332 



NA TURE 



[January 31. 1895 



If ihis rinR discharge passes through air at different pressures, 

 the colour of the discharge chaoges very considerably. The 

 firet bulb I put in was at fairly high pressjre, about ,',7 of a 

 mill i met re or s'». I will now put in another at a lower pressure, 

 and then one at a still lower pressure. Mr. .Vewall, who has 

 been working at the spectra of these ilischirJe*, fin Is that at 

 the pressure in ihe first bulb the spectrum is due to nitrogen ; 

 at the second stage it is due to mercury vapour ; the bulb was 

 pumped by a mercury pump, so that there is in the bulb a 

 certain quanii y of mercury vapour. 



The apple-gteen colour in the more highly exhausted bulb is 

 due 111 some compound of sulphur, which has got into the bulb 

 from the sulphuric acid used to dry the gas. .Mr. Newall finds 

 that if the ordinary discharge from a coil between electrodes is 

 taken in such a bulb, there is no trace of this sulphur spectrum. 

 He has also found that when the bulb is at a pressure inier- 

 mediate between what I may call the mercury and the sulphur 

 stage, when the mercury and sulphur lines are both visilile, 

 these sets of lines come from dilVerent layers, the sulphur lines 

 coming from a layer nearer the surface than the other. 



If we take the discharge through a bulli containing osygen, 

 jrou will see that the ring discharge is succeeded by a bright 

 glow : at first the colour is somewhat opaque, but gradually gets 

 more transparent and changes colour. This gives a continuous 

 spectrum crossed by a few bright lines. If we take the discharge 

 through cyancgen, you see that the glow is even more persistent 

 than the oxygen, though it is not so bright ; all the gases which 



Fig *. 



show this glow belong to the class of substances which poly, 

 merise — that is, whose molecules can combine with each other. 

 I imagine that what lakes place in bulbs filled with these sub- 

 stances is that the discharge produces a polymeric modification, 

 and that this gradually returns to its original stale, and » hilc 

 doing so gives out a pinsphorescent ligh'. It is in accordance 

 with this that al a high temperature where ozone canno: exist 

 a discharge through an oxygen bulb dcics not show any glow. 



I said at the beginning of this discourse that gases were exceed- 

 ingly good conductors of electricily. I will now endeavour to 

 show an cxperimcnl which proves that statement. The apparatus 

 which I shall use for this purpose is a slight modification of 

 the one I have used for prorliicing the ring discharge ; the only 

 difference is that in ihe wire connecting the two c latiiigs of the 

 jars there are two loops instead of one. In one of these loops 

 an exhausted bulb is placed to serve as a kind of galvanometer ; 

 the brightness of the discharge is an indicalinn of the strength 

 of the current flowing round the c >il. Il I plac; a seco n' 

 conductor in Ihe other loop, cuTent- will be started in i;, imd 



f>»rt of the en;rgy of the dii,.hirge will be absorbed j this will 

 cave less energy available for the bulb in the first, so that the 

 discharge in this bulb will be dimmer. The effect produced on 

 Ihe di<chiri;cwill depend upon the conductivity of the substance 

 placed in the second loop. 



The effect ii no*, directly proportional to the con lucliviiy ; in 

 fad, a perfect conductor would not produce any diminution, 



NO. 13 ifS, VOL. 51] 



nor would an absolute non-conductor ; for a given period, and 

 with apparatus of given dimensions, there i.s a certain c indiic- 

 tivity which gives a maximum effect : ihis foil 'ws easily from 

 ihe theory ol induction of currents, but at this laie period in 

 ihe evening I will take a shorter course and prove it by an 

 experiment. 



I put a piece rf br.ass in this loop, and you see it produces but 

 a small efTtci U|ion the Iprightness o( the discharge. Instead of 

 brass I now insert a plumbago crucib'e, which, ihough a con- 

 ductor, is not nearly s j good a one as the brass, and you sec the 

 discharge in the indicating bulb is completely stupped. 



I will now place in the second loop an exliaiisted bulb ; you 

 see it pr.iduces a decided diminution in the intensiiy of the dis- 

 chatge in the galvanometer bulb. I now replace the bulb by 

 another of the same size containing dilute sulphuric acid ; yon 

 see il does not produce nearly so large an eflccl as the exhausted 

 bulb : this might be due, as we have seen, to the sulphuric acid 

 being either too good or too b.ad a conductor. I can >how that 

 it is the latter by putting a bulb in filled wuh a stronger solu- 

 tion, which has a higher conductivity than the weak solution ; 

 if the smallness of the ertect produced by the weak acid were 

 due toils being a better conductor than the g.as, then increasing 

 the conductivity would still further diminish the efl'ect of the 

 acid ; yiu see, on the contrary, that the strong acid produces a 

 distinctly greater effect than Ihe weak, hence the rarefied gas in 

 the bulb is a better conductor even than the strong electrolyte. 

 Let us con<^ider for a moment the molecular c ■r.ducnviiies of 

 the two substances, the raicfied gas and the electrolyte. The 



Fig 5, 



pressure of the gas is about ,,',,. of a millimetre, «hilc in the 

 electrolyte there are sufTicient molecules of the acid to produce, 

 if they were in Ihe gaseous state, a pressure of more than 100 

 atmospheres ; thus the conductivity of the gas eslimaled per 

 molecule is about 10 million times that of the acid, this is greater 

 than the molecular conductivity of even Ihe best conducting 

 mel.ils. 



If the pressure of the gas is diminished below a certain point, 

 the conductivity begins to diminish. I have here an experi- 

 ment which I hope will show this. The apparatus (Fig. 5) con- 

 sists of two bulbs, one outside the other ; the inner bulb con- 

 tains air at a low pressure, while the space between the two 

 bulbs is a very high vacuum conlaining practically nothing but 

 a litlle mercury and iis vapour. The amount of mercury vapour 

 in this space is, at the leniperauire of the room, exceedingly 

 small, but as the apparatus is heated the vaoour pressure in- 

 creases, and we are thus able to produce a fairly wide range of 

 pressu.e in :hc space lelween Ihe bulb-. The outer sphere is 

 ■ui'rourdcd by the coil connecling the outer coalings of the two 

 I.eyden jars. When ihc space between the bulbs is a conductor, 

 the alternating currents circulating in the coil will induce 

 in Ihis conduciur curicnis whose inductive effect is opposite 

 to Ihal of the currents in the coil ; and in ihis case Ihis layer 

 will screen off from the inner bulb the electromotive force due 

 lo the alternating currents in Ihe coil. If, on the other hand, 

 the space between Ihe bulbs is a non conductor, Ihe inner bulb 

 will be exposed to the full ellecl of these forces. We now try 



