April 30, 1885J 



NATURE 



605 



adopt some form of classification of variable stars. The 

 following arrangement suggested by Prof. Pickering will 

 suit our purpose ; but at the same time it should be re- 

 marked that links of association may sometimes be dis- 

 covered between individual members of different classes 

 in respect of some of their characteristics. It is probable, 

 too, that after all some stars must remain unclassed. 



" Class I. Temporary stars, or those which shine out 

 suddenly, sometimes with great brilliancy, and gradually 

 fade away. Examples : Tycho Brahe's star of 1572 ; 

 new star in Corona 1S66. 



"Class II. Long-period variables, or those undergoing 

 great variations of light, the changes recurring in periods 

 of several months. Examples : o Ceti and x Cygni. 



"Class III. Stars undergoing slight changes according 

 to laws as yet unknown. Examples : a Orionis and a 

 Cassiopeia;. 



" Class IV. Short-period variables, or stars whose light 

 is continually varying, but the changes are repeated with 

 great regularity in a period not exceeding a few days. 

 Examples : Lyrae and 8 Cephei. 



" Class V. Algol stars, or stars which, for the greater 

 portion of the time, undergo no change in light, but every 

 few days suffer a remarkable diminution in light for a 

 few hours. Examples : (3 Persei (Algol) and S Cancri." 



The temporary or new stars form a remarkable class of 

 stars, which blaze unexpectedly into view and then 

 gradually decline. A striking object of this class was 

 Tycho Brahe's star of 1572, which attained such a 

 brilliancy as to be visible by day. It was not, however, 

 till the year 1S66 that a clue was found to the probable 

 nature of these outbursts, when the examination of the 

 spectrum of the new star which appeared in Corona 

 Borealis in May of that year by Dr. Huggins suggested 

 the view that in these cases the outburst is due to the 

 liberation of large volumes of gas, which enwraps the star 

 in a naming envelope which gradually burns itself out. 



The most recently-observed star of this type has a 

 curious history. On September 24, 1876, the late Dr. 

 Schmidt discovered, in the constellation Cygnus, a new 

 star of the 3rd magnitude, which soon began to fade. 

 Like the star T Corona? it had a double spectrum. In 

 September, 1S77, when the star had fallen to 105 mag., 

 an examination of its spectrum at the Earl of Crawford's 

 observatory showed that the continuous spectrum had 

 disappeared and that the star's light was monochromatic. 

 In fact, to all appearance the star had become a minute 

 planetary nebula ! 



The distinguishing characteristic of stars of this type, 

 namely the temporary character of their phenomena, 

 sharply marks them off from the variables of all the other 

 classes, in which the changes recur with greater or less 

 regularity. A connecting-link, however, may perhaps be 

 found in the remarkable variable, U Geminorum, dis- 

 covered by Mr. Hind in 1 S55. It has a very irregular 

 period, which ranges between 70 and 126 days, during 

 about three-quarters or more of which time it remains 

 fluctuating about a minimum magnitude of I4'5- It rises 

 rapidly to maximum (at the maximum of February, 1877, 

 at the rate of over three magnitudes in twenty-four hours), 

 and then, at first gradually and then more rapidly, falls 

 to minimum again. Its colour has generally been de- 

 scribed as bluish-white (though it has been noted ruddy), 

 and a curious, ill-defined or hazy appearance has been 

 noticed by several observers which would suggest the 

 possibility of bright lines being found in its spectrum, a 

 suspicion which has not as yet been confirmed. 



Class II. includes by far the greater number of known 

 variable stars. Many of these are highly coloured, 

 showing tints of red or orange of various degrees of 

 intensity, and among them are to be found stars having 

 fine banded spectra of Secchi's types III. and IV. The 

 regularity with which they go through their changes is of 

 various degrees, and varies even in the same star at 



different times, while in some cases there is evidenced a 

 tendency to form subsidiary maxima or minima on the 

 main light curve. In some instances also the magnitude 

 touched by the same star at maximum or minimum is 

 subject to fluctuations, and this apparently quite inde- 

 pendently of the degree of regularity with which the 

 changes are gone through in respect of time. In two 

 stars at least of this class — Mira Ceti and R Geminorum — 

 bright lines have been observed in their spectra. 



It is perhaps to be regretted that a separate class has 

 not been formed for variable stars having a double period, 

 with two equal or nearly equal maxima and two unequal 

 minima, of which /3 Lyras is the type. A star of this order, 

 with a period of about 70 days, R Sagittae, included in 

 Class II. (though with an expression of doubt) in Prof. 

 Pickering's list, seems to call for special remark. It was 

 discovered by Mr. Baxendell in 1859, and his observa- 

 tions have shown first the approach to equality and then 

 the reversal of the principal and secondary minima. The 

 equalisation of the minima was also observed by Prof. 

 Schonfeld, and their reversal by Mr. Chandler in 

 America. The phenomenon thus exhibited is a remark- 

 able one, though perhaps not unique, as something similar 

 appears to have been noticed by Prof. Argelander and 

 Prof. Schonfeld in the case of R Scuti. 



Turning to Class III., a point should be mentioned in 

 regard to one of the examples of the class a Orionis. Ob- 

 serving the star in March, 1866, Dr. Huggins noticed that 

 "a group of lines and shading, as if of fine lines" had 

 disappeared from its spectrum, the star at the time being 

 at its maximum brilliancy. Six years later, however, 

 Dr. Vogel, at Bothkamp, failed to detect any change of 

 this character. 



Passing to Class IV. we have, in one of the examples, 

 /3 Lyrae, a star presenting points of singular interest. As 

 has been already mentioned, its period of 1 2 '9 days is a 

 double one, with two equal maxima and two unequal 

 minima, and Herr E. von Gothard, of the Hereny 

 Observatory, has discovered that its spectrum is also 

 variable. Herr von Gothard has also observed the 

 D 3 line (showing that Helium has a home in other suns 

 than ours), and the lines of hydrogen as bright lines, and 

 has further {Astr. Nach., No. 2651) found them to vary 

 in intensity in a period of about seven days. Further 

 observation is required before any decided opinion can 

 be expressed as to the relation between the variation of 

 the spectrum and the variation of the star's light, but a 

 comparison of Herr von Gothard's observations with the 

 predictions of an ephemeris seems to suggest (though the 

 evidence is not quite conclusive) that the bright lines are 

 at their brightest when the star is near a minimum. 



The stars of Class V., of which Algol is an example, 

 form a group of variables of a well-marked type. The 

 general features of their changes are fairly represented by 

 the supposition of an eclipsing satellite. But in the case 

 of U Cephei, a star of the group discovered a few years 

 ago by Ceraski, a new feature is introduced which some- 

 what complicates the theory. Its period has been shown 

 with some degree of probability to be a double one, with 

 slightly unequal minima. Another curious fact which has 

 been observed in regard to the star is that, as it falls 

 below the 8th magnitude, its light becomes decidedly 

 ruddy (indicating absorption as well as eclipse ?), the 

 ruddy colour being lost as the star rises to the 8th magni- 

 tude again, when it regains its ordinary brilliant bluish- 

 white hue. It is only fair to remark that in Prof. 

 Pickering's view the suggestion as to the duplicity of the 

 star's period should be at present received with caution. 



This brief review will suffice to show that any attempt 

 to answer the question — What is a variable star? — in- 

 volves the examination of a multiplicity of phenomena. 

 At the same time, the causes presumably at work may 

 be grouped broadly under two heads — geometric and 

 chemico-physical. We have seen that in the case of the 



