Feb. 7, 1889] 



NATURE 



347 



example, attributes all tlie red tints to this cause) ; but 

 both in this section and those that follow, many con- 

 siderations are urged against the view held by Prof. 

 Kiessling that the development of the primary glow is 

 chiefly due to diffraction, while the secondary glow is as 

 confidently asserted to be due to reflection. One of 

 the principal objections to the reflection hypothesis in 

 explanation of both the ordinary as well as the present 

 ■extraordinary development of the purple glow, is its 

 limitation at first to a narrow band, a fact which cannot 

 be explained by absorption, and which is equally at 

 variance with Fresnel's law of reflection from small 

 globular dust, which would be equal in all directions. On 

 the lamina, and particularly the vitreous lamina assump- 

 tion, however, it is intelligible, since the maximum reflec- 

 tion would then be like that from the sea, in the vertical 

 plane through the sun and the eye. 



Moreover, the richly coloured and prolonged secondary 

 glows, which were the most characteristic feature of the 

 Krakatab twilights, arc shown by Mr. Archibald, in 

 Section IV:, when dealing with t'leir secular duration, to 

 have reached a distinct minimum when the large diffrac- 

 tion corona round the sun, from Prof. Ricco's observa- 

 tions,' appeared at its greatest brilliancy ; while the curve 

 of their duration on Plate xxxix., representing Dr. 

 Riggenbach and Mr. Clark's observations, shows that 

 they never again reached the same brilliancy or duration 

 as in the two or three months immediately succeeding 

 their first appearance in Europe. Both these facts aid the 

 ■conclusion arrived at by Mr. Russell, and indorsed by 

 Prof. Kiessling, that they were reflections by the haze 

 stratum of the primary glows. But if these were reflec- 

 tions, the question naturally arises, why not the primary 

 also? and until more effective arguments are brought 

 against this view, as well as Prof Ricco's objections to 

 Prof. Kiessling's theory of diffraction alone, which are 

 detailed in Section I. {c), p. 250, Mr. Russell's view of 

 the origin of both glows seems to be the inore probable 

 as well as reasonable of the two. The haze stratum 

 appears to have been capable of exerting two influences : 

 one, diffraction of the sun's rays by its smallest particles, 

 which, with the absorption and diffraction usually effected 

 by the dust and vapour present in the lower atmosphere, 

 -caused the horizontal tinted layers ; the other, reflection by 

 its larger particles or laminx of the horizontal layers, parti- 

 cularly of the lowest red one, when the earth's shadow had 

 arrived at about 25 above the western horizon, and into a 

 position whence the maximum reflective effect could be 

 seen unmasked by a diffusely illuminated bac'^ground. 



The question of the blue and green coloration of the 

 sun is next discussed by Mr. Archibald, particularly with 

 reference to its intrinsic characteristics and physical 

 origin. In Section VH., in which the distribution of the 

 twilight glows and the blue suns on their first circuit of 

 the globe is compared, it is shown that the mean limit 

 •of the band of coloured suns was about 11" north and 

 south of the latitude of Krakatab right round the equator, 

 while that of the glows lay 5" beyond this on either side. 

 Along the latitude of KrakataT) the colours were mostly 

 white or silvery, and in one or two cases coppery. The 

 colours thus evidently depended on the density of the 

 stream, the glows appearing on its borders or fringes 

 where it was less dense. A similar relation to density 

 appears from a study of the diurnal changes with varying 

 solar altitude, the sun appearing to change from blue 

 near the zenith, throughgreen, to yellow, or disappearance 

 ■on the horizon. No direct physical explanation of such 

 a phenomenon appears forthcoming, since, according to 

 the physical laws enunciated by Lord Kayleigh and Prof. 

 Stokes, the diffraction of light by particles of the same 

 order of magnitude as a wave-lengih tends to sift out the 

 shorter blue,and preserve the longer red waves of light. Re- 

 peated reflections by s nail particles tend to the same result. 



' Section I. V.), p. 241. 



The explanation proposed by Mr. Norman Loclcyer in 

 his article in the Times attributed the blue colour of the 

 sun to differences in absorbing power of particles of 

 different sizes, the larger particles being supposed to 

 transmit more of the rays near the red end of the spec- 

 trum ar.d the smaller those of shorter wave-length. The 

 difference of size has been shown by Prof. Kiessling to 

 be inoperative so far as scattering is concerned, since for 

 particles whose magnitude is of the same order as that of 

 a wave length of light, the only case to which Prof. 

 Stokes's law applies, the intensity of the scattered blue 

 rays will be always sixteen times that of the red 

 rays. It can therefore only be explained as an effect of 

 absorption, due to some particular absorptive property 

 of the materials which composed the haze. The phe- 

 nomenon of a blue or green sun has been observed under 

 natural conditions, many of vvli,ich are quoted, and in 

 most cases where the air was filled with fine dust from a 

 great variety of sources. It has also been artificially 

 reproduced by Prof. Kiessling with dust-filled air and 

 vapour of water, and particularly of sulphur. Several 

 accounts are given in Section V. of blue suns seen in 

 connection with former eruptions, and .Mr. Whymper's 

 observation during an eruption of Cotopaxi is conclusive 

 as to the ability of the finest volcanic ejecta to cause sucl;i 

 an appearance. The problem which still awaits solu- 

 tion is. What was the precise nature of the particles or 

 gases which produced the absorption ? It seems probable 

 that they were metallic sulphides. 



Mr. Archibald next deals with the sky haze and its 

 peculiar effects, more particularly on astronomical defini- 

 tion. Here again it seems to have possessed a selec- 

 tive absorption of the red rays, for in two separate lunar 

 eclipses, 1884 and 1885, the usual coppery tint of the 

 moon was conspicuously absent. He then passes on 

 to the peculiar large corona round the sun and moon, 

 which was first observed by Mr. Bishop at Honolulu 

 on September 5, and which, though less striking than 

 the twilight glows, was, if anything, more uncommon, 

 more constant, and more prolonged in duration. It was 

 a true diffraction corona with a reddish border, and of 

 almost exactly the same size as the ordinary ice-halo, viz. 

 45° in diameter. It lasted from September 5, 1883, up 

 to October 15, 1886, since which date it has entirely dis- 

 appeared. Its diameter has afforded an approximate 

 determination of the mean radius of the smaller dust- 

 particles composing the haze, which Mr. Archibald calcu- 

 lates to be o 00006 of an inch. Its connection with, an^ 

 independence of, the glows is discussed at length, but we 

 have not space to refer to it. 



In Section II., Mr. Russell gives a list of the first 

 appearances of all the optical phenomena from the be- 

 ginning of 1883 to the end of 1S84, from which date the 

 local duration of the glows is carried on by Mr. Archibald 

 in Section IV. up to the end of 1885. 



In Section III., Mr. Russell works out the geographical 

 distribution of the optical phenomena, including blue 

 suns and glows, up to the end of 1883, by which time 

 they had virtually covered the whole earth. The general 

 conclusion is^ that the phenomena all propagated them- 

 selves (with the exception of a narrow offshoot towards 

 Japan) at first due west from Java, at a rate of about 76 

 miles an hour right round the earth parallel to the equator, 

 and in a band symmetrically disposed for 16' on either 

 side of the latitude through Krakatab. A second circuit 

 with wider limits, 30' north and south of KrakataT), was 

 traced at the same rate, after which the motion became 

 indistinguishable. They then gradually spread in latitude, 

 and ultimately the haze which caused them appears to 

 have invaded our latitudes, like the anti-trade, from south- 

 west to north-east. These circumstances may be best 

 realized from a survey of Mr. Russell's maps, especially 

 that showing the successive limits of the appearances for 

 the first nine days succeeding the eruption. 



