. 

 But this resulted in rupture of the shell. The dissolving action of 

 kerosene is slight, so that it was possible to make one run with the 

 pressure transmitted directly to the carbamide by kerosene. This 

 gave good values for the equilibrium values of pressure and tempera- 

 ture, but not consistent values for A;;, presumably partly because of a 

 very slight dissolving action. In the final arrangement the carbamide 

 was hammered into compact cylindrical forms and placed loosely 

 in detached hunks beneath the surface of mercury, and prevented 

 from rising to the surface by a clip of obvious design. Carbamide is 

 quite unusual with regard to its distortion during a transition. The 

 growth of one modification at the expense of another takes place in 

 such a direction that a cylindrical block increases in diameter. This 

 increase of thickness may develop considerable pressure, as shown 

 by the bursting of the nickel steel shell. Even the blocks which 

 were placed loose beneath the surface of mercury had, after several 

 transitions, so swelled laterally as to tightly fill the shell, just as if 

 they had been melted into position. This lateral growth, besides 

 being very inconvenient because of the rupture that it may produce, 

 may also produce very appreciable error in the measurements of Av, 

 because of internal strains. Because of this effect the measurements of 

 Av of the first runs were irregular and had to be discarded, and with 

 the final arrangement, loose hunks under mercury, the apparatus had 



21 Liubarvin, Ber. D. Chem. Ges., 3, 305. 



