of Lie/Jit Radiations emitted by Iron Vapour. 17 



The distinguishing feature of the groups in this class is 

 that they, unlike the groups in Classes I. and II., converge 

 towards the ultra-violet. In regard to the first group, it is 

 possible that it continues beyond 4737, but my observations 

 in this part of the spectrum have been greatly impeded by 

 the presence of a strong continuous spectrum, which, com- 

 bined with the low dispersion of my spectrograph, would 

 naturally obliterate all but the strongest lines. To try and 

 bring in further lines by extrapolation is almost hopeless, in 

 view of the fact that the wave-lengths are possibly uncertain 

 to within 0'05 of a unit, and, further, the number of 

 remaining spark or arc lines in that region is so great that 

 coincidences could always be found. 



The last four lines of the second group are close doublets 

 both in arc and spark. But on a photograph taken of this 

 region by M. de Watteville and myself with a high dis- 

 persion, I noticed that in the air-coal gas cone only one of 

 the components is brought out. As I have now not access 

 to this document, I am not sure that it is the more refrangible 

 component. 



In addition to the above two groups, there is a quartet 

 with the head line at 4022. 



Groups in the spectrum of self-induction spark. 



w i ti Oscillation 



wave-length. -^ A,. A,,. 



5 Frequency. l 1 



f4903"32 20394-3 



J 4938-83 20247-7 



1 4966-10 20136-5 



14985-26 r7? 20059-1 



J 4938-83 20247-7 \™ % 354 



1 4966-10 20136-5 "i* 33"8 



... - 77'4- 



The frequency differences between the components of the 

 above group are nearly the same as those for group /3 of 

 Class L, and it is not unlikely that there is some relation 

 between the two. 



In addition to this group, there is another quartet with 

 head line at 5400, likewise converging towards the red. 



It is quite possible that these two groups are really con- 

 stituted of temperature lines and belong to Class I. 



Although all the flame lines had been eliminated from the 

 spectrum of the spark, it seems probable that the remaining 

 spectrum is still a mixture of lines of different character. 

 For this reason no special attention has been devoted to this 

 class for the moment. 



Phil. Mag. S. 6. Yol. 33. No. 193. Jan. 1917. C 



