210 APPLIED SCIENCE 



Fortunately, joints can now be tested apart from the rest of the 

 cable. In old times when a joint had been made the whole cable 

 was tested ; if the leak from the new joint was inconsiderable 

 in comparison with the loss from the whole cable, perhaps some 

 hundred miles long, the joint was supposed to be good, although, 

 perhaps, it may have allowed a greater loss in its few inches 

 of length than occurred from some miles of sound cable. A 

 bad joint seldom does more than this at first, but in time it 

 becomes brittle, cracks, leaves the sound gutta-percha at each 

 side, and, finally, allows the water free access to the strand. 

 Joints of this character have been found in considerable number 

 in old cables, and especially in the old 1857-58 Atlantic 

 Cable. Some of these present an appearance of extraordinary 

 carelessness, even the copper strands being imperfectly joined. 

 It is almost certain that the final failure of the 1858 Atlantic 

 Cable was due to one of these joints in which the copper was 

 imperfectly joined ; the wires were pulled asunder when the 

 cable was being laid, they came together again when the strain 

 was removed, but the points of contact soon were oxidised, and 

 all communication ceased. Mere loss of insulation hardly ever 

 entirely stops signals. 



The test now employed shows whether a joint is as good as 

 any equal length of the wire, and all joints which do not reach 

 this standard are mercilessly cut out. First the joints to be 

 tested are allowed to soak in water for twenty-four hours, 

 then they are placed in an insulated trough of water con- 

 nected with a Ley den jar of large surface, the cable is charged 

 with a powerful battery, and a little electricity leaks out 

 through the joints into the insulated trough. If the joint is 

 good, this leakage is so small that the current produced by it 

 could not be shown by the most sensitive galvanometer, but 

 after a minute or two minutes, the insulated trough and Leyden 

 jar will be charged by the gradual accumulation of electricity 

 which has slowly leaked through the joint. If this be now 

 discharged through a galvanometer, it will produce a sensible 

 effect, and can be measured. In fact, the leak which was too 

 small to be directly perceptible is not only perceived, but its 

 amount ascertained by measuring the quantity which accumu- 

 lates from it in a given time. The test is due to Messrs. 



