168 Messrs. W. De La Rue and H. W. Miiller on [Jan. 6, 



Telegraph Engineers, and also in October last a written one to the Aca- 

 deruie des Sciences of Paris *, wherein we have stated that the length 

 of the spark in air appears to be in the direct ratio of the square of the 

 number of cells. 



Having completed the 2400 cells, and charged them up in a single day, 

 they were exactly in the same condition as to electromotive force and 

 internal resistance; consequently they afforded the means of testing 

 the truth of the law of the length of spark in a manner more effica- 

 cious than had hitherto obtained, the more especially as by the use of 

 paraffin corks and other precautions we had obtained an excellent insu- 

 lation. 



Our assistant, Mr. Pram, has constructed a discharger which permits 

 of the accurate measurement of the distance of the terminals to read to 

 XuxjTT °f an iuch, and by estimation to the tenth of that quantity. The 

 nut through which the screw ( ^-of-an-inch thread) carrying one of the ter- 

 minals works is divided into two parts, which are separated by a spiral 

 pressure-spring so as to prevent shake. In making measurements the ter- 

 minals are separated to a greater quantity than the anticipated striking- 

 distance, and gradually approached until the spark passes ; the discharger 

 is then detached from the battery, and, after reading the scale, connected 

 up with a separate battery of 10 cells, with a detector-galvanometer in 

 circuit. The terminals are again approached until the motion of the 

 galvanometer indicates contact between them ; the scale is again read, 

 and the length of spark obtained by the difference between the first and 

 second reading. 



Rod chloride, 600 cells had a striking-distance of 0*0033 in. 

 1200 „ „ 0*0130 „ 



1800 „ „ 0-0345 



55 



2400 „ „ 0-0535 



Taking as the unit 600 cells, the spark of which was 0*0033, the length 

 of spark of 1200, 1800, 2400 would, according to theory, be that number 

 multiplied by the square of 2, 3, 4 respectively. 



600 cells, striking-distance 0-0033 in. 



1200 „ 0-0033 x 4 0-0132 „ 



1800 „ 0-0033 x 9 0-0297 „ 



2400 „ 0-0033 x 16 0-0528 „ 



which numbers agree nearly with those obtained by experiment. 



The length of the spark is much influenced by the shape of the ter- 

 minals ; those which we frequently employ consist of a point as one ter- 

 minal and a plane for the other. Hitherto we have used copper terminals, 

 making the point and the plane alternately positive and negative by 

 * Comptes Rendus, no. 16, p. 686, & no. 17, p. 746 (1875). 



