94 Miss H. Brooks on the Damping of the 



was necessary to use only one detector-needle. The accuracy 

 of the experiments depends to a very great extent upon the 

 nature of this needle. Prof. Rutherford (Trans. Roy. Soc. 



1896) has shown that the effect on a magnetized steel needle 

 differs in anv given circuit according to the length of the 

 needle, the hardness of the steel, and the thickness of the wire 

 used in its construction. An examination of a needle after it 

 has been partially demagnetized by a discharge, shows a 

 surface-layer magnetized in a direction opposite to the internal 

 magnetization. Since a leyden-jar in general gives several 

 complete oscillations before it is greatly damped down, it 

 would be expected that the surface-layer of a uniformly 

 magnetized steel needle would be either completely demag-. 

 netized or show evidence of several oscillations in opposite 

 directions. The effect may be explained when the demag- 

 netizing force of the ends of a needle on itself is taken into 

 account. The first half-oscillation that tends to demagnetize 

 the needle has the demagnetizing force of the ends of the 

 needle assisting it, while the return oscillation has it in 

 opposition. The return oscillation will not therefore be able 

 to remagnetize entirely the surface-layer already affected, but 

 a thin layer will be left in the interior ; this layer is added to 

 with every oscillation, until the final effect will be that the 

 surface of the needle will be magnetized in the opposite 

 direction to the interior. The shorter the needle, the greater 

 is the demagnetizing force of the ends upon it. The length 

 found by experiments with a number of needles of different 

 lengths to be most suitable for the conditions of the investi- 

 gation was about 1^ centim. 



A most important point to be considered is the thickness 

 .of the wire. The detector did not consist of a single wire, 

 but of several of the same length made up in the form of a 

 compound magnet, and the several wires insulated from one 

 another by paraffin-wax to prevent eddy-currents. With a 

 rapidly alternating current, each effect lasts for so short a 

 time that the needle does not become demagnetized to any 

 depth, and the outer layer then acts as a "metallic screen for 

 the interior. In such a rapidly alternating field as the 

 leyden-jar circuit under consideration, where the alternations 

 are of the order 10 6 per second, the demagnetization is confined 

 to a very thin layer on the surface. Thick wires are affected 

 to a less depth than thin ones. Erskine (Wied. Ann. vol. lxiL 



1897) has found that the screening effect on the needle in the 

 solenoid B, where the first, third, and fifth half-oscillations 

 demagnetize the needle, is less than on that in A, where the 

 ,even oscillations are in opposition to the original magneti- 



