above the photosphere is as great as 4,000 miles. The bright lines are identified as belonging to iron, titanium, chromium, hydrogen and other elements. The origin of some of the lines is unknown.

Although no other time may be so favorable for the study of the reversing layer as at total eclipses, the chromosphere and prominences may nevertheless be well studied on any clear day.

In connection with the eclipse of 1868 Janssen and Lockyer each independently discovered that by spectroscopic means the light of the chromosphere and prominences may be so separated from that of the sky as to become visible without an eclipse. The light from the region. just outside the sun's limb is composed of skylight and the light of the solar atmosphere. Each is about equally bright. When this combined light is passed through a prism, that due to the sky is spread out into a continuous surface, thus becoming much fainter, while that due to the chromosphere or prominence, from its gaseous nature, is collected into bright bands, which thus surpass the skylight in intensity and may be seen or photographed. This line of work has been

greatly extended by different scientists, notably by Hale, of this country, who, by a device known as the spectro-heliograph, has succeeded in making, without an eclipse, photographs showing all the prominences surrounding the sun and the details of the solar surface at the same time. These photographs are made in monochromatic light. They represent what would be seen if the eye were sensitive to light of the wave-length of the K line only. Figures 3 and 4 show a great eruptive prominence photographed by Professor Hale, March 25, 1895. The interval between the two photographs was 24 minutes, during which time the prominence was thrown upward from a height of

135,000 miles to 281,000 miles. This implies a velocity of at least 100 miles per second.

At times of total eclipse it is perhaps possible to obtain better photographs showing finer details than can be made under other conditions. Figure 5 is an enlargement of a photograph made at the eclipse of 1900, by Professor E. E. Barnard, assisted by Mr. G. W. Ritchey. It shows a mass of prominences at the southwest quadrant of the sun. Along the irregular limb of the moon, which appears black, is seen the ragged storm-tossed surface of the chromosphere, of increasing depth toward the right owing to the moon's position at the instant of the exposure. Thrown up from this are the vast fantastic masses of the prominences or 'red flames.' They remind us of pictures which show the effects produced by the explosion of submarine torpedoes. The larger mass at the left rises to the height of 60,000

FIG. 5. SOLAR PROMINENCES. ECLIPSE OF MAY 28, 1900. PHOTOGRAPHED WITH A TELESCOPE OF 6 INCHES APERTURE AND 62 FEET FOCUS, BY PROFESSOR BARNARD AND MR. RITCHEY.

miles. This photograph was made with a telescope of only six inches aperture and six and a half feet focal length, a small instrument compared with some which have been used at recent eclipses. The writer has seen no other photograph of prominences, however, which, in delicacy of detail, surpasses the one here shown.

The single feature of a total eclipse which can be seen and studied only at such times is the corona. In early ages small mention was made of the corona. Apparently the dread of impending evil overwhelmed man, and prevented careful observations. As fear disappeared and scientific interest grew, attention was drawn to the 'red flames,' and at nearly the same time to the beautiful halo of light which has been fittingly named the 'corona.' Since that time the favorable moments of totality have been too few to clear up the mystery of its nature. Reasoning from the methods which have made the study of the chromosphere and prominences possible without an eclipse, various attempts have been also made to thus observe and photograph the corona. The simplest way would be by direct vision or photography. There is no doubt but that, if we could remove for a moment the earth's atmosphere, whose glare interferes with our vision, we

should be able to see the chromosphere, prominences and corona without any artificial aid. The brightness of the inner corona is about the same as that of the ordinary sky near the sun. If then one could find

FIG. 7. SOLAR CORONA. ECLIPSE OF 1930. NEAR SUNSPOT MINIMUM. MADE BY MR. C A. R. LUNDIN AT SOUTHERN PINES, N. C.

a locality where the sky was extraordinarily clear, he might hope, by placing a shield in front of the sun itself, to see these fainter features. The writer of this article made an attempt several years ago in this

way on the summit of El Misti, Peru, at an elevation of 19,200 feet. At this altitude one-half the earth's atmosphere is below the observer and that which remains is of extraordinary clearness. Photographs were made of the region immediately about the sun, using an opaque disc to protect the plate from the sun's direct image. The true corona did not appear upon the plates. Other methods promised better results, such as the use of monochromatic light, presumably that of the line 'K 1474.' Experiments in this line have been carried on by Professor Hale with skill and enthusiasm on the summit of Pike's Peak, on Mount Etna and elsewhere, but without success. He has also attempted to solve the difficulty by a study of the heat, using the bolometer. Recent investigations given below explain the failure

FIG. 3. SOLAR CORONA. ECLIPSE OF 1893. NEAR SUNSPOT MAXIMUM. MADE BY PROFESSOR J. M. SCHAEBERLE, LICK OBSERVATORY.

of this method. The polarization of the coronal light also suggests a method which has not yet yielded successful results. Although the future may furnish the solution, none of the attempts yet made has been successful, and for the present our only knowledge of the corona must be obtained from what can be learned during the brief moments of total eclipses. Good photographs of the corona can be easily and rapidly made and if an abundance of these were alone necessary our knowledge would be well advanced. The general features of the corona have a certain permanence. Comparatively slight changes are known to take place during the three or four hours while an eclipse is passing over the surface of the earth. There may be, however, finer details than are shown on the best photographs yet obtained, which

would give witness to more rapid changes. From year to year large changes in the form of the corona occur and these appear to be associated with the sun-spot period. This is a natural inference, especially since the solar prominences are thus associated. This is well shown. by a comparison of the form of the corona in 1889 and 1900, which occurred near the sun-spot minimum, with the form in 1893, which was near sun-spot maximum. These are given in Figures 6, 7 and 8. The equatorial streamers and the divergent polar streamers are much more pronounced at the time of sun-spot minimum. At maximum the corona is more nearly circular. The polar streamers are beautifully shown in Figure 9, a photograph made by the eclipse party, which was under the direction of Secretary Langley, of the Smithsonian Institution. The true nature of the corona and the complex changes which it undergoes are unknown. The spectroscope is the magician's wand which science generally uses to reveal the constitution of unknown objects, but in this case the revelation is only partial. In 1869 Professor Young found the spectrum to be characterized by a bright line in the green, which he identified as Kirchhoff's line 1474. The unknown substance which produces this line has been given the name 'coronium.' There are also other less conspicuous bright lines. When the name 'helium' was assigned to the origin of certain lines in the solar spectrum, no such terrestrial substance was known. Later it was found by Ramsay. A similar issue for coronium would be very acceptable. The corona also yields a faint continuous spectrum, in which Janssen and others have reported certain dark lines of the solar spectrum. This signifies, that in addition to luminous gases, giving a spectrum of bright lines, the corona contains some substance, like a cloud, which is capable of reflecting ordinary sunlight. A part of the light appears to be polarized. It is thought by some observers that there is also a bright continuous spectrum free from dark lines. If true, this would imply a three-fold origin to the coronal light. For the explanation of the corona we have the diffraction theory of Hastings, the mechanical theory of Schaeberle, the magnetic theory of Bigelow, and others. The complete solution of the problem is of the greatest difficulty and of the greatest importance. At the eclipse of 1900 some experiments with that remarkable instrument, the bolometer, appear to throw new light on this subject. These experiments were made by Secretary Langley's chief assistant, Mr. C. I. Abbott, who reached the following conclusions:

These observations indicate not only that the coronal radiation is very slight, but that the apparent temperature of the inner corona is below 20° C. For it will be noticed that the bolometer lost heat by radiation to the corona, as evidenced by a negative deflection. Hence, when we consider its visual photometric brightness at the point where the bolometric measures were taken,