SAUNT)ERS. ARC SPECTRA OP THE ALKALI METALS. 445 



positions, even though the amount of vapor in the arc is small, in which 

 case they are separated from each other by an absolutely clear space on 

 the negative. The conclusion seems unavoidable from these images that 

 we have here to deal with no reversal at all, but with two lines ; a strong 

 one at 4003.2, much broadened towards the red, and a weaker one at 

 4G01.6, broadened toward the violet, neither of them being ordinarily 

 reversed. If this view be adopted, the s])ectrura of lithium shows 

 another analogy to that of sodium, for, in the latter spectrum, imme- 

 diately beside the pair in the first subordinate series which is homologous 

 to Li 4602 lies a faint pair broadened toward the violet in a similar 

 manner. 



The 4602 group presents the same aspect in the spark spectrum as in 

 the arc. 



Sodium. 



In the spectrum of sodium, the new series of Lenard has been success- 

 fully photographed and measured ; a new term in the red has been 

 added and a faint haze at 4372 was detected on one photograph which is 

 doubtless the sixth member of the series. The pair at 4472 was so 

 diffuse that the lines could not be seen separately ; the setting was made 

 on the middle of it. The writer's measurements differ considerably 

 from those of Lehmann and of Konen and Hagenbach on several lines ; 

 a repetition of the measurements led to the same values. Konen and 

 Hagenbach give a line at 4973 which does not appear with certainty on 

 any of the present photographs; nor does there seem to be any companion 

 to 4660. The pair at 7410 is exceedingly faint, and may possibly not 

 belong to sodium. It has, however, approximately the same separation 

 as the Lenard series pair near by, and, like that pair, the line of greater 

 wave-length is slightly the stronger. If this is a member of another 

 series, it would seem likely that the other lines are too faint to have 

 been observed, unless, possibly, it could be grouped with the pair at 

 5670 and the lines at 4975 and 4660. An extremely faint group was 

 observed at 8210, but it was impossible to determine just what it was; 

 this was the greatest wave-length which was photographed on these 

 films. 



The lines 5100, 4820, and 4730 mentioned by Lenard could not be 

 found on the photographs, nor were eye observations with a plane 

 grating any more successful. 



It may be worth noting that the wave-number differences for Lenard's 

 series (III) are as follows: 14.72, 14.77, 18.2, and 18.7. As the first 



