POLYMORPHISM AT HIGH PRESSURES. 



lOS 



curve I-II, working from lower to higher temperatures, II can be 

 carried only an inappreciable distance into the region of I at tempera- 

 tures above 100°. That is, if I has been recently present in the appara- 

 tus, so that II carries in its crystalline structure a nuclear memory 

 of the structure of I, then II will change to I as soon as pressure is 

 manipulated so that I is the absolutely stable form. But if I has been 

 melted, and the liquid forced to crystallize to II, the sojourn in the 

 liquid state now effaces all memory of I, and II will show very little 

 tendency to change to I on being carried into its region of stability, 

 but will melt to the liquid instead, if the temperature is not too far 

 below the triple point. If, however, temperature is too far below the 

 triple point (at 101.5° in an actual case), II wnll spontaneously change 

 to I before the melting curve can be reached. 



It was fortunate that I could obtain a very satisfactory measure- 

 ment of the change of volume when II changed to I at 101.5° at a 

 point 1800 kgm. distant from the transition line. This evidently 



20° 60° 100° 

 Temperature 



Figure 7. Acetamide. The observed differences of compressibility be- 

 tween the two solid phases. 



gives, by a method independent of that usually used, the difference of 

 compressibility between I and II. It turns out that II, the phase with 

 the smaller volume, is on the average O.O5I2 cm.Vgm. per kgm. more 

 compressible than I between 3800 and 5600 kgm. The difference cf 

 compressibility between I and II along the transition line was also 

 determined in the usual way from the difference of slopes of the iso- 

 thermals above and below the transition point. This method shows 

 that II is more compressible than I over the entire length of the tran- 

 sition line, the difference being pronouncedly greater at the lower 

 temperatures. The observed values for Aa, which show a gratifying 

 regularity, are reproduced in Figure 7. At 101.5° the agreement by 

 the two independent methods, O.O5I2 against O.O5I4, is unexpectedly 

 good. The small difference may quite possibly be due to the differ- 

 ence of pressure range. 



