WILLIAM SIEMENS, F.R.S. 97 



iliM'rrent times. The tests of the first portion, though appearing 

 rather low in insulation, proved otherwise regular, and tend rather 

 ablish a general decrease of the insulating properties of the 

 gutta percha, than to point to any local defect. In the second 

 p" rt ion, however, a fault suddenly appeared, when the cable was 

 roiled on board the Rangoon in August, 1861, indicated chiefly 

 by a fluctuating resistance owing to electrolytic action. If several 

 days could have been allowed to the Government electricians to 

 observe and gradually develop this fault, it might have been cut 

 out before the ship left the Thames ; but such a delay was deemed 

 unnecessary, inasmuch as the fault, if any, would certainly increase 

 during the outward voyage, and the defective piece could be removed 

 at Malta before the laying of the cable was commenced. On test- 

 ing the same cable at Malta, it was found that its insulation had 

 actually gone down from 120 millions of units to 3 millions of 

 units ; but it rose again to 70 millions of units after the faulty 

 piece had been cut out. This piece has since been stripped of the iron 

 sheathing, when a place about an inch in length appeared, where 

 the gutta percha had been deeply indented, apparently by some 

 weight falling upon it during the manufacture of the sheathing, 

 which did not expose the copper conductor to sight, but which 

 had assumed the character of a fault in coiling the cable from 

 the tanks into the ship's hold. 



On comparing the insulation of the cables after being laid down, 

 with the insulation observed shortly before on board ship, there is 

 a decided improvement after submersion. This is partly due to the 

 pressure upon the cables, the insulation improving 2 per cent, on 

 an average, per 100 Ibs. of pressure upon the square inch, and 

 partly to the lower temperature at the bottom of the sea. The 

 latter circumstance influences the tests, by the introduction of a 

 factor, the value of which it is difficult, if not impossible, to ascer- 

 tain, owing to the changeable nature of gutta percha ; thus pre- 

 venting a direct comparison between the last tests on board ship 

 and those after submersion. All that could be expected was, there- 

 fore, to find an adequate increase of insulation in testing the sub- 

 merged cable, as compared with the last test on shore ; and that 

 the electrification, both by positive and negative currents, should 

 show the regularity due to a homogeneous condition of the insu- 

 VOL. n. H 



