POLYMORPHISM AT HIGH PRESSURES. 133 



tion is to be possible at all. It is therefore improbable that the phase 

 found above is the same as that mentioned by Gossner. There must 

 be at least three modifications. From the meagre data at hand a 

 phase diagram in its general features like that of Agl does not seem 

 improbable. 



The difference of compressibility can be determined in the usual way 

 from the difference of slope of the isothermals above and below the 

 transition. All five determinations gave fairly consistent results. 

 The low temperature form is less compressible, and a fair average for 

 the difference is O.Oe^ cm.^ per gm. per kgm. The measurements were 

 not accurate enough to give the variation of this along the transition 

 line. Using this value for Aa, we may find as average values over the 

 entire range for A/3, 0.00006, and for ACp, 0.34 kgm. cm. (0.0014 gm. 

 cal.). The high temperature form is the more expansible and has the 

 higher specific heat. The relations of the two phases, so far as the 

 signs of Aa, A|3, and ACp go, is exactly that of water and ordinary ice. 



Ammonium Iodide. — This substance has a new modification, which 

 is striking because of the large change of volume, about 14%. The 

 transition line runs steeply, with pronounced concavity toward the 

 temperature axis, and crosses the line of atmospheric pressure at about 

 — 17.6°. It should, therefore, be comparatively easy to study this 

 form under atmospheric conditions. 



In all, three different sets of runs were made. The first set was for 

 purposes of exploration. No transition at 20° was found between 

 12300 and 3500 kgm., and at 200° the only transition found up to 

 12300 was at approximately 2000. The second set included the 

 measurements A\ath the high pressure apparatus, and included all 

 points between 20° and 200°. The material was from Hoffmann and 

 Kropff, dried in vacuum before use, and hammered cold into the steel 

 shell. Pressure was transmitted directly by kerosene. The third 

 run with the low pressure apparatus, by the method of varying 

 pressure at constant temperature, gave the point at 0°. The sub- 

 stance for this run was obtained from Eimer and Amend, and was used 

 directly without any preparation. It contained slight traces of 

 moisture, and judging from its light yellow color, it could not have 

 been quite so pure as the previous sample. 



The experimental values of pressure and temperature and of the 

 change of volume are shown in Figure 22, the calculated values of 

 the latent heat and the change of internal energy in Figure 23, and the 

 numerical values are collected in Table XI. 



At 200° the velocity of transition is very high, but it becomes less at 



