538 



NA TURE 



[April 7, 1892 



rapidly, and in a few moments it will become too hot to 

 hold. The " fracture " of the metal is totally different 

 before and after the molecular change, which is the 

 cause of this evolution of heat, has taken place. The 

 change, moreover, takes place in the solid metal, and is 

 not due to the release of the latent heat of fusion. The 

 mass, solid as it appears to be, must be the scene of an 

 internal struggle between the molecules in the effort to 

 attain a state of equilibrium, and this conflict is but a 

 type of the action that takes place in many metals and 

 alloys which are of vast industrial importance. 



Time will only permit me to deal with three cases of 

 the action of high temperatures on atoms and molecules 

 of metals. In the first case, the arrangement of the 

 atoms in the molecule of a metal, iron, is disturbed, and 

 the result is of great industrial importance. In the 

 second case, the atoms of a metal, gold, appear to com- \ 

 bine with those of another metal ; and the result, while it 

 is mainly of interest in connection with the history of 

 science, has nevertheless an important bearing upon art. 

 The third case relates to the molecular bombardment which 

 takes place when a small quantity of metal is dissolved 

 in a mass of a metallic solvent, and is of interest in con- I 

 nection with modern views both as to osmotic pressure \ 

 and solution generally. j 



(i) The pyrometric couple is inserted in the centre of ' 

 a little mass of steel, which is being slowly raised to a 

 bright red heat ; when the flame is withdrawn, the spot of 

 light will return towards the zero end of the scale, falling 

 slowly until a temperature of 655° is reached, and then ] 

 there will be an abrupt and prolonged arrest. The metal , 

 has never been near its melting-point, and the evolution | 

 of heat must be due to a molecular change in the solid 

 metal. In the case of this particular sample of steel, 

 the evolution of heat is mainly the result of a change in 

 the relation between the carbon and the iron ; but by 

 laboratory experiments and careful chronographic records, 

 Osmond has shown that, in the case of certain varieties 

 of steel, it can be demonstrated that what here appears 

 as a single change, attended by an evolution of heat, is 

 really an exceedingly c©:Tiplex one. I have shown that 

 it occurs in the purest iron the chemists can prepare by 

 electrolysis, and I agree with Osmond in believing that 

 the change which occurs in pure iron at 855° is a molecu- 

 lar one, independent of the presence of impurity If the 

 mass of steel (Fig. 5, a) be heated again and allowed to 



SCREEN 



Fig. 



■cool, you will observe that the point of " recalescence " 

 appears to be that at which the iron regains its magnetic 

 property ; ^ for a magnetized needle, b, is attracted at the 



I The temperature at which these molecular changes take place in iron and 

 steel was first demon -t rated to an audience in my Newcastle lecture, 1889; 

 but my friend Prof. Reinold, of the Royal Naval College, first arranged an 

 experiment for lecture purposes which showed the magnetic change 

 simultaneously with the thermal one. 



NO. I 171, VOL. 45] 



moment the arrest of the spot of light on the pyrometer 

 scale marks the temperature at which the change occurs, 

 and at that precise moment a second spot of light from a 

 mirror, mounted on the magnetic needle, will rapidly 

 move away from its zero. 1 have elsewhere ^ dwelt on 

 the importance of the molecular change in iron and steel, 

 and can now only summarize the significant facts. 



It is unnecessary to point to the extreme industrial 

 importance of the property steel possesses, of becoming 

 hard when it is quenched from redness in a fluid which 

 will abstract its heat with more or less rapidity. 



The changes which take place at 855' and 650'' have to 

 be arrested, as it were, by rapidly cooling the mass of steel ; 

 and if this is done, the steel will be more or less hard ac- 

 cording to the rapidity with which the progress of the 

 molecular change has been stopped. It is, however, 

 useless to attempt to harden steel if the temperature of 

 the mass has fallen below 650°. In " oil hardening" or 

 cooling a large mass of steel, like the " A " tube of a gun, 

 which may be 30 feet long, great care should be taken to 

 insure that the temperature of the mass is as uniform as 

 possible ; for, if part of the mass is hotter than 650'', while 

 part is colder, the oil will really be cooling a mass of steel 

 which is itself passing through various stages of complex 

 molecular change, and the operation of " hardening " 

 arrests, as it were, the atoms in the midst of a conflict 

 incidental to their attempt to group themselves into one 

 or other of the molecular modifications of iron. By cool- 

 ing a mass of steel which is not at uniform temperature, 

 stresses of great complexity and intensity are set up, 

 stresses that may greatly reduce the effective strength of 

 the gun.^ The result is told in failures, by which many 

 lives have been sacrificed ; but I need hardly say that the 

 Director-General of Ordnance is fully sensible of the 

 national importance of studying the behaviour of iron and 

 steel at high temperatures, and, at Dr. Anderson's sug- 

 gestion, the Institution of Mechanical Engineers ap- 

 pointed a Committee, and have intrusted me with a large 

 portion of the inquiry. 



In the next experiment, Fig. 6, a bar (a) of steel, iinch 



in section and 18 inches long, was heated to bright redness 

 and fixed firmly at one end ; a weight of about 2 pounds 

 is rapidly hung to the free end, a light pointer is added to 

 magnify the motion of the bar, and the thermo-junction 

 is rapidly introduced into a small hole drilled in what is 

 arranged to be the hottest part of the bar. The bar is 

 not softest at a red heat ; it remains perfectly rigid until it 

 has cooled down to dull redness, and the temperature, as 

 measured by the spot of light from the galvanometer, 

 shows that " recalescence " has occurred. At that mo- 



' Report to the Instliulion of Mechanical Engineers, Proceedings, 

 1891, p. 543. 

 ^ " Internal Stresses in Cast Iron and Steel," by Nicholas Kalakoutsky, 



