220 Dr. D. K. Morris on the Magnetic Properties 



carefully burning out all the hydrocarbons in one layer of 

 insulation, leaving, so to speak, nothing but the asbestos 

 ' ash/ a white and very brittle covering ; and then cautiously 

 over-covering this with a layer of ordinary asbestos paper. 



In the first experiments, the leakage between the heating- 

 coil (with 10 volts across its terminals) and the secondary 

 winding was sufficient, when the ring was at 1000°, to drive 

 the astatic needle of the sensitive galvanometer included in 

 the secondary circuit almost round to its stops; and though 

 the needle could be easily brought back artificially to its 

 zero position again, the leakage was uncertain enough in 

 amount to render ballistic observations at this temperature 

 quite out of the question. The introduction of the above- 

 described improvement in the asbestos insulation reduced this 

 leakage to about y^^h P ar ^ °f ^ s former amount ; but it 

 still represented a not very constant zero error of several 

 centimetres. 



With the object of still further reducing the leakage, mica 

 was tried as insulation between the layers. This plan 

 ultimately succeeded very well ; but the difficulty of applying 

 a reliable layer of mica insulation to a surface of " double 

 curvature/' such as that of a small ring, without the assistance 

 of any .adherent whatever, appeared at first insurmountable. 

 Latterly, however, it was found that if the mica insulation 

 for the " doubly-curved " parts was applied in the form of 

 small suitably-shaped pieces, previously heated in a Bunsen 

 flame * and then bent or continuously broken by pressing 

 over a rounded edge, then, by putting each piece of mica 

 into place as the coil of insulated wire above it was wound, a 

 reliable layer of mica insulation could be obtained without 

 much trouble. 



Mica was also used in the later ring-magnets both for the 

 insulation of the iron strip and for that of the thermometer- 

 wire, in a way about to be described ; it was not only put 

 between the turns of the iron strip, but was also placed next 

 to it, beneath the asbestos covering ; since any carbon deposited 

 from the latter would, if in contact with the glowing iron, 

 combine with it and alter its character. 



It is easily shown that no appreciable error in the measured 

 resistance of the iron or thermometer- wire would arise from a 

 conductivity of the insulation sufficient to give rise to even a 

 large deflexion of the ballistic galvanometer needle. 



Construction of Ring-Magnets. — The rings containing 



* This diminishes the tendency of the mica to split, and seems to do 

 so by burning ofF the raw cut ed^es. 



