Variable Stars

( Originally Published Early 1900's )

A LISTLESS observer of the stars will regard them as always preserving their brilliancy, be it much or little, unchanged, but such is not the case with all of them ; a certain number vary from time to time in their light, and are therefore called " variable " stars. The number of those of which it may be said with certainty that they undergo periodical changes of brilliancy amounts to nearly 300 ; but it is probable that as many again may be regarded as possibly subject to fluctuations of light. In the absence of absolute standards for comparison, the systematic study of variable stars is a matter involving much patience on the part of the observer and much refinement in his procedure. Were the number of observers endowed with the requisite patience and experience much increased, there is no doubt that large additions would soon be made to our lists of variable stars. This department of astronomy is entirely modern, for the ancients have left us merely a few vague statements of stars having disappeared, and we can seldom determine with adequate precision the places occupied by them.

Professor Young has made some remarks on the method of observation to be resorted to in the case of variable stars, which it may be useful to quote here. He says :—" There is no better way than that of comparing the star by the eye, or with the help of an opera-glass, with surrounding stars of about the same brightness at the time when its light is near the maximum or minimum ; noting to which of them it is just equal at that moment, and also those which are a shade brighter or fainter. It is possible for an amateur to do really valuable work in this way, by putting himself in relation with some observatory which is interested in the subject. The observations themselves require so much time that it is difficult for the working force in a regular observatory to attend to the matter properly, and outside assistance is heartily welcomed in gathering the needed facts. The observations themselves are not specially difficult, require no very great labour or mathematical skill in their reduction, and, as has been said, can be made without instruments ; but they require patience, assiduity, and a keen eye."

One of the most celebrated of the periodical stars is o Ceti, otherwise known as Mira (the " wonderful " star), which latter name has been given to it precisely because it undergoes such remarkable changes. Its period is 331d. 8h.; that is to say, it goes through its changes 12 times in about 11 years. At its maximum brightness it sometimes rises to the 2nd mag., remaining thereat for about a fortnight ; it then diminishes during about 3 months, becomes invisible except in large telescopes for 5 months, and finally requires another 3 months to regain its pristine maximum brilliancy. These may be taken as average intervals, but it does not always diminish or increase by the same gradations ; its maximum brightness is not always the same, nor are the intervals of time between maximum and maximum always identical. These irregularities were studied very carefully by Argelander, who came to the conclusion that the period of 331d. of Mira itself varies in about 88 of such periods, with the result that the single periods gradually lengthen and shorten alternately to the extent of 25 days one way or the other. Moreover, it is not improbable that the irregularities in the star's maximum brilliancy are also periodical, and that at every 11th maximum the star attains to a greater degree of brilliancy than its usual maximum brilliancy. This supposition would explain the fact that whilst the naked-eye visibility of Mira generally extends to about 18 weeks, it was in 1859-60 so observed during 21 weeks, whilst in 1868 the term was but 12 weeks. It was in this year that Heis noted the maximum magnitude to be only the 5th, whilst in 1888 it attained to about the 3rd mag., and therefore of course remained a naked-eye object for a much longer period. The discovery of the variability of this star dates back as far as 1596. On August 13 of that year David Fabricius noted a certain star in Cetus to be of the 3rd magnitude. In the following October he failed to find it. In 1603 Bayer, in preparing his " Star Atlas," assigned the Greek letter o to a star in Cetus occupying the spot where Fabricius's star had disappeared He noted it to be of the 4th mag., but being either ignorant of, or neglectful of, the earlier observations of Fabricius, he lost the chance of being able to claim the honour of discovering the star's variability. The spectrum of Mira is a remarkable one of Secchi's IIIrd Type, in which bright lines have been seen.

Perhaps Algol (3 Persei) may be regarded as, after Mira Ceti, the second most remarkable variable in the heavens, or, at any rate, in the Northern hemisphere, as it is second also in point of date of discovery. The fact of its variability was noticed by Montanari in 1669, was con-firmed by Maraldi in 1694, and investigated half a century later by a Saxon farmer named Palitzch, celebrated for his early detection of Halley's Comet in 1758. But it was Goodricke, in 1782, who first determined in full detail the changes of brilliancy which Algol undergoes. It commonly shines as a star of mag. 21; from that it descends to about A. Pickering, from photometric measures at Harvard College, finds that the star's light diminishes during 4h. 23m. before minimum. When the minimum is reached, then 5h. 37m. pass before the star regains its normal maximum. It remains at this for about 2d. loh. The most rapid changes take place during about loo minutes before and too minutes after the epoch of mini-mum. Pickering suggests that the range of variability is less than is commonly stated, and does not exceed one whole magnitude. The period in which the entire series of changes take place is about 2d. 20h. 48m., and is thought by Chandler to have diminished by 8s. since Goodricke's time ; but to talk about 8s. in such a connection is a refinement of precision which savours of affectation.

Another naked-eye variable handy, by reason of its position and magnitude, for observers in the Northern hemisphere is a Cephei. Its period is 5d. 8h. 47m. counting from minimum to minimum, and its range from about mag. 3 3/4 to mag. 4 3/4. The interval between maximum and maximum is not equally divided by the minimum phase, for it takes longer for the star to pass from its maximum to its minimum than it does to regain its maximum after a minimum. The former transformation occupies 3d. 19h. but the latter only rd. 14h. The variability of

Cephei was discovered by Goodricke in 1784, and the whole period is put at 5d. 8h. 48m.

n Aquile and ß Lyre may also be mentioned as short period variables, which on that account, and because they are visible to the naked eye, are specially suitable for observation by amateurs in England.

n Aquile varies from about mag. 31 to mag. 41 in a period of about 7d. 4h. 14m., but this period itself seems variable. The star is yellow in colour and yields a spectrum of Secchi's IInd Type.

ß Lyre is remarkable as having a double maximum and a double minimum, which together make up a main period of 2d. 21h. 47m. The variations take the following form :—Starting from a maximum when the star is of mag. 31, it descends to its first minimum of mag. 4 ; it then rises to the same maximum as before, but in descending to the next minimum it goes down to mag. 4 1/2. Argelander ascertained that B Lyre resembles Mira Ceti as regards the circumstances of its period in other words, that its period is itself variable ; that down to 184o the period was increasing, but that after 184o it began to decrease, and was decreasing at the time when Argelander made this remark in 1866. Pickering has propounded the idea that this star is a "surface of revolution" or a spheroid in form and differently luminous in different parts, and that the epoch of minimum light represents a time when the darker portion at one of the ends is presented to the earth. This seems to be one of those far fetched fancies which can neither be proved nor disproved. The variability of B Lyre was discovered by Goodricke in 1784.

I have reserved to the last that which is perhaps the most remarkable, as certainly it is the most erratic, o an the prominent variable stars, Argūs. Unfortunatey it is not visible in the Northern hemisphere. Halley, on his return from St: Helena, as far back as 1677, frequentiv expressed doubts as to the constancy of the light of the stars in Argo. Though he seems only to have based his conclusions upon Ptolemy's statements of star magnitudes, yet these were generally so accurate that when discrepancies were found to exist between modern and ancient records the idea at once suggested itself that there had been actual change rather than mistakes of observation. Palley, in 1677, rated r Argūs as of the 4th magnitude. In 1751 La Caille noted it as of the 2nd magnitude. In the next half century it diminished, for Burchell, during his residence and travels in South Africa, between 1811 and 1815, saw it as of the 4th magnitude. Fallows, in 1822, at the Cape, and Sir T. M. Brisbane, between 1822 and 1826, in New South Wales, saw it as of the 2nd magnitude. In the following year, that is, on Feb. 1, 1827, Burchell, then at St. Paul's, in Brazil, saw it as of the 1st magnitude, and almost as bright as a Crucis ; but within a year, that is, by Feb. 20, 1828, it had decreased to the 2nd magnitude, and as such was entered by M. J. Johnson and Taylor in their respective catalogues between 1829 and 1833. Sir John Herschel, who started observations at the Cape in 1834, found it then and for several years afterwards to be something between mags. 1 and 2 but nearer 2. It seems to have remained stationary, or nearly so, for well-nigh 3 years, but on December 16, 1837, on resuming work after an interval, Sir John was startled to find it had become one of the very brightest stars of the 1st magnitude, excelling all belonging to that category except Sirius and Canopus. Sir John Herschel's account of it will bear quoting: '' Its light was nearly tripled. . . . It very decidedly surpassed Procyon, which was about the same altitude, and was far superior to Aldebaran. It exceeded a Orionis, and the only star (Sirius and Canopus excepted) which could at all be compared with it was Rigel, which it somewhat surpassed. From this time its light continued to increase. On December 28 it was far superior to Rigel and, could only be compared with a Centauri, which it equalled, having the ad-vantage of altitude, but fell somewhat short of it as the altitudes approached equality. The maximum of brightness seems to have been obtained [attained] about January 2, 1838, on which night both stars being high and the sky clear and pure, it was judged to be very nearl) matched indeed with a Centauri, sometimes the one, sometimes the other, being judged brighter, but on the whole a was considered to have some little superiority. After this the light began to fade." Sir John then goes on to narrate the incidents of the declension of the star's light. His own observations ceased in April, 1838, but the star even then remained bright enough to be compared with Aidebaran. From other sources we learn that the diminution of light went on for 5 years, but that even in March, 1843, the star's lustre continued equal to that of an ordinary 1st magnitude star. At about that time a new outburst took place, and according to the observations of Mackey at Calcutta, and Maclear at the Cape, n Argūs surpassed Canopus and scarcely fell short of Sirius in brilliancy. This lasted more or less through 1844, when a decline in its brilliancy set in. This proceeded, however, very slowly, because in February, 185o, Lieut. Gilliss, then in Chili, reported n Argūs as being nearly as bright as Canopus but of a reddish-yellow colour, somewhat darker than Mars. In 1856 it was still of the 1st mag., but a steady decline was evidently in progress. Hence we find that in 1858 it was rated at mag. 2 1/2 by Powell ; in 186o at mag. 3 by Tebbutt : in 1861 at mag. 4i by Abbott ; in 1863 at mag. 5 by Ellery; in 1867 at mag. 6 by Tebbutt. During the next to years it fell to mag. 7, and in March, 1886, was rated at mag. 7 1/2 by Finlay at the Cape of Good Hope. This appears to have been the lowest point, for by May, 1888, the light had increased by fully half a magnitude, so that apparently it is on its way towards another maxi-mum, which perhaps may be expected within the first decade of the 20th century. From the foregoing account it is, however, clear that we do not possess sufficient information to assign with any reasonable degree of accuracy a period to n Argūs, though Wolf has suggested 46 years, and Loomis 67 years. Schonfeld, however, thinks that the star has no regular period at all. At any rate the maximum stage seems very complicated and to consist of 3 maxima which jointly occupy 25 years of the period whatever that may be. During this subperiod, the changes may perhaps be regarded as restricted to the 1st and 2nd magnitudes, and this subperiod may perhaps be assumed to fall something like in the mid-interval between every 6th or 7th magnitude minimum of the star.

n Argūs is in the field with the celebrated "Great Nebula in Argo," and some remarkable circumstances bearing alike on the star and on the nebula will come under consideration in a later chapter in which the nebula will be described.

The reader who has followed me thus far in trying to pick up some ideas about the peculiarities of the stars called "variable" will very likely wish now to put the question, " What is a variable star? " It is impossible to answer such an inquiry with any confidence. It seems, however, likely that the variability of the stars which are known to be variable may be due to one of two causes, one of which applies to one class of star and the other to another class. It is generally accepted by astronomers that Algol is a type of a small number of stars which owe their peculiarity to a cause quite different from that which applies to the vast majority of these objects. The idea was started by Pigott in 1783, and has met with much acceptance, that the periodical fluctuations in the light of Algol are due to the revolution round it of an opaque satellite smaller than itself yet large enough to eclipse partially the primary. With respect to the general run of variables it is thought that we may draw some inferences respecting them from what we know of the physical constitution of the sun and of what happens in or upon that luminary. Now we know that from time to time, and ac-cording to a period which is recognised to amount to about 11 years, dark spots of various sizes and shapes and of different depths of shade break out upon the sun. The solar spots which we are accustomed to see, even the very largest of them, are too small relatively to the size and brilliancy of the sun to cause any measurable depreciation in the aggregate of the sun's light, but let us suppose it were otherwise, and that every 11 years masses of spots so extensive as to represent one-half or even one-fourth of the apparent surface of the sun, burst forth, we should then have the great centre of our system converted from a permanently bright star into a variable star. I speak of our sun as a bright star because probably it represents for us on the earth neither less nor more than what Sirius or other bright stars * represent to the inhabitants of other worlds in far-off regions of space. If we could travel from the earth a long way towards Sirius we should probably find Sirius to grow into what we should without reservation call a sun, whilst our sun would deteriorate into what we now call a star.

So much for the possible circumstances of those stars which undergo periodic changes of light. But this explanation, even if accepted so far, does not meet the case or those temporary outbursts of stellar light which we considered in Chapter XI. (ante). Here again, however, solar history may be brought in. It is now quite recognised as a fact that the red flames seen during total eclipses of the sun are outbursts of glowing hydrogen gas emanating from the interior of the sun ; nay, more, that such emanations of burning hydrogen are constantly occurring on the sun. Now in the case of the temporary star in Corona Borealis which became visible in 1866, Huggins's observations tended to show that there happened in that star a sudden and extraordinary outburst of glowing hydrogen, which by its own light, as well perhaps as by heating up the whole surface of the star, caused the unwonted in-crease in its brilliancy which then took place. These ideas find confirmation in other directions, but it seems hardly within the design of this work to go further into details of this character.

There are, however, some miscellaneous facts connected with variable stars which are too interesting to be passed over. For instance, it is an undoubted fact that the vast majority of the variable stars are red or reddish in colour ; and so general is this rule that whenever a new star is found it is a safe presumption to start with that if its colour is red it has hitherto escaped observation be-cause of its being variable. Hind has noticed that variable stars when at minimum often appear hazy or foggy, on which Arago suggested the idea that the diminution of brilliancy might be due to the interference of clouds. It is an undoubted fact that in the case of red variable stars as they diminish in brilliancy they deepen in colour, whilst as their light increases their hue becomes paler.

An experienced American observer, Chandler, has evolved a connection between the colours and periods of variable stars. He not only subscribes to the opinion that variable stars are generally red, but he finds that the more red they are the longer their periods. Of 112 variables whose colours and periods are fairly well established, classifying them in groups having periods of under loo days, of too days but under zoo, and so on up to over 400, he finds that whilst of those under too days barely one-half are red, of those over too days three-fourths are red ; whilst those over 400 days are all red. His statistics arranged in another form show that whilst the periods of the white and yellow stars average t25 days, the periods of the red stars, average 288 days, and of the intense red 477 days.

Espin has arrived at some curious statistics concerning the distribution of variable stars in the heavens, and also concerning their periods. He finds that they are especially numerous in a zone of the heavens inclined 15° or 20° to the equator ; that this zone crosses the preceding side of the Milky Way on the N. side of the equator, and the following side of the Milky Way on the S. side of the equator ; that the northern portion of the zone is not many degrees broad and is clearly marked, but that the southern portion is split into 2 streams of stars and that the place where this occurs is near the place where the Milky Way is also divided ; that hereabouts the variables seem connected with the Milky Way, often occurring in the gaps and constantly on the edges of the gaps, but rarely in the centre of the star sprays from the Milky Way ; whilst the northern stream of variable stars is sharply defined by itself and seems unconnected with the Milky Way. Espin adds that with one or two exceptions all the temporary stars have appeared in the region where the Milky Way and the variable star zone are both broken into two streams and that stars which do not belong to the above-named zone are chiefly the bright and short-period variables. Espin's statistics in detail are too elaborate for embodiment in these pages in their entirety, but some further general conclusions are of sufficient interest and importance to be reproduced. Writing in 1882 he found that the variables then known readily fell into two classes : (I) those with periods of less than 70 days ; and (2) those with periods of more than 135 days ; there being none with periods between 71 and 135 days. Of the former group it might be said that they were in colour white or red in tolerably even numbers, and large in magnitude ; whilst the latter group were mainly red and small in magnitude.

Some other conclusions which he arrived at were that if the variation of light be small in extent, or if the star be bright, the period will probably be short ; on the other hand, where the period ranges from 135 days up to 420 days the number of stars increases with the length of the period ; also, that between a range of 1 magnitude up to 6 magnitudes the number of stars increases with the variation in magnitude. These rules seem, however, to fail where the stars have periods of more than 420 days, or where the range extends beyond 6 magnitudes.

The foregoing statistics are based upon only a minority of the known variables, and therefore cannot yet be put forward as disclosing a series of general laws. Nevertheless, they are sufficiently interesting and pronounced to de-serve attention now, as well as to encourage further inquiry in the future.

The following classification of variable stars has met with some acceptance in America, and therefore it may be given here, but it is open to the objection that it assumes that temporary stars are merely long-period variables. which at present is, at the best, an assumption

(1) Stars showing slow continuous change.

(2) Stars exhibiting irregular fluctuations of light : alternately brightening up and becoming dim without any apparent law.

(3) Temporary stars, which blaze out suddenly and then disappear.

(4) Periodic stars of the type of o Ceti, usually of long period.

(5) Periodic stars of the type of ß Lyrae, of short period.

(6) Periodic stars of the type of Algol, in which the variation of light is such as would result from some intervening body eclipsing the primary star.

It is evident from all that has gone before that variable stars form a very interesting branch of observational astronomy.