TJie Theory of Tem])orary Stars. 7 



5. — Bearing of the Second Stage Phenomena on 

 Theory. 



1. If we desire to express these phenomena in terms of the 

 collision tiieory as usually presented, we must proceed as 

 follows : — 



The sources of the bright and dark bands possessed, before 

 and during maximum, little or no relative velocity in the line of 

 sight ; two days later they were moving in the same direction — 

 towards the sun — but with different speeds ; a fortnight later 

 they were moving in opposite directions with the enormous 

 relative velocity of 1800 kilometres per second; and after 

 another fortnight both came to rest. Meanwhile, another body 

 concerned in the phenomena was travelling slowly away from the 

 sun with uniform speed. 



I cannot imagine how such a series of performances is possible ; 

 neither collision nor orbital motion fits the facts. But it does 

 seem to me that we have here as strongly marked a case of 

 Pickering's experiinentum criicis as Pickering himself could desire ; 

 consequently, as far as Nova Fersei is concerned, we must 

 abandon all such explanations of the spectrum as are based on 

 the assumption of relative motion of different bodies.^ 



2. Wilsing's hypothesis fits the facts very well. The 

 \ariations, both in magnitude and direction, of the linear 

 displacements are readily explicable as due to pressure changes ; 

 we need only suppose tliat in the earlier stages the pressure of 

 the incandescent gases was comparatively low, then rapidly 

 increased, remained almost stationary for a time, and finally fell 

 off; while the fact, attested by all the observers, of the multiplicity 

 of the maxima in both bright and dark bands is readily explained 

 by the very probable supposition of different pressures in 

 diflferent parts of the gaseous layer — ^pressures, moreover, which 

 can in all cases be estimated in numbers of atmospheres by no 

 means unreasonably great. 



1 Had the distance between the bright and dark bands remained constant, the deduc- 

 tions would have been different. 



