209. Stellar colors.—We have already seen that one star differs from another in respect of color as well as brightness, and the diligent student of the sky will not fail to observe for himself how the luster of Sirius and Rigel is more nearly a pure white than is that of any other stars in the heavens, while at the other end of the scale aOrionis and Aldebaran are strongly ruddy, and Antares presents an even deeper tone of red. Between these extremes the light of every star shows a mixture of the rainbow hues, in which a very pale yellow is the predominant color, shading off, as we have seen, to white at one end of the scale and red at the other. There are no green stars, or blue stars, or violet stars, save in one exceptional class of cases—viz., where the two components of a double star are of very different brightness, it is quite the usual thing for them to have different colors, and then, almost without exception, the color of the fainter star lies nearer to the violet end of the spectrum than does the color of the bright one, and sometimes shows a distinctly blue or green hue. A fine type of such double star is ߠCygni, in which the components are respectively yellow and blue, and the yellow star furnishes eight times as much light as the blue one.

The exception which double stars thus make to the general rule of stellar colors, yellow and red, but no color of shorter wave length, has never been satisfactorily explained, but the rule itself presents no difficulties. Each star is an incandescent body, giving off radiant energy of every wave length within the limits of the visible spectrum, and, indeed, far beyond these limits. If this radiant energy could come unhindered to our eyes every star would appear white, but they are all surrounded by atmospheres—analogous to the chromosphere and reversing layer of the sun—which absorb a portion of their radiant energy and, like the earth's atmosphere, take a heavier toll from the violet than from the red end of the spectrum. The greater the absorption in the star's atmosphere, therefore, the feebler and the ruddier will be its light, and corresponding to this the red stars are as a class fainter than the white ones.210. Chemistry of the stars.—The spectroscope is pre-eminently the instrument to deal with this absorption of light in the stellar atmospheres, just as it deals with that absorption in the sun's atmosphere to which are due the dark lines of the solar spectrum, although the faintness of starlight, compared with that of the sun, presents a serious obstacle to its use. Despite this difficulty most of the lucid stars and many of the telescopic ones have been studied with the spectroscope and found to be similar to the sun and the earth as respects the material of which they are made. Such familiar chemical elements as hydrogen and iron, carbon, sodium, and calcium are scattered broadcast throughout the visible universe, and while it would be unwarranted by the present state of knowledge to say that the stars contain nothing not found in the earth and the sun, it is evident that in a broad way their substance is like rather than unlike that composing the solar system, and is subject to the same physical and chemical laws which obtain here. Galileo and Newton extended to the heavens the terrestrial sciences of mathematics and mechanics, but it remained to the nineteenth century to show that the physics and chemistry of the sky are like the physics and chemistry of the earth.211. Stellar spectra.—When the spectra of great numbers of stars are compared one with another, it is found that they bear some relation to the colors of the stars, as, indeed, we should expect, since spectrum and color are both produced by the stellar atmospheres, and it is found useful to classify these spectra into three types, as follows:

Type I. Sirian stars.—Speaking generally, the stars which are white or very faintly tinged with yellow, furnish spectra like that of Sirius, from which they take their name, or that of ߠAurigÆ (Fig.124), which is a continuous spectrum, especially rich in energy of short wave length—i.e., violet and ultra-violet light, and is crossed by a relatively small number of heavy dark lines corresponding to the spectrum of hydrogen. Sometimes, however, these lines are much fainter than is here shown, and we find associated with them still other faint ones pointing to the presence of other metallic substances in the star's atmosphere. These metallic lines are not always present, and sometimes even the hydrogen lines themselves are lacking, but the spectrum is always rich in violet and ultra-violet light.

Since with increasing temperature a body emits a continually increasing proportion of energy of short wave length (§118), the richness of these spectra in such energy points to a very high temperature in these stars, probably surpassing in some considerable measure that of the sun. Stars with this type of spectrum are more numerous than all others combined, but next to them in point of numbers stands—

Type II. Solar stars.—To this type of spectrum belong the yellow stars, which show spectra like that of the sun, or of Pollux (Fig.125). These are not so rich in violet light as are those of TypeI, but in complexity of spectrum and in the number of their absorption lines they far surpass the Sirian stars. They are supposed to be at a lower temperature than the Sirian stars, and a much larger number of chemical elements seems present and active in the reversing layer of their atmospheres. The strong resemblance which these spectra bear to that of the sun, together with the fact that most of the sun's stellar neighbors have spectra of this type, justify us in ranking both them and it as members of one class, called solar stars.

Type III. Red stars.—A small number of stars show spectra comparable with that of aHerculis (Fig.134), in which the blue and the violet part of the spectrum is almost obliterated, and the remaining yellow and red parts show not only dark lines, but also numerous broad dark bands, sharp at one edge, and gradually fading out at the other. It is this selective absorption, extinguishing the blue and leaving the red end of the spectrum, which produces the ruddy color of these stars, while the bands in their spectra "are characteristic of chemical combinations, and their presence ... proves that at certain elevations in the atmospheres of these stars the temperature has sunk so low that chemical combinations can be formed and maintained" (Scheiner-Frost). One of the chemical compounds here indicated is a hydrocarbon similar to that found in comets. In the white and yellow stars the temperatures are so high that the same chemical elements, although present, can not unite one with another to form compound substances.

Most of the variable stars are red and have spectra of the third type; but this does not hold true for the eclipse variables like Algol, all of which are white stars with spectra of the first type. The ordinary variable star is therefore one with a dense atmosphere of relatively low temperature and complex structure, which produces the prevailing red color of these stars by absorbing the major part of their radiant energy of short wave length while allowing the longer, red waves to escape. Although their exact nature is not understood, there can be little doubt that the fluctuation in the light of these stars is due to processes taking place within the star itself, but whether above or below its photosphere is still uncertain.212. Classes of stars.—There is no hard-and-fast dividing line between these types of stellar spectra, but the change from one to another is by insensible gradations, like the transition from youth to manhood and from manhood to old age, and along the line of transition are to be found numberless peculiarities and varieties of spectra not enumerated above—e.g., a few stars show not only dark absorption lines in their spectra but bright lines as well, which, like those in Fig.48, point to the presence of incandescent vapors, even in the outer parts of their atmospheres. Among the lucid stars about 75 per cent have spectra of the first type, 23 per cent are of the second type, 1per cent of the third type, and the remaining 1 per cent are peculiar or of doubtful classification. Among the telescopic stars it is probable that much the same distribution holds, but in the present state of knowledge it is not prudent to speak with entire confidence upon this point.

That the great number of stars whose spectra have been studied should admit of a classification so simple as the above, is an impressive fact which, when supplemented by the further fact of a gradual transition from one type of spectrum to the next, leaves little room for doubt that in the stars we have an innumerable throng of individuals belonging to the same species but in different stages of development, and that the sun is only one of these individuals, of something less than medium size and in a stage of development which is not at all peculiar, since it is shared by nearly a fourth of all the stars.

213. Star clusters.—In previous chapters we have noted the Pleiades and PrÆsepe as star clusters visible to the naked eye, and to them we may add the Hyades, near Aldebaran, and the little constellation Coma Berenices. But more impressive than any of these, although visible only in a telescope, is the splendid cluster in Hercules, whose appearance in a telescope of moderate size is shown in Fig.135, while Fig.136 is a photograph of the same cluster taken with a very large reflecting telescope. This is only a type of many telescopic clusters which are scattered over the sky, and which are made up of stars packed so closely together as to become indistinguishable, one from another, at the center of the cluster. Within an area which could be covered by a third of the full moon's face are crowded in this cluster more than five thousand stars which are unquestionably close neighbors, but whose apparent nearness to each other is doubtless due to their great distance from us. It is quite probable that even at the center of this cluster, where more than a thousand stars are included within a radius of 160", the actual distances separating adjoining stars are much greater than that separating earth and sun, but far less than that separating the sun from its nearest stellar neighbor.

An interesting discovery of recent date, made by Professor Bailey in photographing star clusters, is that some few of them, which are especially rich in stars, contain an extraordinary number of variable stars, mostly very faint and of short period. Two clusters, one in the northern and one in the southern hemisphere, contain each more than a hundred variables, and an even more extraordinary case is presented by a cluster, called Messier5, not far from the star aSerpentis, which contains no less than sixty-three variables, all about of the fourteenth magnitude, all having light periods which differ but little from half a day, all having light curves of about the same shape, and all having a range of brightness from maximum to minimum of about one magnitude. An extraordinary set of coincidences which "points unmistakably to a common origin and cause of variability."

214. NebulÆ.—Returning to Fig.136, we note that its background has a hazy appearance, and that at its center the stars can no longer be distinguished, but blend one with another so as to appear like a bright cloud. The outer part of the cluster is resolved into stars, while in the picture the inner portion is not so resolved, although in the original photographic plate the individual stars can be distinguished to the very center of the cluster. In many cases, however, this is not possible, and we have an irresolvable cluster which it is customary to call a nebula (Latin, little cloud).

The most conspicuous example of this in the northern heavens is the great nebula in Andromeda (R.A. 0^h 37^m, Dec.+41°), which may be seen with the naked eye as a faint patch of foggy light. Look for it. This appears in an opera glass or very small telescope not unlike Fig.137, which is reproduced from a sketch. Fig.138 is from a photograph of the same object showing essentially the same shape as in the preceding figure, but bringing out more detail. Note the two small nebulÆ adjoining the large one, and at the bottom of the picture an object which might easily be taken for another nebula but which is in fact a tailless comet that chanced to be passing that part of the sky when the picture was taken. Fig.139 is from another drawing of this nebula, although it is hardly to be recognized as a representation of the same thing; but its characteristic feature, the two dark streaks near the center of the picture, is justified in part by Fig.140, which is from a photograph made with a large reflecting telescope.

A comparison of these several representations of the same thing will serve to illustrate the vagueness of its outlines, and how much the impressions to be derived from nebulÆ depend upon the telescopes employed and upon the observer's own prepossessions. The differences among the pictures can not be due to any change in the nebula itself, for half a century ago it was sketched much as shown in the latest of them (Fig.140).

215. Typical nebulÆ.—Some of the fantastic forms which nebulÆ present in the telescope are shown on a small scale in Fig.141, but in recent years astronomers have learned to place little reliance upon drawings such as these, which are now almost entirely supplanted by photographs made with long exposures in powerful telescopes. One of the most exquisite of these modern photographs is that of the Trifid nebula in Sagittarius (Fig.142). Note especially the dark lanes that give to this nebula its name, Trifid, and which run through its brightest parts, breaking it into seemingly independent sections. The area of the sky shown in this cut is about 15 per cent less than that covered by the full moon.

Fig.143 shows a very different type of nebula, found in the constellation Cygnus, which appears made up of filaments closely intertwined, and stretches across the sky for a distance considerably greater than the moon's diameter.

A much smaller but equally striking nebula is that in the constellation Canes Venatici (Fig.144), which shows a most extraordinary spiral structure, as if the stars composing it were flowing in along curved lines toward a center of condensation. The diameter of the circular part of this nebula, omitting the projection toward the bottom of the picture, is about five minutes of arc, a sixth part of the diameter of the moon, and its thickness is probably very small compared with its breadth, perhaps not much exceeding the width of the spiral streams which compose it. Note how the bright stars that appear within the area of this nebula fall on the streams of nebulous matter as if they were part of them. This characteristic grouping of the stars, which is followed in many other nebulÆ, shows that they are really part and parcel of the nebula and not merely on line with it. Fig.145 shows how a great nebula is associated with the star ?Ophiuchi.

Probably the most impressive of all nebulÆ is the great one in Orion (Fig.146), whose position is shown on the star map between Rigel and ?Orionis. Look for it with an opera glass or even with the unaided eye. This is sometimes called an amorphous—i.e., shapeless—nebula, because it presents no definite form which the eye can grasp and little trace of structure or organization. It is "without form and void" at least in its central portions, although on its edges curved filaments may be traced streaming away from the brighter parts of the central region. This nebula, as shown in Fig.146, covers an area about equal to that of the full moon, without counting as any part of this the companion nebula shown at one side, but photographs made with suitable exposures show that faint outlying parts of the nebula extend in curved lines over the larger part of the constellation Orion. Indeed, over a large part of the entire sky the background is faintly covered with nebulous light whose brighter portions, if each were counted as a separate nebula, would carry the total number of such objects well into the hundreds of thousands.

The Pleiades (PlateIV) present a case of a resolvable star cluster projected against such a nebulous background whose varying intensity should be noted in the figure. A part of this nebulous matter is shown in wisps extending from one star to the next, after the fashion of a bridge, and leaving little doubt that the nebula is actually a part of the cluster and not merely a background for it.

Fig.147 shows a series of so-called double nebulÆ perhaps comparable with double stars, although the most recent photographic work seems to indicate that they are really faint spiral nebulÆ in which only the brightest parts are shown by the telescope.

According to Keeler, the spiral is the prevailing type of nebulÆ, and while Fig.144 presents the most perfect example of such a nebula, the student should not fail to note that the Andromeda nebula (Fig.140) shows distinct traces of a spiral structure, only here we do not see its true shape, the nebula being turned nearly edgewise toward us so that its presumably circular outline is foreshortened into a narrow ellipse.

Another type of nebula of some consequence presents in the telescope round disks like those of Uranus or Neptune, and this appearance has given them the name planetary nebulÆ. The comet in Fig.138, if smaller, would represent fairly well the nebulÆ of this type. Sometimes a planetary nebula has a star at its center, and sometimes it appears hollow, like a smoke ring, and is then called a ring nebula. The most famous of these is in the constellation Lyra, not far from Vega.216. Spectra of nebulÆ.—A star cluster, like the one in Hercules, shows, of course, stellar spectra, and even when irresolvable the spectrum is a continuous one, testifying to the presence of stars, although they stand too close together to be separately seen. But in a certain number of nebulÆ the spectrum is altogether different, a discontinuous one containing only a few bright lines, showing that here the nebular light comes from glowing gases which are subject to no considerable pressure. The planetary nebulÆ all have spectra of this kind and make up about half of all the known gaseous nebulÆ. It is worthy of note that a century ago Sir William Herschel had observed a green shimmer in the light of certain nebulÆ which led him to believe that they were "not of a starry nature," a conclusion which has been abundantly confirmed by the spectroscope. The green shimmer is, in fact, caused by a line in the green part of the spectrum that is always present and is always the brightest part of the spectrum of gaseous nebulÆ.

In faint nebulÆ this line constitutes the whole of their visible spectrum, but in brighter ones two or three other and fainter lines are usually associated with it, and a very bright nebula, like that in Orion, may show a considerable number of extra lines, but for the most part they can not be identified in the spectrum of any terrestrial substances. An exception to this is found in the hydrogen lines, which are well marked in most spectra of gaseous nebulÆ, and there are indications of one or two other known substances.217. Density of nebulÆ.—It is known from laboratory experiments that diminishing the pressure to which an incandescent gas is subject, diminishes the number of lines contained in its spectrum, and we may surmise from the very simple character and few lines of these nebular spectra that the gas which produces them has a very small density. But this is far from showing that the nebula itself is correspondingly attenuated, for we must not assume that this shining gas is all that exists in the nebula; so far as telescope or camera are concerned, there may be associated with it any amount of dark matter which can not be seen because it sends to us no light. It is easy to think in this connection of meteoric dust or the stuff of which comets are made, for these seem to be scattered broadcast on every side of the solar system and may, perchance, extend out to the region of the nebulÆ.

But, whatever may be associated in the nebula with the glowing gas which we see, the total amount of matter, invisible as well as visible, must be very small, or rather its average density must be very small, for the space occupied by such a nebula as that of Orion is so great that if the average density of its matter were equal to that of air the resulting mass by its attraction would exert a sensible effect upon the motion of the sun through space. The brighter parts of this nebula as seen from the earth subtend an angle of about half a degree, and while we know nothing of its distance from us, it is easy to see that the farther it is away the greater must be its real dimensions, and that this increase of bulk and mass with increasing distance will just compensate the diminishing intensity of gravity at great distances, so that for a given angular diameter—e.g., half a degree—the force with which this nebula attracts the sun depends upon its density but not at all upon its distance. Now, the nebula must attract the sun in some degree, and must tend to move it and the planets in an orbit about the attracting center so that year after year we should see the nebula from slightly different points of view, and this changed point of view should produce a change in the apparent direction of the nebula from us—i.e., a proper motion, whose amount would depend upon the attracting force, and therefore upon the density of the attracting matter. Observations of the Orion nebula show that its proper motion is wholly inappreciable, certainly far less than half a second of arc per year, and corresponding to this amount of proper motion the mean density of the nebula must be some millions of times (10¹⁰ according to Ranyard) less than that of air at sea level—i.e., the average density throughout the nebula is comparable with that of those upper parts of the earth's atmosphere in which meteors first become visible.218. Motion of nebulÆ.—The extreme minuteness of their proper motions is a characteristic feature of all nebulÆ. Indeed, there is hardly a known case of sensible proper motion of one of these bodies, although a dozen or more of them show velocities in the line of sight ranging in amount from +30 to -40 miles per second, the plus sign indicating an increasing distance. While a part of these velocities may be only apparent and due to the motion of earth and sun through space, a part at least is real motion of the nebulÆ themselves. These seem to move through the celestial spaces in much the same way and with the same velocities as do the stars, and their smaller proper motions across the line of sight (angular motions) are an index of their great distance from us. No one has ever succeeded in measuring the parallax of a nebula or star cluster.

The law of gravitation presumably holds sway within these bodies, and the fact that their several parts and the stars which are involved within them, although attracted by each other, have shown little or no change of position during the past century, is further evidence of their low density and feeble attraction. In a few cases, however, there seem to be in progress within a nebula changes of brightness, so that what was formerly a faint part has become a brighter one, or vice versa; but, on the whole, even these changes are very small.

219. The Milky Way.—Closely related to nebulÆ and star clusters is another feature of the sky, the galaxy or Milky Way, with whose appearance to the unaided eye the student should become familiar by direct study of the thing itself. Figs.148 and149 are from photographs of two small parts of it, and serve to bring out the small stars of which it is composed. Every star shown in these pictures is invisible to the naked eye, although their combined light is easily seen. The general course of the galaxy across the heavens is shown in the star maps, but these contain no indication of the wealth of detail which even the naked eye may detect in it. Bright and faint parts, dark rifts which cut it into segments, here and there a hole as if the ribbon of light had been shot away—such are some of the features to be found by attentive examination.

Speaking generally, the course of the Milky Way is a great circle completely girdling the sky and having its north pole in the constellation Coma Berenices. The width of this stream of light is very different in different parts of the heavens, amounting where it is widest, in Lyra and Cygnus, to something more than 30°, although its boundaries are too vague and ill defined to permit much accuracy of measurement. Observe the very bright part between ߠand ?Cygni, nearly opposite Vega, and note how even an opera glass will partially resolve the nebulous light into a great number of stars, which are here rather brighter than in other parts of its course. But the resolution into stars is only partial, and there still remains a background of unresolved shimmer. Fig.150 is a photograph of a small part of this region in which, although each fleck of light represents a separate star, the galaxy is not completely resolved. Compare with this region, rich in stars, the nearly empty space between the branches of the galaxy a little west of Altair. Another hole in the Milky Way may be found a little north and east of aCygni, and between the extremes of abundance and poverty here noted there may be found every gradation of nebulous light.

The Milky Way is not so simple in its structure as might at first be thought, but a clear and moonless night is required to bring out its details. The nature of these details, the structure of the galaxy, its shape and extent, the arrangement of its parts, and their relation to stars and nebulÆ in general, have been subjects of much speculation by astronomers and others who have sought to trace out in this way what is called the construction of the heavens.220. Distribution of the stars.—How far out into space do the stars extend? Are they limited or infinite in number? Do they form a system of mutually related parts, or are they bunched promiscuously, each for itself, without reference to the others? Here is what has been well called "the most important problem of stellar astronomy, the acquisition of well-founded ideas about the distribution of the stars." While many of the ideas upon this subject which have been advanced by eminent astronomers and which are still current in the books are certainly wrong, and few of their speculations along this line are demonstrably true, the theme itself is of such grandeur and permanent interest as to demand at least a brief consideration. But before proceeding to its speculative side we need to collect facts upon which to build, and these, however inadequate, are in the main simple and not far to seek.

Parallaxes, proper motions, motions in the line of sight, while pertinent to the problem of stellar distribution, are of small avail, since they are far too scanty in number and relate only to limited classes of stars, usually the very bright ones or those nearest to the sun. Almost the sole available data are contained in the brightness of the stars and the way in which they seem scattered in the sky. The most casual survey of the heavens is enough to show that the stars are not evenly sprinkled upon it. The lucid stars are abundant in some regions, few in others, and the laborious star gauges, actual counting of the stars in sample regions of the sky, which have been made by the Herschels, Celoria, and others, suffice to show that this lack of uniformity in distribution is even more markedly true of the telescopic stars.

The rate of increase in the number of stars from one magnitude to the next, as shown in §187, is proof of another kind of irregularity in their distribution. It is not difficult to show, mathematically, that if in distant regions of space the stars were on the average as numerous and as bright as they are in the regions nearer to the sun, then the stars of any particular magnitude ought to be four times as numerous as those of the next brighter magnitude—e.g., four times as many sixth-magnitude stars as there are fifth-magnitude ones. But, as we have already seen in §187, by actual count there are only three times as many, and from the discrepancy between these numbers, an actual threefold increase instead of a fourfold one, we must conclude that on the whole the stars near the sun are either bigger or brighter or more numerous than in the remoter depths of space.221. The stellar system.—But the arrangement of the stars is not altogether lawless and chaotic; there are traces of order and system, and among these the Milky Way is the dominant feature. Telescope and photographic plate alike show that it is made up of stars which, although quite irregularly scattered along its course, are on the average some twenty times as numerous in the galaxy as at its poles, and which thin out as we recede from it on either side, at first rapidly and then more slowly. This tendency to cluster along the Milky Way is much more pronounced among the very faint telescopic stars than among the brighter ones, for the lucid stars and the telescopic ones down to the tenth or eleventh magnitude, while very plainly showing the clustering tendency, are not more than three times as numerous in the galaxy as in the constellations most remote from it. It is remarkable as showing the condensation of the brightest stars that one half of all the stars in the sky which are brighter than the second magnitude are included within a belt extending 12° on either side of the center line of the galaxy.

In addition to this general condensation of stars toward the Milky Way, there are peculiarities in the distribution of certain classes of stars which are worth attention. Planetary nebulÆ and new stars are seldom, if ever, found far from the Milky Way, and stars with bright lines in their spectra especially affect this region of the sky. Stars with spectra of the first type—Sirian stars—are much more strongly condensed toward the Milky Way than are stars of the solar type, and in consequence of this the Milky Way is peculiarly rich in light of short wave lengths. Resolvable star clusters are so much more numerous in the galaxy than elsewhere, that its course across the sky would be plainly indicated by their grouping upon a map showing nothing but clusters of this kind.

On the other hand, nebulÆ as a class show a distinct aversion for the galaxy, and are found most abundantly in those parts of the sky farthest from it, much as if they represented raw material which was lacking along the Milky Way, because already worked up to make the stars which are there so numerous.222. Relation of the sun to the Milky Way.—The fact that the galaxy is a great circle of the sky, but only of moderate width, shows that it is a widely extended and comparatively thin stratum of stars within which the solar system lies, a member of the galactic system, and probably not very far from its center. This position, however, is not to be looked upon as a permanent one, since the sun's motion, which lies nearly in the plane of the Milky Way, is ceaselessly altering its relation to the center of that system, and may ultimately carry us outside its limits.

The Milky Way itself is commonly thought to be a ring, or series of rings, like the coils of the great spiral nebula in Andromeda, and separated from us by a space far greater than the thickness of the ring itself. Note in Figs.149 and150 how the background is made up of bright and dark parts curiously interlaced, and presenting much the appearance of a thin sheet of cloud through which we look to barren space beyond. While, mathematically, this appearance can not be considered as proof that the galaxy is in fact a distant ring, rather than a sheet of starry matter stretching continuously from the nearer stellar neighbors of the sun into the remotest depths of space, nevertheless, most students of the question hold it to be such a ring of stars, which are relatively close together while its center is comparatively vacant, although even here are some hundreds of thousands of stars which on the whole have a tendency to cluster near its plane and to crowd together a little more densely than elsewhere in the region where the sun is placed.223. Dimensions of the galaxy.—The dimensions of this stellar system are wholly unknown, but there can be no doubt that it extends farther in the plane of the Milky Way than at right angles to that plane, for stars of the fifteenth and sixteenth magnitudes are common in the galaxy, and testify by their feeble light to their great distance from the earth, while near the poles of the Milky Way there seem to be few stars fainter than the twelfth magnitude. Herschel, with his telescope of 18 inches aperture, could count in the Milky Way more than a dozen times as many stars per square degree as could Celoria with a telescope of 4 inches aperture; but around the poles of the galaxy the two telescopes showed practically the same number of stars, indicating that here even the smaller telescope reached to the limits of the stellar system. Very recently, indeed, the telescope with which Fig.140 was photographed seems to have reached the farthest limit of the Milky Way, for on a photographic plate of one of its richest regions Roberts finds it completely resolved into stars which stand out upon a black background with no trace of nebulous light between them.224. Beyond the Milky Way.—Each additional step into the depths of space brings us into a region of which less is known, and what lies beyond the Milky Way is largely a matter of conjecture. We shrink from thinking it an infinite void, endless emptiness, and our intellectual sympathies go out to Lambert's speculation of a universe filled with stellar systems, of which ours, bounded by the galaxy, is only one. There is, indeed, little direct evidence that other such systems exist, but the Andromeda nebula is not altogether unlike a galaxy with a central cloud of stars, and in the southern hemisphere, invisible in our latitudes, are two remarkable stellar bodies like the Milky Way in appearance, but cut off from all apparent connection with it, much as we might expect to find independent stellar systems, if such there be.

These two bodies are known as the Magellanic clouds, and individually bear the names of Major and Minor Nubecula. According to Sir John Herschel, "the Nubecula Major, like the Minor, consists partly of large tracts and ill-defined patches of irresolvable nebula, and of nebulosity in every stage of resolution up to perfectly resolved stars like the Milky Way, as also of regular and irregular nebulÆ ... of globular clusters in every stage of resolvability, and of clustering groups sufficiently insulated and condensed to come under the designation of clusters of stars." Its outlines are vague and somewhat uncertain, but surely include an area of more than 40 square degrees—i.e., as much as the bowl of the Big Dipper—and within this area Herschel counted several hundred nebulÆ and clusters "which far exceeds anything that is to be met with in any other region of the heavens." Although its excessive complexity of detail baffled Herschel's attempts at artistic delineation, it has yielded to the modern photographic processes, which show the Nubecula Major to be an enormous spiral nebula made up of subordinate stars, nebulÆ, and clusters, as is the Milky Way.

Compared with the Andromeda nebula, its greater angular extent suggests a smaller distance, although for the present all efforts at determining the parallax of either seem hopeless. But the spiral form which is common to both suggests that the Milky Way itself may be a gigantic spiral nebula near whose center lies the sun, a humble member of a great cluster of stars which is roughly globular in shape, but flattened at the poles of the galaxy and completely encircled by its coils. However plausible such a view may appear, it is for the present, at least, pure hypothesis, although vigorously advocated by Easton, who bases his argument upon the appearance of the galaxy itself.225. Absorption of starlight.—We have had abundant occasion to learn that at least within the confines of the solar system meteoric matter, cosmic dust, is profusely scattered, and it appears not improbable that the same is true, although in smaller degree, in even the remoter parts of space. In this case the light which comes from the farther stars over a path requiring many centuries to travel, must be in some measure absorbed and enfeebled by the obstacles which it encounters on the way. Unless celestial space is transparent to an improbable degree the remoter stars do not show their true brightness; there is a certain limit beyond which no star is able to send its light, and beyond which the universe must be to us a blank. A lighthouse throws into the fog its beams only to have them extinguished before a single mile is passed, and though the celestial lights shine farther, a limit to their reach is none the less certain if meteoric dust exists outside the solar system. If there is such an absorption of light in space, as seems plausible, the universe may well be limitless and the number of stellar systems infinite, although the most attenuated of dust clouds suffices to conceal from us and to shut off from our investigation all save a minor fraction of it and them.