POLYMORPHISM AT HIGH PRESSURES. 145 



apparently unstable, so that it was not possible to make any accurate 

 measurements. About all that I could do was to determine the general 

 cliiiracter and location of the transition. Two runs were made. 

 The first was an exploring run; it showed the existence of a new modi- 

 fication with a transition near 11000 kgm. at 100°. No transition 

 was found at room temperature to beyond 12000. Presumably, 

 therefore, the transition is of the ice type. The second run, for pur- 

 poses of measurement, verified the existence of the transition, and gave 

 for the more accurate coordinates at 100°, 11480 kgm., with a change 

 of volume of about 0.0035 cm.^ per gm. Temperature was then 

 lowered to 86°, and the transition pressure found to be at least as high 

 as 11800, verifying that the transition is of the ice type. The appara- 

 tus was now left over night. The next morning at 128°, instead of one 

 transition point, two were found, at 9600 and 8900 kgm. The transi- 

 tion line has split, and there are three modifications. But a disquieting 

 fact was that the total change of these two new transitions was less 

 than half that found for the single transition the day before. These 

 two transitions were verified by repetition. Temperature was now 

 raised to 141°, and again two transitions found at still lower pressures, 

 at very roughly 8000 and 8700. The apparatus was again allowed to 

 rest, this time over Sunday. On Monday morning temperature was 

 raised to 90° and pressure to 12500, preparatory to a run. It took an 

 unusually long time to reach pressure equilibrium, the pressure con- 

 tinuing to drop. After approximate equilibrium had been reached, 

 no trace of the transition was to be found within 1800 kgm. of the 

 previous location, and again on raising temperature to 100° the transi- 

 tion had entirely disappeared. 



The evident explanation of these effects is that (NH4)2KP04 has 

 three modifications, with transition coordinates approximately as 

 given. But in the region in which the transition runs the substance is 

 itself chemically unstable, and gradually changes to some other sub- 

 stance with decrease of volume. The fact that the change of volume 

 is a decrease shows that the instability has been brought about by the 

 high pressure and not by the high temperature. The change is not 

 accompanied by change of color, nor setting free of NHj, to judge by 

 the lack of odor. 



It is unfortunate that the decomposition does not allow closer 

 study of these transitions, becau^ they are of a rather rare type. 

 At the triple point it seems that three lines, all of the ice t>T)e, come 

 together. The only p^e^^ous example of this is KHSO4. 



It is quite possible that the decomposition is a splitting up into two 



