Haktley — On the Constitution of Electric Sparks. 371 



brown to foxy-red. It was observed that the iron punctured a 

 smaller point than most other metals. 



Cobalt. — Cobalt gives a dark-brown deposit, quite different 

 from that of iron, both in colour and general appearance, being 

 much larger around the perforations. The fluxed mica is some- 

 times coloured pale blue. 



Nickel. — Nickel is deposited as oxide of the same colour as that 

 of cobalt, but the quantity is smaller, and not distributed over so 

 large a space. 



In every case the negative electrode was below, and the dis- 

 charge took place in one direction, namely, from below upwards ; 

 so that it is evident from the deposits being always very much 

 more abundant on the under surface of the mica, that — (1) each 

 spark carries from the negative electrode a charge of metallic 

 vapour ; (2) that the vapour is most abundant in the centre of 

 the spark ; (3) that where the vapour condenses the quantity of 

 vapour decreases as the distance from the centre increases ; but 

 that the vapours of a metal like gold, which is not oxidized, is 

 carried to a considerable distance over a circular area surrounding 

 the body of the spark ; (4) that the vapour of an oxidizable metal 

 is oxidized by the surrounding air. 



In the case of arsenic, the abundance of the vapour was greater 

 than the air in the spark was capable of oxidizing, but such vapour 

 as escaped outside was oxidized. There is, however, good reason 

 for believing that the column of incandescent matter is a mixture 

 of metallic vapour and air, and not an arc of metallic vapour sur- 

 rounded by incandescent air. 



The Reason why Sparks ultimately take a definite Direction uhen 

 discharged over the Surface of a Non-conductor. — It has been ob- 

 served that when electric sparks are discharged at the surface 

 of a film of glass or of mica, that the discharge lor a time takes 

 no definite direction, but is in all directions round a central point 

 distributed in minute flashes. After an interval a dense spark 

 flashes along two or three times consecutively in one direction ; 

 it then either passes over the edge of the film, and so traversing 

 the opposite side reaches the positive electrode, or it perforates the 

 non-conductor at some point in its path. Each spark deposits 

 some metallic vapour, and if several sparks intersect the path of 

 some other spark, then a deposit of metallic particles is formed, 



