190 Proceedings of the Royal Society of Edinburgh. [Sess. 
Similarly, to get the components along PB — 
For rectangle (1) use — l y . 
(2) „ -I, 
'» V (3) „ +h/. 
» » (4) „ +!*• 
Add these, and we get the resultant along PB. Let us call it F 2 , and also 
let the resultant field at P be H. 
H= 
We have then to evaluate ffUdxdy. 
This integral, however, is unmanageable, and I have made an approxi- 
mate calculation as follows : — Suppose a quarter of the cross-section of the 
strip to be divided up into sixteen rectangles similar to the strip. Find 
the values of F 1 and F 2 at the centre of each of the rectangles, and obtain 
3 
the value of H at each of these points, then finally calculate the arith- 
metical mean of the values of H. In this way I obtained an approximation 
from sixty-four points for the whole strip. 
To eliminate the residual effects of stress, the wire was, in all the 
experiments described in this paper, demagnetised after the load had been 
put on. It was also placed at right angles to the earth’s field. 
Iron. 
A wire from the same coil as was used in the experiment in the second 
paper was taken. First, the modulus was determined at the temperature 
of the room, and was found to be the same as in that experiment, the 
total load being 1 kilo. The current was next gradually increased up to 
