42 THE ADDRESSES, LECTURES, ETC., OF 



found that when a good cable is submerged, the electrical resistance 

 or its insulator must increase permanently three-fold ; it must 

 further increase on account of the lowness of the temperature at 

 the bottom. For this double reason the insulating conductor is 

 under the most favourable circumstances that could be imagined, 

 when submerged in a deep ocean. This increase of insulation 

 under great hydrostatic pressure does not affect all materials alike. 

 India-rubber-covered wire, when subjected to the same pressure of 

 300 atmospheres, decreased in insulation, although it immediately 

 resumed its original high degree when the pressure was released. 

 The change of conductivity by pressure is therefore not reducible 

 to any general law, but applies to each individual substance 

 differently, but it is very important to know in dealing with 

 cables, what changes we may have to expect in submerging 

 them. 



Having tested the conductor at the insulating works, it is 

 transferred to the sheathing works, where it is again tested, both 

 during the time of its being covered, and afterwards when it is 

 put into tanks. The usual way of testing it while it is being 

 covered, is to put the conductor into circuit with a battery and an 

 alarum, which latter is made to ring the moment the continuity 

 is broken. Mr. Schwendler, an electrician connected with me, 

 has lately devised a plan by which a continued test for insulation 

 may be applied, and by which an alarum is made to sound the 

 moment the insulation is injured. This is of more importance 

 than the mere continuity test, inasmuch as in the process of cover- 

 ing a cable with iron the soft gutta-percha, or india-rubber, is far 

 more likely to be injured than the conductor itself. 



I may also mention here another investigation which has lately 

 been made by Mr. Schwendler, which is to determine the best 

 resistance, or the resistance producing the maximum of magnetic 

 moment in the galvanometer employed in testing with the Wheat- 

 stone diagram. I have in the early part of the lecture said that 

 in order to weigh accurately with Wheatstone's diagram, the 

 branches of the system should be as nearly as possible equal. But 

 no rule has as yet been established to determine, if you have to 

 deal with resistances of a certain average value, what resistance 

 on the galvanometer would produce the maximum effect. Mr. 

 Schwendler finds that the resistance of the galvanometer should 



