1864. | Farrparrn on Submarine Telegraph Cables. 637 
clearly the entrance of water; and only in one or two instances was 
any defect in the cable detected, beyond question, by the gradual loss 
of insulating power in the specimen under trial. To inadequacy of 
time were added manipulative difficulties; such as the making of a 
packed joint which should hold tight against so enormous a pressure 
as 10,000 Ibs. per square inch, and also the variable hygrometric con- 
dition of the atmosphere. 
The earlier and preliminary experiments were made with a simple 
double pith ball electrometer suspended from one of the exposed ends 
of the cable. This method, however, did not allow of sufficient accu- 
racy in the measurement and regulation of the charge and the rate of 
loss, to afford satisfactory results. 
The following method was then adopted, The core was placed in 
a steel cylinder, £, & (Figs. 4 and 5), with the ends projecting. This 
Ml 
Hh 
7 
c 
= 
My 
il 
eal 
qi! 
( - 
| 
| 
al 
i 
| 
| 
| 
Tig) 
= = 
i] 
cylinder was bored out to seven-eighths of an inch diameter, and at 
either end a pair of strong brass glands, «, «, were fitted, so as to com- 
press round the core the vulcanized india-rubber packing, p, Pp, by the 
| aid of the bolts and nuts, 3,8. The compression thus applied indented 
_ the core to a greater or less degree (Fig. 6) at each of the points where 
FIG.6. 
VOL. I- BUS 
