OCTOBEB, 1902.] 



KNOWLEDGE 



227 



away at about 45"^ .... an orange tinted haze exteudiui^ 

 aliout 45" from the moon was also seen on several nights 

 towards tliy middle of Decemlier." There can be no dotibt 

 that '• Bisho[)'s Ring" was a diffraction elieel due to an 

 immense nuantity of dust particles of an extreme minute- 

 ness driven up by the great explosion to a great height in 

 our atmosphere and slowly subsiding. For as late as 

 1887 the ring still remained, though it could then be only 

 traced as a peculiar white haze to a distance of about lO'^ 

 from the sun. 



The third result of the eruption of Knikatoa was the 

 occurrence of "Afterglows" at sunset and " Foreglows " 

 at sunrise. These were distinguished from normal sunsets 

 and sunrises in that they diftVred from them iu the time 

 of tlu'ir appearance and the place or quarter in which they 

 were formed ; in their periodic action or behaviour ; in the 

 nature of the glow, which was both intense and yet 

 lustreless ; in the regularity of their colouring ; in the 

 colours themselves, which were impure and not of the 

 spectrum ; and, lastly, in the texture of the coloured 

 surfaces, which were neither distinct cloud of recognised 

 make nor yet translucent mediums. The regularity 

 of their colouring was particularly striking. " Four 

 colours in particular have been noticeable in these 

 afterglows, and in a fixed order of time and place— orange, 

 lowest and nearest the sundown ; above this, and broader, 

 green; above this, broader still, a variable red, ending in 

 being crimson ; above this, a faint lilac. The lilac dis- 

 ajipears ; the green deepens, spreads, and encroaches on 

 the orange ; and the red deepens, sjireads, and encroaches 

 on the green, till at last one red, varying downwards from 

 crimson to scarlet, or orange, fills the west and south."* 



These magnificent afterglows reappeared, but on a 

 diminished scale, after the Martinique eruption of May, 

 llil)2. A valuable letter by Prof. A. S. Herschel appeared 

 in Nature for July 24th, 1902, describing the afterglow of 

 June 26th. The sun set about 8h. 25m., and a quarter of 

 an hour later a long low belt of sky in the N.W. had 

 grown orange-yellow, whilst the ruddiness of the sicy in 

 the east had l)y the same time risen nearly to the zenith. 

 Between the two there lay a white tract about 30° in 

 width, which was gnidually invaded and at last quite 

 occupied by the advancing ruddy colour from the east. 

 About 8h. .55m. from the zenith down to 30^ above the 

 jilace of sunset, and for 40° or 50° on either side of the 

 vertical line through it, was a broad expanse of rich, rose- 

 coloured, lake-red light. This red glow sank rather 

 rapidly in height, and by 9 p.m. it had subsided into a 

 brighter glow near the horizon. 



The colour effects of an ordinary sunset are due to the 

 depth of atmosphere through which the sunlight reflected 

 to us from the clouds has pas.sed before illuminating them. 

 In their passage the rays of short wave-length suffer a 

 greater scattering, and are therefore lost to us in a greater 

 proportion than those of greater wavedength, with the 

 efifect of producing a golden or ruddy glow. The afterglow 

 would appear to be due to the intervention of a reflective 

 stratum at a greater height than that of ordinary clouds. 

 In 1883 this was no doubt composed of the finely-divided 

 volcanic dust to which " Bishop's Ring " had been due, 

 suspended high iu our atmosphere, and no doubt the 

 afterglows of 1902 owed their origin to a similar cause. 

 The height of that stratum can be inferred from the 

 depression of the sun below the horizon at the time when 

 the further boundary of the glow is setting, or is at some 

 definite altitude, or from the elevation of the zones of its 

 greatest brightness. In this way Prof. Herschel found 

 the elevation of the reflective stratum on June 26th as 



* Gerard Hopkins : Mature, January 3rd, i88i, p. 223. 



about five miles ; whilst on June 28th, he estimated " the 

 sun's parting illumination of the sky to rosy colour " at not 

 much more than 30 minutes after sunset, and the height 

 of the dust stratum in conseciuence as about 7.'r miles. 



These intervals were very distinctly shorter than those 

 remarked in 1883 and 1884. The primary glow after 

 Krakatoa averaged 54 minutes after sunset until its 

 disajipearance on the horizon ; the secondary averaged [)6 

 minutes. If the secondary glow be due to direct sun- 

 light, the average height of the stratum causing it must 

 have been nearly 40 miles. But if, as appears more 

 probable, the secondary glow was due to the same stratum 

 as that which produced the primary, being in fact a second 

 reflection from it, then the mean value of its height would 

 be about 11 miles. 



On a bright clear evening, as the sun goes down, an 

 interesting phenomenon may often be watched. If the 

 observer turns his back on the sun, he will see iu the east, 

 immediately after sunset, a long dark line spread along 

 the horizon. This darkness, which is indeed " the Shadow 

 of the Earth," thrown upwards as the suu goes down, 

 rises somewhat rapidly, its upper edge, under favourable 

 circumstances, being quite sharply defined. This " shadow 

 of the earth," which may be made out more or less dis- 

 tinctly on any clear evening in a suitable locality, and 

 which is especially easy to watch at sea, revealed itself in 

 a very interesting manner in the case of the luminous 

 night clouds described in the beginning of this chajiter. 

 If the clouds were seen in the earlier portion of the night, 

 that is to say before midnight, then the shadow of the 

 earth covering them little by little would darken them 

 from the top downwards. On the other hand, if they were 

 first seen after midnight, then it was the upper edge that 

 lighted up first, the cloud rising out of the shadow. 



The " Earth's Shadow " came strongly into evidence in 

 the case of the Krakatoa "afterglows." The Eev. S. E. 

 Bishop, writing to Nature, Vol. XXIX., p. 549, says : — 

 " I beg special attention to my former remark of the ' eartU- 

 sliadow sharply cutting off ' the upper rim of the first glow. This 

 was very manifest in the strong heavy fjlows of September, showing 

 clearly that the first glow dirt-ctly rellected the sun's rays, while in 

 the afterglow, which hud no defined upper rim but continued much 

 longer, the liaze reflects only the liglit of the first glow." 

 And, again, in Nature, Vol. XXX., p. 194, he writes : — • 



"In your issue of April 10th (p. 5-19), is the statement by an 

 observer in Australia that the ' red glow was margined by an im- 

 mense black bow stretching across from the north-west to south-east.' 



" I wish to say that the above language almost exactly describes 

 the appearance to wliich I alluded on the same page as ' the earth- 

 shadow cutting off the upper rim of the glow.' The ' black bow ' 

 of the Australian was evidently the shadow of the horizon projected 

 on the haze stratum. In both the above eases the lower surface of 

 tlio haze was evidently well defined, so that, as the horizon inter- 

 cepted the direct rays of the sun, a well-marked shadow moved 

 westward and c'ownward. Above this black rim or bow aj)peared 

 the secondary glow, produced by the reflection of the sun's rays 

 from that portion of the Iiaze surface wliich was directly illuminated. 

 Very often the second glow was more conspicuous and impressive 

 than the fii'st, because it shone agamst the dark sky of night." 



If the secondary glow were due to the reflection of direct 

 sunlight, theu no doubt its upper edge would have been 

 sharply marked off by the earth's shadow, just as was tlte 

 case with the primary glow. But its diffusion as compared 

 with the definiteness of the earlier glow points to its being 

 a reflection of the latter, a view strengthened by the fact 

 that the depression of the sun below the horizon at the 

 sinking of the second glow was as nearly as could be 

 determined double what it was at the sinking of the first. 



Prof. A. Ricco, in addition to these two points, mentions 

 a third. Prom time to time in fine weather the phe- 

 nomenon is presented at sunset of " Crepuscular Rays." 

 These are bi'oad pink ribands of light diverging like the 



