179 CHAPTER XII. VARIABLE DOUBLE STARS. THE light-changes of double stars are commonly of a fitful and indecisive kind. They may affect one or both members of stationary pairs; but visibly revolving stars, as a rule, conspire to vary, if they vary at all. The alternating fluctuations of y Virginis, discoverable only by close attention to the swaying balance of lustre between the components, are in this respect typical. Each may be described as normally of the third magnitude, and each in turn declines by about half a magnitude and recovers within a few days, yet so that the general preponderance during a cycle of several years, remains to the same star. The existence of this double periodicity was recognised in 1851 by M. Otto Struve, who, however, despaired of investigating it with success in a latitude where the stars in question never rise more than 30° above the horizon.' Their circulation is in the most eccentric of ascertained stellar orbits (see fig. 30). The ellipse traversed by y Virginis in 180 years is, in fact, proportionately somewhat narrower than the path round the sun of Encke's comet, so that the stars will in 1926 be separated by fully seventeen times the interval of space between them in 1836, when they merged into a single telescopic object. Their inequalities of light seem to have developed as they approached each other; at least, they first began to be noticed by Struve in 1818, and they at present tend to become obliterated, whether to revive with regained proximity towards the close of the twentieth century, future observations must decide. A spectrum of the Observations de Poulkowa, t. ix. p. 122. Sirian pattern is combined with a perceptible tinge of yellow in their light. Relative variability is in 44 Boötis still more marked than in y Virginis. But here a fundamental disparity between the components is seldom and temporarily abolished. Noted by Herschel as considerably unequal in 1781, they appeared to him perfectly matched in 1787. And it may be noted that they had in the interim passed periastron. Struve observed, June 16, 1819, a difference between them of two magnitudes, which had sunk to half a magnitude in 1833. Argelander found them precisely equal June 6, 1830; Dawes perceived, April 27, 1841, a slight advantage on the side of the usually smaller star; while the superiority of its companion was recorded by M. Dunér at Lund as ranging, during the years 1869 to 1875, from 0.4 to 1.3 magnitudes. Since their changes are often simultaneous, though not always in the same direction, their combined variability has never been conspicuous. The stars of 44 Boötis complete their rounds in a highly eccentric orbit in a period of 261 years.2 Their tints, varying from yellow and sky-blue to white and dull grey, cannot be without influence upon their photographic magnitudes, which were determined at Paris in 1886 to be 53 and 6. Their joint light, though of the same spectroscopic quality, has then only one-twelfth the intensity of that of y Virginis. The component stars of Boötis when photometrically measured at Harvard College in 1883 were of 44 and 4.8 magnitudes, but the order of their brightness has been at least three times reversed during a century of observation.3 Their period of revolution must be of enormous length. From 1796 to 1841 they appeared fixed; then a very slow wheeling movement became perceptible, accompanied by a diminution of distance, and it now taxes the powers of the best telescopes to divide them. Their spectrum is of the Sirian type. Lund Observations, 1876, p. 74. 2 Doberck's elements. 3 Harvard Annals, vol. xiv. p. 458; Observations de Poulkowa, t. ix. p. 143; Dunér, Mésures Micrométriques, p. C8. 'Crossley, Handbook of Double Stars, p. 299; Tarrant, Jour Liverpool Astr. Society, vol. v. p. 77. An analogous object is a Piscium, made up of a fourth and a fifth magnitude star at 3" distance, and revolving in a period unlikely to be much less than two thousand years. The larger certainly varies in light, and perhaps also in colour, the smaller certainly in colour, and perhaps also in light.' An observation made by Mr. Tebbutt in New South Wales, August 22, 1887, gave a unique proof of the relative variability of a close double star in Virgo (OΣ 256). At its occultation by the moon on that night, the chief part of the light went out with the disappearance of the reputed lesser star, the component which had of late passed for its primary remaining still for a few moments separately but dimly visible. Similar but less marked reversals had already been noticed by O. Struve and Dembowski in this slowly circulating pair.3 Variability, as we have said, affects both or neither of two stars so intimately united that their orbital movements have become apparent after a comparatively short lapse of time. A possible exception, however, to this rule is met with in & Cygni. This beautiful and delicate pair was discovered by Sir William Herschel in 1783, but in 1802 and 1804 he totally failed, with improved optical means, to see the eighthmagnitude companion. His son was equally unsuccessful under the best atmospheric conditions in 1823, and Sir James South and Gambart in 1825. It emerged to view, however, with Struve's nine-inch Fraunhofer in 1826, and has since been rarely missed. An occultation of one star by the other, postulated to account for the telescopic singleness of the pair between 1802 and 1826, was by their subsequent movements decisively shown not to have taken place, and the alternative hypothesis of a temporary loss of light in the small component was, almost of necessity, adopted. Yet it has received no strong countenance from recent observations. M. Dunér acquired the conviction from seven years of experience that the visibility of an object at all times difficult Harvard Annals, vols. xi. p. 112, xiv. p. 433; Flammarion, Catalogue, p. 12 depends entirely upon the state of the air; and Mr. Burnham seems to be of the same opinion.2 Changes of colour in this satellite star are nevertheless patent. Struve found it of an ashen shade from 1826 to 1833; in 1836 of a bright red. It has since generally appeared blue; but Dunér saw it once olive, though otherwise always red; and intervals of greyness are also on record. The computed orbits have hitherto failed to represent the movements of the system with any degree of accuracy; but Mr. Gore's with a period of 377 years may prove more successful. Relative variability has recently been detected by M. Flammarion in y Arietis; it is, or has been, also present in Serpentis, 38 Geminorum, T Boötis, & Arietis, and many other couples, most if not all of which give spectra of the Sirian type. Their agreement in the possession of this particular quality of light is the more remarkable from its being the badge in solitary stars of exceptional emissive stability. Everywhite star,' so far known to be variable, has proved also to be compound, and those of the Algol type are so far from making an exception to this rule, that they are among the most rapid of possibly existing binaries. Besides these, we are acquainted with only two Sirian stars, & Orionis and S Monocerotis, which have had periods of lightchange assigned to them. The first is a wide double star of dubious variability; the second is the leading member of a straggling cluster, and was discovered by Winnecke in 1867 to change from 49 to 54 magnitude in 3d. 10h. 38m. Of two close attendants, the inner and brighter, at a distance of 2"-8, has been thought to be in slow circulation, but the point is still unsettled. The system has no appreciable proper motion. A star situated near a Virginis is the only object characterised by a spectrum of the first type known to undergo extensive intrinsic fluctuations of light. From a comparison of observations going back to the tenth century, when Al Sûfi 1 Mésures Micrométriques, p. 118. 2 Westwood Oliver's Astronomy for Amateurs, chap. vii registered it as of fifth to sixth magnitude, Schmidt ascertained, in 1866, its irregular variability from the fifth to the eighth.' The anomaly of such changes in a Sirian star was brought more into harmony with other examples by Burnham's division of Y Virginis,' at Chicago in 1879, into two nearly equal components, less than half a second (0'47) apart.2 A subsequent observation gave no satisfactory evidence of alteration, either in brightness or position, during the intervening two years. This is not surprising, since accesses of light-change in double stars are very generally followed by long intermissions. Nor could orbital motion, although presumably in progress, be expected to become so quickly apparent. The study of its laws, and of the varying magnitudes of the members of this singularly interesting system, ought not to be neglected by the possessors of great telescopes. The fluctuations of U Tauri (observed by the late Mr. Baxendell, 1865 to 1871), like those of Y Virginis, seem for the present suspended. The twofold nature of this object, which is situated quite close to a variable nebula, was detected by Mr. Knott, December 4, 1867. Each component is of 9.7 magnitude, and they lie 44" asunder. No recent observations of them that we are aware of have been made. The variotinted pairs & Cephei and B Cygni both belong to the class treated of in the present chapter. The former is the well-known short-period variable with which we have earlier become acquainted; the latter changes slowly and almost imperceptibly between 3.3 and 3.9 magnitude. The satellite is, in each case, exempt from the suspicion of instability. Not so the fifth magnitude attendant of a Herculis. The elder Struve considered that it declined occasionally to one-sixth its normal brightness; and Father Secchi also perceived irregularities, which have, however, for many years past ceased to be noticeable. The green hue of this star 8 7 Astr. Nach. No. 1597; Harvard Annals, vol. xiv. p. 456. 2 Observatory, vol. iii. p. 192. Mems. R. Astr. Society, vol. xlvii. p. 190. See ante, p. 132. Mensura Micrometrice, p. 97. Ibid. vol. xliii. p. 78. 6 Klein, Astr. Nach. No. 1663. Atti dell'Accad. Pont. t. vii. p. 62. |