40 
PROFESSOR W. A. TILDEX ON THE SPECIFIC HEATS OF METALS AND 
in the case of col)alt and nickel (see Appendix to Bakerian Lectfre), is apimrently 
nntenal )le. 
Plotting the specific heats in Table II. against absolute temperatures, the curves 
shown in fio’. 2 are obtained, from which it is obvious that unless some remarkable 
change in the specific heats of silver and platinum occurs helow — 182'’ C. the curves 
rejiresenting atomic heats cannot meet at the absolute zero. 
It will be observed that the influence of rise of temperature on the specific heat is 
in the inverse order of the atomic weights of the metals compared, being greatest in 
the case of aluminium and least in the case of platinum. This appears to he generally 
true and is supported l)y the experiments of Behn (‘Wiedemann’s Ann.,’vol. 6G, 
]). 237). It appears, therefore, that the usual application of the law of Dflong and 
Petit to the rectification of atomic weights is a rough empirical rule vdiich, setting- 
aside boron, carbon, silicon, and beryllium, is oidy available v-hen the specific heats 
have been determined at comparatively low temperatures, usually, and most con¬ 
veniently, between 0° and 100° C. 
What mechanical properties of the metals are concerned in affecting the value of 
the specific heat is not known. The work done in expansion has apparently very 
little to do with it. Lead and platinum, for example, the atomic weights and specific 
heats of which are near together, have very different coefficients of exjiansion, that of 
lead being nearly ten times as great as that of platinum. I have, however, on the 
suggestion of Professor Perry, thought it of some interest to determine the specific 
heat of the remarkable nickel steel which is said to have a smaller dilatation than 
that of any other metal. The sample used was found by analysis to contain 35‘92 
per cent, of nickel, practically 36 per cent., with 'll of carbon and about ‘30 of 
manganese. The mean specific heats observed at four widely separate temperatures 
show that there is decidedly an increase with rise of temperature to an extent about 
the same as in the case of nickel itself 
Range of temperature. 
]Mean specific heat of 
nickel steel. 
- 182 to + 1°5 
•0947 
15 „ 100 
•1204 
15 „ .360 
• 1245 
15 „ GOO 
•1258 
Evidence as to the cause of the difference lietween two such metals as nickel and 
silver has been sought by making comparative experiments with the two metals in 
the form of sulphide. These compounds were prepared by precijiitation by hydrogen 
sulpliide from solufions of the sulphate and nitrate respectively, and subsequent 
fusion of the dry sulpliide with an excess of sulphur. If the differences between the 
metals are due to peculiai-ities of the atoms of each, similar differences would be 
