( '^3] ) 



solul)ili(_v of conieiilito in iron could nol Ix^ pursued hoyoiid =fc 1000° 

 and Bakhuis Roozkboom tliought that (his was owing (o the deconi- 

 posilion of cementite. With supply of heat this conversion would 

 viz. take place : 



cementite —^ niartensite -f- graphite, 

 in consequence of which under ± 1000'' two more two-phase regions 

 would be added to that of martensite -|- graphite, viz. a region for 

 martensite -\- cementite and a region for cementite -{- graphite. 



So under ± 1000° cementite would be able to occur as a stable 

 phase and the formation of cementite would have to take place the 

 more completely below this temperature as the system cools more 

 slowly. Practice, however, teaches just the contrary, so tliat the 

 assumption of cementite appearing as stable phase in the system 

 iron-carbon, has been resolutely rejected by men of practice. Hence 

 I think I may assume with Charpy, Benedicks and others that when 

 cementite occurs in the system iron-carbon, this is always metastable. 



On this assumption, the T, A'-diagram, as may be considered 

 known, becomes as is schematically represented in the subjoined figure. 



