August 1, 1891.] 



KNOWLEDGE. 



157 



not vary in tint. If, for instance, Mitchell bad hit the 

 truth with his conjecture in 1767, that Antares, Aldebaran, 

 and their fellows glowed with the ruddy light of decaying 

 incandescence, they could obviously be subject only to 

 changes proceeding in the uniform direction of the still 

 further deepening and darkening of their fires. But if 

 such variations as Von Zach anticipated do actually occur, 

 then the reasons prohibitive of them must necessarily be 

 invalid. As a matter of fact, do they occur ? A great 

 deal of evidence, recently accumulated, tends to show 

 that they do. If this be so, and the conditions producing 

 redness in stars should prove to be fluctuating, it becomes 

 evidently inadmissible to make colour, directly or in- 

 directly, a test of relative standing. What comes and goes 

 cannot be the badge of what is permanent. 



Stellar hues depend mainly, perhaps exclusively, upon 

 the composition and extent of stellar atmospheres. Un- 

 veiled stellar photospheres would probably all agree in the 

 display of a bluish tinge ; but on this point inference and 

 conjecture are our only guides ; exxserience keeps aloof. 

 What is thoroughly ascertained, on the other hand, is 

 that the light which we receive from the stars is a residuum ; 

 it is what gets through the siftmg apparatus, by various 

 makes and modes of which they are all, without exception, 

 surrounded. The resulting colours differ, just because the 

 sifting apparatus is of various makes and modes. The 

 more powerfully it acts, moreover, the redder, as a general 

 rule, are the transmitted rays, because the quick, short, 

 blue and violet vibrations get predominantly entangled and 

 stopped in gaseous envelopes of normal constitution. A 

 corresponding effect is produced m our own atmosphere ; 

 hence, the sun at the zenith is nearly white, but shows at 

 the horizon, where selective absorption exerts all its 

 strength, a more or less reddened disc. 



More or less. The effect is not always the same. Often 

 barely tinged with orange, at certain times our luminary 

 disappears \-ividly glo-ning in crimson or scarlet. In other 

 words, the setting sun is a colour-variable. And why? 

 Plainly because the state and composition of our atmosphere 

 are not always the same. The proportion of water-vapour 

 held in suspension is one inconstant element ; another is 

 the amount of floating dust-particles subsiding in still air, 

 or ceaselessly bome onwards by untiring upper currents. 

 And this may help to illustrate the cause of stellar 

 fluctuations in hue. It is, imquestionably, to be found 

 in changes of temperature, of chemical constitution, of 

 electrical condition, or of all three together (for they are 

 pretty sure to be mutually interdependent) taking place in 

 the glowing vaporous envelopes determinant of star- 

 coloiu's. 



Their fluctuations must then be evident in the spectrum, 

 as well as to the eye. Spectroscopic change, indeed, is only 

 a rc-attirmation, under a different form, of colour-change ; 

 the prism analyzing what the eye perceives. Colour sums 

 up, in one subjective impression, the whole series of facts 

 separately stated in each star-spectrum. The subjective 

 impression needs accordingly to be tested and interpreted 

 by the statement of fact. Its differences from time to 

 time, or from person to person, are indicative merely ; 

 they may sometimes be admitted as assertions, but they 

 must be unusually well established before they can amount 

 to demonstrations of objective change. 



It has long been known that the ruddy tinge of periodical 

 stars, like Mira, lightens and deepens with their gain and 

 loss of brightness ; and it has lately been ascertained 

 that remarkable spectral appearances accompany and 

 explain these transitions. In other cases, spectra have 

 boeii observed to vary independently of any noticeable 

 concomitant effect upon either light or hue. Some ex- 



ceptional objects, again, show, it would seem, marked 

 changes of colour apart from changes of brightness ; and 

 these alone are properly entitled " colour- variables." Such 

 of them as are in systemic connection with neighbouring 

 bodies display a wide prismatic diversity, and both in 

 themselves, and by their instabiUty, give rise to highly 

 intricate considerations ; while the phases of solitary 

 colour-variables consist entirely in the paling, or even total 

 disappearance of their customary red hue. 



Ostensible instances of this kind are numerous; a few 

 are fairly well-authenticated. Take the eighth magnitude 

 star in Virgo (148 Schjellerup), numbered 352 in the 

 " E spin-Birmingham Catalogue of Red Stars." It shows 

 a finely developed banded spectrum of the thn'd type, and 

 shines by Duner's estimate in his catalogue of such objects, 

 with a deep red yellow hght. Mr. Espin noted its " fine 

 pale red" tint. May 12th, 1885, described by Lord Rosse 

 in 1841 as " scarlet," by d'Arrest in 1866 as " dark 

 red." These, however, are but trifling discrepancies com- 

 pared with the alleged total absence of colom* from the 

 star on May 8th, 1874. Its whitness at that date is 

 attested by one — unfortunately only one — observation of 

 Bu-mingham's. Any recent notes of its colour which may 

 happen to have been made, would be of special interest 

 for purposes of comparison with the older records. Those 

 relating to the star 90 Schjellerup (Espin-Birmingham 

 221) are still less mutually reconcilable. But here there 

 is just a possibility — though a remote one — that two 

 different stars, one white, the other orange, may be in 

 question. No shadow of doubt, however, overshadows the 

 identity of a star in Orion (63 Schjellerup) numbered 1888 

 in the " Copenhagen Catalogue for 1864," which appeared 

 red in 1863, and is now no longer so. Schjellerup perceived 

 in 1870 that the hue he had originally been struck with 

 had vanished. A 7'5 magnitude star in Scutum Sobieski 

 (Schjellerup 214, Espin-Birmingham 544) may with some 

 confidence be pronoiinced variable in colour. Of its red- 

 ness, first perceived by Schjellerup, Birmingham could see 

 no trace in repeated observations made in the years 1872-4 ; 

 twice, indeed, he qualified the object as actually hJiu\ But 

 it had reverted, on September 29th, 1889, according to 

 Mr. Espin, to a "pale orange red "; and since its spectrum 

 is a banded one of the third type, it may claim to rank 

 normally — its occasionally blanched aspect notwithstand- 

 ing — as a red star. 



It must here suffice to add one further example from the 

 southern skies. It is that of r Velorum. Dr. Gould, at 

 Cordoba, found the colour of this bright star (5-3 magni- 

 tude) so pronounced as in part to baffle estimates of its 

 brightness. But judging from two observations of my own 

 at the Capo, it had completely lost this quality in the 

 autumn of 1888, and my eyes are rather exti-a-sensitive to 

 the warmer tints of the spectrum. An examination of the 

 star by Mr. Tebbutt in New Soutli Wales, March 3rd, 

 1891, showed it " to be very slightly tinged with red," 

 though, with a larger instrument (an eight-inch equatorial), 

 three days later, the tinge seemed more decided — •' very 

 decided," indeed, "compared with the other smaller, but 

 white stars in the same field of view."'- The recovery of 

 colour, however, is still evidently very imperfect. Mr. 

 Tebbutt, one is glad to learn, does not intend to lose 

 sight of this interesting object. 



Up to tlie present, suspected colour-variables have been 

 unaccountably neglected, notwithstanding the great im- 

 portance both of verifying and of investigating their 

 changes. For they supply what Bacon called " migratory 

 instances," which, in the course of their jourueyings from 



* Journal of the British Astrouojuii-til Association, vol. I., p. 423. 



