644 



MATURE 



[February ii, 191 5. 



NOTES ON STELLAR CLASSIFICATIONS 

 III. 



BEARING in mind what has been said about the 

 orig-in of the special class of bright line stars, 

 I g'ive a diagram showing- that they occur exactly 

 where we should expect to find them if the sug- 

 gested origin is a sound one. 



Alni 



Alni 



Tau 



Pol 

 Ald 



Ant 



B 

 B 



A 



F 



G 



K 



M N 



AcM XX 



Alq xkk 



Mark 



Sin 



Phoc 



Arct 



Pisc 



Ant 



M N 





Piso 



Fig. s-^Position of the bright 

 line stars at the top of the 

 temperature curve. 



Fig. 6. — Position of the 

 bright line variables at 

 the base of the ascend- 

 ing arm of the tempera- 

 ture curve. 



I retura to that class of variable stars the 

 changing luminosity of which, according to my 

 hypothesis, is due to the heat and light produced 

 by collisions between meteor swarms (Fig. 6). 



Of these stars, in which the bright radiations 

 have been recorded at maximum, one only, 

 o Ceti, is included among the eighteen classified at 

 Kensington as Antarian ; 225 stars classified as Md 

 (Antarian) are given on the Harvard List.'' These 

 are chiefly variables of long period. Prof. 

 Pickering has obtained forty photographs showing 

 bright lines, and ten stars have been discovered 

 by means of this peculiarity of the spectrum.^ 

 Near minimum, when the star is faint, the bright 

 lines are not seen. 



There can be no doubt that all are physically 

 similar, and that the position of all is at the base 

 of the ascending arm of the temperature curve. 

 That they are found here and nowhere else is an 

 important test of the view that we have to deal 

 with nebulous clouds consisting of discrete 

 meteorites, as in comets, the gases from which 

 produced by the heat supplied by collisions fill the 

 interspaces at maximum and superadd the bright 

 radiations to the fluting and line absorption seen 

 at all times with varying intensity. 



We next pass from these long-period variables 

 with smooth curves, sometimes called " Mira 

 variables," to another class with short periods, and 

 with curves not so smooth, called "Cepheid vari- 

 ables." 



The diagram Fig. 7 shows that these occur in 

 the greatest number at a stage higher up the 

 ascending branch of the temperature curve than 



* Continued from p. 619. 



7 Annals, vol. Iv., pt. i. 



8 Astr. Nachrichten, vol. clxxxvii., p i. 



the Miras. So far bright lines have not been 

 recorded at maximum, but it has been noted that 

 the position of maximum light energy moves 

 towards the violet ^ as the star brightens. Their 

 shorter periods, taken in connection with the fact 

 that they must be more condensed than the Miras 

 judging from their position on the curve, explains 

 this difference of behaviour — there are no inter- 

 spaces. 



The Cepheid variables are not the only class 

 without bright lines. I have already (Bull. I.) 

 referred to the "eclipsing variables." Of these 

 there are two classes, represented by ^ Lyrae and 

 Algol, in which the shape of the curve is very 

 different. This difference seems to arise from 

 the fact that in the case of the Algols we are 

 dealing with true discs, while in the case of the 

 Lyras this stage has not been yet reached. These 

 conditions are indicated by the position of the 

 Lyras on the ascending and of the Algols on the 

 descending arm of the curve. 



Among the Piscian stars at the base of the 

 descending arm we find a number distinguished by 

 their irregular variability. In these we deal with a 

 star rapidly dimming, surrounded by vapours of 

 carbon giving absorption flutings, and of iron 

 giving a multitude of absorption lines. The 

 changes in the density of these vapours, to say. 

 nothing of the possible formation of scoriae, are 

 sufficifent to explain the variation and irregularity 

 in the light given out by these bodies. There 

 are no interspaces to produce bright lines. 



It is singular that some inquirers seem to regard 

 the Antarian and Piscian stars (M and N) as being 

 similar and so treat them as convertible terms. 



Alni 



AcH 



Alq 



Mark 



Sir 



Proc 



Arct 



Pisc 



Fig. 7.— Position of the Cepheid 

 variables halfway up the 

 ascending arm of the tem- 

 perature curve. J 



while the fact is that the only similarity is their 

 temperature. The following table will indicate 

 their dissimilarities : — • 



Antarian. 



Least condensed 



Titanium absorption flutings 



Regular variables 



Bright lines . . 



Piscian. 



Most condensed 

 Carbon absorption flutings 

 Irregular variables 

 No bright lines 



NO. 2363, VOL. 94] 



9 Albrect in Astra^iysical Journal, xxv., 332. 



