374 THOMAS GRAY ON THE EFFECT OF PERMANENT ELONGATION 
The second column gives the weight on the wire when the resistance was 
measured ; the third column gives the percentage elongation, being the sum 
of the permanent and elastic elongations in each case ; the fourth column gives 
the percentage increase of specific resistance from the beginning; the fifth 
column gives the density in those cases where it was determined; the sixth 
column gives the number of the experiment in the order of their performance 
for each specimen. 
The mean of the first four experiments on Japanese copper gives 0°55 per 
cent. increase of specific resistance per 1 per cent. increase of length. This 
number is increased to 0°61 if the fifth experiment be included. The fifth 
experiment was made for the purpose of finding the effect of elastic elongation, 
and the change of density due to stretching. The only difference in the 
performance was the occasional removal of the weights, and a more rapid 
stretching than in the previous cases. Whether either of these causes may 
have influenced the result I am unable without more investigation to say. 
The experiment on English copper given in Table II. shows an increase of 
0°61 per cent. per 1 per cent. increase of length, which corresponds with the 
mean of the five experiments on Japanese wire. This wire was much thicker 
than the Japanese, the former being No. 15 B.W.G., and the latter 22 B.W.G. 
The variation of copper due to stretching seems very nearly proportional to 
the amount of stretch. 
In the case of the iron wires, the variation was least for the specimen of 
English iron (Table III), and greatest for a very hard strong specimen of 
Japanese iron (Table V.). Probably the difference is mostly due to the greater 
elastic elongation in the harder specimens. The percentages may be roughly 
taken as about 0°73, 0°79, and 0°82 increase of specific resistance per 1 per 
cent. increase of length. 
One of the curves for the Table IV. is somewhat irregular ; the irregularity 
is perhaps due to the fact that the wire was heated in a BuNSEN flame before 
experimenting on it. Generally the curves of results for iron show convexity 
upwards on the whole, but have a marked hollow towards the middle portion 
which runs through all the curves. 
The German silver wires gave the smallest variation of all that were tried, 
being about 0°54 per cent. per 1 per cent. for the first 4 or 5 per cent. of 
elongation, and rather less as the elongation increases. One of the experi- 
ments on German silver shows considerable irregularity ; this must have been 
due, I think, to want of homogeniety in the wire, but possibly in part to error 
of observation. 
The result of Table I. No. 5 shows a variation of 2°57 per cent. per 1 
per cent. for elastic elongation in this specimen of copper. On the other hand, 
the English copper shows a variation after it has been stretched 7 per cent. of 
