151 
The mode of manufacture of the 8 samples of steel leads 
one to believe that there is a regular increase of carbon and 
other impurities from No. 8 to 1. The writer hopes shortly 
to analyse these wires, as the resistance of some of the 
samples is most remarkable, being no less than 3 times that 
of pure iron. 
No attempt has been made to reduce the electric observa- 
tions by calculation to some fixed temperature, as it seemed 
improbable that samples with such varied resistances would 
increase in the same proportion as pure iron. 
Table B shows us that annealing diminishes the electrical 
resistance of bright steel wire some 1 per cent, while 
hardening and tempering increases the resistance of an- 
nealed steel about 5 per cent. Now we may observe that 
the greatest increase of electrical resistance, viz,, 8T per 
cent, and of breaking strain resulting from hardening and 
tempering, is found in piano steel, thus clearly showing that 
any process like wire drawing or tempering, which makes 
the particles of steel more rigid and difficult to separate, 
increases the electrical resistance. 
The torsion tests vary a good deal among themselves, but 
it is interesting to observe, from table B, that the mean 
breaking strain of hardened and tempered steel is to that 
of the same steel annealed in the inverse ratio of the 
number of twists in 8 inches. The electrical resistance of 
the samples of steel is seen from table B to be usually 
inversely proportional to the number of twists in 8 inches. 
It is generally allowed that the heat conductivity of 
metals is nearly proportionate to their electric conductivity. 
If this relation between the thermal and electric conduc- 
tivity holds good for different qualities of iron and steel, then 
