6i8 SCIENCE PROGRESS 



rendered luminous in the laboratory by an electric spark 

 discharge. This type of spectrum was observed in the case 

 of Nova Persei (1901), Nova Geminorum (191 2), Nova Aquilse 

 (191 8), and Nova Cygni (1920). It is most likely a character- 

 istic phase in the spectral changes of all new stars. In the 

 first three stars named their light was waning when the spectrum 

 was observed. In the case of Nova Cygni (1920) this same 

 type of spectrum was photographed by Dr. Lockyer, also when 

 the star was approaching its maximum of luminosity. 



This spectrum substantially matches that of the star 

 a Cygni. Now the spectrum of a Cygni is very similar to that, 

 which is not infrequently observed, in outbursts of metallic 

 vapours and gases on the sun's surface, in the neighbourhood 

 of sun-spots. Such outbursts are indicative of an eruptive 

 release of immense stores of internal energy. The sudden ap- 

 pearance of new stars, their rapid rise to luminous maximum, 

 and the metallic vapours involved in the early spectrum are 

 all analogous to a cataclysmic eruption, though on a greatly 

 enhanced scale, on the solar surface. 



At this stage, too, in the progression of spectral changes, the 

 spectrum extends far into the ultra-violet, which indicates a 

 very high degree of temperature. In the case of Nova Aquilae 

 (191 8), as observed at Stonyhurst, the ultra-violet spectrum 

 embraced the whole series of hydrogen absorptions, and ex- 

 tended well beyond them. Here again we have a connecting 

 link with the spectrum of the lower layers of the sun's atmo- 

 sphere, which are revealed in particular at times of total solar 

 eclipse. Probably when the new star is near the period of 

 greatest brightness, or at any rate a very few days after the 

 appearance of the absorption solar chromospheric spectrum, 

 a series of bright radiation bands begin to appear. This bright 

 band spectrum is for the most part the exact replica and 

 counterpart of the absorption spectrum. The brightest bands 

 are those due to hydrogen, and to iron vapour excited in the 

 electric spark discharge, and these bright broad bands always 

 occupy positions just less refrangible, or to the red side of the 

 accompanying absorptions. By far the brightest of these bands 

 is the red radiation of hydrogen. This is the cause of the 

 visual red or orange colour of the star, and of the change of 

 colour from white to red, observed by the early observers, 

 Tycho Brahe and Kepler. 



These bright bands, which increase in intensity as the star 

 begins to fall in luminosity, are of great breadth, extending 

 over many wave-lengths in the spectrum, on each side of the 

 normal positions of the radiations. In the case of Nova 

 Aquilae (191 8), for instance, they covered about fifty units of 

 wave-length. The effect of pressure on an incandescent 



