596 EEPOBT — 1891. 



chemistry is far in advance of metallurgic chemistry. I cannot, as yet, state what 

 is the atomic grouping in the brilliantly-coloured gold-aluminium alloy, AuAl^, 

 which I have had the good fortune to discover, but, in it, the gold is probably 

 present in the same state as that in which it occurs in the purple of cassius. 



Much valuable information on the important question of allotropy in metals has 

 already been gathered by Pionchon, Ditte, Moissan, Le Chatelier, and Osmond, 

 but reference can only be made to the work of the two latter. Le Chatelier con- 

 cludes that in metals which do not undergo molecular transformation the electrical 

 resistance increases proportionally to the temperature. The same law holds good 

 for other metals at temperatures above that at which their last change takes place, 

 for example in the case of nickel above 340°, and in that of iron above 850°. 



It is probable that minute quantities of foreign matter which profoundly modify 

 the structure of masses of metal also induce allotropic changes. In the case of 

 the remarkable action of impurities upon pure gold I have suggested that the 

 modifications which are produced may have direct connection with the periodic 

 law of Mendel^eff, and that the causes of the specific vaiiations in the properties 

 of iron and steel may thus be explained. Tlie question is of great industrial im- 

 portance, especially in the case of iron ; and Osmond, whose excellent work I have 

 already brought before the members of this Association in a lecture delivered at 

 Newcastle in 188'J, has specially studied the influence upon iron exerted by certain 

 elements, lie shows that elements whose atomic volumes are smaller than that 

 of iron delay, during the cooling of a mass of iron from a red heat, the change of 

 the /3, or hard variety of iron, to the a, or soft variety. On the other hand, 

 elements whose atomic volumes are greater than that of iron tend to hasten the 

 change of /3 to a iron. It is, however, unnecessary to dwell upon this subject, as 

 it was dealt with last }-ear in the Address of the President of the Association. 



It may be added that the recent use of nickel-steel for armour plate and the 

 advocacy of the use of copper-steel for certain purposes, is the industrial justifica- 

 tion of my own views as to the influence of the atomic volume of an added element 

 on the mechanical properties of iron, and it is remarkable that the two bodies, silicon 

 and aluminium, the properties of which when in a free state are so totally different, 

 shotild, nevertheless, when they are alloyed with iron, afl'ect it in the same way. 

 Silicon and aluminium have almost the same atomic volumes. 



The consequences of allotropic changes which result in alteration of structure 

 are very great. The case of the tin regimental buttons which fell into a shapeless 

 heap w'hen exposed to the rigorous winter at St. Petersburg is well known. The 

 recent remarkable discovery by Ilopkinson of the changes in the density of nickel- 

 steel (couiainiug 22 per cent, of nickel) which are produced hy cooling to —30°, 

 affords another instance. This variety of steel, after being frozen, is readily 

 magnetizable, although it was not so before ; its density, moreover, is permanently 

 reduced by no less than 2 per cent, by the exposure to cold ; and it is startling to 

 contemplate the effect which would be produced by a visit to the arctic regions of 

 a ship of war built in a temperate climate of ordinarj' steel and clad with some 

 three thousand tons of such nickel-steel armour ; the shearing which would result 

 from the expansion of the armour by exposure to cold would destroy the ship. 

 Experimental compound armour-plates have been made faced with 25 per cent, 

 nickel-steel, but it remains to be seen w^hether a siinilar though lessened efl'ect 

 would be produced on the steel containing 5 to 7 per cent, of nickel, specially 

 studied by J. Riley, the use of which is warml}' advocated for defensive purposes. 

 Further information as to the molecular condition of nickel-steel has w'ithin the 

 last few weeks been given by Mercadier, who has shown that alloying iron with 

 25 per cent, of nickel renders the metal isotropic. 



The molecular behaviour of alloys is indeed most interesting. W. Spring has- 

 shown, in a long series of investigations, that alloys may be formed at the ordinary 

 temperature, provided that minute particles of the constituent metals are submitted 

 to great pressure. W. Ilailock has recently given strong evidence in favour of 

 the view that an alloy can be produced from its constituent metals with but slight 

 pressure if the temperature to which the mass is submitted be above the melting- 

 point of the alloy, even though it be far below the melting-point of the most easily 



