seems entirely too great to let us conclude that the blue stars in general could have been formed from planetary nebulæ such as the 150 planetaries now known to us.

For the hypothesis that the stars in general have evolved from irregular nebulæ a vastly stronger case can be made.

In a priceless possession of astronomy, the Draper catalogue of stellar spectra, Harvard College Observatory has classified the spectra of a great many thousands of the brighter stars. They have been arranged in a sequence, running from the so-called extremely blue stars through the yellow stars to the red stars, which can be readily described. The main divisions are illustrated in Fig. 29. Each main class has ten subdivisions, but we need not dwell upon details. The dark lines in these spectra indicate the presence of certain vapors and gases of the chemical elements in the outer strata of the stars. In the Class B stars the helium lines rise to their maximum strength near the middle of the B subgroups and sink to insignificance in the later B or earlier A subdivisions, and the helium lines are not found at all in Class F and the later types. The hydrogen lines are fairly prominent in the Class B stars, but they increase to maximum intensity in Class A, and then de

crease continuously throughout the remaining groups. The hydrogen lines are very inconspicuous in Class M, or red stars. The magnesium lines go up to a maximum in Class A and down to disappearance in the F's and G's. Some of the metallic lines, such as titanium and iron, usually begin to show in the later subdivisions of Class A, other metallic lines first enter upon the scene in Class F, and they increase in numbers and prominence up to a maximum in the red stars. The calcium lines are weak or entirely wanting in the Class B's, and they increase in intensity as we pass down through the series, until they become the most prominent features in the spectra. We can not change the arrangement of these spectral classes without throwing the sequence of development of the spectral lines into hopeless disorder. This sequence is thought to represent the order of stellar evolution. A Class B star, according to that hypothesis, is a comparatively young star. It should develop, in the course of long ages, into an F, a G, a K, an M, and so on, to its final destination of darkness and invisibility.

Some of the Class B stars contain bright lines of hydrogen. Alcyone (Fig. 30) and Pleione, in the Pleiades, contain both bright

and dark lines of hydrogen. There are hundreds of stars whose spectra contain a wide variety of bright lines and dark lines. Now and then a star's spectrum consists almost wholly of bright lines, as in the case of Eta Carinæ (Fig. 31). The bright lines in stellar spectra tell us that their stars contain extremely extensive atmospheres of the gases and vapors hydrogen, helium, and so on. A close relationship exists between the Class B stars, with and without bright lines, and a class of stars found extensively in the Milky Way structure whose spectra, containing many bright lines, are known as Wolf-Rayet spectra, after the discoverers of the first few stars in the class. Now, as Wright has shown, the central stars in the

planetary nebulæ are of the Wolf-Rayet type (see Fig. 32, B.D+30° 3639) in nearly all cases, and in other cases their spectra are closely related to Class B spectra (see Fig. 32, N.G.C. 2932). It has been shown, also chiefly by Wright, that the spectra of the nebulous parts of the planetary nebulæ have many points of connection with the spectra of their central stars. The spectra of the planetary nebula and the spectra of those large extended nebulæ (see Fig. 32, Orion, etc.) which give bright lines are essentially alike in character. The nebulium lines of the nebulæ have never been found to exist in any true star, no matter what its class, and we leave them out of account in this discussion; they do not in

fluence our present question one way or the other. Aside from nebulium, the hydrogen and helium lines are the most prominent ones in the bright-line nebulæ. They are the most prominent ones in the nuclei of the planetary nebulæ. They are prominent in the isolated Wolf-Rayet stars, in the Class B stars containing bright lines, and in the

FIG. 33. The Ring Nebula in Lyra, photographed by Curtis with the Crossley Reflector of the Lick Observatory.

Class B stars containing only dark lines. Helium has never been found in any yellow or red stars, except feebly in a certain stratum of our sun's atmosphere, and that is due to our ability to observe our sun somewhat in detail; and the hydrogen is relatively feeble in the later classes of the Harvard sequence.

The continuous spectra of stars as arranged in the Harvard sequence decrease continuously in the relative strength of the blue, violet and ultra-violet regions as we pass from the central stars of the planetary nebulæ, which are wonderfully strong in ultra-violet light, down through the blue stars with both bright and dark lines,

through the blue stars containing dark lines only, and on through the yellow stars to the red stars. The central star in the ring nebula in Lyra (Fig. 33) is invisible to many inexperienced observers, even in the great Lick telescope. Yet it is so rich in violet light, and especially in ultra-violet light, that it can be photographed in the great reflecting telescopes with an exposure of only two or three seconds. Wright has recently made the interesting discovery that the continuous spectra of many planetary nebulæ are remarkably strong in a long stretch of the ultra-violet region. The sequence of decreasing richness in blue and violet light extends unbroken from the nebulæ and nebular nuclei down to the red stars.

The spectra of the stars are indicative of conditions existing in their surface strata. We do not know definitely just what conditions produce or accompany certain types of spectra, but surface temperature seems to be a prime influence, and there are some reasons for thinking that electrical conditions may also be extremely important. Now all determinations of the surface temperatures of the stars make the extreme blue-violet stars the hottest of all, with the effective temperatures decreasing continuously as we pass from the nuclear stars of the planetary nebulæ down through the Harvard sequence to the red stars. Fabry and his associates in Marseilles have recently arrived at the result, by physical methods, that the temperature of the Orion nebula is very high, vastly higher than the temperatures of the red and yellow stars,

6 Huggins has called attention to a reduction in the brightness of the ultra-violet spectrum of the first-magnitude star Vega (Class A) by virtue of what seems to be an absorption band many hundred Angstrom units in width. It is not yet known to what extent the ultra-violet spectra of blue stars in general may be affected by such an absorption band.

and comparable with the temperatures of the blue stars. There again is a sequence with the nebulæ and the blue stars at one end and the red stars at the other.

Fowler has called attention to the remarkable facts that certain lines of helium, etc., requiring the most powerful electric discharges at command to produce them in the laboratory-super-spark lines, he calls them are found in the greatest relative strength in the gaseous nebulæ; next in order of strength in the bright-line or WolfRayet stars, which class includes the central stars of the planetary nebulæ; and in lesser strength in the first subdivision of the Class B stars, that is, in the bluest of the blue stars containing dark lines only; secondly, that the lines in the spectra of the later Class B stars, of the Class A and Class F stars are prevailingly those produced under the less intense conditions of the ordinary electric spark; and, thirdly, that the lines in the yellow-red stars are prevailingly the arc lines which indicate relatively weak electrical conditions. That is another sequence of conditions running from the nebulæ down to the red stars.

There are still other sequences running harmoniously through the Harvard classification. For example, the velocities of the stars in their travels within the stellar system increase as we pass by spectral classes from the large bright-line nebulæ and the very blue stars down through the yellow stars to the red stars. The distances apart of the two components of double stars, and consequently their periods of revolution

FIG. 34. The Irregular Nebulæ near Gamma Cassiopeia, photographed by Curtis with the Crossley Reflector of the Lick Observatory.

[The immensely over-exposed image of the brilliant star Gamma Cassiopeia is at the lower lefthand corner. The lines radiating from the center of the star image are diffraction effects produced within the telescope. The upper right half of the figure contains much nebulosity, the brighter angles of two nebular structures pointing approximately toward the star Gamma.]

around each other, increase consecutively as we pass from the blue stars to the red.

It might be said that, so far as all these stellar sequences are concerned, the course of evolution could begin with either the red stars or the blue stars and proceed to the other end of the sequence. Not to mention several very weighty objections to the assumption that the red stars are effectively young and the blue stars effectively old, we submit the case on the evidence of the nebulæ. The planetary nebulæ, the irregular nebulæ, the Wolf-Rayet stars, the Class B stars with bright lines, and the Class B

The sky is divided into eight zones of equal areas, with the boundaries of the zones parallel to the galactic plane. The first line of figures contains the numbers of stars of the different spectral classes in the one eighth of the sky around the north pole of the Galaxy; the fourth and fifth lines the numbers in the zones containing the Galaxy, one zone adjoining the Milky Way on the north and the other on the south; and the last line the numbers of stars of the different classes in the eighth of the sky around the south pole of the Galaxy. If the stars of the different classes were uniformly distributed over the sky, the eight numbers in each vertical column would be equal. It is seen that the Class B stars are prevailingly in the Milky Way, and that the red stars of Class M are about uniformly distributed over the sky, though they are all in our stellar system.

The Class B stars and the stars containing bright lines are where the planetary and irregular nebulæ exist. Going further into detail wherever there is a great nebulous region either in, or near, or outside of the Milky Way you will find the Class B and earlier types of stars abnormally plentiful; and the chances are fairly strong that

some of the stellar spectra will contain bright lines. This is true of great regions in the Milky Way, it is true of the Orion and Pleiades regions, which we see at some distance outside of the Milky Way structure, though they are doubtless within our system. If you see a wisp of nebulosity near a bright star, look up the star's spectrum and you will probably find it an early Class B, as in the case of Gamma Cassiopeiæ, a second magnitude star, with nebulous structure near it (Fig. 34), whose spectrum contains both bright and dark lines of hydrogen and helium (Fig. 35). If you see an isolated bright star apparently enmeshed in an isolated patch of nebulosity, such as the one shown in Fig. 37, and the books say the star (BD-10°4713) is yellow, or of Class G, communicate your suspicions that the books are mistaken about that star's spectrum to Professor Pickering, and he will probably reply that the star is in reality a very blue one, of early Class B. That is what happened a fortnight ago about this particular nebula and the star near its apparent center. If you find a red or yellow star of normal type, do not look for a nebula in apparent contact with it. Nebulæ and red stars do not coexist. You will find about the same number of red stars in the Milky Way that are visible in similar areas far from the Milky Way. You will find an occasional red star in the region of the Orion nebula and of other large nebulæ, but red stars will not appear there in greater numbers than their approximately uniform distribution over the sky requires.

The connection between the nebulæ and the bright-line stars, and between the nebulæ and the early Class B stars is close, both as to their types of spectra and as to their geometric distribution.

Do the nebulæ form stars or do the stars form nebulæ, or both? There is abundant