486 Professor H. C. H. Carpenter [March 7, 



from 850° C. in water at 15° C. This curve gives the temperature 

 at the centre of the bar at any given moment. It will be observed 

 that it does not fall for the first fraction of a second ; that the rate of 

 fall starts slowly, then increases rapidly to a maximum, afterwards 

 decreasing more and more slowly until it reaches the temperature of 

 the surrounding water. From our present point of view the important 

 thing to notice is that the time from the moment of immersion until 

 the temperature had dropped to 500° C. was Ij second. As the 

 diameter of the bar increases, the rate of chilling is of course 

 diminished, and no doubt in the case of a rectangular turning tool of 

 l|-inch diameter several seconds would be required for the same drop 

 in temperature at the centre. As McCance has pointed out, the 

 perfect theoretical conditions for quenching are "that a specimen 

 heated uniformly has its surface suddenly cooled to a lower tempera- 

 ture and kept at that lower temperature without alteration until it 

 has once more obtained uniformity, but at the lower temperature ; 

 the temperature of the bar changes then in a manner depending on 

 its thermal properties and dimensions, and the rate of change thus 

 obtained cannot be exceeded between similar temperature limits. 

 Practical quenching depends on how far these conditions are satisfied, 

 which really resolves itself into the simple question : How constant 

 can the surface of the specimen be kept at the lower temperature 

 after immersion in the liquid, and what properties must the liquid 

 possess to fulfil this purpose best ? " McCance accepts Benedicks' 

 conclusion that vaporization plays the most important part in the 

 quenching power of a liquid, and the fact that water with its high 

 latent heat of vaporization is the best quenching liquid known accords 

 with this view. 



What happens when a uniformly heated bar is plunged into water 

 at 0° 0. is somewhat as follows : — The layers of water in immediate 

 contact with the bar are rapidly heated to their boiling-point and 

 steam is formed ; this expands outwards and causes a fresh layer to 

 come in contact with the surface, and so the process goes on. The 

 surface of the bar alternates then between 0° C. and 100° C, the 

 steam acting as a carrier of heat to the body of the liquid ; this goes 

 on until the supply of heat is insufficient to form steam, when the 

 transference of heat and the cooling of the bar take place by con- 

 vection only ; and at this stage of the process a high conductivity in 

 the liquid is an actual disadvantage, since it retards convection. A 

 low viscosity is wanted to enable the steam to travel outwards with as 

 little resistance as possible, wliile a high specific heat ensures that the 

 temperature changes of the liquid are small. 



We have next to consider certain fundamental properties of iron 

 and its alloys with carbon which have been established by modern 

 research, and without a knowledge of which any discussion of theories 

 of hardening w^ould be unintelligible. These })roperties have been 

 for the most part investigated during the last fifty years by a small 



