the Spark-Length of the Wimshurst Machine. 67 



so that finally the electrodes were held in position with sealing- 

 wax. The gases were in each case dried by passing them over 

 pumice moistened with sulphuric acid. Faraday (" Experimental 

 Researches in Electricity," Series 12, 1425 — 1454) gives the 

 phenomena of the brush discharge in various gases and some 

 of these were tried with the following results : 



(1) Coal-gas. In this gas the brush is difficult to obtain. 

 With terminals 7 cms. apart, a point of light on the positive ball 

 was seen on taking a spark from the negative pole and sometimes 

 a very faint brush. There was no spark discharge produced, but 

 when the gas was flushed out and the air admitted the spark 

 discharge could be easily obtained by Expt. 1. 



With a gap of 4 cms. I was rarely able to make a discharge 

 between the terminals and that passing to the finger synchronise, 

 but on the admission of air it was easy enough. 



(2) Carbon-dioxide. The brush is also difficult to obtain in 

 this gas-. It was easier to make the sparks synchronise at 4 cms. 

 or thereabouts than with coal-gas but it does not require so great 

 a potential difference between the terminals to produce spark 

 discharge and this might partly compensate for the poorness of 

 the brush. 



(3) Nitrogen. The poles were 4"5 cms. apart ; on driving 

 the machine so slowly that no spark passed under ordinary 

 conditions, then a spark taken from the negative side produced 

 a very bright spark discharge ; the same effect could also at times 

 be obtained by taking a spark from the other terminal. A brush 

 is formed in this gas very easily. 



Conclusion. 



It seems that a spark taken from a terminal as in Expts. 1 and 

 2 produces a surge or wave which is carried over to the other 

 terminal by means of the jars, a sudden change in the potential 

 of the negative pole producing a corresponding change in the 

 positive, because if the jars are absent or even insulated from each 

 other none of these phenomena is produced; changing the size 

 of the terminal moreover shows that the effect is always greatest 

 when the spark is taken from the larger ball and the disturbance 

 carried over to the lesser; around the smaller the electromotive 

 intensity is greatest and this is greatly increased by the sudden 

 change produced by the wave. For a ball of a given size there is 

 a given electromotive intensity at which a brush is formed, and 

 therefore a brush discharge is easiest formed on the smaller 

 terminal; this by produciug dissociated ions puts the gas in 

 a condition in which it is electrically weaker, and if the distance 



