tlirough Rarefied Gases. 359 



18. DifFerences ill thickness, curvaturej mobility, are easily 

 observed. These clitierences are strikingly shown by the colour 

 of the dili'erent stratifications. Just as one and the same 

 stratification may have different colours in different parts, the 

 different stratifications of the same discharge may vary in 

 colour. 



In a tube filled with hydrogen I saw, for instance, the first 

 stratification blue, the second pink, the third half blue and 

 half pink. A variation in the pressure, as I have said, varies 

 the colour of each stratification ; but each stratification shows 

 a different variation. 



19. I shall not here enter into various combinations Avhich 

 may thus be obtained ; but it is clear that the statement 

 that one and the same gas shows always the same colour if 

 rendered luminous by the electric discharge is wrong. The 

 discharge may show a great yariety in colour for the same 

 gas, even if the cross section of the tube be constant through- 

 out. 



Different spectra must correspond to the different colours, 

 f-hough in some cases such a difference may only be due to 

 difference in the relative intensity of the different parts of the 

 spectra. 



It follows that for the same gas, the same pressure and cross 

 section of the tube, and intensity of discharge, the spectrum 

 may yet be different. 



20. One is tempted, of course, to look for the cause of these 

 differences in the irregularity of the tube or the impurities. 

 I shall give, further on, the proof that the real cause lies deeper, 

 and that there is a kiw regulating these differences. 



If we number each stratification, calling the one next to the 

 negative light "1," we can say that the character of each 

 stratification is a function of its number. 



21. The stratification of the positive light is, in all cases 

 which I have examined, the better defined the nearer it is to 

 the negative side of the positive light. This is true for what- 

 ever the shape of the tube may be. The gases which I have 

 examined are air, hydrogen, nitrogen, carbonic oxide, carbonic 

 acid, aqueous vapour, alcohol, ether, and metallic vapours. 



22. The stratification need not extend along the whole of 

 the positive light; in the same gas the column of light is 

 stratified to the greater length the wider the tube is. 



23. If a tube is composed of various parts, each part behaves 

 like a separate tube which has its electrodes at the two points 

 of entrance of the current. I shall call these points secondary 

 poles. Behind each secondary negative pole the stratifications 

 begin again with fresh definition, even if they are almost uu- 



