CHAPTER XIISPECIAL PROBLEMS IN STELLAR ATMOSPHERES
THE greater part of the present work has dealt with the discussion and interpretation of the normal spectral sequence,to,and the main features of the series have been satisfactorily attributed to thermal ionization at high temperatures. Such a discussion must naturally be the first step in the analysis of the stellar atmosphere. When the more general results of observation have been reduced, in some measure, to an orderly system, it becomes possible to consider special problems involving stars or groups of stars, which lie outside the system, or which, though included in the system, display definite abnormalities.
The special problems of stellar spectroscopy are very numerous. We may mention the novae, the Classstars,stars that show emission lines, the problem of-star classification and the peculiarstars,stars that display both giant and dwarf spectral characteristics, the classification of thestars, the apparent splitting into three groups of the spectral sequence for temperatures below 4000°, the problem of the c-stars, the Cepheid variables with their variable spectra, and the variables of long period.
It is not possible, in a work like the present, to touch upon many of these subjects, and the writer has selected for brief discussion three upon which she can contribute new material: the problem of thestars, the classification of thestars, and the interpretation of the spectra of the stars that display the c-characteristic.
THE STARS OF CLASS
The statistically negligible class containing thestars is placed, at the present stage of investigation, at the top of the stellar sequence. These spectra indicate higher excitation thanthose of any other class, and ionization theory distributes their temperatures between 25,000° and 40,000°. Their spectra are among the most puzzling encountered in the whole stellar sequence, and theory has hitherto been unsuccessful in suggesting the conditions that produce them.
A hundred and forty non-Magellanicstars[439]are enumerated in the Draper catalogue, and in addition a small number of apparently faintstars should probably be transferred to Class,as Victoria has already done[440]for a group of stars in Monoceros. Thestars have a very definite distribution; they lie either very near the galactic plane, or in one of the Magellanic clouds, or they constitute the nuclei of planetary nebulae.
Thestars other than the nuclei of planetaries have high intrinsic luminosities, but the material is insufficient for a satisfactory estimate of the absolute magnitudes of the non-Magellanicstars; various indications point to a value at least as high as -4. For the Magellanicstars, absolute magnitudes as great as -6.7 have been derived.[441]The measured parallaxes of the planetary nebulae, however, give for the nuclei absolute magnitudes[442]in the neighborhood of +8. The wide difference in absolute magnitude can merely be pointed out; it has never received adequate explanation.
The masses are presumably very high for thestars, though but few have been accurately measured. The star B.D. 6°1309, a spectroscopic binary reported by J. S. Plaskett,[443]has a minimum mass eighty times that of the sun, and the stars 29 Canis Majoris andOrionis also appear to be very massive.[444]
The spectra of the stars of Classdiffer widely among themselves, but they are signalized by the lines of ionized helium, which are normally observed only in this class and in the nebulae. In addition, the atoms of H, He, Mg+, C++, 0++, N++, and Si+++ are represented.The atmospheres of these stars are thus in a state of high ionization, which is attributed to high temperature, in harmony with the work already outlined in previous chapters. The spectra of the stars of Classhave been described by W. W. Campbell,[445]Miss Cannon,[446]Wright,[447]H. H. Plaskett,[448]J. S. Plaskett,[449]and the writer,[450]and the material upon which the present discussion is based will be found in the papers quoted.
Many of thestars, such asCanis Majoris, H.D. 150135, H.D. 159176, H.D. 199579, H.D. 164794, H.D. 167771, H.D. 165052, give a pure absorption spectrum, containing the Balmer series of hydrogen, the lines of Si+++, the Pickering and “4686” lines of He+, and the N++ line at 4097. The stars are mentioned in order of increasing ionization, with He+ rising in intensity. The other lines mentioned are clearly beyond their maximum, and fall progressively in strength. The stars mentioned probably represent successive steps in a sequence with rising temperature, connecting directly with Class,and ranging from 25,000° to 35,000°.
This sequence ofstars would form a simple and explicable series if it were an isolated group. There are, however, other stars, with spectra so similar to those of the series just quoted that there can be no doubt of a close relationship—they display absorption lines due to the same elements with about the same relative intensities—but emission lines tend to occur in various parts of the spectrum. 29 Canis Majoris,Muscae, andPuppis have absorption spectra which resemble those in the sequence just quoted, but at 4650 and 4686 there are emission lines or “bands.” The bright lines are so wide and diffuse, inMuscae, as to be blended together at their edges, while they are sharp and clear in the other two stars. Between the two series just quoted—pure absorption stars and absorption starswith some bright lines, comesCircini, a star which has all the characteristics of the first group, and also shows faint emission on the red edge of the absorption lines at 4650 and 4686.
There are other stars, such asCephei,Persei, S Monocerotis, H.D. 152408, H.D. 112244, andOrionis, that have absorption spectra such as were described for the first group, but which display faint emission lines at the red edge of the hydrogen and helium lines. It is obvious that all the stars so far enumerated may legitimately be classed together, but that there is a very universal tendency for emission lines and bands to appear in them. This tendency is so marked in the stars that are still to be mentioned as to constitute their most salient feature.
In the subgroup of thestars which are collectively designated the Wolf-Rayet stars, the emission lines are the most conspicuous characteristic of the spectrum. The best known and brightest star of this group isVelorum, which possesses an extremely complex spectrum, made up of an absorption spectrum similar to that ofCircini, and a large number of wide “emission bands.” An analysis of the spectrum ofVelorum has been published by the writer;[451]all the stronger lines of H, He, He+, C+H-, O++, N++, Si+++, and Mg+ are represented in the spectrum, and a comparatively small number of lines remains unidentified.
Otherstars that have spectra in which the emission lines are the prominent feature are H.D. 151932, H.D. 92740, H.D. 93131, H.D. 152270, H.D. 156385, and H.D. 97152. All of these stars, excepting the last, have also absorption spectra displaying the lines of H and He+. The lines of N++ and Si+++ are absent, and these stars are therefore probably at the extreme high-temperature end of the sequence.
The question of absorption in the Wolf-Rayet spectrum is a difficult one, because the bright lines show up before any other feature of the spectrum, while an appreciable continuous background is necessary before absorption can be detected. The detection of absorption linesin many stars, such as H.D. 152270, where no absorption had previously been recorded, has resulted from a general survey of spectra that had received exposures sufficient to bring out the continuous background. The writer has been led to the opinion that absorption is a common, if not universal, feature of all the Wolf-Rayet stars, except those classed at Harvard as.This subclass has bright bands that do not coincide with those of the otherstars, and among them absorption lines appear to be exceptional.
It is perhaps to be expected that absorption should normally occur among the Wolf-Rayet stars, as it does among the other classes. In all other stars, the bright lines that appear are the abnormal feature, and are superposed on a normal continuous spectrum crossed by absorption lines. Spectra consisting of bright linesonlydo not occur elsewhere, excepting for the gaseous nebulae. The gaseous nebulae have, presumably, no photosphere, and the continuous background that they sometimes display is probably the result of reflected and transformed starlight; absorption lines appear normally to accompany the existence of a photosphere.
It is clear that thestars are a very complex group. Those that have pure absorption spectra can be arranged in a series immediately preceding;and those that show a similar absorption spectrum, with faint superposed emission, presumably also fit into the sequence. When astar shows emission lines (asCassiopeiae does) it is placed in theclass appropriate to its absorption spectrum, with the additional designation “e” to indicate the abnormality, and the same procedure appears to be equally satisfactory for thestars.
As emission predominates more and more, the spectrum resembles those of the normal members of the sequence less and less. If a star has an absorption spectrum it can always be assigned a place in the sequence, and this method of arrangement appears to be logical. But it is clear that the sequence so formed is no longer physically homogeneous. The stars that have no absorption lines, although some of them have obvious affinities with stars that have absorption spectra, havemoreover no place in a sequence formed on the basis of absorption intensities.
It is, of course, possible to devise a self-consistent scheme for the arrangement of a limited number of thestars, and such a scheme is, for many purposes, both desirable and convenient. It is, however, exceedingly hard to know where the division should be drawn between “absorption” and “emission” stars. Perhaps the most satisfactory plan is to treat allstars as asequence, with special comment for the large number of them that require it.
THE CLASSSTARS
(a)The Balmer Lines.—The spectra of thestars are dominated by the Balmer series of hydrogen, which, with the exception of theandlines in the cooler stars, are stronger atthan any other line seen in the stellar sequence. The maximum of the Balmer series has been stated[452]to occur at,and this value was used by Fowler and Milne[453]in calibrating their temperature scale based upon ionization theory. It is in accordance with theory that the subordinate lines of hydrogen, with ionization potential 13.54 volts, and excitation potential 10.15 volts, should have their maximum at about 10,000°.
The position of the maximum can be placed elsewhere by the use of special stars in estimating the line-intensities. The intensity of the hydrogen lines is in fact unusually difficult to determine, as they differ from star to star in width, wings, and probably also in central intensity. Using a series of individual stars, Menzel[454]placed the maximum at,with the note that “on the average the lines seem to be strongest in Classesto,but the mean intensity is often greatly exceeded in certain,and evenstars.” A general study of the Classspectra confirms the statement that the mean intensity at maximum is often exceeded for individual stars in other classes, and the writer is inclined to be of the opinion that no significant maximum can be derived from alimited number of estimates. The maximum given in the Henry Draper Catalogue is the product of the examination of an enormous number of very short dispersion plates, and is entitled to a greater weight than any other. In the estimation of such strong lines, the width and especially the wings are likely to affect the estimates extensively, and short dispersion plates probably reduce the difficulty, and permit of the greatest possible accuracy. It must be emphasized that the maximum given in the Henry Draper Catalogue cannot be superseded by measures made on an arbitrary selection of stars, such as is used when stars bright enough to be photographed with (say) two objective prisms are discussed, for it is a generalization from the most complete data hitherto examined, or to be examined for some time to come. The non-homogeneity of theclasses, presently to be discussed, includes wide variations in the widths of the hydrogen lines, and renders unnecessary any attempt to correct the hydrogen maximum at,which appears to be of a statistical nature.
(b)Metallic Lines and Band Absorption.—The metallic lines, which become so conspicuous in intensity in the later classes, appear in the types immediately succeeding,and increase progressively in strength as cooler classes are approached. In general, all the related lines belonging to any one element appear at the same class, although sometimes the fainter components of metallic multiplets are not seen until the stronger components have attained a considerable intensity. For example the weaker lines of an element that is seen atmay not appear till.The disparity in intensity between the components of a multiplet is usually not so great at appearance as at maximum. The relative strengths of unblended lines conform at maximum with the laboratory intensities, to an extent that raises questions as to the degree of saturation[455]of the more intense components. It seems that none of the metallic lines, excepting those of calcium, are greatly oversaturated, even at maximum, to judge from the relative intensities of related lines at that point.
It is within thestars that the first signs of band absorption appear. Theband is seen in somestars, and a drop in the continuous spectrum of Vega around 4160 has been ascribed[456]to the cyanogen band headed at 4215. Similar “band” absorption can be traced in other stars of Class,and is even seen as early as.Identification of the cyanogen band headed at 3885, which always accompanies the 4125 band, would confirm the attribution to cyanogen, but the violet band does not seem to have been observed. The wings of H,which are often wide and conspicuous, render it difficult to trace anything of the nature of band absorption near 3885 for anstar.
(c)The Classification ofStars.—Several lines of evidence have indicated that the classes into which thestars have been divided are not physically homogeneous, and the problem of their classification is one of the future tasks of astrophysics. It is hoped that the writer can in the future make a more complete discussion of the question than is here desirable, and therefore the present treatment is to be considered merely suggestive and tentative. The material quoted is slight, and must be increased before conclusions can be justified.
It has been suggested[457]that a one-dimensional arrangement will not suffice for the classification of thestars. The spectra have been classified, at least for the hottertypes, by the strength of theandlines of Ca+. With the dispersion used in making the Henry Draper Catalogue, these lines constitute the conspicuous difference between one spectrum and another, and are the obvious criterion of class. If the spectra are classed by the strength of these lines alone, the classification is of course quite unambiguous, and for a one-dimensional sequence of spectra it would have been ideal. That the classes so formed arenothomogeneous[458]indicates that some second variable must be described in a satisfactory classification, and that the strength of no one line could have been used with any greater success than that of theanddoublet. Further, practical difficulty in duplicating the classifications has been caused by the fact that theandlines are so far into the violet that they do not appear at all on many slit spectra, such as those used at Mount Wilson, and when other criteria are chosen for classification, it is likely that the results will deviate somewhat from those of the Henry Draper Catalogue.
By analogy with what is observed in other types, it has been suggested that the range in line-sharpness that is found within a given class among thestars is an effect of absolute magnitude, and the sharpness of the hydrogen lines has indeed been used at Mount Wilson[459]as a quantitative measure of luminosity. From an analysis of the widths of hydrogen lines made by Miss Fairfield,[460]it appears that the line sharpness may be used to distinguishstars of the highest luminosity from those of the lowest, but that it cannot be used for the accurate estimation of absolute magnitude between those limits.
The special problem of classifying thestars is only in its initial stage. That the present system is inadequate is certain, but as yet no satisfactory alternative has been proposed. The direction in which work should be pursued is, in this instance, probably the study of the differences between individual spectra. As the problem appears to hinge on the presence of abnormalities within a given class, it is of especial importance to examine the frequency, magnitude, and nature of these abnormalities.
(d)Silicon and Strontium Stars.—There are among thestars two small groups of especial interest—the so-called “silicon” and “strontium” stars. These occur chiefly in,,,and are distinguished by the unusual intensity of the lines 4128 and 4131 of ionized silicon, and the lines 4077 and 4215 of ionized strontium. Such stars are regarded in the Henry Draper Catalogue as definitely abnormal, and are individually mentioned in the Remarks. The strontium stars in classes later thanareapparently ordinary high-luminosity stars, and the line-intensity is involved in the well known absolute magnitude effect.
The absolute magnitudes of the strontium stars have been supposed, on general grounds, to be very high, but an examination of the proper motions indicates that this is perhaps not even generally true. The well knownstarCircini is apparently a dwarf,[461]andEquulei has similar spectral peculiarities and proper motion. Examples might be multiplied, but there is not enough material at present available for a full discussion, and from what has already been said it is evident that the strontium stars constitute no ordinary absolute magnitude problem, although the condition that produces strong strontium lines in some dwarf stars may be something, like low surface gravity, that also prevails in stars of high luminosity.
There are too few parallaxes, proper motions and radial velocities for significant statistical treatment of the silicon stars, and still less material for the strontium stars; but the galactic distributions of both classes indicate that their absolute magnitudes are at least not extremely high. There does not at present appear to be sufficient justification for the statement that these stars are “distinctly brighter than the average.”[462]Their brightness would rather seem to be about the same as that of a normalstar.[463]
The silicon and strontium stars raise spectroscopic difficulties that differ somewhat in the two cases. Most of the silicon stars occur at or near,where the Si+ lines are normally at maximum intensity. On the other hand, Sr+ has its maximum at Classor,but the intensity in suchstars asOphiuchi andMicroscopii is as great as it is in these types. The strontium problem illustrates the general conceptions underlying the methods of estimating line-intensities, and will therefore be discussed in slightly more detail.
Abnormal intensity of a spectrum line can be attributed to (1) blending, (2) unusual conditions, or (3) abnormal abundance. These conditions will be discussed in order.
(1)Blending.—Blending, excepting where lines are spectroscopically resolved, can only be detected indirectly, by examining the behavior of other lines belonging to the same spectral series as the line in question. If the relative intensities of all the lines in the series are the same as those found in the laboratory, and if changes of intensity from class to class affect all the lines of a series equally, it may be inferred that blending is not a serious disturbing factor and that the abnormal intensity is due to other causes. The close correlation between the stellar intensities of 4215 and 4077, the components of the principal doublet of ionized strontium, in the different spectral classes, leaves little doubt that these lines are effectively unblended in thestars, although the difference of spectral class for the maximum of the two lines (,)and the presence of a solar iron line at 4215, suggest a blend for the latter in stars cooler than about.It is also to be remarked that the head of a “cyanogen” band falls at 4216.
(2)Abnormal Conditions.—Abnormal conditions permit of no direct observational test, but it would be anticipated that they would also affect other lines to a degree greater than is observed. The change of temperature that would be required to raise the Sr+ lines to their maximum strength at(10,000°) would be a fall of about 5000°, which is quite inadmissible, for the resulting change of spectrum would produce astar. The required change of pressure is also too great to be possible: this subject cannot profitably be discussed here, and reference should be made toChapter X. The existence of a strontium cloud has been suggested[464]by analogy with the “calcium cloud,” and might possibly provide an explanation, as it would furnish a low temperature for the strontium without unduly lowering the temperature for the star in general. The observation of stationary strontium lines would materially strengthen this argument,but they have not so far been recorded. The fact that the strontium stars are scattered, and not concentrated in any one part of the sky, reduces the probability of this suggestion.
(3)Abnormal Abundance.—Abnormal abundance has been progressively abandoned as an explanation of the various phenomena of stellar spectra, and that it is the true interpretation of strontium peculiarities seems somewhat unlikely. For the silicon stars, unusual abundance is probably an untenable hypothesis, since the great strength of the Si+ lines is apparently not accompanied by increase in the silicon line, which should presumably occur if pure abundance is the cause of the increased strength of the ionized silicon lines. Abnormal strength of silicon in the cooler stars, doubly ionized silicon in the earlystars, or triply ionized silicon in thestars, has not been observed, and it is not very probable that, if silicon is unevenly distributed in the universe, the irregularity would be revealed in stars at one temperature only.
Such considerations point to the problem of the silicon and strontium stars as one involving the atom and its energy supply, rather than an abnormal distribution of the element in question. It is likely that the problem of classifying thestars will be elucidated by a more detailed study of the silicon and strontium lines. The behavior of strontium appears, however, in some cases, to warrant the description of “abnormal,” and it may be that the first step in thestar problem will be the elimination from the general classification of spectra such as those of the strontium stars. The present writer inclines to the belief that the silicon and strontium stars will be included in the normalstar classification, when such a one is satisfactorily devised.
(e)Peculiar ClassStars.—Among thestars there are three which appear to be of special interest.
The starAndromedae, designatedin the Draper Catalogue, has been shown to display enhanced lines of manganese, broadly winged, and of unusual strength.[465]
The starCanum Venaticorum has been the subject of extensive work.[466][467]The chief point of interest concerning it is the occurrence of lines ascribed to the rare earths.[468]The spectra of these elements are so rich in lines that spurious coincidences are certain to occur, but comparisons with the spectrum ofCygni and of the chromosphere suggest that the strongest lines of europium and terbium are indeed represented. From general ionization principles it would appear that enhanced spectra are probably involved, but until series relations are known it is not possible to discuss the subject further.
The super-giant, or c-star,Cygni, Class,has probably greater possibilities for the stellar spectroscopist than any other star, as its spectrum is peculiarly rich in fine sharp lines, many of which are unidentified,Cygni is representative of a large class of stars, but it is the only one of them that has an apparent magnitude bright enough to render it readily accessible. The spectrum has been tabulated by Lockyer[469]and by Wright.[470]At the temperatures concerned, the doubly enhanced lines of the metals are to be anticipated, and it is probable that many of the faint unidentified lines in the spectrum of this star are those of twice ionized metallic atoms. The strongest doubly enhanced lines of the metals fall, as is well known, in the ultra-violet,Cygni contains the lines of theseries of neutral helium, the most persistent lines of the element, and this is significant in view of the extremely low pressure that is assigned to the atmosphere of the star on the basis of absolute magnitude.
THEC-STARS
The c-characteristic was first used by Miss Maury[471]to designate stars, found in several spectral classes, that have marked spectral peculiarities. The intensities of some of the metallic lines, chieflythose of ionized atoms, are greatly strengthened for the class, and other lines, chiefly those of the neutral atom, are weakened. Theband becomes markedly discontinuous, and heavy blends at 4072, 4077, become conspicuous. The spectrum of a c-star is unmistakable in appearance.
The foregoing tabulation contains a list of the lines that are strongly enhanced in the spectra of the c-stars. Successive columns give the approximate wave-length, taken from Miss Maury’s original list, thelaboratory wave-length of the line with which the stellar line is identified, the atom, and the series relations.
The preponderating characteristic lines are clearly those of ionized iron and ionized titanium.[472]All the strong lines of these atoms are found in the c-star spectrum, and they are there stronger than in any other class. It is found that spectra possessing the c-character have in general unusually sharp and narrow lines. It is probable that the lines in the spectrum of a c-star are actually stronger, as well as sharper, than the corresponding lines in a star of the same class and lower luminosity.[473]