( 533 ) 



difïieult to show that tliis contradiction is only apparent, and that a 

 very obvious exphiiiation may he given for it. 



Already in my theory of allotropy ^) 1 pointed out that if we can 

 get different substances as new phases from an homogeneous phase 

 .dependent on the method of working, it will in general have to be 

 assumed that the kinds of molecules which give rise to the format ioji 

 of these ditferent substances were already present in the homogeneous 

 phase beforehand. --' -r .■:, ^ - : ir.-' ;;" ■-.'-. 



., • As from a definite liquid mixture of iron and graphite graphite can 

 deposit in case of slow cooling, and cementite in case of rapid cooling, we 

 shall have to assume that not only carbon, but by the side of it 

 also iron carbide is present in the liquid mixture, which substances 

 form the following internal equilibrium with the iron in the homo- 

 geneous liquid phase : 



3Fa + C^Fe3C. 



Let us assume that we start from the state P and cool the homo- 

 geneous liquid phase, then the depositing of graphite will begin at 

 L'g when the cooling takes place slowly. It follows from this that 

 at this temperature the internal equilibrium of the liquid phase is 

 such that the carbon has just reached its saturation concentration, 

 and not the iron-carhide. 



On decrease of temperature the internal equilibrium will be shifted 

 in consequence of two influences. If we consider these influences 

 separately, we observe that in the first place the homogeneous equi- 

 librium is shifted to the left by the decrease of the carbon concen- 

 tration, in which therefore the iron carbide dissociates more and 

 more, and in the second place the decrease of temperature will bring 

 about a shifting. 



. If the formation of iron-carbide in the liquid is endothermic, the 

 decrease of temperature will also bring about a dissociation of the 

 iron-carbide, so that in this case it may be expected, though not 

 with perfect certainty, that the homogeneous phase will not be 

 saturated with respect to iron-carbide at any temperature. ^) 



If the formation of iron-carbide in the liquid was exothermic, it 

 would be possible on the contrary, that the influence of the temperature 

 begins to predominate, and at a given temperature the liquid became 

 safui'a'e not only with respect to graphite, but also with respect to 



1; These Próc. XII, p. 763. : vMn-.;r, - •" 



• 2) The contrary, however, remains still possible, for the decrease of the solu- 

 blHly might prevail, and then saturation with regard to the iron-carbide might 

 occur. 



36 



Proceedings Royal Acad. Amsterdam. Vol. XIV. 



