260 PROGRESS OF METEOROLOGY IN 1889. 



frangible radiatious at sunrise, diminisbiug towards midday, then in- 

 creasing towards sunset, but not reaching, in homologous hours after 

 noon, the same values as in the morning. Tbe figures show to what 

 extent the light is bluer than the direct sunlight, and the light of tbe 

 sky at Montpellier. 



Polarization of sTcylir/ht. — Mr. J. C. McConuel has made observations 

 with a polarimeter at St. Moritz, Thersis, and Davos, and derives the 

 following results: 



(1) Tbe polarization of sky-light is weakest at midday, and is greater 

 tbe nearer tbe sun to the horizon. 



(2) Snow-covered ground diminishes polarization, and in general the 

 brighter the ground is illuminated, the weaker tbe polarization. 



(3) Polarization is greater at high altitudes than at sea-level. {Phil. 

 Mag., 1889, p. 81.) 



Twilight phenomena and the Krahatoa eruption. — Prof. J. Kiessling 

 has published the results of his studies on the twilight phenomena 

 accompanying the Krakatoa eruption in a quarto volume of IGO pages, 

 illustrated with colored plates, charts, and wood cuts. 



The author first inquires in what particular the optical phenomena of 

 1883-'86 differ from ordinary twilight, and linds that it is essentially the 

 intensity and frequent repetition which distinguish the one from the 

 other. Tbe extension of this optical phenomena over the world after tbe 

 Krakatoa erui)tion is represented by four charts on which are entered 

 the places and times of observation. Kiessling's experiments teach 

 that the colored suns observed after the eruption are produced by dense 

 clouds of smoke and dust, tbe single particles of which may be of quite 

 different sizes. He was able to obtain all colors from reddish brown to 

 violet, except the green colors were not pure, but appeared with a yel- 

 lowish tint. Colored rings are seen around tbe sun only when tbe par- 

 ticles are approximately of equal size, and tbe more nearly equal the 

 size, tbe brigbter tbe coloring. This forces tbe conclusion that tbe rings 

 are ijbeuomena of diffraction. In the discussion of twilight i)henomena, 

 Kiessling separates the colored horizoual bands from the after-glows 

 (purpurlicht). Tbe latter are phenomena of the same kind as Bishop's 

 ring; tbe former are due essentially to absorption. Of special interest 

 is tbe question vA^bether the particles that produced the abnormal phe- 

 nomena were smoke, dust, or water. It is i)0ssible to exclude dust 

 (staub) at once, since dust particles would not possess tbe necessary 

 uniformity of size; and tbe assumption that such uniformity was 

 attained by gradually falling is inadmissable, because the i)benomena 

 were seen immediately after the eruption. As between smoke and 

 water i)articles it is not easy to descriminate, and probably both were 

 engaaed in tbe production of tbe pbeuomena. 



Report of Royal Society on the Kralatoa Eruption.— The committee ap- 

 pointed by tbe Ivoyal Society to report upon tbe eruption of Krakatoa 

 have linished their work, and published a quarto volume of 490 pages, 



