Prominences are merely portions of the chromosphere shot up to great heights above the sun. Since we are able to photograph the prominences without waiting for an eclipse, would it not be possible to obtain a photograph of the chromosphere, even of the portion that lies closest to the photosphere, the reversing layer, which produces the flash spectrum? Young many years ago observed the brighter chromosphere lines without an eclipse. Evi dently it would be quite possible to photograph the flash spectrum provided that during the ex posur.e the slit could be held pointed at the reversing layer, and the photosphere not be allowed to encroach. An idea can be obtained of the accuracy demanded in pointing the slit to the proper region on the sun when it is realized that the sun is 866,000 miles in diam eter, while the reversing layer is less than 400 miles in thickness. To photograph the flash spectrum without an eclipse, there is hence needed a large image of the sun, and some con venient guiding device. Hale and Adams, using a very ingenious arrangement and the 60-foot tower telescope, have obtained the only photo graphs of the flash spectrum without an eclipse. Using a dispersion of 1 mm equal 0.9 Ang stroms, excellent photographs have been ob tained which show about as many lines as se cured by Mitchell with an accuracy of wave length about the same as that obtained by him.
Success, however, has not followed the at tempts to photograph the corona in full sun light without an eclipse. This was tried as early as 1885 by Sir William Huggins. In 1893 and 1894, Ricc6 at Mount Etna, and Hale at Pike's Peak, attempted to photograph the corona, both by direct photography and also by the spec troheliograph, which has been so successful in obtaining photographs of the prominences. Failure to reach results, turned Hale's attention to the bolometer joined up with a sensitive gal vanometer, and many experiments were made at Yerkes Observatory between the years 1895 and 1900 (Astrophysical Journal, I, 372, 1900). The cause of the failures is shown from the measures at the eclipse of 1908 by Abbot with the aid of the bolometer. The work of Abbot ( 'Sun,' p. 132, 1911) shows that the daylight sky as seen from sea-level even 20° from the sun is fully 10 times as bright as the corona in its brightest parts close to the sun's limb. Close to the sun's edge the daylight sky is so in tensely brilliant compared with the corona, that no method yet attempted, even when tried at a mountain top to diminish the atmospheric glare, has had the slightest amount of success.
Remarkable work has been done by Evershed in measuring the radial motion of matter form ing sun-spots. With the better equipment at Mount Wilson, St. John has carried out Ever shed's work more in detail. His researches (Astrophysical Journal, XXXVIII, 341, 1913) have been supplemented by the determination of heights by Mitchell from observations at the time of an eclipse. These results show that the different displacements of the Fraunhofer lines at the edges of sun-spots seem to find their simplest explanation in movements of the solar vapors tangential to the solar surface with velocities varying with the elevation : large out flowing displacements are taken by vapors that exist at low heights while large inflowing mo tions are taken by vapors like calcium and hydrogen which exist at great heights. This excellent work of St. John s gives a very clear insight into the physical constitution of sun spots. The lighter materials of the chromo sphere which cause the darkest lines in the sun's spectrum, hydrogen, helium, calcium and sub stances forming the stronger enhanced lines, flow in to the spots, the higher the elevation above the photosphere the greater the speed of the in-flow. This in-flow of matter is partly bal anced by the out-flow of matter from the spot, matter at the least heights above the photo sphere flowing with the greatest velocities. The spot vortex, therefore, resembles in appearance that of an inverted water-spout seen in a ter restrial storm. In addition to giving great in formation regarding sun-spots, St. John's work coupled with eclipse observations has given the means of measuring the depths of the layers of interpreted only by assuming differences of elevations in the outer chromosphere. We therefore seem to have more information con cerning what is happening at different elevations in the sun's atmosphere than what is taking place in the earth's atmosphere in spite of the great improvements in methods of observing meteorological phenomena.
Motion in the Line of. One of the greatest triumphs of astrophysics has been the measurement of radial velocity, or motion in the line of sight, in linear units of miles or kilometres per second. The Doppler-Fizeau principle is easy to understand. Light is a wave phenomenon. When the source of light and the observer are approaching each other, either by a motion of the object, or the observer, or both combined, then the number of waves which reach the eye per second is increased, while if the observer and object are moving from each other, fewer light waves reach the eye per sec ond. If more waves are met, the wave-length must be shortened. The position of a line in the spectrum is known by its wave-length, and, if in consequence of motion, the wave-length is shortened, the position of the line is shifted toward the region of short wave-lengths, the violet. The measurement of the amount of the shift gives the means of determining the velocity in the line of sight. Consequently, a shift toward the violet means a motion of the object toward the observer (the motion taking place either in the object, or in the observer, or both), while a shift toward the red means a motion increasing the distance between ob server and object. This is the Doppler Fizeau principle. All are familiar with this same principle in sound, in the sudden drop in the pitch of the whistle of a locomotive as a train going in the opposite direction is met and passed. The motion in the line of sight corre sponding to a shift PA of a spectral line of wave-length X is given by the formula A?' A = •-• Where \IA is the velocity of the star relative to the observer, and is the velocity of light. If the velocity of light is measured in miles per second (186,300), the velocity is known in miles per second. In scientific work veloci ties are usually determined in kilometres per second, when the velocity of light (in round numbers) is 300,000. One can easily see, there fore, the accuracy necessary in the measure ment of spectrograms in order to attain knowl edge of radial motions of stars to a given de gree of refinement. If, for instance, we wish to know the velocity in the line of sight ac curately within one kilometre per second, it will be necessary to determine wave-lengths with a precision of one part in 300,000. For a line of wave-lengths X 4500, it is necessary to measure the shift in wave-lengths within 0.015 Angstroms. This accuracy is reached in the best work done with our best spectrographs, when the spectrum of the star being invcsti gated consists of fine, clean-cut lines. With a slit spectrograph a comparison spectrum is placed alongside the stellar spectrum. From the known wave-lengths of the comparison lines the wave-lengths of the stellar lines may be found, and these, compared with the normal wave-lengths, give the shift due to the motion of the star in the line of sight. The grandest application is seen in Keeler's proof (Astro physical Journal, I,.416 1895) of the meteoric constitution of Saturn s rings. If meteoric, the linear motion of the rings will be greatest nearest the planet and decrease outward; if solid the rings will rotate as a whole, all par ticles having the same angular motion, the linear speed increasing from centre to circum ference. With solid rings and a slit placed across the planet's equator, the lines on the side moving toward us would be shifted toward the region of short wave-lengths, the shift being proportional to the linear motion, while for the side moving away from us the lines would be shifted in the opposite direction. Thus, on account of the gradual increase in linear motion from centre to circumference, the lines would be gradually shifteoi, in the complete spectrum having the effect of lines slightly inclined. Such, however, is not the appearance of the lines photographed by Keeler, and these can be ex plained only under the assumption that the rings are a collection of small satellites, giving, therefore, a direct confirmation of the mathe matical theory of Maxwell.