DISCHARGE OF ELECTRICITY. 237 



enced in getting the discbage to pass tbroiigii an ordinary vacnuni-tnbe 

 wbei) tbe pressure is very low is not altogetber due to tbe difficulty of 

 getting" tbe electricity from tbe electrodes into tbe gas, but tbat it also 

 occurs in tubes witbout electrodes, tbougii in tbis case the critical 

 pressure is very nni(;b lower tban wben tbere are electrodes. In otber 

 words, we see tbat as tbe state of tbe bulb sipproacbes tbat of a per- 

 fect vacuum its iiisulatiug power becomes stronger and stronger. Tbis 

 result is continued by several otber experiments of a different kind, 

 which will be described later. 



Tbe discbarge presents a perfectly continu<ms appearance, with no 

 sign of striation, of wbicb I have never observed any trace on any of 

 these discharges, though I must have observed many thousands of 

 them under widely different conditions. 



Action of a ma<j)iet on the discharfic — Tbe discharges wbicb take 

 place in these tubes and bulbs are produced by })eriodic currents, so 

 tbat the discharges themselves are periodic, and the luminosity is pro- 

 duced by currents j^assiug in opposite directi<ms. As tbis is the case, 

 it seemed possible that the uniformity of the luminosity seen in the 

 discharge was due to tbe super-position of two stratified discharges in 

 opposite directions, the places of maximum luminosity in the one 

 fitting into those of minimum luminosity in the other. Since these 

 discharges are in opposite directions, they will be pushed opposite 

 ways wben a magnetic force acts at right angles to them, the dis- 

 cbarges in opposite directions can thus be separated by the application 

 of a magnetic force and examined separately. In tbe experiment wbicb 

 was tried with this object, a scpiare tube was used placed outside the 

 primary, the tube at one or two places being blown out into a bulb so 

 as to allow of tbe wider separation of the constituent <liscbarges. 

 Wben one of these bulbs was i)lace(l in a magnetic field where tbe 

 force was at right angles to tbe discharge, the luminous discbarge 

 tlirougb the bulb was divided into two portions wliicli were dri\en to 

 oi)posite sides of the bulb; each of these portions was of uniform lumi- 

 nosity and exhibited no trace of striation. It was noticed, howevei", in 

 making this experiment that tbe discharge seemed to have much greater 

 (lifiiculty in ])assing through tbe tube when the electromagnet was on 

 tban wben it was oft. Tbis observation was followed up by several 

 other experiments, and it was found thai tbe discbarge is r(>tarded in 

 a nu)st remarkable way by a magnetic force acting at right angles to 

 the line of discharge. Tbis effect is most strikingly shown when the 

 discbarge passes as a ring through a s])herical bulb. If such a bulb 

 is placed near a strong electro-magnet, it is easy to adjust tin' length of 

 spark so tbat when tbe magnet is oft' a brilliant discbarge passes 

 through the bulb, while when tbe magnet is on no discharge at all can 

 be detected. The action is very striking, and the explanaticui of it 

 wliicb seems to fit in best with tbe jdienomena 1 have observed, is that 

 the discbarge through tbe rarefied gas does not rise to its full intensity 



