S7K WILLIAM SIEMENS, F.R.S. 95 



upper horizontal broken line, indicate the average derived from the 

 olist-rvnl resistances. Diagrams B B, Plate 7, give the curves of 

 chiirge, discharge, and loss per minute, the ordinates of the latter 

 curve, when the original charge is taken as unity, expressing in 

 fractions the amount of current passed through the dielectrfc during 

 the minute. The different diagrams, severally marked C and D, 

 iv| in-sent the results obtained at the Sheathing Works of Messrs. 

 <i lass, Elliot, and Co.; C represents the insulation, and D the 

 chiirge, &c., as in the previous cases, but with the difference that 

 there is no test under pressure, and that the abscissae represent 

 time. The diagram marked E shows the tests of insulation on 

 board ship, and that marked F gives a comparison between the 

 average resistances at 75 Fahr. observed at the Gutta Percha 

 Works, after the sheathing, and finally on board ship shortly 

 before submersion. The abscissas represent in this case the length 

 of cable. The diagram marked G gives the insulation actually 

 observed during a certain period after submersion. 



The results obtained from the cables composing the section 

 between Alexandria and Benghazi, are shown in a similar manner 

 in Plate G. The diagrams of insulation resistance in Plate 7 

 illustrate the state of a portion of this section (Cable No. 5), 

 which having been in a defective tank at the Sheathing Works, 

 exposed to atmospheric influence, began to show symptoms of 

 spontaneous heating. The insulation-tests fell low, while the 

 resistance of the copper showed, by the ratio of its increase, that 

 the mass of the cable must have attained a temperature of 

 80 Fahr. Upon this, the cable was cut into three pieces, 

 which were marked 5 a, 5 b, and 5 c, and tested separately, when 

 it appeared that 5 a, the portion which was at the bottom of the 

 tanks, had suffered most, while the two remaining portions still 

 showed the marked effects of a gradual transition from a wet to a 

 dry state. 



The three portions of this cable, after being coiled over into 

 another tank, and being kept under water, gradually returned to a 

 normal state of tests, except 5 a, which never returned to its former 

 high state of insulation, and which for some time gave rise to 

 doubts, whether the effect of the heating had not rendered it unfit 

 for use. 



The ordinates in the diagram, connected by a black line, repre- 



