104 BELL SYSTEM TECHNICAL JOURNAL 



volved: an allotropic change of the iron from gamma to alpha and a 

 precipitation of the carbide FesC. This matter is more fully discussed 

 in a recent paper ^ to which those interested are referred. 



Troostite is not like martensite in respect to habit of formation. 

 It does not assume fully the old austenitic crystalline symmetry. It 

 seems to have a new crystalline orientation of its own. 



Troostite develops along grain boundaries and within the grains. 

 It may develop as a spine or branch along a crystallographic plane. The 

 nodules may be roughly spherical masses; they may be semi-spherical 

 masses, the flat side being bounded by a crystallographic plane or a 

 grain boundary, or they may be rounded but irregular shaped masses. 

 In any event, whether in ball-shaped masses or some constricted form, 

 the small fan-shaped grains are found radiating from a nucleus of 

 growth. This nucleus in most cases can be identified as an inclusion, 

 a void, or a sharp corner in a grain boundary. Thus the tendency is 

 for reorientation of the iron to occur when nodular troostite develops. 



It is well known that a slow rate of cooling promotes more troostite. 

 Rapid cooling results in more needles or the constituent we call marten- 

 site. Evidently when the rate of cooling is favorable the freshly trans- 

 formed alpha iron is given time to reorient itself and does do so by 

 growing about some convenient inclusion or other body. Thus nodular 

 troostite develops. Whether the carbide is held in solid solution in the 

 freshly transformed alpha iron seems to be a matter of speculation. 



A small specimen of steel weighing less than ten grams, heated to 

 1000° C. in a vacuum for a suitable length of time, and quenched in ice 

 and brine will contain almost innumerable troostitic bodies, many of 

 them very small, and some quite large. The larger ones perhaps are 

 a few ten-thousandths of an inch in diameter and from this dimension 

 the troostitic particles decrease in size to the vanishing point of present 

 microscopic resolution which is around 200 atom diameters. 



The specimen itself will have a hardness on the Rockwell scale of 

 about C-65. Nevertheless the troostitic bodies have been clearly re- 

 solved to show the presence of fully laminated pearlite. So that in a 

 small specimen of steel quenched from a high temperature in a very 

 effective cooling bath, one finds not only the needle constituent mar- 

 tensite but nodules of troostite containing fan-shaped grains of fully 

 stratified pearlite. 



The question naturally arises as to whether the steel in its transition 

 from austenite to pearlite first develops a needle structure (martensitic) 

 and then this in turn is replaced by a nodular (troostitic) one. 



Some light ^ was thrown on this angle of the problem by a high power 

 examination of an iron carbon allov. The carbon content was 2.65 



