;o5 



NA TURE 



[January 26, 1893 



by supposing that matter is thrown out by the sun normally at 

 its surface, and that its projection is turned on one side by the 

 rotation of the sun. Mr. Schaeberle, of the Lick Observatory, 

 has mathematically studied this theory, and on applying it to 

 eclipses already observed, has been able to report that the cur- 

 vature of the structures has always conformed with the theory. 



We must also examine the photographs taken with the 

 longest exposure, to determine whether the dark parts which 

 sometimes separate the luminous structures, and which we can 

 trace to the base of the corona, are entirely destitute of light. 

 The existence of these rifts, as the English call them, is difficult 

 to explain, if we suppose that the coronal atmosphere entirely 

 surrounds the sun, for in that case we should always see, pro- 

 jected on the plane perpendicular to the line of sight, the 

 coronal matter all round the sun. According to Prof. Hastings, 

 the presence of these rifts is a confirmation of his theory which 

 ascribes the corona to diffraction, and not to the existence of a 

 material mass. 



Is the aspect of the corona quickly modified, or are the 

 changes which we notice from one eclipse to another effected 

 slowly? Hitherto we have never proved the difference 

 between the photographs of the same eclipse taken at several 

 hours' interval, and at stations very distant from each other. 

 The English astronomers thought of testing this question in 

 December, 1889, and the two English expeditions sent, one to 

 the Salut Isles, and the other to the west coast of Africa, were 

 furnished with photographic outfits as identical as possible, in 

 order to obtain, at an interval of two and a half hours, com- 

 parable negatives of the corona. Unfortunately the complete 

 failure of the expedition on the African coast did not permit 

 the carrying out of this programme ; otherwise it is very pro- 

 bable they would have proved no sensible difference between 

 the negatives at the two stations, for photographs show that the 

 corona of December 22, 1889, was almost identical with that of 

 January I of the same year. We may say, then, that during 

 the year 1889, a year of quietude on the solar surface, the 

 appearance of the corona did not appreciably change. 



However, the experiment attempted by the English expedi- 

 tions needs to be repeated ; if not to study the internal move- 

 ments of the corona, to obtain identical photographs at two 

 different angles, which would enable us, with the aid of the 

 stereoscope, to judge of the 7-<?//<?/"of the corona. 



Photographs of a total eclipse will not only inform us as to the 

 structure of the corona, but will permit us to, measure its actinic 

 brightness. We can estimate the relative intensity of different 

 parts of the corona by superposing several photographs, made 

 on the same scale, but obtained with very different photo- 

 graphic actions. The outlines of each image would give lines 

 of equal actinic intensity of the corona. The absolute intensity 

 may be measured by comparing the opacity of the image on the 

 plate with the opacities produced on the same plate from a 

 source of standard light. For this purpose the plates destined 

 for photograph y are previously exposed on their borders to a 

 standard light for varying periods of time. When these plates 

 are developed, a scale of tones which allows: a comparison of 

 opacities is obtained at the same time as the image of the 

 phenomena. 



The spectroscopic examination of the corona confronts us 

 with still more complex and more interesting problems. When 

 we keep the slit of the spectroscope on the crescent of the sun, 

 which narrows more and more ia proportion as the moon ad- 

 vances, the spectrum darkens and the dark lines become less 

 and less apparent ; then all at once the field of sight is covered 

 with an infinite number of bright lines, which seem to be sub- 

 stituted for each dark line of the Fraunhofer spectrum. This 

 phenomenon only lasts two or three seconds. Such is the re- 

 markable observation made by Prof. Young in 1870. In the 

 preceding year he had tried to prove this transformation of 

 dark into bright lines, but failed because he had arranged the 

 slit of his spectroscope as a radius to the sun, which gave the 

 bright lines too little length to be perceptible. With a tan- 

 gential slit, however, the lines were long enough to be easily 

 recognised. 



Prof. Young's observations revealed to us the existence 

 round the sun of a layer of incandescent vapours, of relatively 

 low temperature, which, conformably to Kirchoff's theory, pro- 

 duce by their absorbing power reversal of the lines of the 

 solar spectrum. It is very probable that the vapours to which 

 the reversal is due are not all situated in the atmosphere which 

 Prof. Young has revealed to us, and which has a thickness of 



only 1000 kilometres. If it were so, absorption must be 

 infinitely more intense at the edge of the sun than it is at the 

 centre. Nevertheless, the borders of the sun show no trace of 

 this abnormal absorption. The observations of M. Janssen in 

 1867 showed this, and it is also proved by photographs of the 

 spectrum of the sun which I took at the annular eclipse of 1890 

 at Canoe in the island of Crete. 



It is probable that the reversal of lines is produced in a series 

 of layers whose total thickness is great enough to make the 

 difference of absorption between the centre and limb of the sun 

 inappreciable. According to Prof. Lockyer the sun should be 

 surrounded by concentric layers of vapours arranged in order 

 of density, which, according to his own expression, envelope 

 the sun like " the leaves of an onion." Prof. Young's layer of 

 vapours would comprehend only some of these layers. This 

 hypothesis seems confirmed by the observation made by M. 

 Trepied in 1882 : although he saw "a veritable rain of bright 

 lines in the spectrum," he proved that the coincidence of the 

 bright and dark lines was not complete. 



Prof. Lockyer's theory involves also, as another consequence, 

 the unequal length and width of these bright lines ; indeed, 

 the layer nearest to the sun should give short lines correspond- 

 ing to the thickness of this layer, and as the temperature here 

 must be very high the lines should be rather wide. The follow- 

 ing layer being seen by projection, and having a thickness equal 

 to the two layers, should give lines twice as long ; moreover, 

 this second layer being cooler than the preceding the lines 

 should be narrower. The same reasoning applies to the suc- 

 ceeding layers, so that we ought to find, soon after the beginning 

 of the total eclipse, short and wide lines, then long and narrow 

 ones. The observations of 1882 confirmed these predictions, 

 and English astronomers wished to repeat the experiment in 

 1886. Unfortunately the observations of Father Perry and 

 Mr. Turner were made under conditions too unfavourable for 

 us to draw any certain conclusion from them. To fully eluci- 

 date the question it is necessary to obtain instantaneous photo- 

 graphs of these bright lines. The experiment was indeed 

 attempted by English observers in 1883, but they seem to have 

 obtained no result. Prof. Lockyer proposes during the 

 approaching eclipse to photograph these lines as well as the 

 bright lines of the corona. He intends to use not only an 

 analytic spectroscope, but a prismatic object glass. This 

 apparatus will give the monochromatic images of the corona, 

 that is to say, the kind of rings corresponding to each elemen- 

 tary radiation emitted by the coronal light. 



When we turn the spectroscope towards the corona itself we 

 observe a continuous spectrum, crossed by a bright green line 

 which does not belong to any known element. This line, near 

 the line E, corresponds to the division 1474 of Kirchoff's scale, 

 and was at first believed to coincide exactly with a line of iron ; 

 but in 1876 Prof. Young was able to separate the line 1474 with 

 powerful dispersion, and proved that one of its two components 

 belongs to iron while the other belongs to the coronal matter. 

 This line 1474 has been shown in every total eclipse, but its in- 

 tensity has been very variable and seems always to have followed 

 the fluctuations of solar activity. Thus in 1878, a period of 

 maximum spots, the green line wa- so faintly visible that it 

 escaped all observers except two. On the other hand, in 1882, 

 when the solar activity was almost at its maximum, the green 

 line was visible to within 40' from the limb. However, we must 

 remember that these estimations made by different observers, 

 observing with very dissimilar instruments, are scarcely com- 

 parable, and trustworthy evidence can only be obtained from pho- 

 tographs of the spectrum of the corona. The new plates sensitive 

 to the green will no doubt allow the line 147410 be photographed 

 in the approaching eclipse. 



It would be interesting to know whether the intensity of the 

 green line varies with the brilliancy of the different parts of the 

 corona, whether it is completely wanting in the rifts, whether it 

 extends further than the visible corona, whether it has the same 

 width in its whole extent, &c. These observations can only be 

 made by associating with the spectroscope a telescope serving 

 as a finder, in which cross wires have been arranged to indicate 

 at each moment towards what part of the corona the slit is 

 directed. The spectrum is observed with one eye while with 

 the other the corona is examined. This is the arrangement 

 which M. Janssen has always adopted in his spectrum observations 

 of the corona. 



If the terrestrial atmosphere is loaded with water vapour, we 

 must expect a general diffusion of the coronal light, and this is 



NO. 



213, VOL. 47] 



