a a i a ak 
ON MOLECULAR PHENOMENA IN MAGNETISED IRON. 155 
and again between 1,000° and 1,050° C., considerable absorption of heat 
taking place at these temperatures. The lower of these temperatures 
corresponds with the recalescent point and loss of magnetic susceptibility 
in steel. Mr. H. Tomlinson, as already noticed, has also observed two 
critical points in iron, one about 550° C. and the otber at 1,000° C.,! when 
a sudden change occurs in the viscosity of this metal. No doubt these 
are the same points as those observed by Pionchon, for the variable com- 
position of the iron also, and errors in determination of these high tempe- 
ratures are probably sufficient to account for the differences observed. 
The amount of heat given out during recalescence we have estimated 
from the observed expansion of the metal that occurs during recalescence. 
Taking Pionchon’s determination of the specific heat of iron at a red 
heat, the heat liberated in the recalescence of a specimen of iron would 
thus appear to be somewhat over 100 times as much as would raise 
the same mass of iron 1° C. Dr. Hopkinson,? from the length of the 
break in the time curve of cooling, has estimated that the heat liberated 
in the recalescence of hard steel is equal to 173 times that liberated 
when the same material falls 1° C. The amount of recalescence in hard 
steel, as already stated, is considerably greater than that in iron. 
IX. Here it may be mentioned that the hardening of steel by sudden 
quenching in water cannot be produced unless the metal be raised to the 
temperature of recalescence.? Brinnell’s researches have shown that the 
carbon in steel is in two different conditions above and below the reca- 
lescence, and by sudden quenching the so-called ‘hardening carbon’ is 
preserved in the condition in which it exists at a high temperature. At 
a high temperature it appears to be simply free carbon mixed with or 
dissolved in the iron; at the temperature of the air the researches of 
Miiller, Abel, and Osmond and Werth, have shown that in ordinary steel 
carbon is combined with the iron in the form of a compound, having the 
definite composition Fe,C. 
X. We must now consider the general cause of these phenomena. 
The secretary of this Committee long since suggested it was probably to 
be found in the carbon present in the iron, as recalescence was most 
marked in those specimens of iron and steel which contained larger per- 
centages of carbon, and this cause Osmond has now, we think, satisfac- 
torily established. 
Recalescence in steel Osmond attributes to the chemical combination 
of the iron with the carbon present in a free state, and which has been 
liberated by heat. Thus the point of recalescence is that at which iron 
_ ¢arbide, Fe,C, forms; a body which is stable at ordinary temperatures 
but decomposed, with absorption of heat (producing the chilling effect 
observed on heating) at a red heat. Now the heat of combination we 
find to be about 3,000 calories per gram of carbon present in the iron, as 
deduced from Hopkinson’s estimate of the amount of heat liberated 
during recalescence ; further experiments on this part of the subject are 
necessary, and we hope to make them shortly. 
Recalescence in iron Osmond attributes to an allotropic change 
which he believes iron to undergo at a temperature of about 750° C. 
Below this temperature iron exists in one molecular state, which Osmond 
) Phil. Mag., February 1888. 
? Phil. Trans., May 1889. 
’ J. H. Brinnell, Jernkontoret’s Annalen, 1885, and Stahl und Eisen, Nov. 1885. 
Independently observed by one of us, Newall, Camb. Phil. Soc., Jan. 1888. 
