28 THE SCIENTIFIC PAPERS OF 



workmen at the moment of danger, and which, moreover, do nob 

 relieve the cable from retarding strain at the brake, should be 

 discarded, and the cable be made to possess in itself all the 

 requisite degree of buoyancy and strength. For this purpose the 

 conducting wire or wires should be as light as possible consistent 

 with good conducting power, a combination of properties which 

 seems to point to the newly discovered metal, aluminium, as likely 

 ultimately to supersede copper. The insulatiug covering of gutta- 

 percha increases the bulk without adding to the weight of the 

 cable, being nearly of the same specific gravity as sea-water, 

 it improves both the mechanical and electrical properties of the 

 cable, and the only limit to its desirable thickness is its expense. 

 The principal weight, and all the available strength of the cable 

 reside in its sheathing, which should be made of a material com- 

 bining strength with lightness, and also with hardness, to resist 

 the crushing and tearing action of the brake wheel ; and there can 

 be no doubt that steel wire combines these qualities in the highest 

 degree, nor do I think it would be much dearer than iron if power 

 of suspension was taken for the basis of calculation. 



It can easily be shown, by the simple rule given above, regard- 

 ing the strain upon the cable in leaving the vessel, that an iron- 

 sheathed cable cannot, under the most favourable circumstances, 

 be laid in water of more than three miles in depth, without a 

 certainty of rupture taking place, whereas a steel covered cable 

 might be laid with -reasonable safety to a depth of five or six 

 miles, which depth is, I believe, rarely exceeded in any ocean. 



Respecting the paying-out machinery, I have to notice Messrs. 

 Newall and Co.'s apparatus, consisting of a solid centre, and heavy 

 rings to form a double cone for guiding the cable safely out of the 

 hold, and the brake, which latter should be made as light as 

 possible, to avoid jerks upon the cable, and should indicate the 

 variable strain put upon it, to harmonise its speed with that of the 

 vessel. 



In order to ensure continuity of the electric conductor in a 

 cable, a strand of several copper wires is now generally adopted, 

 instead of a single wire, which latter is found to be very liable to 

 break. This simple but useful plan was, I believe, first thought 

 of and acted upon by myself, having ordered some gutta-percha 

 coated strand, for experiment, from the Gutta-percha Company 



