342 POPULAR LECTURES AND ADDRESSES. 



solder all along the interstices. On testing this 

 splice, I found that, although it would bear within 

 10 Ibs. or 20 Ibs. of the full breaking- weight of the 

 wire, yet in every case the wire broke at the 

 splice. This was precisely in accordance with 

 theory. The sudden change of area of section 

 between the long cylindrical wire, and the thicken- 

 ing produced by the solder, is an essential element 

 of weakness, of a character well known to engineers. 

 Inevitably, if the wire is of uniform character, it 

 breaks close beside the solder. To avoid this 

 weakening of the wire, an exceedingly gradual 

 commencement of the force by which one piece of 

 wire pulls the other must be attained. The obvious 

 way of attaining this is by a very long splice. A 

 splice of two feet long I have found quite sufficient ; 

 but three feet may be safer. The two pieces of 

 wire to be spliced are first prepared by warming 

 them slightly and melting on a coating of marine 

 glue to promote surface friction. About three feet 

 of the ends so prepared are laid together and held 

 between finger and thumb at the middle of the 

 portions thus overlapping. Then the free foot and 



