18 Dr. E. N. da ( J. Andrade on the Electrical 



fig. 1 show this effect. The abscissae are the distance of the 

 mid-line of the streak from the positive electrode; the elec- 

 trodes were 1 cm. apart, and the streak roughly 4 mm. broad. 



Fig. 1. 



6 



ICU 



Position of the streak -*» 



The ordinates are the current in arbitrary units (galvano- 

 meter deflexion) ; in order to show the great increase when 

 the streak of vapour is in good contact with the negative 

 electrode the curve (A) is given ; the point on the extreme 

 right corresponds to contact. The ordinates arc , l „ the 

 scale of the curve (a). Thus to obtain an estimate of some 

 absolute measure of the conductivity this method cannot lie 

 used, as the glowing negative electrode produce- a very 

 disturbing effect. 



To avoid the complication caused by the interaction of the 

 glowing platinum and the vapour, experiments were made 

 withan electrode kepi artificially cool. The electrode was 

 made of flattened brass tube, through which a constant 

 circulation of cold water was kept up. Tin- electrode was 

 wrll i n the dame, but, uevertheless, kept quite cool : it was 

 earthed, and the other electrode kepi at a known positive 

 potential by means of a battery o\' small accumulators, one 



pole of which earthed, and 'the other connected to the 

 electrode through a galvanometer. This electrode was a 



