TRANSACTIONS OP THE SECTIONS. 255 



its exposed end, and y the other portion of the cable, and z' its ends' resistance, then 

 if x+z+y+z' = S, the resistance of a perfect wire, it is evident that z+z' offer little 

 or no appreciable resistance, and the locality of the fault is immediately known. 

 When, however, z or z' offer resistance, the value of z may be approximated either 

 by measuring the amount of electrostatic charge of the cable, or by measuring the 

 resistance, first with negative and then with positive currents. 



It sometimes happens that one of the exposed ends of the conductor gets entangled 

 with the iron outer wires of the cable. This is to be sought for ; and if such be the 

 case, it offers no appreciable resistance : this is immediately ascertained by connect- 

 ing the conductor of the cable through a delicate galvanometer to the iron outer 

 covering, when, if the copper wire at the fault be not in contact with the iron wires, 

 an electric cm-rent will be found to flow through the galvanometer. The electro- 

 motive force of this current should be tested, and it will generally be found equal to 

 an iron copper pair charged with sea-water. If the conductor touch the outer iron, 

 there is no current through the galvanometer. If there be the current, — 



1st. Measure the resistance of #-f-z with negative current, an<l note whether it 

 varies in amount. 



2nd. Measure the resistance as before, but witb a positive current, and note how 

 it varies. If it vary much, especially with the negative current, it indicates that the 

 fault offers much resistance. 



3rd. Make an artificial fault, or rather several faults, that behave like the cable, 

 with a like resistance to that of the cable, and with the same battery power. Having 

 made such a fault that resembles as nearly as possible the cable with positive and with 

 negative currents of various powers, measure its resistance, and subtract that amount 

 from x-\-z, and that will indicate the distance of the fault. The positive current de- 

 composes the sea-water and its salts, oxygen and chlorine are set free and com- 

 bine with the copper wire at the fault, forming a coating offering considerable re- 

 sistance to the passage of the current. In this way the resistance of a fault may 

 often be very considerably increased. If it can, it shows that the surface exposed 

 at the defect is small, and offers considerable resistance even with a negative current. 

 A negative current covers the exposed wire with hydrogen, which keeps it clean and 

 in good contact with the water, unless the aperture admitting the water be very small, 

 and located in shallow water, when the hydrogen will sometimes expel the water and 

 so increase the resistance. 



The next plan of ascertaining the resistance of the fault, is by measuring the in- 

 duction or statical charge and discharge. The author detailed several plans of doing 

 this approximately, and indicated how an apparatus might be made to effect this 

 perfectly, and which he had tried on a small scale with perfect success. He then 

 showed how he had tested for the faults in the Atlantic Telegraph Cable, and pointed 

 out the utter impossibility of the great fault being in the Valencia harbour, and which 

 was proved by two distinct modes of testing. The exposed copper wire at the fault, 

 formed with the iron outer covering a voltaic element of copper and iron. He con- 

 trived with this battery alone to measure the resistance of the cable and the fault. 

 Much depends on the skill of the manipulator in choosing those plans most suited 

 for the occasion. 



After detailing several curious defects, he showed that when the defect in a cable 

 was small and immersed in clay-mud, the fault might often be sealed up b) r a posi- 

 tive current so completely as to enable the conductor to be used. In this way he 

 had sealed up one of the Orfordness Scheveningen Cable, which was defective, and 

 thus kept it working above eighteen months ; and when it got bad, it was again and 

 again sealed up by strong positive currents. He stated that he had used some of 

 the plans described for the last twelve years, and had rarely, if ever, found a greater 

 error in the estimated distance than 5 per cent, of the cable tested. The plans de- 

 tailed, as far as the author was concerned, were original, save No. 2, which was partly 

 borrowed from the Abbe Moigno's treatise on Electric Telegraphs. 



