( 534 ) 



ccmcntite, so that then at loioer temperatures cementite instead of 

 graphite would deposit as stable solid phase. From the fact that 

 this does not happen, however, we must not conclude to the endo- 

 thermic nature of the discussed reaction, for first of all it would be 

 possible that the said phenomenon failed to appear because the for- 

 mation of carbide was too feebly exothermic, and secondly it might 

 be that the said phenomenon could be i-ealised only in metastable 

 siafe in consequence of the too early appearance of the eutectic 

 point. So the question whether the formation of iron-carbide in the 

 liquid phase is exothermic or endothermic cannot be decided from 

 the observed phenomena, nor is it indispensable for our subsequent 

 reasoning to assume either the one or the other. Up to now it has 

 been assumed that the liquid mixture P was slowly cooled ; let us 

 now, however, consider the case that the cooling takes place rapidly, 

 then it is possible that no depositing of graphite takes place, and 

 cementite deposits as solid substance. And now it is the question 

 how to account for this. 



But for the hypothesis made by me on the existence of an internal 

 equilibrium this phenomenon would be just as puzzling as it is now 

 easy to explain. If, namely, the deposition of graphite fails to appear, 

 then the internal equilibrium in the liquid phase will lie more to 

 the right at the same temperature in consequence of the greater 

 carbon concentration than in case the deposition of graphite takes 

 place, and this renders it possible that the liquid phase becomes 

 super-saturate also with respect to cementite and that this can 

 deposit at Lc. 



The reason why the liquid which is supersaturate with respect to 

 graphite and cementite, will deposit not graphite, but cementite, is 

 in connection with the fact that the undercooled liquid shall be soon 

 more supersaturate with respect to cementite as to graphite '). So 

 we see from this that the assumption of an internal equilibrium 

 in the liquid can give a very plausible explanation of the discussed 

 phenomenon. 



In the theory of allotropy I have further made the very obvious 

 supposition that when in case of coexistence of a liquid phase with 

 a mixed crystal phase or in case of coexistence of two mixed crystal 

 phases there is reason to assume that one of the coexisting phases 



1) I touch here npon a frequently observed phenomenon, from which Ostwald 



wanted to derive a "Gese/s der Umwandhmgsstufen". This important question 



being the subject of a subsequent communication, I shall not discuss it any 

 further here. 



