340 Professor Liveing, On Differences 



capillary both spectra are bright at first, but as the pressure is 

 reduced the lines G, F, &c. become relatively brighter and the 

 second spectrum weaker, until the latter entirely disappears from 

 the capillary, and is seen only at the bright spot on the anode. 

 At the higher pressures (7 mm.) the bright spot on the anode 

 shews only the second spectrum without a trace of G or F. The 

 behaviour of the two channelled spectra of nitrogen is precisely 

 similar. When the dark space is sufficiently developed the 

 kathode glow shews none of the brilliant channellings in the 

 orange which are characteristic of the positive spectrum, but 

 they appear plainly at the head of the first striation, and as 

 the exhaustion proceeds are driven further and further from the 

 kathode, until the capillary shews only the spectrum of the nega- 

 tive glow, and the orange and yellow channellings are seen only 

 on the anode. 



In order to see these appearances the discharge must be 

 continuous in one direction, and therefore, if an induction coil 

 be used, there must be sufficient resistance in the circuit to stop 

 the passage of the indirect induced current. A short air gap 

 will serve this purpose very well without making the discharge 

 discontinuous. A condenser cannot be used in connexion with 

 the secondary circuit, since it produces an oscillatory discharge. 



Pure oxygen produces no luminous positive column and no 

 striations, as was observed by Morren, and the whole of the 

 light emitted by it in a tube at low pressure under the influence 

 of a continuous electric discharge appears to be that of the nega- 

 tive glow. Schuster has carefully observed the spectrum of 

 oxygen, but my observations, made with a continuous current 

 from an induction coil giving a spark in air 7 cm. long, do not 

 accord with his in all respects. In the first place I have failed 

 to see the continuous spectrum and the narrow spark throughout 

 the tube, when the exhaustion first reaches the stage at which 

 the discharge will pass through the tube 1 . The first luminous 

 indication of a discharge that I see is a thin sheath of luminosity 

 around the kathode and nothing more. As the exhaustion pro- 

 ceeds the glow at the kathode spreads out, and in a darkened 

 room a very faint orange-coloured light is seen in the capillary, 

 but the rest of the tube is dark, and so is the anode if all carbon 

 dioxide has been well removed. A trace of carbon dioxide shews 

 itself by producing a slight whitish cloud on the anode, and is 

 almost always seen after a time, so that it is only in a tube from 

 which carbon dioxide has been well exhausted, and then only 

 for a short time after being refilled with pure oxygen, that the 

 anode shews no cloud. By further exhaustion the negative glow 

 becomes more diffuse and in the capillary becomes quite bright. 

 1 "Spectrum of Metalloids," Phil. Trans. R. 8., Part I., 1879, Exp. 8. 



