CONDUCTIBILITT OF MEECUEY AND AMALGAMS. 
835 
It will no doubt be remembered that we also experimented upon bars which, instead of 
being composed of cubes, were made of two longitudinal bars of copper, juxtaposed and 
soldered to two other longitudinal bars of either tin, zinc, or lead, and that all these 
bars conducted heat as if they had been entirely composed of pure copper, and had 
not contained half their bulk of tin, zinc, or lead. We have made, since those results 
were published, a great number of experiments with the hope of throwing some light 
on the above interesting fact, but we regret to say without success, We have, however, 
noticed a result which deserves to be recorded; it is, that a bar composed of 2 of 
bismuth and 2 of antimony, juxtaposed and soldered together, is the only one which 
conducts heat in the same ratio as indicated by theory ; for example, — 
No. 3 bar 
2 bars of Bismuth 
2 bars of Antimony 
Found. 
3-90 
Calculated. 
3-63 
We have also ascertained that the fine film of solder existing between the blades 
exerts no influence whatever on the proportion of heat conducted by the compound 
bars, for we have — 
No. 4 bar soldered' 
2 of Copper 
2 of Zinc 
gives 26’85. 
No, 5 bar, in which gutta percha was employed f 2 of Copperl . 
to keep together the four small blades . .12 of Zinc 26 3 . 
alloys (see page 833). It has been pointed out to us that in the case of the bars composed of dilferent 
metals placed_ end to end, the theory of the conduction of heat leads to the following simple result : — the 
resistance of the whole bar, midtiplied by its length, is equal to the sum of the specific resistances 
of the separate metals multiplied by their respective lengths, the resistance being measured by the reci- 
procal ol the conductibdity. This gives for the conductibilities of the bars Nos. 1 and 2, the reciprocal of 
the mean of the reciprocals of the conductibilities of the two component metals. Taking the numbers 
given in the former paper, copper (rolled) 26-95, antimony (mean of two) 6-485, bismuth 1-95, we thus find 
lor bar No. 1, 10-45, for No. 2, 3-64, which do not greatly differ from the numbers given by experiment. 
