Condition of Iron by Heating to Redness and Cooling. 477 



gravity upon tempering. This decrease is the greatest in the 

 case of steel and white cast iron, smaller with the iron wire (c), 

 and smallest with the cast iron and the electrolytic iron. But 

 whilst this decrease in density is beyond doubt with all the 

 kinds of steel examined, the iron wire, and the grey and white 

 pig-iron, the mean values found for the electrolytic iron must 

 be received with caution ; since, of the whole fifteen values of 

 8 r — Si, nine are positive and six negative. Hence I consider 

 it not improbable that a larger number of measurements 

 (which was not possible on account of the gradual diminution 

 in the size of the piece) would give a mean value approaching 

 zero. I therefore regard the existence of a difference in den- 

 sity of electrolytic iron, after slow cooling or rapid cooling, as 

 not proved. The different varieties of iron are classified, as 

 we know, according to the quantity of carbon they contain, 

 which may be present either in the free condition or chemically 

 combined ; for the percentage of carbon determines, if not 

 exclusively yet to a great extent, the properties of any par- 

 ticular kind of iron. Hence I had no doubt that both the 

 sign and magnitude of the change in density produced by 

 tempering would be dependent upon the percentage of carbon. 

 The observations indeed do not contradict this assumption, 

 but they by no means determine the law of the connexion. It 

 results from the measurements made with cast iron and with 

 white pig-iron, that the change in density depends less upon 

 the total percentage of carbon than upon the quantity of che- 

 mically combined carbon. For in the case of white pig-iron 

 containing much combined carbon, the change was ten times 

 as great as with grey pig-iron, in which the carbon was con- 

 tained chiefly as graphite. 



Whilst with white pig-iron and steel (substances containing 

 the largest percentage of combined carbon) the density dimi- 

 nished upon tempering, it increased with the three kinds of 

 iron wire examined — substances which contained only a small 

 percentage of combined carbon, The third kind of iron wire, 

 obtained by drawing down one of the first two kinds, behaved in 

 the opposite way; which may possibly be explained by a change 

 of free carbon into combined carbon, caused by the operation 

 of drawing. If so, smaller quantities of chemically combined 

 carbon in iron produce an increase in density, and larger 

 quantities a decrease, when the red-hot wire is plunged into 

 cold water. 



This assumption does not appear to be negatived by the 

 two observations, which alone, so far as I know, have shown 

 an increase in density of steel upon tempering. For, accord- 

 ing to a table given in Scheerer's Metallurgies Styrian steel, 



