206 MR. S. IttDWELL ON CHANGES PRODUCED BY MAGNETISATION IN 



It is clear that if such experiments could be made with iron rings instead of rods 

 several of these objections would at once disappear. The field due to a current through 

 a wire wrapped evenly round the ring would be very approximately uniform and easily 

 calculable, while the ends, with their attendant disadvantages and uncertainties, would 

 no longer exist. I have from time to time during the last two years made attempts in 

 this direction, but the practical difficulties encountered turned out to be unexpectedly 

 great, and it is only recently that I have succeeded in surmounting them. 



To bring the strong fields required within the compass of a battery of reasonable 

 size, it was essential that the rings should be small. On the other hand, if they were 

 too small, the exceedingly minute changes in their diameters due to magnetisation 

 would not be measurable. Ten centimetres had been fixed upon for the length of the 

 rods used in the old experiments, as being the shortest which was advantageously 

 possible ; but a ring 10 cm. in diameter would require more than three times the 

 number of turns of wire necessary for the rods to give an equal field, and the increased 

 resistance thus introduced would demand a correspondingly larger battery. It was 

 finally determined to give the rings a mean diameter of 6 cm., but with this small 

 diameter the number of turns of wire per centimetre which could be properly 

 wound on was less than in the case of the straight solenoid. The strongest field yet 

 obtained with such a ring, vising a battery of 30 GROVE'S cells, was 450 C.G.S. units. 

 When the same battery is employed with the straight solenoid the strength of the 

 field at the centre reaches 840 units. 



Again, the troublesome and misleading effects of heat are much less easily avoided 

 in a ring than in a straight rod. The importance of securing a steady temperature 

 may be inferred from the fact that, taking the coefficient of expansion of iron to be 

 '0000 122, the elongation accompanying a rise of temperature of a single degree 

 is more than eighty times greater than that which it is desired to measure. Several 

 rings were prepared before the difficulty was surmounted of sufficiently protecting the 

 iron from the heat generated in the surrounding coil by strong currents. Unless an 

 interval of at least half a second intervenes between the closing of the circuit and 

 the commencement of the consequent heat expansion, no satisfactory observation 

 can be made. 



In order to get rid of the permanent magnetism which remains after every obser- 

 vation, a modification was adopted of the device employed by Professor EWING, and 

 described by him in 'Phil. Trans./ vol. 176, p. 537. For the variable resistance it 

 was found advantageous to use a long German silver wire, wound in a close spiral 

 upon a wooden cylinder, instead of a tube containing sulphate of zinc solution. 

 Contact with the wire was made by a transverse spring which was capable of sliding 

 from end to end of the cylinder. The arrangement is indicated in fig. 1. It will be 

 seen that the German silver spiral AB acts to the coil wound upon the ring S as a 

 shunt, the resistance of which gradually diminishes from 26 ohms* to nothing as the 



* The resistance of the whole sp'ral. 



