740 



Professor W. A. Tilden 



[May 13, 



cannot be properly expressed in tlie words of the discoverers. For, 

 although the exact values of the atomic weights of these two elements, 

 cobalt and nickel, are not known, it is certain that they are not so far 

 apart as would be implied by these values for the specific heats. 



Two other examples of somewhat similar kind are shown by gold 

 and platinum, copper and iron. 



For the gold I naturally applied to my colleague, Professor 

 Eoberts- Austen. The platinum I prepared from ordinary foil, by re- 

 solution and re-precipitation as ammonio-chloride, and subsequent 

 heating. Both metals were fused into buttons before use. The 

 atomic heats come closer together than those of Co and Ni. 



Copper and iron differ considerably in melting point, but both 

 at the temperature of 100° are far removed from even incipient 

 fusion. The copper was prepared from pure sulphate by electrolysis, 

 the iron by reduction of pure oxide in pure hydrogen. Notwith- 

 standing all our care, it was disappointing to find it contained • 1 per 

 cent, of carbon, the source of which I am at a loss to explain. This 

 iron is purer than any examined by Eegnault or Kopp. 



Specific Heats op Gold and Platinum. 



SrEciFic Heats of Copper and Iron. 



Fused. 



20° 

 Copper (pure) S. G. ^^ 8-522. 



•09248 

 •09241 

 •09205 

 •09234 



Aritb. mean 

 Atomic heat 



• 09232 

 5-83 



15° 



Iron, S. G. — r^, 7 '745, contains 

 Q-Ol per cent, copper. 



11022 

 ■11037 



Aritb. mean .. 

 Atomic beat , . 



•110.30 

 0^13 



Tlie differences observed between cobalt and nickel, and between 

 gold and platinum, are manifestly not due to allotropes or to differ- 



