268 
THE TELEGRAPH 
units. If the fault has no resistance at b, the measure from c 
to b will be 400—150 or 250 units. 
But we must endeavour to eliminate the resistance of this 
fault. Now, the resistance of c to e = 280 units ; therefore 
(400 + 150) - 280 
AB = . - -— = 135 
and 
(400 + 280 - 150 
bc = --- = 265 
each of these quantities divided by five shows that the fault is 
twenty-seven miles from A, and fifty-three miles from b. 
A 172 B 55 K d 
This method would be perfect if the resistance of the fault 
were really constant and did not vary during the testing ; but 
these faults vary much, and especially during the testing, in 
consequence of the polarizing effects of the current; and from 
this it happens that, except in the case of great faults with a 
small resistance, the method is defective. There are, how¬ 
ever, other more exact methods, and of these we may explain 
to our readers the following. 
Defects of insulation in submarine lines are usually due to 
a perforation of the gutta-percha. The opening gradually 
enlarges from the effects of the current, although polarization 
