the globular star-clusters—structure of these objects—variability of stars in the cluster—telescopic resemblance of a cluster to a nebula—resolution of a nebula—supposition that all nebul? may be clusters—a criterion for distinguishing a nebula and a cluster—dark lines on a bright background characterise the structure of a star—bright lines on a dark background characterise the structure of a nebula—characteristics of the spectrum of a true nebula and of a resolvable nebula—spectra of the sun and capella—spectra of the nebula in orion and of a white star compared—number of lines in a nebular spectrum—criterion of a nebular spectrum—spiral nebula not gaseous—solar spectra during an eclipse—bearing on the nebular theory—herschel’s work—the objection to the theory—the objection removed in 1864.
there is perhaps hardly any telescopic object more pleasing or more instructive than a globular cluster of stars when viewed through an instrument sufficiently powerful to do justice to the spectacle. there are several star-clusters of the class designated as “globular.” the most famous of these, or, at all events, the one best known to northern astronomers, is found in the constellation of hercules, and is for most purposes sufficiently described by the expression, “the cluster in hercules.” the genuine lover of 53nature finds it hard to withhold an exclamation of wonder and admiration when for the first time, or even for the hundredth time, the cluster in hercules is adequately displayed in the field of a first-class telescope.
fig. 9.—the cluster in hercules.
(photographed by dr. w. e. wilson, f.r.s.)
54in fig. 9 is a photograph of this celebrated object, which was taken by dr. w. e. wilson, f.r.s., at his observatory at daramona, in ireland. the picture has been obtained from an enlargement of the original photograph taken with the telescope in mr. wilson’s observatory. it is, however, precisely as nature has given it, except for this enlargement. you will note that towards the margin of the cluster the several stars are seen separately, and in many cases with admirable distinctness. we do, however, occasionally find two or more stars so close together that their images overlap; and, indeed, in the centre of the cluster the stars are so close together that it is impossible to differentiate them, so as to see them as individual points of light. we need have no doubt, however, that the cluster is mainly composed of separate stars, although the difficulties interposed by our atmosphere, added to the necessary imperfections of our appliances, make it impossible for us to discriminate the individual stars.
in looking at a star group of this particular kind the observer may perhaps be reminded of a swarm of bees in flight from the hive, for the stars in the cluster are, on a vast scale, apparently associated in the same way as the bees, on a small scale, are associated in the swarm. we may also compare the stars in the cluster to the bees in the swarm in another respect. each bee in the swarm is in incessant movement. there can be no doubt that each star in a globular cluster is unceasingly changing its position with reference to the others. the distance by which the cluster is separated from the earth renders it impossible for us to see those movements, at all events 55within those narrow limits of time over which our observations have as yet extended; but the laws of mechanics assure us that the mutual attraction of the stars in this cluster must give rise to incessant movements, and that this must be the case notwithstanding the fact that the relative places of the stars in the cluster show no alteration that can be recognised from one year’s end to another.
i may, however, mention that though there may be no movements in these stars great enough to be observed, yet the brightness of some of them shows most remarkable fluctuations. the investigations of professor bailey and other astronomers have, indeed, disclosed such curious variability in the brightness of some of these stars that if it were not for the exceedingly high authority by which this phenomenon has been guaranteed we should, perhaps, almost hesitate to believe so startling a fact. it has, however, been most certainly proved that many of the stars in certain globular clusters pass through a series of periodical changes of lustre. the period is a very short one as compared with the periods of better known variable stars, for in this case twenty-four hours are more than sufficient for a complete cycle of changes, and it not infrequently happens that in the course of a single quarter of an hour a star will lose or gain brightness to the extent of a whole magnitude. the phenomenon referred to is at the present moment engaging the careful attention of astronomers; but it offers a problem of which, indeed, it is not at present easy to see the solution.
our immediate concern, however, with the globular star-clusters relates to a point hardly of such refinement 56as that to which i have just referred; it is one of a much more elementary nature. the photograph in the figure may be considered to represent the cluster in hercules as it would be seen with a telescope of very considerable visual power, for the object would assume a different appearance in a telescope which was not first class. the perfection of a really powerful instrument is tested by its capability of exhibiting as two separate points a pair of stars which are excessively close together, and which in an instrument of inferior power cannot be distinguished, but seem fused into a single object. the defining power of a telescope—that is to say, its capability for separating close double stars—is increased with the size of the instrument, always granting, of course, that there is equal optical perfection in both cases. it follows that the more powerful the telescope the more numerous are the stars which can be seen separately in a globular cluster.
if, however, a small telescope be used, or a telescope which, though of considerable size, has not the high optical perfection that is demanded in the best modern instruments, then adjacent stars are not always to be seen separately. it may be that the telescope, on account of its small size, cannot separate the objects sufficiently, or it may be that the imperfections of the telescope do not present the star as a point of light, but rather as a more or less diffused, luminous disc. in either case it may happen that a star overlaps other stars in its immediate neighbourhood, and consequently an object which is really a cluster of separate stars may fail altogether to present the appearance of a cluster.
57i have been alluding to something which, as every astronomer knows, is of practical importance in the observatory. like every one else who has ever used a telescope, i have myself seen the cluster of hercules with just the same misty appearance in a small telescope that an undoubted nebula possesses in the very finest instrument. it is, accordingly, sometimes impossible, merely by observation with a small instrument, to distinguish between what is certainly a cluster of stars and what is certainly a nebula. it has indeed not infrequently happened that an observer with a small telescope has discovered what appeared to him to be a nebula, and he has recorded it as such; and yet when the same object was subsequently examined with an instrument of greater defining power the nebulous character has been seen to have been wrongly attributed. the object in such a case is proved to be nothing more than a cluster of stars, of which the individual members are either intrinsically faint or exceedingly remote; it certainly is not a mass of that fire-mist or gaseous material which alone is entitled to be called a nebula.
it is therefore a question of importance in practical astronomy to decide whether objects which appear to be nebul? are really entitled to the name, or whether the nebulous appearance may not be an optical illusion. the operation by which an object previously deemed to be a nebula is shown by the application of increased telescopic power to be a cluster of stars is commonly known as the resolution of a nebula. about fifty years ago the mighty six-foot reflecting telescope of lord rosse, and other 58great instruments, were largely employed on this work. it was, indeed, at that time held to be one of the special tasks which came most legitimately within the province of the big telescopes, to show that the so-called nebul? of earlier observers were resolvable into star-clusters under the superior powers now brought to bear upon them.
the success with which this process was applied to many reputed nebul?, which were thereby shown to be not entitled to the name, led not unnaturally to a certain conjecture. it was admitted that certain objects which had successfully resisted the resolving powers of inferior instruments were forced to confess themselves as mere star-clusters when greatly increased telescopic power was brought to bear on them; and it was conjectured that similar success would attend the attempts to resolve still other nebul?. it was even supposed that every object described as a nebula could only be entitled to bear that designation provisionally, only indeed until some telescope of sufficient power should have been brought to bear on it. it seemed not unreasonable to surmise that every one of the so-called nebul? is a cluster of stars, even though a telescope sufficiently powerful to effect its resolution might never be actually forthcoming.
i do not, indeed, believe that this opinion as to the ultimate resolvability of all nebul? could have been shared by those who had much practical experience in the actual observation of these objects with the great telescopes, for the particular classes of nebul? which in telescopes of superior powers resolved themselves into groups of stars had a characteristic appearance. after 59a little experience the observer soon learned to recognise those nebul? which promised to be resolvable. the object might not indeed be resolvable with the powers at his disposal, but yet from its appearance he often felt that the nebula would be probably resolved if ever the time should come that greater powers were applied to the task.
it is easy to illustrate the question at issue by the help of the photograph of the cluster in hercules in fig. 9. each of the stars is there distinct, except where they are much crowded in the centre. if, however, the photograph be examined through one of those large lenses which are often used for the purpose, and if the lens be held very much out of focus, the stars will not be distinguishable separately, and the whole object will be merely a haze of light. this illustration may help to explain how the different optical conditions under which an object is looked at may exhibit, at one time as a diffused nebula, an object which in better circumstances is seen to be a star-cluster.
the astronomer who was fortunate enough to have the use of a really great telescope would not fail to notice that, in addition to the so-called nebul? already referred to, which were presumably resolvable, there were certain other objects, generally characterised by a bluish hue, which in no circumstances whatever presented the appearance of being composed of separate stars. we now know for certain that these bluish objects are not clusters of stars, but that they are in the strictest sense entitled to the name of nebul?, and that they are gaseous masses or mists of fire-cloud. the full demonstration of this important point was not effected until 1864.
60the fact that so very many of the nebul? were resolved led not unreasonably to the presumption that all the nebul? would in due time also yield. but there were many who could not accept this view, and there was a long discussion on the subject. at last, however, the improvements in astronomical methods have cleared up the question. sir w. huggins has shown that there are two totally distinct classes of nebul?, or rather of so-called nebul?. there are certain nebul? which can be resolved, and there are certain nebul? which cannot. a nebula which can be resolved would be a veritable cluster of stars, and is not really entitled to the name of nebula; a nebula which cannot be resolved would be entitled to the name, for it is a volume of gas or of gaseous material which is itself incandescent. we have been provided with a beautiful criterion by which we can decide to which of these classes any nebulous-looking object belongs.
the spectroscope is the instrument which discriminates the two different classes of objects. this remarkable apparatus, to which we owe so much in every department of astronomy, receives the beam of light from the celestial body. the instrument then analyses the light into its component rays, and conducts each one of those rays separately to a distinct place on the photographic plate. when the photograph is developed we find on the various parts of the plate the evidence as to the class of rays which have entered into the composition of the light that has been submitted to this very searching form of examination.
the light which comes from a star or any star-like body, including the sun itself, may first be described. that light, after passing through the spectroscope and 61having been conducted to the photographic plate, will produce a picture of dark lines on a bright background; this is, at least, the spectrum which a star generally presents. there are, indeed, many types of stellar spectra, for there are many different kinds of stars, and each kind of star is conveniently characterised by the particular spectrum that it yields. if the star be one of small magnitude, then the lines in its spectrum may be detected, but only with great difficulty. it not infrequently happens that the photograph of the spectrum of such a star will show no more than a continuous band of light without recognisable lines; and this is what occurs in the case of a resolvable nebula, where the stars are so closely associated that the spectrum of each separate star cannot be distinguished. the spectrum of a resolvable nebula is merely a streak of light, which is the joint effect of all the spectra. the spectrum is then too faint to show the rainbow hues which present such beautiful features in the spectrum of a bright star, as they do in the spectrum of the sun itself.
i give, in the adjoining figure (fig. 10), portions of the photographs of two spectra of celestial objects. they have been taken from the atlas of representative stellar spectra in which sir william and lady huggins have recorded the results of their great labours. two spectra are represented in this picture, the uppermost being the spectrum of the sun, while the lower and broader one is the spectrum of the bright star capella. it has not been possible within the limits of this picture to include the whole length of these two spectra, and it must therefore be understood that the photographs given in the atlas are each about five times as long as the parts which are here reproduced.
62
fig. 10.—sun and capella.
sun above. capella below.
(sir william and lady huggins.)
but the characteristic portions of the spectra selected are sufficient for our present argument. it will be noted, first of all, that there is a singular resemblance between the details of the spectrum of the sun and those of the spectrum of the star. no doubt the breadth of the stellar picture in the lower line is greater than that of the solar picture in the upper line; but this point is not significant. the breadth of the spectrum of the sun could easily have been made as wide or wider if necessary. the breadth is immaterial, for the character of a spectrum is determined not by its breadth, but by those lines which cross it transversely. it will be seen that there are here a multitude of lines, some being very dark, and some so faint as to be hardly visible. both spectra exhibit every variety of line, between the delicate marks which can barely be seen and the two bold columns on the right-hand side of the picture.
the characteristic of the spectrum is given by the number, the arrangement, the breadth, the darkness, and the definiteness of the lines by which it is crossed, and the first point that we note is the remarkable resemblance in these different respects between the two 63spectra. the lines are practically identical, at least so far as those parts of the spectrum represented in this picture are concerned. we have thus a striking illustration of the important fact, to which we have so often to make allusion, of the general resemblance of the sun to the stars. not only do we know that if the sun were removed about a million times as far as it is at present its light would be reduced to that of a star, but that the star capella transmits to us light consisting essentially of the same waves as those which enter into a beam of sunlight. no more striking illustration of the analogy between the sun and a star can be found than that which is given in this photograph from the famous observatory at tulse hill.
but it must not be inferred that because the spectra of sun and star are like each other, they are therefore absolutely identical. there are many lines and details to be seen on the actual photographic plate which are too delicate to be reproduced in such copies as it is possible to make. when a close comparison is made on the actual plate itself of the lines in the solar spectrum and the lines in the spectrum of capella, it is observed that, though they are the same so far as the more important lines are concerned, yet that there are many lines found in the spectrum of capella which are not found in the spectrum of the sun.
fig. 11.—spectrum of nebula in orion and
spectrum of a white star.
(sir william huggins, k.c.b.)
the contrast between the spectrum of a nebula properly so called and the spectrum of a star is well illustrated by the accompanying picture (fig. 11), in which sir w. huggins exhibits the photograph of the spectrum of the nebula in orion in comparison with the spectrum of a star. the uppermost of the two is the spectrum of the star. it will be noted that this 64spectrum is very different from that which we have already seen in capella. instead of a vast multitude of lines resembling the lines of the solar spectrum, the spectrum of a star of the type here represented, of which we may take sirius as the most striking example, exhibits but a few lines. we regard them as one system of lines, for we know they are physically connected. they are all alike due to the presence of a single element in the star, that element being in fact hydrogen. but though the spectra of capella and sirius are so totally different, the differences relate only to the distribution of the lines, and to their number, darkness, and width. in both cases we observe the characteristic of the light from an ordinary bright star, namely, that the spectrum is composed of a bright band with dark lines across it. it ought, perhaps, to be mentioned here that there are certain very special stars which do exhibit some bright lines in addition to a more ordinary spectrum; this is especially the case in the new stars which occasionally appear. thus in the case of the new star which appeared in perseus, in 1901, there were several remarkable bright lines. this most interesting object will be referred to again in a later chapter.
65widely different from the spectrum of any star whatever is the lower of the two spectra which are shown in the figure. this lower spectrum is that of the great nebula in orion. at once we see the fundamental characteristic of a nebula; its spectrum exhibits five bright lines on a dark field. i do not say that the great nebula in orion has not more than five lines; there are indeed many others, for sir william huggins has himself pointed out a considerable number, and the labours of other observers have added still more; but the five lines here set down are the principal lines. they are those most easily seen; the others are generally extremely delicate objects arranged in groups of five or six. but the lines which this picture shows are quite sufficient to exhibit that fundamental characteristic of the nebular spectrum, namely, a system of bright lines on a dark field. i may further mention that certain lines in the spectrum indicate the presence of the element hydrogen in the great nebula in orion, and we owe to dr. copeland the interesting discovery that the remarkable element helium is also proved to exist in the nebula.
the pictures, at which we have been looking, will suffice to make clear the criterion, which astronomers now possess, for deciding whether an object which looks nebulous is really a gaseous nebula, or ought rather to be regarded as a star-cluster. if the object be a star-cluster, then the spectrum that it gives will be the resultant of the spectra of the stars, and this will be a continuous band of light. if the stars are bright enough, it may be that dark lines can be detected crossing the spectra, but in the case of the clusters it will be more usual to find the continuous 66band of light so faint that the dark lines, even if they are there, are not distinguishable.
if, on the other hand, the object at which we are looking, not being a cluster of stars, is indeed a mass of glowing gas, or true nebula, then the spectrum that it sends us is not the continuous spectrum such as we expect from the stars. the spectrum which the nebula proper transmits to the plate is said to be discontinuous. in some cases it is characterised by only a single bright line, and in others there may be two, or three, or four bright lines, or, as in the case shown in fig. 11, the number of bright lines may be as many as five. it may indeed happen, in the case of some exquisite photographs, that the number of lines in the spectrum of the nebula will be increased to a score or possibly more. there may also be faint traces of a continuous spectrum present, this being due to the stars scattered through the object, from which perhaps even the most gaseous nebula is not entirely free. but the characteristic type of nebular spectrum is that in which the bright lines, be they one, or few, or many, are separated by intervals of perfect darkness. when it is found that the spectrum of a nebula can be thus described, it is correct to say that the nebula is truly a gaseous object.
in the lists given by scheiner in his interesting book, “astronomical photography,” the number of gaseous nebul? is set down as seventy-three. of course no one pretends that this enumeration is exhaustive. it claims to be no more than a statement of the number of nebul? which have been proved, by observations made up to the present, to be of a gaseous description. seeing that there are, as we have already 67stated, many scores of thousands of nebulous-looking objects, it is probable that the number above given is not more than a small fraction of the number of gaseous nebul? actually within reach of our instruments.
it may, however, be assumed that more than half the objects which are called nebul? are not of the gaseous type. this is a point of some importance, which appears to follow from the facts stated by professor keeler in connection with his memorable researches with the crossley reflector. in a later chapter we discuss important questions connected with what are called spiral nebul?. we may, however, here record that no spiral nebul? have as yet been pronounced gaseous. professor keeler assures us that, of the one hundred and twenty thousand nebul? which he estimates to be within reach of the crossley reflector, far more than half are of the spiral character. if, then, we assume that the spectra of spiral nebul? are always continuous, it seems to follow that less than half the nebulous contents of the heavens possesses the discontinuous spectrum which is characteristic of a gaseous object.
we are not entitled to assume that a nebula, or reputed nebula, which shows a continuous spectrum, must necessarily be a cluster, not merely of star-like bodies, but of bodies with masses comparable with those of the ordinary stars. our argument does most certainly suggest that the body which yields a continuous spectrum is not a gaseous body; but it may be going too far to assert that therefore it is a cluster of stars in the ordinary sense. we do often find true nebul? and star-clusters in close association. the nebula in the pleiades (fig. 13) is an example.
68it may be desirable to add a few words here as to the physical difference between a continuous spectrum and a discontinuous spectrum. the light from a body, known to be gaseous, shows through the prism the discontinuous spectrum of bright lines upon a dark background. if, on the other hand, a solid be raised to incandescence, such, for instance, as a platinum wire heated white-hot by an electric current, or a cylinder of lime submitted to an oxyhydrogen blowpipe, then the spectrum that it yields is continuous. all the colours of the rainbow, red, orange, yellow, green, blue, indigo, violet, are shown in such a spectrum as a continuous band of light, though the band is not crossed by dark lines. it would therefore appear that the continuous spectrum is characteristic of an incandescent solid, and the discontinuous spectrum of a glowing gas. but here it may be urged that the sun presents a difficulty. we so often refer to the spectrum of the sun as continuous, that it might at first appear as if the spectrum of the sun resembled that produced by radiation from a solid body. but, as is well known, the sun is not a solid body. even if the sun be solid at the centre, it is certainly far from being solid in those superficial regions called the photosphere, from which alone its copious radiation is emitted. if the sun is not a solid body, how comes it to emit a radiation characterised in the same way as the radiation from a white-hot solid? why does the solar spectrum not exhibit features characteristic of radiation from an incandescent gas? the point is well worthy of attention; it finds an explanation in the nature of the photosphere from which the sun’s radiation proceeds.
69the photosphere, though not, of course, to be described as a solid body, does not most certainly, so far as its radiation is concerned, behave like a gaseous body. in the glowing clouds of the photosphere the carbon, of which they are composed, is not in the gaseous form; it has passed into solid particles, and it is these particles, in the highest condition of incandescence, which emit the solar radiation. although these particles are sustained by the gases of the sun, and are associated in aggregations which form the dazzling clouds of the photosphere, yet each one of them, in so far as its individual radiation is concerned, ought to be regarded as a solid body. the radiation from the sun is, therefore, essentially not the radiation from an incandescent gas; it is the radiation from a glowing solid. this is the reason why the solar spectrum is of the continuous type.
fig. 12.—solar spectra with bright lines and dark lines during eclipse.
(photographed by captain hills, r.e.)
by the kindness of captain hills, r.e., i am able to show a photograph (fig. 12) containing two spectra taken during a recent eclipse, which will serve as an excellent illustration of the different points which we 70have been discussing. it is, indeed, true that neither of the spectra, here referred to, belongs to nebul?, whether genuine gaseous objects or not. both of the spectra in captain hills’ picture are actually taken from the sun. the conditions under which these spectra were obtained make them, however, serve as excellent illustrations of the different types of spectra. we are to notice that the upper band, which contains what is called the “flash” spectrum, exhibits bright lines on a dark background. see, for instance, the two lines so very distinctly marked, which are indicated by the letters h and k. these lines are very characteristic of the solar spectrum, and it may be mentioned that they are indications of the presence of a well-known element. these lines prove that the sun contains calcium, the metal of which common lime is the oxide. it is, indeed, the presence of this substance in the sun which gives rise to these lines. we shall refer again to this subject in a later chapter.
as the upper of the two spectra exhibit h and k as white lines on a dark background, so the lower represents the same lines as dark objects on a white background. these photographs give illustrations of spectra of the two different classes which provide means of discriminating between a genuine nebula and an object which, though it looks like a nebula, is not itself gaseous.
fig. 13.—the nebula in the pleiades (exposure 10 hours).
(photographed by dr. isaac roberts, f.r.s.)
but, it will be asked, how can the spectra of the two distinct types both be obtained from the sun? the explanation of this point is an interesting one. the lower of the two is the ordinary solar spectrum; it is a continuous spectrum showing dark lines on a bright field. the upper spectrum, which shows bright lines 72on a dark field, is produced by a small part of the sun just at the moment when the eclipse is total. the circumstances in which that picture was secured will explain its character. the moon had completely covered that dazzling part of the sun which we ordinarily see, but a region of intensely glowing gaseous material in the sun’s atmosphere was too high above the surface to be completely hidden by the moon. the spectrum of this region, consisting of the characteristic bright gaseous lines, is here represented. the ordinary light of the sun being cut off, opportunity was thus afforded for the production of the spectrum of the light from the glowing gas, and we see this spectrum to be of the nebular type.
and now we may bring this chapter to a close by calling attention to the very important bearing which its facts have on the nebular theory. it is essential for us to see how far modern investigation and discovery have tended either to substantiate or refute that famous doctrine which traces the development of the solar system from a nebula. to do this it is necessary to contrast the knowledge of nebul?, as it exists at present, with the knowledge of nebul? as it existed in the days of kant and laplace and herschel.
we assuredly do no injustice to kant or to laplace if we say that their actual knowledge of the nebulous contents of the heavens was vastly inferior to that possessed by herschel. there is not a single astronomical observation of nebul? recorded by either kant or laplace; it may be doubted whether either of them ever even saw a nebula. their splendid contributions to science were made in directions far removed from those of the practical observer, who passes long 73hours of darkness in the scrutiny of the celestial bodies. herschel, on the other hand, was pre-eminently an observer. his nights were spent in the most diligent practical observation of the heavens, and at all times the nebul? were the objects which received the largest measure of his attention, with the result that the knowledge of nebul? received the most extraordinary development from his labours. earlier astronomers had no doubt observed nebul? occasionally, but with their imperfect appliances only the brighter of these objects were discernible by them. the astonishing advance made by the observations of herschel is only paralleled by the advance made a hundred years later by the photographs of keeler.
but it must be remembered that though herschel observed nebul?, and discovered nebul?, and discoursed on nebul? in papers which to this day are classics in this important subject, yet not to the last day of his life could he have felt sure that he had ever seen a genuine nebula. he might have surmised, and he did surmise, that many of the objects he set down as nebul? were actually gaseous objects, but he knew that many apparent nebul? were in truth clusters of stars, and he had no means of knowing whether all so-called nebul? might not belong to the same category.
it was not till nearly half a century after sir william herschel’s unrivalled career had closed that the spectroscope was invoked to decide finally on the nature of these mysterious objects. that decision, which has been of such transcendent importance in the study of the heavens, was not pronounced till 1864. in that year sir william huggins established 74the fundamental truth that the so-called nebul? are not all star-clusters, but that the universe does contain objects which are most certainly gigantic volumes of incandescent gases.
this great achievement provided a complete answer to those who urged an objection, which seemed once very weighty, against the nebular theory. it must be admitted that before 1864 no one could have affirmed with confidence that any genuine nebula really existed. it was, therefore, impossible for the authors of the nebular theory to point to any object in the heavens which might have illustrated the great principles involved in the theory. the nebular theory required that in the beginning there should have been a gaseous nebula from which the solar system has been evolved. but the objector, who was pleased to contend that the gaseous nebula was a figment of the imagination, could never have been effectively silenced by kant or laplace or herschel. it would have been useless for them to point to the nebula in orion, for the objector might say that it was only a cluster of stars, and at that time there would have been no way of confuting him.
the authors of the nebular theory had, in respect to this class of objector, a much more difficult task than falls to its modern advocate. the latter is able to deny in the most emphatic manner that a gaseous nebula is no more than an imaginary conception.
the famous discovery of sir w. huggins has removed the first great objection to the nebular theory.
