366 Scientific Proceedings, Royal Dublin Society. 



pass upwards, carrying a charge of metallic vapour. The direc- 

 tion of the current should not be reversed. Under these circum- 

 stances those metals which are present in very minute proportions, 

 such as traces of impurities, are recognized by their longest lines 

 becoming visible at the position of the negative electrode only. 



Silver has thus been detected in what was supposed, from its 

 mode of preparation, to be pure copper, zinc in cadmium, cadmium 

 in zinc, copper in aluminium, and magnesium in graphite. It 

 might be supposed that the negative supplies vapour at the highest 

 temperature, and that a portion of this becomes dissipated or 

 oxidized, perhaps even condensed, before the spark touches the 

 upper electrode ; but this is not the explanation of the fact which 

 I propose to give. It will be dealt with later on in connexion 

 with the oscillation of the electric discharge. I have elsewhere 

 proved that photographs produced under similar conditions always 

 yield the same metallic spectrum. 



On the Path of the Sparks. — Sparks of great intensity, which 

 pass from one metallic point to another, are distinguished by their 

 brightness and the loud noise produced during their passage. A 

 rapid succession of sparks causes a noise which is sometimes almost 

 deafening, and the light is apparently continuous and very bril- 

 liant. When single sparks are photographed, there are seen on 

 or below the edges of the electrodes minute bright points, and 

 joining the electrodes is an irregular path described by the in- 

 candescent vapour. Two sparks in rapid succession do not take 

 the same path, though they may pass from and to the same points. 

 The most curious effect is caused by a short succession of sparks ; 

 they form a brush or bundle of what has the appearance of threads 

 twisting round each other, much in the same way as the fibres 

 from an untwisted piece of twine. Each of these is a stream of 

 metallic vapour, and between and around them is the air heated 

 to incandescence. 



The Shape of the Sparks. — It has been stated that the sparks are 

 irregular in their course, and that they seldom, if ever, pass from 

 and to the same point ; accordingly it is a matter of interest to 

 ascertain whether the shape of their section is flat, square, circular, 

 or irregular. In order to prove this, sparks were passed through 

 paper, tin-foil, and very thin films of mica, the electrodes being 

 gold. The perforation caused by a single spark in mica or tin- 



