366 



KNOWLEDGE. 



Septkmuer. I'tll. 



Soon, however, the h\-drogen escapes and becomes 

 free, and forms an ensphering shell many diameters 

 of that of the brilliant nucleus of hea\\' matter. This 

 brilliant gas shell, as it is expanding in all directions, 

 is moving at e\er\- angle to the line of sight, both 

 from and towards us. Hence what would, were the 

 gas in a fixed position, be a bright line, becomes an 

 exceedingly broad band, indicating at its extreme 

 edgesatomic \-elocities of a thousand miles awa\- from 

 and towards us. Hence we now have a continuous 

 spectrum covered with brilliant bands of the 

 h\drogen series. But the part of the shell of 

 h_\drogen that is immediateh- in front of the 



atmospheric coiulitions do not seem to have been 

 satisfactorv at South Kensington, when these stages 

 of Nova Persci occurred ; but they show themselves 

 beautifully in the Sttmyhurst spectograms. Father 

 Sidgreaves,the eminent astronomer of Stonyhurst,has 

 stated : — " That he has no doubt at all but that Nova 

 Persei actually was the third body, torn from a pair of 

 colliding suns."' He is quite opposed to the opinion 

 ex))ressed b\" some astronomers in The Times, that 

 the phenomena of No\'a Persei might be thought to 

 be explained b\- the idea of a dense body passing into 

 a nebula. He sa\'S the light could not grow up so 

 suddenK. It would take an ordinarv star fully a 



Time. 



Figure 9. Deduced Light curve of the Third Body. 



brilliant nucleus, must absorb its isochromatic ra\s 

 and produce re\-ersion. The hydrogen coming our 

 way has its wave length shortened, that is, displaced 

 towards the violet ; consequently each one of these 

 bright bands will have a dark band edging it, on the 

 side towards the violet. Many observatories all o\-er 

 the world obtained spectrograms of No\a Persei, when 

 it was exhibiting this spectrum. Those of South 

 Kensington and Stonyhurst were both singularly per- 

 fect, and examples are given here (see F"igures 7 and 8), 

 and it will be seen that observation exactly corresponds 

 with deduction. The next deduced change in the 

 spectrum of the third body will be that as the speed 

 of expansion of the ensphering shell does not appre- 

 ciabh' diminish, the ratio of the portion of the sphere 

 that is in front of the absorption nucleus must get to 

 be extremely small. Consequenth" the dark lines 

 edging the bright bands, although the\' will not alter 

 their position, w ill become exceedingl}' thin and hair 

 like. Later on, the tenuity of the gas shell will be 

 so extreme that it will possess no absorbing power 

 and will disappear. .\ little later on the vast 

 ensphering shells of hydrogen will become so rare, 

 that molecular encounter will diminish so much that 

 the lessened luminosit\ will cease to photograph 

 itself, and so the bright bands will die out. The 



thousand years to pass from the outside to the 

 centre of the No\'a Persei nebula, whereas Nova 

 Persei rose to a maximum in less than two da\'s. 

 This rapidit}- of formation renders it certain that if 

 Nova Persei was the result of any kind of collision, it 

 must have been the collision of compact bodies. 

 There are fulh' a dozen other points of coincidence 

 between deduced and observed spectral phenomena, 

 but there are twd that are extremely interesting and 

 about which we must say a few words. For various 

 physical reasons, already hinted at, I assumed that 

 the ends of the s})indle would consist of elements 

 having atomic w eights ajjiiroximately that of iron and 

 titanium, and because iron was a \er\' abundant 

 element in nature, I thought that the spindle ends 

 would largeh- be made up of iron. Of course, the 

 speed of the iron molecules would be much less than 

 that of hxdrogen, but as hxdrogen was not escaping 

 from the ends of the siiindle, the iron would not 

 have its kinetol lessened, and because its light would 

 not pass in front of the nucleus, save under verj' 

 exceptional conditions, the lines of iron would not 

 have a shadow band on the end towards the violet, 

 and would be moderateh' broad. \et not as broad 

 as hydrogen. Figure 7 (given on page J65) shows 

 [lart of a South Kensington spectrogram, a set of 



