102 Prof. A. F. Sundell's Researches 



considerable pressure, together with the hydrogen and air 

 lines ; in hydrogen tubes the lines 408, 405, and (at high 

 exhaustion) 483* feebly, were also seen. 



Under the circumstances named, the air began to be 

 luminous at a pressure of 10-12 millim.t; the end with the 

 positive coating became luminous somewhat sooner (at 

 14 millim.) than the whole tube. It was, however, possible 

 to produce a feeble luminosity of the tube at a considerably 

 higher pressure (about 50 millim.) by pushing a conductor 

 along the tube in contact with the wall until a spark passed 

 between the conductor and the one coating. An unexpected 

 phenomenon was observed at a pressure of about 8 millim., 

 inasmuch as the light appeared to be stratified in a peculiar 

 manner. As I propose to examine this stratification more 

 particularly, I describe it here only briefly. The first quarter 

 from the positive coating was strongly luminous, the intensity 

 diminishing slightly towards the end of this portion. The 

 second quarter also began with a strong luminosity, which 

 became somewhat weaker towards the middle of the tube. 

 The third quarter was distinctly stratified ; in one case there 

 were only three or four luminous stratifications, which oscil- 

 lated tolerably rapidly and appeared like luminous balls ; in 

 another experiment I observed a number of thinner layers. 

 Only towards the beginning of the fourth quarter was there 

 a thicker layer : all the stratifications oscillated rapidly {. 

 The last quarter next the negative coating was always strongly 

 luminous without stratification §. The narrow spiral glass 

 tube which made connexion with the pump was also generally 

 strongly luminous, whether the coating at this end was negative 

 or positive. 



The spectroscope showed a large number of bands. With 

 narrow bands the reading was taken for the middle ; with 

 broader bands for both edges. The following wave-lengths 

 are the arithmetic means of various measurements within the 

 limits of pressure (0*2 to 1*2 millim.) between which the 



* E. Wiedemann has investigated the spectroscopic behaviour of 

 mercury-vapour mixed with other gases at high temperatures, Wied. 

 Ann. v. p. 547 (1878). Compare also H. W. Vcgel, Berlin. Monats- 

 berichte, 1879, p. 536. 



t All measurements of pressure are given in columns of mercury 

 at about 20° C. ; for details of the method of measuring pressure, I refer 

 to the paper already cited on the air-pump {Acta Soc. Scient. Fennicce, 

 xv. p. 169). 



\ Similar balls of light {Glimmlichtkiigehi) have been observed under 

 certain conditions by Reitlinger and v. Urbanitzky in short and wide 

 tubes (Wied. Ann. xiii. p. 673). 



§ I have observed a similar stratification in a hydrogen-tube. 



