NA rURE 



[November 19, 1908 



W'ith two single-prism spectrographs, specially 

 designed by Prof. Campbell, three photographs of 

 the coronal spectrum were obtained, the slits being 

 placed east and west across the centre of the solar 

 image. With the first, a good spectrum of the 

 extreme inner corona, extending from \ 3550 to K 5390, 

 was obtained during an exposure of 3m. 40s. Both 

 the line at A 5303 and an apparently new line at 

 A 36or3, also obtained by Prof. Lewis, are recorded 

 strongly, each extending to 3' from the west, and 2' 

 from the east, limb ; the line at A 3987'o is faintly, 

 and that at a 423 i's is easily, visible, whilst another 

 apparently new, and faint, line shows at A 36255. 

 The absorption lines are seen most readily in the 

 regions lying between 10' and 20' from the limb, but 

 are quite faint in the outer, and apparently absent 

 from the inner, corona. A comparison of the continu- 

 ous spectrum with solar spectra taken with the same 

 instrument at Mount Hamilton indicates, again, a 

 lower temperature for the former, although, in 

 the absence of further details as to the similar 

 treatment of the photographs considered, this 

 evidence is not so conclusive as that educed by Prof. 

 Lewis. 



With the second single-prism spectrograph, Cramer 

 isochromatic plates were used and good spectra, ex- 

 tending from A 3600 to A 6000, obtained, the general 

 features being the same as in the preceding spectrum. 

 Measures of the green line give its wave-length as 

 530I-4 + 0-5. 



Owing to the inter-diffusion of the bright-line, the 

 continuous, and the absorption spectra, these coronal 

 spectrograms are rather difficult to interpret, but it 

 appears certain that the radiations of the outer corona, 

 of, say, more than 20* from the moon, are either not 

 recorded at all or are masked by the diffused 

 spectrum of the brighter inner corona. The propor- 

 tion of the light radiated by the coronal particles at 

 this eclipse appears to have been relatively great as 

 compared with the amount of sunlight diffused by 

 them, but whether this is due to variability in the 

 corona or to the interference of light clouds at the 

 previous eclipses is a question which Prof. Campbell 

 and Dr. .Albrecht leave for future consideration. To 

 those observers who had had previous eclipse experi- 

 ence, the outer corona on this occasion appeared dis- 

 appointingly faint. 



It is unnecessary here to enter into the numerous 

 details given by Prof. Abbot concerning his bolometric 

 apparatus, but several important modifications of that 

 previously employed had been made for this eclipse, 

 chiefly with the object of restricting the operative 

 radiations during each observation to that definite, 

 small area of the corona that was under actual obser- 

 vation. Preliminary observations of the relative 

 brightness of the sun, of the sky at different distances 

 from the sun, and of the moon were made before 

 eclipse day, and the results are shown below. Tests 

 were also made of the quality of the light from each 

 of these sources, the criterion being the proportional 

 transmissibility of the various rays by an ashphaltum 

 screen. For sun, sky and moon the proportions of 

 light transmitted were o'2g to o'37, 0*25 and o'^o 

 respectively, the difference sky-moon being due, pre- 

 sumably, to the size and nature of the reflecting 

 particles. These results lead to the inference that if 

 the brightness of the corona be due merely to reflected 

 ordinary sun rays, the transmissibility of the ravs will 

 be greater if the reflecting particles be of the 

 grossness of those composing the lunar surface, 

 and less if they are minute like the molecules of 

 gases. 



During the eclipse, observations were made with 



NO. 2038, VOL. 79] 



the sun in six different positions in regard to the 

 cross-wires of the finder, the latter being crossed to 

 form angles of 75° and 105°. In positions i. and ii. 

 the moon's image was tangent in one of the obtuse 

 angles, and in iii. in one of the acute angles; this 

 meant that in i. and ii. the bolometer was central on 

 points 4' of arc beyond the extremities of a lunar 

 diameter inclined 52°-5 to the E. and W. diameter ; 

 whilst in iii. the distance of the moon's limb was 

 about 12' of arc. In positions iv. and v. the wires 

 were i''5 of arc beyond the extremities of the moon's 

 diameter, east and west respectivel}', and in vi. they 

 intersected at the centre of the dark moon. The ob- 

 servations showed that the transmissibility of the 

 inner coronal radiations, which passed through a 

 screen transmitting light of less wave-length than 

 3 n, was o'364 ; on the same day the mean transmissi- 

 bilitv of the solar radiations was found to be o'332. 

 The whole of the reduced results are collected in the 

 following table, taking the sun's brightness as 

 10,000,000 : — 



Brightness 



Sun, near zenith (Flint Island) io,oco,ooo 



Sky, 20° from sun ,, ... ... ... 140 



,, distant from sun ,. ... ... 31 



,, average ,, ... ... ... 6j 



,, ,, previously measured at Mount Wilson 15 



Corona, positions iv. and V. 13 



i- ., ii 4 



Moon, about zenith di'itance 50' (Flint Isl.ind)... 12 (?) 



In positions iii. and vi. the galvanometer showed no deflection. 



Several significant facts appear from these figures, 

 of which two call for special mention. The first is 

 that at sea-level (Flint Island) the sky is some four 

 tiines as bright as observed at an altitude of 1800 

 metres (Mount Wilson), thus showing the profound 

 importance of establisliing our observatories in the 

 higher altitudes. .Secondly, the brightness of the sky 

 at 20° from the sun is to that of the corona as 140 :i3; 

 nearer the sun the ratio would be enormously in- 

 creased, and it therefore appears that to obtain photo- 

 graphs of the corona without waiting for an eclipse 

 is unpromising, to say the least. 



Summing up the evidence now available. Prof. 

 .-\bbot believes that the best explanation of the 

 character of the coronal radiation is that it is, in the 

 main, due to reflected sunlight. The apparent 

 absence of polarisation need not weaken this supposi- 

 tion, for, owing to the fact that they receive light 

 from the solid angle of a whole hemisphere of the 

 sun, the particles of the inner corona would be 

 partially polarised in many directions, and therefore 

 exhibit no definite polarisation in any one. The 

 observed modification of the reflected solar spectrum 

 may be due to the fact that the bright emissions from 

 the heated particles close to the sun are sufticient U> 

 obliterate the absorption lines ; further out the 

 particles would be cooler — perhaps mainly solid and 

 liquid — and the emission spectrum would then 

 diminish, the absorption spectrum, as a consequence, 

 appearing gradually. But as the light is still reflected 

 sunlight only its intensity would diminish and its 

 character and transmissibility would remain the same, 

 as the observations show that they do. 



Whilst recognising fully the difficulty of accounting 

 rigidly for all the observed phenomena. Prof. Abbot 

 concludes that in all probability the brightness of the 

 corona is mainly due to reflected ordinary sunlight, 

 diluted to some extent by the emissions from incan- 

 descent particles, and possibly, also, by some small 

 amount of " luminescence " such as produces the 

 aurora. 



William E. Rolston. 



