devote a very small amount of time to the subject. As you doubtless know, the essential feature of a star spectroscope is the prism or train of prisms by which the star light is divided into its constituent parts. After passing through the prisms the light of the star is spread out into a long band, which shows all the colors of the rainbow, beginning

Photographed with the two-foot reflecting telescope of the Yerkes Observatory (Ritchey).

with red at one end and passing through orange, yellow, green and blue, to violet at the other. This band is crossed by lines, and the problem of the spectroscopist is to interpret the meaning of these lines. If the lines are dark he knows that the light of the star after originat

ing in an interior incandescent body has passed through a mass of cooler vapors, and that during its transmission some of the light has suffered absorption. If, on the other hand, the lines are bright, he knows that the region where they are produced is hotter than that lying below. Thus a single glance at the spectrum of a star is sufficient to give important information regarding the physical condition of its atmosphere.

But the spectral lines are able to tell a far more complete story of stellar conditions. If their exact position in the spectrum can be measured it becomes possible to determine the chemical composition of the star's atmosphere. And here the spectroscopist may be said to have the advantage of the archeologist, in that the key to stellar hieroglyphs is a master key, capable of interpreting not merely the language of a single people or a single age, but of laying bare the secrets of the most distant portions of the universe and applying with equal force to the primitive and to the most highly developed forms of celestial phenomena. If we take a piece of iron wire and turn it into vapor in the intense heat of an electric arc lamp we find that the light which the glowing iron vapor emits, when spread out into a spectrum by a prism, consists of a series of lines characteristically spaced and always occupying the same relative positions. In the same way every other element when transformed into vapor by a sufficiently intense heat emits characteristic radiations, consisting of groups of lines. occupying definite positions in the spectrum. It is thus easy to see how the presence of iron vapor can be detected in the atmosphere of Sirius or in that of the sun. In the spectrum of each of these stars we find a group of lines occupying the same relative positions as the lines furnished by the iron vapor in an electric arc. Hydrogen gives an even more characteristic group of lines, which grow closer and closer together as we pass from the red end of the spectrum toward the violet. This group occurs in the spectra of thousands of stars and serves as an important guide in determining a star's place in a general scheme of stellar evolution.

The practical means of carrying out this method of research may be illustrated by a reference to the stellar spectroscope employed with the 40-inch Yerkes telescope. The spectroscope is rigidly attached to the lower end of the telescope tube. The image of a star formed by the 40-inch lens passes into the spectroscope through a slit about one one-thousandth of an inch wide. After analysis by a train of three prisms an image of the resulting spectrum is formed by a suitable lens upon a photographie plate. In making the photograph it is only necessary to keep the image of a star exactly on the slit throughout the exposure, which may occupy from one minute to several hours, the duration depending upon the brightness of the star.

We have seen that a single glance at the spectrum of a star is sufficient to give us important information as to the structure of its atmosphere, while a study of the position of the lines tells what chemical elements are present. We might go on to consider how the width and sharpness of the lines, together with shifts in their position toward the red end of the spectrum, furnish the means of estimating the density of the vapors and the pressure to which they are subjected. The relative intensities of certain lines also serve as a clue to the temperature. Thus in the spectrum of magnesium there is a pair of lines, one of which is the stronger at the temperature of the electric spark, while the other is the stronger at the lower temperature of the electric arc. In the spectra of certain stars the greater intensity of the first line indicates that the temperature is high and approximates that of the electric spark, while in other stars the relative intensities are reversed, indicating that the temperature is lower and corresponds more closely with that of the electric arc. In addition to all this, certain easily measurable changes in the position of the spectral lines are known from Doppler's principle to indicate motion of the star in the direction of the earth. Thus if the lines are shifted toward the red with reference to their normal position, and if we have evidence that the shift is not due to pressure, we may conclude that the distance between the earth. and the star is increasing, while if the lines are shifted toward the violet we conclude that the distance between the earth and the star is decreasing. As the earth's motion is known, the velocity of the star in the line of sight can therefore be accurately determined.

After this glance at the methods employed by the spectroscopist, we may return to a further consideration of the stages of stellar evolution. We have seen that the long continued action of gravity tends to produce condensation of a cosmical cloud. The constellation of Orion. contains many examples of stars in this early stage of development. As the mass condenses its temperature rises, and corresponding with this rise in temperature and in the density of the vapors which constitute the star we find characteristic changes in the spectrum and also in the star's color. Such a brilliant white or bluish-white star as Sirius or Vega may be taken as representative of the next stage of stellar development. Here the broad bands of hydrogen, which constitute a beautiful series expressible by a simple mathematical formula, serve as the chief mark of distinction. The conditions are not yet ripe for the marked development of metallic lines, though doubtless the numerous elements which constitute the sun and which for the most part are familiar to us on the earth, are present in such stars, though they are not revealed through a study of the spectrum. It is true that evidence exists of the presence of iron and a few other substances, but the lines are thin and few in number and would be overlooked in a VOL. LX.-20.

casual examination of the spectrum. The period for their greatest development has not yet arrived. The light gas hydrogen, reaching far above the white-hot mass of condensed vapors which constitutes the nucleus of the star, is at this stage the predominant element, at least so far as we may judge from a study of the light radiation.

An interesting question has arisen regarding the period in a star's life at which the highest temperature is attained. The apparently paradoxical statement of Lane's law that the temperature of a cooling mass of incandescent vapors, instead of falling, actually increases until a certain stage has passed, applies in the present instance. We indeed know that a condensing nebula losing heat by radiation into space will continue to rise in temperature for thousands and even millions of years. A question which has received some discussion of late is with regard to the precise period at which the maximum temperature occurs. Shall we seek it in white stars like Sirius or in yellow stars like the sun, which represents the next well-defined stage of stellar evolution? With an instrument of extraordinary delicacy Professor Nichols has recently measured at the Yerkes Observatory the amount of heat which we receive from Vega and Arcturus. The distance of these stars is so inconceivably great that the quantity of heat which they send to the surface of the earth has hitherto been too small to be detected by the most sensitive instruments. Professor Nichols' radiometer, which in combination with a large concave mirror renders it easy to measure the heat radiated from a man's face 2,000 feet away, proved adequate for the task. He found that Arcturus sends us about as much heat as we should get from a candle six miles away if there were no intervening atmosphere to reduce the candle's intensity. Vega, which to the eye is precisely equal to Arcturus in brightness, was found to send us only half as much heat. If the absorbing atmospheres of Arcturus and Vega were similar in character, it would follow from Professor Nichols' results that Vega, though it sends us less heat, is really the hotter of the two stars. For we know from laboratory experiments that the proportion of long (heat) waves to short (light) waves is greater in the radiation of the cooler of two bodies heated to incandescence. In this case the fact that Arcturus sends the greater amount of heat would be ascribed rather to greater size than to lesser distance, as there is good reason to believe that it is farther from us than Vega.

But unfortunately the dissimilarity of the atmospheres of the two stars renders it uncertain whether such conclusions can safely be drawn. This is particularly true in view of the fact that Sir William Huggins concludes from his spectroscopic studies that the highest stage of stellar temperature is reached in stars like Vega, while stars like Arcturus and the Sun have passed the stage of highest temperature and are already well advanced in their decline.

a

While some uncertainty must therefore prevail until further investigations have been completed regarding the exact stage at which the highest stellar temperatures are attained, there can be little doubt as to the path which is followed when through the long continued action of gravitation a young star like Vega develops into a star like the Sun. We are fortunate in possessing examples of a great number of intermediate stages in this orderly progress (Fig. 8). As condensation continues, and as the vapors which constitute the star continue to crowd upon each other, the stellar nucleus becomes denser and denser and the vast atmosphere of hydrogen gradually gives place to a much shallower atmosphere, in which hydrogen is still conspicuous, though it no longer predominates in a very striking manner over the other elements. The spectral lines of such elements as iron, magnesium, sodium and cal

cium, rise into prominence as the hydrogen lines fade. Meanwhile the light of the star undergoes a change of color, completely losing its bluish cast and assuming a distinctly yellow hue. There can be little if any doubt that our own sun once passed through the successive stages which are represented by the spectra shown in Fig. 8. The time which has elapsed since it acquired its present size and density as the result of the condensation of the great nebula in which the earth and the other planets also had their origin, covers many millions of years. It is fortunate for the study of stellar evolution that the stages through which the sun once passed are all exemplified in existing stars, which for unknown reasons began their stellar life at widely different times.

It will be profitable to consider for a moment some of the remarkable phenomena which are presented to us by the sun, not only because of their intrinsic interest, but also because it is perfectly safe to assume that similar phenomena, sometimes on a much greater scale, would be presented by other stars, were they not at so great a distance from the earth as to reduce them to mere points of light, even in the most power