320 Mr. A. K. Chapman on the 



That the form of these curves does not depend upon the fre- 

 quency of reversal of the radial current is demonstrated by the 

 entire agreement of experiments made with twenty reversals 

 per second with those where five or six per second were used. 



In view of the possibility that these peculiar effects in the 

 iron disks might arise through the close proximity of the 

 heavy pole-pieces of the magnet, through the fact that one 

 of them is traversed by a one-centimetre hole, or through 

 some other unnoticed circumstance, it was thought that a 

 repetition of the experiments would be of value, with the 

 disk in a solenoid. For this purpose a suitable solenoid, 

 137 cm. long and 22 cm. in diameter, was constructed. It 

 was wound on a brass tube of 7 cm. internal diameter, and 

 consisted of three sections having respectively 4347, 3450, 

 and 3490 turns of number 10 copper wire, thus giving 

 fields of 40*22, 31*88, and 32*30 lines per square centimetre 

 per ampere. In series with each section was a one-tenth 

 ohm standard resistance immersed in an oil-bath ; the 

 potential drop across these standards could be read with a 

 Wolff potentiometer, thus giving an accurate measure of 

 the current, and hence of the field due to each section. At 

 the higher fields the large currents produced a marked 

 heating of the apparatus, so that it was necessary to allow a 

 liberal stream of water to flow through the inner brass tube 

 in which the disk was placed. The disk and inductance- 

 coils were exactly the same as before except for the fact 

 that they were enclosed in a water-tight brass box. In this 

 way the temperature of the disk was kept constant to within 

 a very small range, even though the solenoid itself became 

 warm. As before, readings on the variable inductance were 

 taken with the magnetic field in both directions, and the 

 determinations of the current flowing in the sections were 

 made both before and after adjusting the balancing inductance. 

 The numbers given are the means of series of readings. 



The solenoid has the advantage of affording a uniform 

 field over the whole disk, together with the possibility of 

 dealing accurately with much lower fields and with much 

 smaller variations in higher fields than is possible with the 

 usual type of magnet. It is obvious, from the figures given 

 for the number of lines per square centimetre per ampere, 

 that it is easy to vary the fields to the greatest nicety and 

 also to determine them with an accuracy more than amply 

 sufficient for this purpose. There is another advantage of 

 great importance in the manipulation. In the arrangement 

 used, there are always coils of very high inductance in series 

 with the galvanometer and, at the same time, in the magnetic 

 field, so that any unsteadiness in the magnetizing current, 



