201

SPECTROSCOPIC OBSERVATIONS OF THE SUN.

[109.

In the case of hydrogen the three characteristic bright bands of its ordinary spectrum become broader and broader as the intensity of the current is gradually increased, until the light becomes nearly continuous; the violet band expands first, whilst the red always preserves a certain definite character.

In the case of sodium, the bright lines at D gradually expand and coalesce, as the intensity of the current is increased; finally, this band expands, and becomes traversed by two black absorption lines coincident in position with the original bright lines.

Oxygen, chlorine, bromine, iodine, and phosphorus furnish only a single spectrum; these spectra are of the second order, and are figured in the memoir above cited. Arsenic and mercury also give spectra of the second order only. The general appearance of the first spectrum of a gas or vapour consists of broad luminous bands, generally nearly equidistant, and very similar one to the other throughout the particular spectrum. These bright bands are crossed by numberless dark lines, as well defined as those of Fraunhofer. The general character of the second spectrum is that of brilliant bands upon a more or less luminous ground. These bright bands have a breadth regulated by the width of the slit of the spectroscope.

Frankland and Lockyer (Proc. Roy. Soc. 1869, xvii. 288) investigated the spectra obtained by passing electric sparks from an induction coil through various gases at different pressures, and they find that under a pressure of 10 atmospheres hydrogen gives a perfectly continuous spectrum; on diminishing the pressure the spectrum gradually breaks up into lines, the red line corresponding to the Fraunhofer line c being the first to become definitely separated. At the atmospheric pressure the red line is sharply defined, but the green band corresponding to Fraunhofer's F line is wide and nebulous. On reducing the pressure (which was done by connecting the tube through which the sparks were passed with a Sprengel's pump and a pressure gauge) the green line became narrower and narrower until at a very low pressure the line was as thin as the red line. The breadth of the red line is also diminished during this process, but not nearly to the same extent as the green one.

(109 a) During the last few years many investigations have been prosecuted, and with great success, to determine the constitution of the sun. Astronomers have known for many years that, during the eclipses of the sun, red flames were observed protruding beyond the edges of the dark moon, and which are only visible during total eclipses, in consequence of the brilliancy of the white light of the sun and the very intense illumination of the earth's atmosphere. Various explanations have been given of these phenomena: by some they were supposed to belong to the sun, and by others to the moon. In the eclipse of 1860 it was

Such spectra of the first order are usually succeeded by those of the second order, when the secondary discharge is made through the tube, whilst at the same time a Leyden jar is used, one of the insulated secondary wires being connected with the outer coating and the other with the inner coating of the jar.

109 a.]

SPECTROSCOPIC OBSERVATIONS OF THE SUN.

205

shown by Mr. Warren de la Rue's photographs that these red protuberances belonged exclusively to the sun, as they were observed to remain stationary whilst the moon passed over the sun. The first systematic examination of these flames by the spectroscope was made during the eclipse of Aug. 18, 1868. Observations were conducted in India by Capt. Haig, Capt. Tanner, Mr. Kero Laxuman (Proc. Roy. Soc. 1868, xvii. 74), and Capt. Herschel (Proc. Roy. Soc. 1868, xvii. 104). Captain Haig saw in the spectrum of two of the red flames "two broad bright bands quite sharply defined, one rose-madder and the other light golden," which were soon lost in the spectrum of the moon's edge just before emergence, which had also two well-defined bright bands (one green and one indigo) about a quarter of the width of the bands in the spectra of the flames."

The spectrum of the prominence seen by Capt. Herschel consisted of "three vivid lines, red, orange, blue: no others, and no trace of a continuous spectrum." The orange line Capt. Herschel considered coincident with the line D, he believed that the blue line corresponded with F, and was even more doubtful with regard to the red line, which he thought might be c, or less likely B.

As early as 1866, Mr. J. Norman Lockyer devised a method by which the red flames might be examined at any time, and without waiting for a total eclipse, an event of rare occurrence and of such short duration as to render obserrations on the protuberances difficult or impossible except under favourable atmospheric conditions. Assuming that the red flames were caused by the light emitted by incandescent gases or vapours, and the spectrum of which might be presumed to consist of bright lines only, Mr. Lockyer thought that, by using a spectroscope of very great dispersive power, it might be possible to elongate the spectrum of the white light of the sun and of the brilliantly illuminated terrestrial atmosphere to such an extent that the intensity of the light would be so much diminished as to enable the observer to see the spectrum of the incandescent gases. The difficulty of obtaining a spectroscope of sufficient power, and other causes, prevented the realization of this ingenious idea until October 20, 1868; but in a short letter of this date, received by the secretary of the Royal Society on October 21 (Proc. Roy. Soc. xvii. 91), Mr. Lockyer says that he had perfectly succeeded in obtaining and observing part of the spectrum of a solar prominence. He established the existence of three bright lines in the following positions:-1, a red line absolutely coincident with c; 2, a green line nearly coincident with F; and 3, a yellow line near D. The last was more refrangible than the more refrangible of the two darkest D lines by eight or nine degrees of Kirchhoff's scale.

The application of this method of observation had, however, been previously made by Janssen. This astronomer had observed the eclipse of August 18 in India, and the following day he succeeded by means of a powerful spectroscope in seeing the spectrum of a prominence. The account of this observation was sent to the Academy of Sciences in Paris, and arrived and was read on the same day as Mr. Lockyer's description of his later observation of October 20: this is perhaps one of the most curious coincidences presented by the annals of modern science, in which two investigators have published the same result at the same time, though working independently, one in England and the other in India. An account of Mr. Janssen's research was published in a letter to Mr. Warren de la Rue, and printed in the Proc. Roy. Soc. 1869, xvii. 276.

It may be well to describe in this place the construction of the apparatus employed by Mr. Lockyer for the investigation of the spectrum of the solar prominences (Phil. Trans. 1869, 425). An equatorial telescope of 6 inches aperture, and 9 feet focal length, is used to condense the light of the sun. From the telescope the eyepiece is removed and a strong support adapted to the

206

SPECTROSCOPIC OBSERVATIONS OF THE SUN.

[1.09 a. tube of the telescope; this support carries a spectroscope containing seven prisins, the slit at the end of the collimator of the spectroscope being placed in the primary focus of the object glass of the telescope so that an image of the sun nearly an inch in diameter is thrown on the slit. By moving the telescope, so as to alter the position of the image, the observer is enabled to examine the light emanating from any portion of the sun, and to place the slit radially or tangentially to the image of the disk. The refraction produced by the seven prisms is so great that the beam of light is deflected through about 270°, so that the observing telescope of the spectroscope is nearly at right angles to the direction of the collimator. When the slit is placed radially to the disk of the sun, so that the edge of the disk is projected on a portion of the slit, a continuous spectrum with black Fraunhofer's lines is seen on looking through the telescope of the spectroscope, and if the atmospheric conditions are very favourable, and there is no haze perceptible round the sun, the spectrum of the sky will scarcely be visible. If it should happen that the image of a prominence falls on the slit the bright lines of its spectrum will be seen as the prolongations of the black lines in the spectrum of the sun's limb. In this manner Mr. Lockyer has shown that the sun is surrounded in every part by a luminous envelope about 800 miles in depth, which he calls the chromosphere, and of which the prominences or red flames are merely elevated portions. The bright lines constituting the spectrum of the chromosphere indicate that this gaseous envelope consists principally of hydrogen, the most prominent lines being a red one corresponding to the black line c, a green line corresponding to F, a blue one corresponding to a black line near G, and a violet one corresponding with h.

By this mode of observation it is obvious that it is only possible to obtain sections, so to speak, of the prominences, but by moving the telescope to a small extent, so as to bring the different parts of the image of the prominence on to the slit, the observer is enabled to obtain valuable information as to the actual form of the protuberance under examination. It was evidently desirable to be able to obtain an actual view of the red flames, and this was accomplished by Mr. Huggins by placing a piece of ruby glass over the slit of the spectroscope in order to absorb the more refrangible rays, and on now opening the slit of the spectroscope to a sufficient extent, the prominence itself was seen. (Proc. Roy. Soc. 1868, xvii. 302.) Mr. Lockyer afterwards found that the red glass might be dispensed with, and that by so placing the telescope that the image of the limb of the sun only just touched the end of the slit when the latter was placed radially, or just touched the edge when it was arranged tangentially, the slit might be safely opened so as to give a beautiful red image of the prominence, and without any apprehension of injury to the eye by the brilliancy of the light. The same fact was independently discovered by Capt. Herschel. The red flames are sometimes of prodigious height; the one observed by Mr. Lockyer on October 20, subtended an angle of 3' or was about 80,000 miles high. The protuberances are undergoing constant and rapid motion, and by the help of the mode of observation above described, the astronomer is able to watch the alterations of the form of these strange phenomena as easily as those of the clouds in the terrestrial atmosphere. At 11.5 A M., on March 14, 1869, Mr. Lockyer made a sketch of a protuberance 27,000 miles in height, and consisting of an irregular-shaped cloud and a long perfectly straight one. At 11.15 the straight portion had entirely disappeared, having been apparently absorbed by the irregular cloud.

Mr. C. A. Young afterwards described in the Boston Journal of Chemistry a still more wonderful exhibition of solar prominences. (Nature, 1871, iv. 488.) At 12.30 on September 7, 1871, Mr. Young saw a prominence about 54,000 miles high, and 100,000 miles long, made up of filaments

109 a.]

SPECTROSCOPIC OBSERVATIONS OF THE SUN.

207

mostly horizontal, and floating above the sun, with its lower surface about 15,000 miles above the chromosphere and connected to it by vertical columns. On again examining the protuberance at 12.55, the whole had been completely blown to shreds; instead of the vertical pillars, masses of detached filaments were now seen 100,000 miles above the sun, and rapidly ascending. In ten minutes the highest portions of the masses were no less than 200,000 miles from the surface of the sun, having travelled upwards at the rate of 156 miles per second. The cloud then rapidly disappeared as it cooled, and at 1.15 scarcely anything was left to indicate where this violent action or explosion took place.

FIG. 86 a.

The shape of the lines of the spectrum of a prominence is a matter of considerable interest, and one that gives important information respecting the condition of the solar atmosphere. When the spectrum is examined with a narrow slit the red hydrogen line is seen to rest on the black absorption line of the limb of the sun; the base of this line is a little, but not much thicker than the black line; as the distance from the sun increases, the line generally becomes thinner and thinner until it disappears (fig. 86 a). The case is very different with the green line corresponding to F; this is often pointed at the end, but on approaching the sun it becomes broader and broader until at the place where it rests on the edge of the solar spectrum it has widened out like the mouth of a trumpet to five or six times the breadth of the black line, which usually is pointed, and disappears in the base of the bright line (fig. 86 b). This thickening of the line is due to pressure (109), for as the edge of the sun is approached the chromosphere becomes

more and more dense, exactly as is the case with the atmosphere of the earth. In cases where violent action is going on in a prominence, the green line is thickened at various points above the surface of the sun, indicating regions of greater pressure (fig. 86 c).

Besides these lines known to be due to hydrogen there is one in the yellow a little more refrangible than the more refrangible of the D lines, and probably the one noticed by Captain Herschel during the eclipse of August, 1868, and which he thought was due to sodium. This line is almost invariably present in the spectrum of the prominences, and was believed by Lockyer to be due to hydrogen; but after numerous trials under very various conditions, Frankland and Lockyer were unable to obtain the faintest indication of it in the spectrum

FIG. 86 b.

of hydrogen. Mr. Lockyer has since observed that, though generally present in the spectrum of the prominences, the line exhibits certain peculiarities indicating that it is not one of those of the hydrogen spectrum. It appears to belong to an element with a vapour density even less than that of hydrogen, and one which has not hitherto been recognised by chemists among the constituents of the surface of the earth; sometimes, but very rarely, an absorption line corresponding to the bright one is seen in the solar spectrum. Occasionally, during periods of intense action on the sun, the vapours of other elements are projected into the chromosphere: on one occasion, on the 17th of April, 1870, Mr. Lockyer saw hundreds of the Fraun

208

SPECTROSCOPIC OBSERVATIONS OF THE SUN.

[109 a. hofer lines beautifully bright at the base of a prominence. The vapours of the elements producing these appearances are only projected to a short distance above the photosphere, and very seldom indeed are seen in clouds floating above the chromosphere. Mr. Lockyer has recognised the spectra of the following metals exhibiting bright lines, sodium, barium, magnesium, iron. One of the most curious points connected with these bright spectra is that only some of the lines are thus seen; and of these, some are longer than others; thus, in the group

known as b, of which three of the lines are due to magnesium, b, and b, are of equal length, while b is much shorter. Fig. 86 d. Frankland and Lockyer have, however, shown that on examining the spectrum of magnesium obtained by means of the spark of the induction coil between magnesium poles, placed at a short distance from one another, the three b lines in the green extend to unequal distances from the metallic points. Of the 460 lines in the spectrum of iron only a few have been observed in the chromosphere, but this is quite in accordance with the laboratory experiments, in which it was found that some of the lines of hy. drogen and nitrogen disappeared when the density was very much reduced, the same effect probably takes place in the case of other bodies, and the low pressure on the surface of the sun is perhaps sufficient to reduce the complicated spectrum of iron to a much more simple one.

One of the most interesting phenomena observed on examining the light emitted by the solar prominences is the change of wave length due to the motion of the incandescent hydrogen. As will be afterwards explained (114), when a luminous object recedes from the observer, the wave length is increased, and consequently the ray becomes less refrangible; on the other hand, when the object approaches the observer, the wave length is diminished and the refrangibility increased. If the rate of motion of the object is not insensibly small in comparison with the velocity of light, the alteration of the refrangibility may be so great as to produce a distinct alteration of the position of the lines in the spectrum. If the motion of the luminous object is in a direction at right angles to a line extending from the object to the observer, of course no change of wave length will FIG. 86 g.

take place; consequently, on observing the spectrum of the chromosphere, or of a prominence rising from the limb of the sun, the bright lines of the spectrum