446 On the Phenomena of A r fir -S' [Apr. 16, 



faded was nothing more than the spectrum of the nebula itself, as 

 might be expected. Owing to the difficulty of making the observa- 

 tions, the apparent variations of the spectrum from day to day may 

 not be real, and it is hopeless to attempt to explain them by a 

 reference to the effects produced by a gradual fall of temperature. As 

 the star only fell two magnitudes during the whole period of spectro- 

 scopic observation, the change of temperature would not be so great 

 as in Nova Cygni, and the variations would not be so well marked. 

 No lines or bands, however, were on any occasion recorded in the 

 spectrum with which we are not familiar in other bodies which, there 

 is evidence to show, are meteoritic swarms. A diagram shows that 

 the spectrum of the Nova, as seen by Copeland, on October 1, can be 

 reproduced by adding the spectrum of hydrogen to that of the 

 nebula. 



It is next pointed out that the theoretical sequence of phenomena 

 in the spectrum of a Nova produced by the collision of two swarms of 

 different densities is in strict accordance with the partial sequences 

 actually observed. 



A discussion of the colour phenomena shows also that in Nov we 

 have to deal with mixed swarms, the colours at certain stages being 

 compound ones. 



In my former paper, I have shown that carbon radiation is one of 

 the chief characteristics of uncondensed .meteor-swarms, and the dis- 

 cussion of the new stars has revealed the fact that carbon is also one 

 of the chief characteristics of their spectra, though modified by other 

 substances. 



The observed changes in magnitudes of Novae are also in accord- 

 ance with the collision theory. The rapid fading away demonstrates 

 most conclusively that small bodies, and not large ones, are in ques- 

 tion. 



The observations with which I have had to deal have often been 

 imperfect, owing to the difficulty of observing 'this class of bodies, 

 and different observers have frequently disagreed with regard to some 

 of the spectroscopic details, but still, as I have endeavoured to show, 

 most of the discrepancies can be reconciled when difficulties of obser- 

 vation are allowed for. 



