WILLIAM .sy/:.!//-;.v.s, F.K.S. 43 



I 



be the same as that of each of the branches of the Wheatstone 

 diagram, to produce the most delicate reading. 



TESTING JOINTS. We next come to a branch of testing which 

 we have not referred to as yet ; that is, testing the joints. The 

 insulated conductor is sent from the insulating works in lengths 

 of one knot ; these have to be joined together before the cable is 

 sheathed, and in making these joints, extreme care has to be used 

 to prevent flaws. The old way of testing joints was to put the 

 freshly made joint into water, to see whether, by doing so, the 

 general insulation of the cable was decreased. But if the cable is 

 very long, then the general decrease of insulation upon it, by even 

 a defective joint is but small. In joining two ends of a cable 

 together, of one mile each, presenting each the enormous resist- 

 ance of say two hundred millions of units, the total resistance 

 would be seriously affected by a joint which was slightly defective, 

 offering a resistance of say four hundred millions of units ; but as 

 the length of the cable increases, its total insulation-resistance 

 must go down ; when its length reaches one hundred miles, the 

 insulation would be measured by two millions of units, in which 

 case the fault of four hundred millions would be inappreciable by 

 the instrument. I think Mr. Whitehouse first suggested to 

 measure the leakage separately, by putting the joint into an insu- 

 lating trough of water, which was connected with one pole of a 

 powerful battery, the second pole of which was connected to the 

 cable with a galvanometer in the circuit to detect leakage. Mr. 

 Latimer Clark has very much improved upon that method, in 

 connecting the insulated bath with a condenser, which condenser 

 charges itself gradually with the leakage through the joint, and is 

 discharged at a certain interval through the measuring instrument. 

 There is only one objection to the latter method, which is, that 

 the condenser itself leaks. It is indeed exceedingly difficult to get 

 a condenser which is as perfectly insulated as the copper wire 

 within a cable. Therefore the charge you get in the condenser is 

 not absolutely the expression of the amount of electricity that 

 leaks through the joint. 



It is true that the charge accumulating in even an imperfect 

 condenser must always be proportionate to the electricity flowing 

 into it, the leakage itself being proportionate to the accumulation. 

 Still, this would be supposing that the condenser itself did hot 



