314 CARNEGIE INSTITUTION OF WASHINGTON. 



noting the temperature (Ar3) at which the precipitation of pro- 

 eutectoid ferrite starts in cooHng down and that (Ac3) at which its 

 reabsorption completes itself in heating up. The results of the 12 

 most important prior determinations of this upper limit have been 

 discussed, and the formulae 



r° = 917°-306XC and T° = 820°- 105.5 XC 



have been reached, in which 7^° = the temperature of this limit in 

 degrees centigrade and C = the percentage of carbon. The results 

 reached with the 19 steels noted above agree closely with this limit. 

 Of the 12 prior sets of determinations, all but 3 harmonize toler- 

 ably with it. The 7 most important prior determinations of the 

 lower limit of the transformation range have been discussed, and the 

 temperature 723° found the most probable. But the existing data 

 are such that these results should be taken as only provisional. The 

 influence of manganese on the temperature of the transformation 

 range has been studied. The data indicate, though not conclusively, 

 that manganese lowers the Ar points, i. e., those noted in falling 

 temperature at the rate of from 25° to 50° per 1 per cent of that 

 metal; that it has but slight effect on the Ac points, i. e., those 

 observed in rising temperature; and that it probably lowers the Ae 

 points or those corresponding to equilibrium, by an amount so small 

 that it is masked by the undercooling. 



Starting from these determinations of the temperature of the trans- 

 formation range, very many experiments have been made to deter- 

 mine the influence of the process of "grain refining," or cooling 

 rapidly from the upper limit of that range and then reheating to a 

 predetermined temperature below that range, and more particularly 

 to learn the influence of specific deviations of the quenching temper- 

 ature from that limit. These experiments are still in progress. 



Many experiments have been made to determine the laws of 

 ''divorcing annealing," called also the ''segregation" and the "dis- 

 entangling" of the constituents of pearhte, the eutectoid of this series 

 of alloys, and to learn the influence of this divorcing on the more 

 useful properties. The results, which are stiU under discussion, show 

 that this divorcing lowers the yield point, and in the case of eutectoid 

 and hyper-eutectoid steels lowers it very greatly, and that (at least in 

 certain steels) it lessens the impact-resistance as usually determined 

 by noting the energy absorbed in the rupture of a notched test-piece 

 by transverse impact. 



The apparently excessive quantity of pro-eutectoid cementite 

 found at certain stages of the transformation of hyper-eutectoid 

 steels, as pointed out in the last report, seems referable at least in 

 part to undercoohng. 



