Chapter III: Recent Solar Eclipses (1)

By observations made during a series of five remarkable eclipses, comprised within a period of eleven years, knowledge of the solar surroundings was advanced nearly to its present stage. Each of these events brought with it a fresh disclosure of a definite and unmistakable character. We will now briefly review this orderly sequence of discovery.

Photography was first systematically applied to solve the problems presented by the eclipsed sun, July 18, 1860. It is true that a daguerreotype,[512] taken by Berkowski with the Königsberg heliometer during the eclipse of 1851, is still valuable as a record of the corona of that year; and some subsequent attempts were made to register partial phases of solar occultation, notably by Professor Bartlett at West Point in 1854;[513] but the ground remained practically unbroken until 1860.

In that year the track of totality crossed Spain, and thither, accordingly, Warren de la Rue transported his photo-heliograph, and Father Secchi his six-inch Cauchoix refractor. The question then primarily at issue was that relating to the nature of the red protuberances. Although, as already stated, the evidence collected in 1851 gave a reasonable certainty of their connection with the sun, objectors were not silenced; and when the side of incredulity was supported by so considerable an authority as M. Faye, it was impossible to treat it with contempt. Two crucial tests were available. If it could be shown that the fantastic shapes suspended above the edge of the dark moon were seen under an identical aspect from two distant stations, that fact alone would annihilate the theory of optical illusion or "mirage"; while the certainty that they were progressively concealed by the advancing moon on one side, and uncovered on the other, would effectually detach them from dependence on our satellite, and establish them as solar appendages.

Now both these tests were eminently capable of being applied by photography. But the difficulty arose that nothing was known as to the chemical power of the rosy prominence-light, while everything depended on its right estimation. A shot had to be fired, as it were, in the dark. It was a matter of some surprise, and of no small congratulation, that, in both cases, the shot took effect.

De la Rue occupied a station at Rivabellosa, in the Upper Ebro valley; Secchi set up his instrument at Desierto de las Palmas, about 250 miles to the south-east, overlooking the Mediterranean. From the totally eclipsed sun, with its strange garland of flames, each observer derived several perfectly successful impressions, which were found, on comparison, to agree in the most minute details. This at once settled the fundamental question as to the substantial reality of these objects; while their solar character was demonstrated by the passage of the moon _in front_ of them, indisputably attested by pictures taken at successive stages of the eclipse. That forms seeming to defy all laws of equilibrium were, nevertheless, not wholly evanescent, appeared from their identity at an interval of seven minutes, during which the lunar shadow was in transit from one station to the other; and the singular energy of their actinic rays was shown by the record on the sensitive plates of some prominences invisible in the telescope. Moreover, photographic evidence strongly confirmed the inference--previously drawn by Grant and others, and now with fuller assurance by Secchi--that an uninterrupted stratum of prominence-matter encompasses the sun on all sides, forming a reservoir from which gigantic jets issue, and into which they subside.

Thus, first-fruits of accurate knowledge regarding the solar surroundings were gathered, while the value of the brief moments of eclipse gained indefinite increase, by supplementing transient visual impressions with the faithful and lasting records of the camera.

In the year 1868 the history of eclipse spectroscopy virtually began, as that of eclipse photography in 1860; that is to say, the respective methods then first gave definite results. On the 18th of August, 1868, the Indian and Malayan peninsulas were traversed by a lunar shadow producing total obscuration during five minutes and thirty-eight seconds. Two English and two French expeditions were despatched to the distant regions favoured by an event so propitious to the advance of knowledge, chiefly to obtain the verdict of the prism as to the composition of prominences. Nor were they despatched in vain. An identical discovery was made by nearly all the observers. At Jamkandi, in the Western Ghauts, where Lieutenant (now Colonel) Herschel was posted, unremitting bad weather threatened to baffle his eager expectations; but during the lapse of the critical five and a half minutes the clouds broke, and across the driving wrack a "long, finger-like projection" jutted out over the margin of the dark lunar globe. In another moment the spectroscope was pointed towards it; three bright lines--red, orange, and blue--flashed out, and the problem was solved.[514] The problem was solved in this general sense, that the composition out of glowing vapours of the objects infelicitously termed "protuberances" or "prominences" was no longer doubtful; although further inquiry was needed for the determination of the particular species to which those vapours belonged.

Similar, but more complete observations were made, with less atmospheric hindrance, by Tennant and Janssen at Guntoor, by Pogson at Masulipatam, and by Rayet at Wha-Tonne, on the coast of the Malay peninsula, the last observer counting as many as nine bright lines.[515] Among them it was not difficult to recognise the characteristic light of hydrogen; and it was generally, though over-hastily, assumed that the orange ray matched the luminous emissions of sodium. But fuller opportunities were at hand.

The eclipse of 1868 is chiefly memorable for having taught astronomers to do without eclipses, so far, at least, as one particular branch of solar inquiry is concerned. Inspired by the beauty and brilliancy of the variously tinted prominence-lines revealed to him by the spectroscope, Janssen exclaimed to those about him, "Je verrai ces lignes-là en dehors des éclipses!" On the following morning he carried into execution the plan which formed itself in his brain while the phenomenon which suggested it was still before his eyes. It rests upon an easily intelligible principle.

The glare of our own atmosphere alone hides the appendages of the sun from our daily view. To a spectator on an airless planet, the central globe would appear attended by all its splendid retinue of crimson prominences, silvery corona, and far-spreading zodiacal light projected on the star-spangled black background of an absolutely unilluminated sky. Now the spectroscope offers the means of indefinitely weakening atmospheric glare by diffusing a constant amount of it over an area widened _ad libitum_. But monochromatic or "bright-line" light is, by its nature, incapable of being so diffused. It can, of course, be _deviated_ by refraction to any extent desired; but it always remains equally concentrated, in whatever direction it may be thrown. Hence, when it is mixed up with continuous light--as in the case of the solar flames shining through our atmosphere--it derives a _relative_ gain in intensity from every addition to the dispersive power of the spectroscope with which the heterogeneous mass of beams is analysed. Employ prisms enough, and eventually the undiminished rays of persistent colour will stand out from the continually fading rainbow-tinted band, by which they were at first effectually veiled.

This Janssen saw by a flash of intuition while the eclipse was in progress; and this he realised at 10 A.M. next morning, August 19, 1868--the date of the beginning of spectroscopic work at the margin of the unobscured sun. During the whole of that day and many subsequent ones, he enjoyed, as he said, the advantage of a prolonged eclipse. The intense interest with which he surveyed the region suddenly laid bare to his scrutiny was heightened by evidences of rapid and violent change. On the 18th of August, during the eclipse, a vast spiral structure, _at least_ 89,000 miles high, was perceived, planted in surprising splendour on the rim of the interposed moon. If was formed as General Tennant judged from its appearance in his photographs, by the encounter of two mounting torrents of flame, and was distinguished as the "Great Horn." Next day it was in ruins; hardly a trace remained to show where it had been.[516] Janssen's spectroscope furnished him besides with the strongest confirmation of what had already been reported by the telescope and the camera as to the continuous nature of the scarlet "sierra" lying at the base of the prominences. Everywhere at the sun's edge the same bright lines appeared.

It was not until the 19th of September that Janssen thought fit to send news of his discovery to Europe. It seemed little likely to be anticipated; yet a few minutes before his despatch was handed to the Secretary of the Paris Academy of Sciences, a communication similar in purport had been received from Sir Norman Lockyer. There is no need to discuss the narrow and wearisome question of priority; each of the competitors deserves, and has obtained, full credit for his invention. With noteworthy and confident prescience, Lockyer, in 1866, before anything was yet known regarding the constitution of the "red flames," ordered a strongly dispersive spectroscope for the express purpose of viewing, apart from eclipses, the bright-line spectrum which he expected them to give. Various delays, however, supervened, and the instrument was not in his hands until October 16, 1868. On the 20th he picked up the vivid rays, of which the presence and (approximately) the positions had in the interim become known. But there is little doubt that, even without that previous knowledge, they would have been found; and that the eclipse of August 18 only accelerated a discovery already assured.

Sir William Huggins, meanwhile, had been tending towards the same goal during two and a half years in his observatory at Tulse Hill. The principle of the spectroscopic visibility of prominence-lines at the edge of an uneclipsed sun was quite explicitly stated by him in February, 1868,[517] and he devised various apparatus for bringing them into actual view; but not until he knew where to look did he succeed in seeing them.

Astronomers, thus liberated, by the acquisition of power to survey them at any time, from the necessity of studying prominences during eclipses, were able to concentrate the whole of their attention on the corona. The first thing to be done was to ascertain the character of its spectrum. This was seen in 1868 only as a faintly continuous one; for Rayet, who seems to have perceived its distinctive bright line far above the summits of the flames, connected it, nevertheless, with those objects. On the other hand, Lieutenant Campbell ascertained on the same occasion the polarisation of the coronal light in planes passing through the sun's centre,[518] thereby showing that light to be, in whole or in part, reflected sunshine. But if reflected sunshine, it was objected, the chief at least of the dark Fraunhofer lines should be visible in it, as they are visible in moonbeams, sky illumination, and all other sun-derived light. The objection was well founded, but was prematurely urged, as we shall see.

On the 7th of August, 1869, a track of total eclipse crossed the continent of North America diagonally, entering at Behring's Straits, and issuing on the coast of North Carolina. It was beset with observers; but the most effective work was done in Iowa. At Des Moines, Professor Harkness of the Naval Observatory, Washington, obtained from the corona an "absolutely continuous spectrum," slightly less bright than that of the full moon, but traversed by a single green ray.[519] The same green ray was seen at Burlington and its position measured by Professor Young of Dartmouth College.[520] It appeared to coincide with that of a dark line of iron in the solar spectrum, numbered 1,474 on Kirchhoff's scale. But in 1876 Young was able, by the use of greatly increased dispersion, to resolve the Fraunhofer line "1474" into a pair, the more refrangible member of which he considered to be the reversal of the green coronal ray.[521] Scarcely called in question for over twenty years, the identification nevertheless broke down through the testimony of the eclipse-photographs of 1898. Sir Norman Lockyer derived from them a position for the line in question notably higher up in the spectrum than that previously assigned to it. Instead of 5,317, its true wave-length proved to be 5,303 ten millionths of a millimetre;[522] nor does it make any show by absorption in dispersed sunlight. The originating substance, designated "coronium," of which nothing is known to terrestrial chemistry, continues luminous[523] at least 300,000 miles above the sun's surface, and is hence presumably much lighter even than hydrogen.

A further trophy was carried off by American skill[524] sixteen months after the determination due to it of the distinctive spectrum of the corona. The eclipse of December 22, 1870, though lasting only two minutes and ten seconds, drew observers from the New, as well as from the Old World to the shores of the Mediterranean. Janssen issued from beleaguered Paris in a balloon, carrying with him the _vital parts_ of a reflector specially constructed to collect evidence about the corona. But he reached Oran only to find himself shut behind a cloud-curtain more impervious than the Prussian lines. Everywhere the sky was more or less overcast. Lockyer's journey from England to Sicily, and shipwreck in the _Psyche_, were recompensed with a glimpse of the solar aureola during _one second and a half_! Three parties stationed at various heights on Mount Etna saw absolutely nothing. Nevertheless important information was snatched in despite of the elements.

The prominent event was Young's discovery of the "reversing layer." As the surviving solar crescent narrowed before the encroaching moon, "the dark lines of the spectrum," he tells us, "and the spectrum itself, gradually faded away, until all at once, as suddenly as a bursting rocket shoots out its stars, the whole field of view was filled with bright lines more numerous than one could count. The phenomenon was so sudden, so unexpected, and so wonderfully beautiful, as to force an involuntary exclamation."[525] Its duration was about two seconds, and the impression produced was that of a complete reversal of the Fraunhofer spectrum--that is, the substitution of a bright for every dark line.

Now something of the kind was theoretically necessary to account for the dusky rays in sunlight which have taught us so much, and have yet much more to teach us; so that, although surprising from its transitory splendour, the appearance could not strictly be called "unexpected." Moreover, its premonitory symptom in the fading out of these rays had been actually described by Secchi in 1868,[526] and looked for by Young as the moon covered the sun in August 1869. But with the slit of his spectroscope placed _normally_ to the sun's limb, the bright lines gave a flash too thin to catch the eye. In 1870 the position of the slit was _tangential_--it ran along the shallow bed of incandescent vapours, instead of cutting across it: hence his success.

The same observation was made at Xerez de la Frontera by Mr. Pye, a member of Young's party; and, although an exceedingly delicate one, has since frequently been repeated. The whole Fraunhofer series appeared bright (omitting other instances) to Maclear, Herschel, and Fyers in 1871, at the beginning or end of totality; to Pogson, at the break-up of an annual eclipse, June 6, 1872; to Stone at Klipfontein, April 16, 1874, when he saw "the field full of bright lines."[527] But between the picture presented by the "véritable pluie de lignes brilliantes,"[528] which descended into M. Trépied's spectroscope for three seconds after the disappearance of the sun, May 17, 1882, and the familiar one of the dark-line solar spectrum, certain differences were perceiving, showing their relation to be not simply that of a positive to a negative impression.

A "reversing layer," or stratum of mixed vapours, glowing, but at a lower temperature than that of the actual solar surface, was an integral part of Kirchhoff's theory of the production of the Fraunhofer lines. Here it was assumed that the missing rays were stopped, and here also it was assumed that the missing rays would be seen bright, could they be isolated from the overpowering splendour of their background. This isolation is effected by eclipses, with the result--beautifully confirmatory of theory--of _reversing_, or turning from dark to bright, the Fraunhofer spectrum. The completeness and precision of the reversal, however, could not be visually attested; and a quarter of a century elapsed before a successful "snap-shot" provided photographic evidence on the subject. It was taken at Novaya Zemlya by Mr. Shackleton, a member of the late Sir George Baden-Powell's expedition to observe the eclipse of August 9, 1896;[529] and similar records in abundance were secured during the Indian eclipse of January 22, 1898,[530] and the Spanish-American eclipse of May 28, 1900.[531] The result of their leisurely examination has been to verify the existence of a "reversing-layer," in the literal sense of the term. It is true that no single "flash" photograph is an inverted transcript of the Fraunhofer spectrum. The lines are, indeed, in each case--speaking broadly--the same; but their relative intensities are widely different. Yet this need occasion no surprise when we remember that the Fraunhofer spectrum integrates the absorption of multitudinous strata, various in density and composition, while only the upper section of the formation comes within view of the sensitive plates exposed at totalities, the low-lying vaporous beds being necessarily covered by the moon. The total depth of this glowing envelope may be estimated at 500 to 600 miles, and its normal state seems to be one of profound tranquillity, judging from the imperturbable aspect of the array of dark lines due to its sifting action upon light.

The last of the five eclipses which we have grouped together for separate consideration was visible in Southern India and Australia, December 12, 1871. Some splendid photographs were secured by the English parties on the Malabar coast, showing, for the first time, the remarkable branching forms of the coronal emanations; but the most conspicuous result was Janssen's detection of some of the dark Fraunhofer lines, long vainly sought in the continuous spectrum of the corona. Chief among these was the D-line of sodium, the original index, it might be said, to solar chemistry. No proof could be afforded more decisive that this faint _echoing back_ of the distinctive notes of the Fraunhofer spectrum, that the polariscope had spoken the truth in asserting a large part of the coronal radiance to be reflected sunlight. But it is usually so drenched in original luminosity, that its special features are almost obliterated. Janssen's success in seizing them was due in part to the extreme purity of the air at Sholoor, in the Neilgherries, where he was stationed; in part to the use of an instrument adapted by its large aperture and short focus to give an image of the utmost brilliancy. His observation, repeated during the Caroline Island eclipse of 1883, was photographically verified ten years later by M. de la Baume Pluvinel in Senegal.[532]

An instrument of great value for particular purposes was introduced into eclipse-work in 1871. The "slitless spectroscope" consists simply of a prism placed outside the object-glass of a telescope or the lens of a camera, whereby the radiance encompassing the eclipsed sun is separated into as many differently tinted rings as it contains different kinds of light. These tinted rings were simultaneously viewed by Respighi at Poodacottah, and by Lockyer at Baikul. Their photographic registration by the latter in 1875 initiated the transformation of the slitless spectroscope into the prismatic camera.[533] Meanwhile, the use of an ordinary spectroscope by Herschel and Tennant at Dodabetta showed the green ray of coronium to be just as bright in a rift as in the adjacent streamer. The visible structure of the corona was thus seen to be independent of the distribution of the gases which enter into its composition.

By means, then, of the five great eclipses of 1860-71 it was ascertained: first, that the prominences, and at least the lower part of the corona, are genuine solar appurtenances; secondly, that the prominences are composed of hydrogen and other gases in a state of incandescence, and rise, as irregular outliers, from a continuous envelope of the same materials, some thousands of miles in thickness; thirdly, that the corona is of a highly complex constitution, being made up in part of glowing vapours, in part of matter capable of reflecting sunlight. We may now proceed to consider the results of subsequent eclipses.

These have raised, and have helped to solve, some very curious questions. Indeed, every carefully watched total eclipse of the sun stimulates as well as appeases curiosity, and leaves a legacy of outstanding doubt, continually, as time and inquiry go on, removed, but continually replaced. It cannot be denied that the corona is a perplexing phenomenon, and that it does not become less perplexing as we know more about it. It presented itself under quite a new and strange aspect on the occasion of the eclipse which visited the Western States of North America, July 29, 1878. The conditions of observation were peculiarly favourable. The weather was superb; above the Rocky Mountains the sky was of such purity as to permit the detection of Jupiter's satellites with the naked eye on several successive nights. The opportunity for advancing knowledge was made the most of. Nearly a hundred astronomers, including several Englishmen, occupied twelve separate posts, and prepared for an attack in force.

The question had often suggested itself, and was a natural one to ask, whether the corona sympathises with the general condition of the sun? whether, either in shape or brilliancy, it varies with the progress of the sun-spot period? A more propitious moment for getting this question answered could hardly have been chosen than that at which the eclipse occurred. Solar disturbance was just then at its lowest ebb. The development of spots for the month of July, 1878, was represented on Wolf's system of "relative numbers" by the fraction 0·1, as against 135·4 for December, 1870, an epoch of maximum activity. The "chromosphere"[534] was, for the most part, shallow and quiescent; its depth, above the spot zones, had sunk from about 6,000 to 2,000 miles;[535] prominences were few and faint. Obviously, if a type of corona corresponding to a minimum of sun-spots existed, it should be seen then or never. It _was_ seen; but while, in some respects, it agreed with anticipation, in others it completely set it at naught.

The corona of 1878, as compared with those of 1869, 1870, and 1871, was generally admitted to be shrunken in its main outlines and much reduced in brilliancy. Lockyer pronounced it ten times fainter than in 1871; Harkness estimated its light at less than one-seventh that derived from the mist-blotted aureola of 1870.[536] In shape, too, it was markedly different. When sun-spots are numerous, the corona appears to be most fully developed above the spot-zones, thus offering to our eyes a rudely quadrilateral contour. The four great luminous sheaves forming the corners of the square are made up of rays curving together from each side into "synclinal" or ogival groups, each of which may be compared to the petal of a flower. To Janssen, in 1871, the eclipsing moon seemed like the dark heart of a gigantic dahlia, painted in light on the sky; and the similitude to the ornament on a compass-card, used by Airy in 1851, well conveys the decorative effect of the beamy, radiated kind of aureola, never, it would appear, absent when solar activity is at a tolerably high pitch. In his splendid volume on eclipses,[537] with which the systematic study of coronal structure may be said to have begun, Mr. Ranyard first generalised the synclinal peculiarity by a comparison of records; but the symmetry of the arrangement, though frequently striking, is liable to be confused by secondary formations. He further pointed out, with the help of careful drawings from the photographs of 1871 made by Mr. Wesley, the curved and branching shapes assumed by the component filaments of massive bundles of rays. Nothing of all this, however, was visible in 1878. Instead, there was seen, as the groundwork of the corona, a ring of pearly light, nebulous to the eye, but shown by telescopes and in photographs to have a fibrous texture, as if made up of tufts of fine hairs. North and south, a series of short, vivid, electrical-looking flame-brushes diverged with conspicuous regularity from each of the solar poles. Their direction was not towards the centre of the sun, but towards each summit of his axis, so that the farther rays on either side started almost tangentially to the surface.

But the leading, and a truly amazing, characteristic of the phenomenon was formed by two vast, faintly-luminous _wings_ of light, expanded on either side of the sun in the direction of the ecliptic. These were missed by very few careful onlookers; but the extent assigned to them varied with skill in, and facilities for seeing. By far the most striking observations were made by Newcomb at Separation (Wyoming), by Cleveland Abbe from the shoulder of Pike's Peak, and by Langley at its summit, an elevation of 14,100 feet above the sea. Never before had an eclipse been viewed from anything approaching that altitude, or under so translucent a sky. A proof of the great reduction in atmospheric glare was afforded by the perceptibility of the corona four minutes after totality was over. For the 165 seconds of its duration, the remarkable streamers above alluded to continued "persistently visible," stretching away right and left of the sun to a distance of at least ten million miles! One branch was traced over an apparent extent of fully twelve lunar diameters, without sign of a definite termination having been reached; and there were no grounds for supposing the other more restricted.

The resemblance to the zodiacal light was striking; and a community of origin between that enigmatical member of our system and the corona was irresistibly suggested. We should, indeed, expect to see, under such exceptionally favourable atmospheric conditions as Professor Langley enjoyed on Pike's Peak, the _roots_ of the zodiacal light presenting near the sun just such an appearance as he witnessed; but we can imagine no reason why their visibility should be associated with a low state of solar activity. Nevertheless this seems to be the case with the streamers which astonished astronomers in 1878. For in August, 1867, when similar equatorial emanations, accompanied by similar symptoms of polar excitement, were described and depicted by Grosch[538] of the Santiago Observatory, sun-spots were at a minimum; while the corona of 1715, which appears from the record of it by Roger Cotes[539] to have been of the same type, preceded by three years the ensuing maximum. The eclipsed sun was seen by him at Cambridge, May 2, 1715, encompassed with a ring of light about one-sixth of the moon's diameter in breadth, upon which was superposed a luminous cross formed of long bright branches lying very nearly in the plane of the ecliptic, and shorter polar arms so faint as to be only intermittently visible. The resemblance between his sketch and Cleveland Abbe's drawing of the corona of 1878 is extremely striking. It should, nevertheless, be noted that some conspicuous spots were visible on the sun's disc at the time of Cotes's eclipse, and that the preceding minimum (according to Wolf) occurred in 1712. Thus, the coincidence of epochs is imperfect.

Professor Cleveland Abbe was fully persuaded that the long rays carefully observed by him from Pike's Peak were nothing else than streams of meteorites rushing towards or from perihelion; and it is quite certain that the solar neighbourhood must be crowded with such bodies. But the peculiar structure at the base of the streamers displayed in the photographs, the curved rays meeting in pointed arches like Gothic windows, the visible upspringing tendency, the filamentous texture,[540] speak unmistakably of the action of forces proceeding _from_ the sun, not of extraneous matter circling round him.

A further proof of sympathetic change in the corona is afforded by the analysis of its light. In 1878 the bright line so conspicuous in the coronal spectrum in 1870 and 1871 had faded to the very limit of visibility. Several skilled observers failed to see it at all; but Young and Eastman succeeded in tracing the green "coronium" ray all round the sun, to a height estimated at 340,000 miles. The substance emitting it was thus present, though in a low state of incandescence. The continuous spectrum was relatively strong; faint traces of the Fraunhofer lines attested for it an origin, in part by reflection; and polarisation was undoubted, increasing towards the limb, whereas in 1870 it reached a maximum at a considerable distance from it. Experiments with Edison's tasimeter seemed to show that the corona radiates a sensible amount of heat.

The next promising eclipse occurred May 17, 1882. The concourse of astronomers which has become usual on such occasions assembled this time at Sohag, in Upper Egypt. Rarely have seventy-four seconds been turned to such account. To each observer a special task was assigned, and the advantages of a strict division of labour were visible in the variety and amount of the information gained.

The year 1882 was one of numerous sun-spots. On the eve of the eclipse twenty-three separate maculæ were counted. If there were any truth in the theory which connected coronal forms with fluctuations in solar activity, it might be anticipated that the vast equatorial expansions and polar "brushes" of 1878 would be found replaced by the star-like structure of 1871. This expectation was literally fulfilled. No lateral streamers were to be seen. The universal failure to perceive them, after express search in a sky of the most transparent purity, justifies the emphatic assertion that _they were not there_. Instead, the type of corona observed in India eleven years earlier, was reproduced with its shining aigrettes, complex texture and brilliant radiated aspect.

Concordant testimony was given by the spectroscope. The reflected light derived from the corona was weaker than in 1878, while its original emissions were proportionately intensified. Nevertheless, most of the bright lines recorded as coronal[541] were really due, there can be no doubt, to diffused chromospheric light. On this occasion, the first successful attempt was made to photograph the coronal spectrum procured in the ordinary way with a slit and prisms, while the prismatic camera was also profitably employed. It served to bring out at least one important fact--that of the uncommon strength in chromospheric regions of the twin violet beams of calcium, designated "H" and "K"; and prominence-photography signalised its improvement by the registration, in the spectrum of one such object, of twenty-nine rays, including many of the ultra-violet hydrogen series discovered by Sir William Huggins in the emission of white stars.[542]

Dr. Schuster's photographs of the corona itself were the most extensive, as well as the most detailed, of any yet secured. One rift imprinted itself on the plates to a distance of nearly a diameter and a half from the limb; and the transparency of the streamers was shown by the delineation through them of the delicate tracery beyond. The singular and picturesque feature was added of a bright comet, self-depicted in all the exquisite grace of swift movement betrayed by the fine curve of its tail, hurrying away from one of its rare visits to our sun, and rendered momentarily visible by the withdrawal of the splendour in which it had been, and was again quickly veiled.

From a careful study of these valuable records Sir William Huggins derived the idea of a possible mode of photographing the corona _without an eclipse_.[543] As already stated, its ordinary invisibility is entirely due to the "glare" or reflected light diffused through our atmosphere. But Huggins found, on examining Schuster's negatives, that a large proportion of the light in the coronal spectrum, both continuous and interrupted, is collected in the violet region between the Fraunhofer lines G and H. There, then, he hoped that, all other rays being excluded, it might prove strong enough to vanquish inimical glare, and stamp on prepared plates, through _local_ superiority in illuminative power, the forms of the appendage by which it is emitted.

His experiments were begun towards the end of May, 1882, and by September 28 he had obtained a fair earnest of success. The exclusion of all other qualities of light save that with which he desired to operate, was accomplished by using chloride of silver as his sensitive material, that substance being chemically inert to all other but those precise rays in which the corona has the advantage.[544] Plates thus sensitised received impressions which it was hardly possible to regard as spurious. "Not only the general features," Captain Abney affirmed,[545] "are the same, but details, such as rifts and streamers, have the same position and form." It was found, moreover, that the corona photographed during the total eclipse of May 6, 1883, was intermediate in shape between the coronas photographed by Sir William Huggins before and after that event, each picture taking its proper place in a series of progressive modifications highly interesting in themselves, and full of promise for the value of the method employed to record them.[546] But experiments on the subject were singularly interrupted. The volcanic explosion in the Straits of Sunda in August, 1883, brought to astronomers a peculiarly unwelcome addition to their difficulties. The magnificent sunglows due to the diffractive effects on light of the vapours and fine dust flung in vast volumes into the air, and rapidly diffused all round the globe, betokened an atmospheric condition of all others the most prejudicial to delicate researches in the solar vicinity. The filmy coronal forms, accordingly, which had been hopefully traced on the Tulse Hill plates ceased to appear there; nor were any substantially better results obtained by Mr. C. Ray Woods, in the purer air either of the Riffel or the Cape of Good Hope, during the three ensuing years. Moreover, attempts to obtain coronal photographs during the partial phases of the eclipse of August 29, 1886, completely failed. No part of the lunar globe became visible in relief against circumfluous solar radiance on any of the plates exposed at Grenada; and what vestiges of "structure" there were, came out almost better _upon_ the moon than _beside_ her, thus stamping themselves at once as of atmospheric origin.

That the effect sought is a perfectly possible one is proved by the distinct appearance of the moon projected on the corona, in photographs of the partially eclipsed sun in 1858, 1889, and 1890, and very notably in 1898 and 1900.[547]

In the spring of 1893, Professor Hale[548] attacked the problem of coronal daylight photography, employing the "double-slit" method so eminently serviceable for the delineation of prominences.[549] But neither at Kenwood nor at the summit of Pike's Peak, whither, in the course of the summer, he removed his apparatus, was any action of the desired kind secured. Similar ill success attended his and Professor Riccò's employment, on Mount Etna in July, 1894, of a specially designed coronagraph. Yet discouragement did not induce despair. The end in view is indeed too important to be readily abandoned; but it can be reached only when a more particular acquaintance with the nature of coronal light than we now possess indicates the appropriate device for giving it a preferential advantage in self-portraiture. Moreover, the effectiveness of this device may not improbably be enhanced, through changes in the coronal spectrum at epochs of sun-spot maximum.

The prosperous result of the Sohag observations stimulated the desire to repeat them on the first favourable opportunity. This offered itself one year later, May 6, 1883, yet not without the drawbacks incident to terrestrial conditions. The eclipse promised was of rare length, giving no less than five minutes and twenty-three seconds of total obscurity, but its path was almost exclusively a "water-track." It touched land only on the outskirts of the Marquesas group in the Southern Pacific, and presented, as the one available foothold for observers, a coral reef named Caroline Island, seven and a half miles long by one and a half wide, unknown previously to 1874, and visited only for the sake of its stores of guano. Seldom has a more striking proof been given of the vividness of human curiosity as to the condition of the worlds outside our own, than in the assemblage of a group of distinguished men from the chief centres of civilisation, on a barren ridge, isolated in a vast and tempestuous ocean, at a distance, in many cases, of 11,000 miles and upwards from the ordinary scene of their labours. And all these sacrifices--the cost and care of preparation, the transport and readjustment of delicate instruments, the contrivance of new and more subtle means of investigating phenomena--on the precarious chance of a clear sky during one particular five minutes! The event, though fortunate, emphasised the hazard of the venture. The observation of the eclipse was made possible only by the happy accident of a serene interval between two storms.

The American expedition was led by Professor Edward S. Holden, and to it were courteously permitted to be attached Messrs. Lawrance and Woods, photographers, sent out by the Royal Society of London. M. Janssen was chief of the French Academy mission; he was accompanied from Meudon by Trouvelot, and joined from Vienna by Palisa, and from Rome by Tacchini. A large share of the work done was directed to assuring or negativing previous results. The circumstances of an eclipse favour illusion. A single observation by a single observer, made under unfamiliar conditions, and at a moment of peculiar excitement, can scarcely be regarded as offering more than a suggestion for future inquiry. But incredulity may be carried too far. Janssen, for instance, felt compelled, by the survival of unwise doubts, to devote some of the precious minutes of obscurity at Caroline Island to confirming what, in his own persuasion, needed no confirmation--that is, the presence of reflected Fraunhofer lines in the spectrum of the corona. Trouvelot and Palisa, on the other hand, instituted an exhaustive, but fruitless search for the spurious "intramercurian" planets announced by Swift and Watson in 1878.

New information, however, was not deficient. The corona proved identical in type with that of 1882,[550] agreeably to what was expected at an epoch of protracted solar activity. The characteristic aigrettes were of even greater brilliancy than in the preceding year, and the chemical effects of the coronal light proved unusually intense. Janssen's photographs, owing to the considerable apertures (six and eight inches) of his object-glasses, and the long exposures permitted by the duration of totality, were singularly perfect; they gave a greater extension to the coronal than could be traced with the telescope,[551] and showed its forms as absolutely fixed and of remarkable complexity.

The English pictures, taken with exposures up to sixty seconds, were likewise of great value. They exhibited details of structure from the limb to the tips of the streamers, which terminated definitely, and as it seemed actually, where the impressions on the plates ceased. The coronal spectrum was also successfully photographed, and although the reversing layer in its entirety evaded record, a print was caught of some of its more prominent rays just before and after totality. The use of the prismatic camera was baffled by the anomalous scarcity of prominences.

Using an ingenious apparatus for viewing simultaneously the spectrum from both sides of the sun, Professor Hastings noticed at Caroline Island alternations, with the advance of the moon, in the respective heights above the right and left solar limbs of the coronal green line, which were thought to imply that the corona, with its rifts and sheaves and "tangled hanks" of rays, is, after all, merely an illusive appearance produced by the diffraction of sunlight at the moon's edge.[552] But the observation was assuredly misleading or misinterpreted. Atmospheric _diffusion_ may indeed, under favouring circumstances, be effective in deceptively enlarging solar appendages; but always to a very limited extent.

The controversy is an old one as to the part played by our air in producing the radiance visible round the eclipsed sun. In its original form, it is true, it came to an end when Professor Harkness, in 1869,[553] pointed out that the shadow of the moon falls equally over the air and on the earth, and that if the sun had no luminous appendages, a circular space of almost absolute darkness would consequently surround the apparent places of the superposed sun and moon. Mr. Proctor,[554] with his usual ability, impressed this mathematically certain truth upon public attention; and Sir John Herschel calculated that the diameter of the "negative halo" thus produced would be, in general, no less than 23°.

But about the same time a noteworthy circumstance relating to the state of things in the solar vicinity was brought into view. On February 11, 1869, Messrs. Frankland and Lockyer communicated to the Royal Society a series of experiments on gaseous spectra under varying conditions of heat and density, leading them to the conclusion that the higher solar prominences exist in a medium of excessive tenuity, and that even at the base of the chromosphere the pressure is far below that at the earth's surface.[555] This inference was fully borne out by the researches of Wüllner; and Janssen expressed the opinion that the chromospheric gases are rarefied almost to the degree of an air-pump vacuum.[556] Hence was derived a general and fully justified conviction that there could be outside, and incumbent upon the chromosphere, no such vast atmosphere as the corona appeared to represent. Upon the strength of which conviction the "glare" theory entered, chiefly under the auspices of Sir Norman Lockyer, upon the second stage of its existence.

The genuineness of the "inner corona" to the height of 5' or 6' from the limb was admitted; but it was supposed that by the detailed reflection of its light in our air the far more extensive "outer corona" was optically created, the irregularities of the moon's edge being called in to account for the rays and rifts by which its structure was varied. This view received some countenance from Admiral Maclear's observation, during the eclipse of 1870, of bright lines "everywhere"--even at the centre of the lunar disc. Here, indeed, was an undoubted case of atmospheric diffusion; but here, also, was a safe index to the extent of its occurrence. Light scatters equally in all directions; so that when the moon's face at the time of an eclipse shows (as is the common case) a blank in the spectroscope, it is quite certain that the corona is not noticeably enlarged by atmospheric causes. A sky drifted over with thin cirrus clouds and air changed with aqueous vapour amply accounted for the abnormal amount of scattering in 1870.

But even in 1870 positive evidence was obtained of the substantial reality of the radiated outer corona, in the appearance on the photographic plates exposed by Willard in Spain and by Brothers in Sicily of identical dark rifts. The truth is, that far from being developed by misty air, it is peculiarly liable to be effaced by it. The purer the sky, the more extensive, brilliant, and intricate in the details of its structure the corona appears. Take as an example General Myer's description of the eclipse of 1869, as seen from the summit of White Top Mountain, Virginia, at an elevation above the sea of 5,523 feet, in an atmosphere of peculiar clearness.

"To the unaided eye," he wrote,[557] "the eclipse presented, during the total obscuration, a vision magnificent beyond description. As a centre stood the full and intensely black disc of the moon, surrounded by the aureola of a soft bright light, through which shot out, as if from the circumference of the moon, straight, massive, silvery rays, seeming distinct and separate from each other, to a distance of two or three diameters of the solar disc; the whole spectacle showing as on a background of diffused rose-coloured light."

On the same day, at Des Moines, Newcomb could perceive, through somewhat hazy air, no long rays, and the four-pointed outline of the corona reached at its farthest only a _single semidiameter_ of the moon from the limb. The plain fact, that our atmosphere acts rather as a veil to hide the coronal radiance than as the medium through which it is visually formed, emerges from further innumerable records.

No observations of importance were made during the eclipse of September 9, 1885. The path of total obscurity touched land only on the shores of New Zealand, and two minutes was the outside limit of available time. Hence local observers had the phenomenon to themselves; nor were they even favoured by the weather in their efforts to make the most of it. One striking appearance was, however, disclosed. It was that of two "white" prominences of unusual brilliancy, shining like a pair of electric lamps hung one at each end of a solar diameter, right above the places of two large spots.[558] This coincidence of diametrically opposite disturbances is of too frequent occurrence to be accidental. M. Trouvelot observed at Meudon, June 26, 1885, two active and evanescent prominences thus situated, each rising to the enormous height of 300,000 miles; and on August 16, one scarcely less remarkable, balanced by an antipodal spot-group.[559] It towered upward, as if by a process of _unrolling_, to a quarter of a million of miles; after which, in two minutes, the light died out of it; it had become completely extinct. The development, again from the ends of a diameter, of a pair of similar objects was watched, September 19 and 20, 1893, by Father Fényi, Director of the Kalocsa Observatory; and the phenomenon has been too often repeated to be accidental.

The eclipse of August 29, 1886, was total during about four minutes over tropical Atlantic regions; and an English expedition, led by Sir Norman Lockyer, was accordingly despatched to Grenada in the West Indies, for the purpose of using the opportunity it offered. But the rainy season was just then at its height: clouds and squalls were the order of the day; and the elaborately planned programme of observation could only in part be carried through. Some good work, none the less, was done. Professor Tacchini, who had been invited to accompany the party, ascertained besides some significant facts about prominences. From a comparison of their forms and sizes during and after the eclipse, it appeared that only the growing vaporous cores of these objects are shown by the spectroscope under ordinary circumstances; their upper sections, giving a faint continuous spectrum, and composed of presumably cooler materials, can only be seen when the veil of scattered light usually drawn over them is removed by an eclipse. Thus all modestly tall prominences have silvery summits; but all do not appear to possess the "red heart of flame," by which alone they can be rendered perceptible to daylight observation. Some prove to be ordinarily invisible, because silvery throughout--"sheeted ghosts," as it were, met only in the dark.

Specimens of the class had been noted as far back as 1842, but Tacchini first drew particular attention to them. The one observed by him in 1886 rose in a branching form to a height of 150,000 miles, and gave a brilliantly continuous spectrum, with bright lines at H and K, but no hydrogen-lines.[560] Hence the total invisibility of the object before and after the eclipse. During the eclipse, it was seen framed, as it were, in a pointed arch of coronal light, the symmetrical arrangement of which with regard to it was obviously significant. Both its unspringing shape, and the violet rays of calcium strongly emitted by it, contradicted the supposition that "white prominences" represent a downrush of refrigerated materials.

The corona of 1886, as photographed by Dr. Schuster and Mr. Maunder, showed neither the petals and plumes of 1871, nor the streamers of 1878. It might be called of a transition type.[561] Wide polar rifts were filled in with tufted radiations, and bounded on either side by irregularly disposed, compound luminous masses. In the south-western quadrant, a triangular ray, conspicuous to the naked eye, represented, Mr. W. H. Pickering thought, the projection of a huge, hollow cone.[562] Branched and recurving jets were curiously associated with it. The intrinsic photographic brightness of the corona proved, from Pickering's measures, to be about 1/54 that of the average surface of the full moon.

The Russian eclipse of August 19, 1887, can only be remembered as a disastrous failure. Much was expected of it. The shadow-path ran overland from Leipsic to the Japanese sea, so that the solar appurtenances would, it was hoped, be disclosed to observers echeloned along a line of 6,000 miles. But the incalculable element of weather rendered all forecasts nugatory. The clouds never parted, during the critical three minutes, over Central Russia, where many parties were stationed, and Professor D. P. Todd was equally unfortunate in Japan. Some good photographs were, nevertheless, secured by Professor Arai, Director of the Tokio Observatory, as well as by MM. Bélopolsky and Glasenapp at Petrovsk and Jurjevitch respectively. They showed a corona of simpler form than that of the year before, but not yet of the pronounced type first associated by Mr. Ranyard with the lowest stage of solar activity.

The genuineness of the association was ratified by the duplicate spectacle of the next-ensuing minimum year. Two total eclipses of the sun distinguished 1889. The first took place on New Year's Day, when a narrow shadow-path crossed California, allowing less than two minutes for the numerous experiments prompted by the varied nature of modern methods of research. American astronomers availed themselves of the occasion to the full. The heavens were propitious. Photographic records were obtained in unprecedented abundance, and of unusual excellence. Their comparison and study placed it beyond reasonable doubt that the radiated corona belonging to periods of maximum sun-spots gives place, at periods of minimum, to the "winged" type of 1878. Professor Holden perceived further that the equatorial extensions characterising the latter tend to assume a "trumpet-shape."[563] Their extremities diverge, as if mutually repellent, instead of flowing together along a medial plane. The maximum actinic brilliancy of the corona of January 1, 1889, was determined at Lick to be twenty-one times less than that of the full moon.[564] Its colour was described as "of an intense luminous silver, with a bluish tinge, similar to the light of an electric arc."[565] Its spectrum was comparatively simple. Very few bright lines besides those of hydrogen and coronium, and apparently no dark ones, stood out from the prismatic background.

"The marked structural features of the corona, as presented by the negatives" taken by Professors Nipher and Charroppin, were the filaments and the streamers. The filaments issued from polar calottes of 20° radius.

"The impression conveyed to the eye," Professor Pritchett wrote,[566] "is that the equatorial stream of denser coronal matter extends across and through the filaments, simply obscuring them by its greater brightness. The effect is just as if the equatorial belt were superposed upon, or passed through, the filamentary structure. There is nothing in the photographs to prove that the filaments do not exist all round the sun.[567] The testimony from negatives of different lengths of exposure goes to show that the equatorial streamers are made up of numerous interlacing parts inclined at varying angles to the sun's equator."

The coronal extensions, perceptible with the naked eye to a distance of more than 3° from the sun, appeared barely one-third of that length on the best negatives. Little more could be seen of them either in Barnard's exquisite miniature pictures, or in the photographs obtained by W. H. Pickering with a thirteen-inch refractor--the largest instrument so far used in eclipse-photography.

The total eclipse of December 22, 1889, held out a prospect, unfortunately not realized, of removing some of the doubts and difficulties that impeded the progress of coronal photography.[568] Messrs. Burnham and Schaeberle secured at Cayenne some excellent impressions, showing enough of the corona to prove its identical character with that depicted in the beginning of the year, but not enough to convey additional information about its terminal forms or innermost structure. Any better result was indeed impossible, the moisture-laden air having cut down the actinic power of the coronal light to one-fourth its previous value.