On the Classification of Stars of the 5 Cepliei Class. 453 



(2) A period of complete vaporisation during which the atmo- 

 sphere is quiescent, bombardment having ceased, and the radiation 

 being too great to permit condensation in the atmosphere. 



(3) A period of cooling during which the atmosphere is disturbed 

 by the fall of condensation products from the outer parts of the atmo- 

 sphere on to the photosphere. 



Evidence that we are not dealing with Composite Spectra. 



Many of the lines in the spectra of stars of the 8 Cephei class 

 coincide with prominent lines in the spectrum of a Cygni, but it is 

 true that many also coincide with lines in the spectra of stars like 

 cc Orionis and Arcturus, which closely resemble the solar spectrum. 

 It seemed possible, therefore, that we might be dealing with the 

 integrated spectra of two stars in close proximity, one having lines 

 resembling those of a. Orionis or Arctnrus, and the other those of 

 cc Cygni. The spectra of all the stars of this sub-group have accord- 

 ingly been very carefully investigated from this point of view. 

 Enlarged glass positives of ol Cygni and Arcturus on exactly the 

 same scale have been superposed, and the integrated spectra photo- 

 graphed, When this integrated spectrum is compared with 7 Cygni or 

 h Cephei, there is a considerable similarity, but the relative intensi- 

 ties of the various lines and the general appearance of certain parts 

 of the spectrum, especially about Gr, are quite different. Again, if 

 there were two bright bodies physically connected in such a star as 

 7 Cygni there must be a revolution and a consequent doubling of the 

 common lines, unless the plane of movement were perpendicular to 

 the line of sight. No signs of such doubling, however, have been 

 detected in any of the eight stars of the sub-group which have so far 

 been recognised, and it is quite improbable that the plane of revolu- 

 tion would be at right angles to the line of sight in every case, and 

 still more so that the two components would have identical spectra 

 in each of the eight systems. 



Spectra of the h Cephei type must, therefore, be taken to repre- 

 sent a particular stage in the orderly development of cosmical 

 bodies. 



'Evidence from Variability. 



Further evidence in favour of placing stars like S Cephei and 

 7 Cygni on the ascending arm of the temperature curve is afforded 

 by the fact that stars of this class present a special form of vari- 

 ability. This variability is similar in kind, but different in degree, 

 to that associated with stars of Group II, such as Mira. The follow- 

 ing table shows that the amount of variation is very much less than 

 that in variables of the Mira type : — 



