168 



Mr. H. Eainage. Abnormal Changes in 



[Nov. 19, 



refrangible wing is lost in the bright line of wave-length 4603*1 . 

 There appears then in the middle portion of the spectrum a broad 

 dark line with wings of unequal extent, and the less refrangible wing- 

 is broader, and, near its more refrangible edge, much stronger than 

 the other wing. The broad dark line extended on such photographs 

 from (1) 4602-72 to 4601*81, (2) 4602-86 to 4602-05, (3) 4602*73 to 

 4601*62, and in two other cases, where there were less differences of 

 intensity between the two wings, from 4603*26 to 4601*61 and from 

 4603*07 to 4601*59. There were indications on some photographs of 

 a bright line near the middle of the broad dark line ; the measure- 

 ments of such a photograph gave : — 



Less refrangible edge of broad dark line 4603 * 40 



Apparent bright line 4602*57 



More refrangible edge of broad dark line 4601 * 31 



The absorption bands of reversals in such a photograph as this 

 are comparatively bright, for there is considerable action on the plate 

 where the images of these bands fall (4 in figure). The effect is prob- 

 ably due tothe superposition of the spectrum of the outer flame upon 

 that of the inner core when both are giving reversed lines. 



Some observations were made at Professor Liveing's suggestion on 

 the arc spectrum of lithium when the arc was formed between carbon 

 electrodes inserted horizontally in a magnesia brick. The arc enclosed 

 in this way is much steadier than in the open. The light which was 

 examined passed out through a horizontal hole perpendicular to the 

 carbons. The differences described above between the spectra of the 

 vapours near the two poles were easily observed by eye observations. 



Photographs of the spark spectra of lithium were taken in the hope 

 of finding a second line which would account for the remarkable 

 differences in the arc spectra, but no second line was found. A piece 

 of metallic lithium was placed in a cup formed in the end of an 

 aluminium rod, and sparks were passed between this and an alumin- 

 ium wire held immediately over the lithium. No Leyden-jar was 

 put into the secondary circuit at first, and it was found that 

 photographs of the blue line were obtained with exposures of 15 

 minutes. The lines were sharp, and showed no signs of reversal. 

 The wave-length of the blue line was found to be 4603*08. The 

 vapour near the electrodes gave broader lines than the vapour a short 

 distance away. The broadening was much greater at' the negative 

 electrode than the positive, and it extended further on the more 

 refrangible side than on the other side. The bright line and the broad 

 reversed line were, in fact, both present in the spectrum of the spark, 

 near the negative electrode. 



It was much more difficult to obtain photographs of the spark spec- 

 trum when a Ley den jar was introduced. Some were obtained from 



