196 Mr. P. V. Bevan on the 



space absorption in the region 6000-6800. This part of the 

 spectrum will be discussed later. The violet pair w.l.'s 4047, 

 4044 next appear and later another channelled space in the 

 blue-violet region. Finally another channelled space in the 

 orange-green appears. The absorption in the region of the 

 lines 7699, 7655 increases, and the lines broaden, soon making- 

 one dark band which spreads until it joins the red channelled 

 space which has in the meantime become a dark band, and 

 finally all light from the red end of the spectrum to about 

 w.l. 6000 is cut off. The blue-violet channelled space also 

 soon becomes a dark band, and this joins the band formed by 

 the widening of the lines 4047, 4044. Figure 1 shows a 

 series of photographs taken with increasing density of vapour. 

 The lines 7699, 7655 have disappeared, the whole space up to 

 about w.l. 6000 being dark. In (1) the blue-violet 

 channelled space is just appearing, The principal series 

 lines are very definite andean be counted up to about n = W 

 (n = 4 for the line 4047). As we go down this series of 

 photographs we notice how the blue-violet channelled space 

 increases — the channellings do not appear Avell in the plate 

 though they are quite evident in the negative. The red 

 absorption, however, changes its limit on the yellow side- 

 very little. The absorption lines also broaden but the 

 broadening is not symmetrical, the line being more diffuse on 

 the less refrangible side. New absorption lines appear on 

 each side of the main lines, and these seem to form parts of 

 channelled space spectra which appear between the primary 

 lines. In (5) the blue-violet channelled space has extended 

 to meet the broadened 4047 line, and all the light is cut off' 

 in this region. The orange-green absorption extends on the 

 more refrangible side but undergoes little change on the 

 other. The sodium lines D 3303 and 2852 appear faintly 

 owing to a certain amount of sodium present in the 

 potassium as impurity. 



The Principal Series Lines. 



These lines are the most defiuite feature of the absorption 

 spectrum, and as has been mentioned above they have been 

 counted up to the line corresponding to w = 26. As ?2 = 3 for 

 the lines 7699, 7665, the author has been able to observe 24 

 of this series. Hitherto the members up tow = 11 have been 

 observed, so that 15 new lines have been added to this list.. 

 The wave-lengths are as follows : — 



