218 
Proceedings of the Royal Society of Edinburgh. [Sess. 
is obtained. The wooden pieces GG carry both the primary and secondary 
coil terminals P and S. There were 4600 turns of No. 30 wire on the 
secondary circuit with a resistance of 117’4 ohms, while the primary 
circuit consisted of 120 turns of No. 17 wire with a resistance of *262 ohm. 
Galvanometer. 
This was of the d’Arsonval type. The resistance of its coils, found by 
Mauce’s method, was 10*92 ohms. 
Test Pieces. 
All the test pieces were turned down to a diameter of 2 cm. They 
were each slightly more than 20 cm. long, with a length of 6*5 cm. in 
the centre which was filed square. When in position in the apparatus 
this square part was wholly within the coil E. 
Method of Conducting Experiments. 
The induction curve for the iron or steel test piece was first obtained 
with the piece in its normal condition. It was then taken out of the 
apparatus, and its dimensions taken. It was next demagnetised, and then 
compressed to 5 tons per square inch. This was carried out in a 50-ton 
Wicksteed testing machine in the Mechanical Engineering Laboratory. 
Pieces of cast iron AA as shown in fig. 2 were made to fit over the square 
part of the test piece. Pressure was then applied by the machine in the 
direction indicated by the arrowheads. It will be seen that this method 
of applying the pressure leaves two of the sides of the square part free to 
expand laterally. The effect of this is shown in fig. 2 by the two shapes 
of these free faces before and after compression. The measurements of 
dimensions given were all taken at the centre, i.e. where the bulging is 
greatest. After this compression to 5 tons per square inch, the dimensions 
were again taken, the piece was inserted into the apparatus, demagnetised, 
and its induction curve then obtained as before. It was again demag- 
netised, taken out of the apparatus, and compressed to 10 tons per square 
inch, and so on as before. 
This was continued till the specimen broke or became too distorted 
to enter the coil. 
Method of Demagnetisation of the Pieces. 
Demagnetisation was necessary to ensure that the tests would be reliable, 
and was carried out both before and after compression, since it is well 
