236 DISCHARGE OF ELECTRICITY. 



whether the additiou of au extra tuiii of wire is beneficial or the re- 

 verse. The brightness of the discharge depends upon the time of the 

 electrical oscillations as well as upon the niagnitudt of the electro-mo- 

 tive force. Thus, in an experiment to be described later, the brilliancy 

 of the discharge was increased by putting self-induction in the leads, 

 which, though it diminished the intensity of the electromotive force, 

 increased the time constant of the system. When the discharge tube 

 was square and the coil C had also to be square it was found most 

 convenient to make it of glass tubing bent into the required form and 

 filled with mercury. When however the discharge was required in a 

 bulb, the primary coil was made of thick gutta-percha-covered copper 

 wire wound round a beaker Just large enough to receive the exhausted 

 bulb. There is sometimes coasiderable difficulty in getting the first 

 discharge to pass through the bulb, though when it has once been 

 started other discharges follow with much less difficulty. The same 

 effect occurs with ordinary sparks. It seems to be due to the splitting 

 up of the molecules by the first discharge; some of the atoms are left 

 uncombined, and so ready to conduct the discharge, or else when they 

 re-combine they form compounds of smaller electric strength than the 

 original gas. AVhen the discharge was loath to start, I found the most 

 effectual way of inducing it to do so was to pull the terminals of the 

 Wimshurst far apart and then, after the jars had got fully charged, to 

 push the terminals suddenly together. In this way a long spark is 

 obtained, which, if the pressure of the gas is such that any discharge 

 is possible, with the means at our disposal will generally start the dis- 

 charge. 



Appearance of the discharge. — Let us suppose that we have either a 

 square tube placed outside a square primary or a l>ulb placed inside a 

 circular coil of wire, and that we gradually exhaust the discharge tube, 

 the jars sjiarking all the time. At first nothing at all is to be seen in 

 the secondary, but when the exhaustion has proceeded until the pres- 

 sure has fallen to a millimeter or thereabouts, a thin thread of reddish 

 light is seen to go round the tube situated near to but not touching the 

 side of the tube turned towards the jyrimary. As the exhaustion i)ro- 

 ceeds still further, the brightness of this thread rapidly increases, as 

 well as its thickness; it also changes its color, losing its red tinge and 

 becoming white. On continuing the exhaustion the luminosity attains 

 a maximum, and the discharge passes as an exceedingly bright and 

 well-defined ring. On continuing the exhaustion, the luminosity begins 

 to diminish until, when an exceedingly good vacuum is reached, no dis- 

 charge at all passes. The pressure at which the luminosity is a maxi- 

 mum is very much less than that at which the electric strength of the 

 gas is a minimum in a tube provided with electrodes and comparable 

 in size to the bulb. The i^ressure at which the discharge stops is ex- 

 ceedingly low, and it requires long-continued pumping to reach this 

 stage. We see from these results that the difficulty which is experi- 



