168 



Mr. J. Norman Lockyer. 



comparison of the continuous radiation spectra, and also from the 

 appearances of the iron lines. Unfortunately, the Kensington photo- 

 graphs of stellar spectra which have so far been obtained do not 

 extend far enough into the ultra-violet to permit a complete investi- 

 gation of the varying appearances of the two ultra-violet enhanced • 

 lines to which I have called attention in an earlier part of this 

 communication. Both of them appear in the spectrum of Sirius r 

 however, which is certainly a hot star, though not one of the 

 hottest. 



Still the facts are sufficient to show that although H and K appear 

 in some of the stars without any traces of the other calcium lines, 

 there can be no hesitation in accepting them as due to calcium, as 

 in all such stars a very high temperature is independently indicated 

 by the great extension of the continuous spectrum into the ultra- 

 violet and by the iron lines previously discussed. 



Magnesium. 



Magnesium furnishes us with very definite indications of the four 

 stages of temperature, and the discussion of its representative lines 

 in stellar spectra is therefore very important. 



In ol Orionis, the b group is strongly developed, while 4481 is 

 absent. In the map, the ultra-violet triplet commencing with 3834 is 

 taken as typical of the second stage spectrum, but the photographs of 

 the spectrum of a. Oricnis which have been obtained do not extend 

 far enough to enable the presence or absence of this triplet to be 

 ascertained. Another line, at 4352 2. which appears at the second 

 stage of temperature is, however, probably present in a Orionis. 



The flame line 4571'3 is also present,* and I have also shown the 

 probable presence of the two flutings of magnesium in the spectrum 

 of a Orionis and similar stars. f In cc Orionis, then, the most effective 

 absorbing magnesium vapour is at a temperature not greatly 

 different from that of the flame, but the presence of 4352*2 indicates 

 that the temperature must be slightly greater. It is important to 

 remember, however, that on the meteoritic hypothesis, different parts 

 of such a swarm as a. Orionis may have widely different temperatures, 

 as some of the collisions may be end-on while others may be mere 

 grazes, so that a mixed spectrum of high and low temperature lines 

 might be expected. 



Taking 7 Cygni, as before, to be considerably hotter than a Orionis,. 

 as indicated not only by the length of continuous spectrum but by 

 the iron and calcium results, there is a considerable change in the 

 lines representative of magnesium. The flutings have quite dis- 

 appeared and the spark line at 4181 appears as a well-marked line. 



* ' Phil. Trans.,' 1893, vol. 184, A, Plate 28. 

 f k Eoy. Soc. Proc.,' rol. 44, p. 54. 



