406 SCIENTIFIC RECORD FOR 1885. 



influences, gives as the origin of these particles the formation of very- 

 minute ice crystals like those which appear to constitute the cirrus clouds. 

 The trouble with this very simple hypotheses is that so many additional 

 hypotheses are required to sustain it. * For some hypothesis mu^ be 

 made as to the cause of the production of frozen clouds at an unusual 

 altitude. Some reason must be assigned for their permanence during 

 months and even years. And the question why optical phenomena, 

 halos, mock suns, &c., which accompany the ordinary existence of ice 

 crystals intheatmosphere do not appear permanently mUst be answered. 



"Although it might be possible to find explanations for all these 

 points of difficulty, it is evident that the i)roposed hypothesis is simple 

 in aijpearance only, and that it meets with difficulties rather than with 

 confirmations in the observed facts. 



" The second hypothesis attributes to these i)articles a cosmic origin, 

 that is to say, the same origin as is assigned to meteors ; the earth is 

 supposed to have passed in its orbital motion through a cloud of small 

 particles, revolving, like herself, about the sun, and to have drawn after 

 her all the particles within her sphere of attraction. This hypothesis 

 seems at first thought a very satisfactory one; the twilight lights, at 

 first so brilliant, have been growing fainter. This accords with the idea 

 that the particles have come from some external source, and have by 

 degrees fallen, in the order of their size, to the earth's surface. The 

 explanation appears all the more plausible since true rains of dust have 

 been observed in different countries. IsTordenskjold made a chemical 

 examination of a sample of this dust, collected in Sweden, and found 

 it to contain iron and nickel, elements that are present in meteorites. 

 Still, this argument is not decisive, for rains of dust, whose origin is 

 certainly terrestrial, have often been observed ; as, for example, the dry 

 fogs, well known in southern Europe, composed of mineral dust brought 

 by the wind across the Mediterranean from the desert of Sahara. This 

 dust has been collected at sea by sailors, and also on the coasts, and 

 has been found to contain silica, feldspar, and mica, elements of the 

 sands of the African desert. 



" Volcanic ashes have been sometimes found in Norway, coming from 

 the eruptions of Iceland volcanoes ; the microscope shows these to con- 

 tain fragments of pumice and even crystals of pyroxene. The cosmic 

 origin of these particles must therefore be sustained by more direct 

 proofs, if it is to be accepted. 



" The third hypothesis connects the diffusion of these clouds of par- 

 ticles with the frightful cataclysm at Java, when 50,000 persons lost 

 their lives through the explosion of Krakatoa. Nearly all the arguments 

 employed by the jiartisans of the cosmic origin apply with equal force 

 to tliis volcanic origin. It explains why the first twilight manifestations 

 appeared directly after the eruption at Java, nothing of the kind having 

 been observed before ; it explains, moreover, the progressive march of 

 the apparition of these lights about the center of explosion. This pro- 



