486 Proceedings of the Royal Society 
by the winds, must undergo a true sorting in their passage through 
the atmosphere, according to the specific gravity of the amorphous 
elements or crystalline constituents. It results from this, that 
according to the points where they are collected, volcanic ashes 
may, although belonging to the same eruption, present differences 
not only with respect to the size of the grains, but also with respect 
to the minerals. 
In this method of transport it is evident that the vitreous 
particles, other things being equal, will be transported farthest 
from the centre. In the first place they are more abundant 
than the other particles, and again they possess in their chemical 
nature and in their structure, conditions which permit the aerial 
currents to take them up and carry them to great distances j they 
consist of a silicate in which the heavy bases are poorly re- 
presented as compared with the other constituent elements ; 
they are filled with gaseous bubbles which lower their specific 
gravity, and at the same time are capable of being broken up into 
the minutest particles. The minerals with which they are asso- 
ciated at the moment of ejection from the crater are not, like them, 
filled with gaseous bubbles ; they do not break up so easily into 
impalpable powder, for they are not porous, and are not in the same 
state of tension as the rapidly-cooled vitreous dust. Finally, many 
of these species are precisely those whose specific gravity is very 
high, on account of the bases entering into their composition. These 
minerals will not then be carried so far from the centre of eruption, 
and in all cases the vitreous particles are the essential ones in the 
atmospheric dusts derived from volcanic ashes. 
We have a beautiful illustration of this in the ashes of Krakatoa. 
In proportion as the ashes are collected at a greater distance from a 
volcano, so are they less rich in minerals, and the quantity of vitreous 
matter predominates. According to a verbal communication from 
Professor Judd, the ashes collected at Japan contain only a relatively 
small proportion of pyroxene and magnetite. 
If we wish to assure ourselves of the nature of an atmospheric 
dust, and, as has lately been frequently attempted in Europe, to 
show that the dust is really from the Krakatoa eruption, it is 
important above all to seek for the presence of vitreous fragments. 
The characters which we have indicated permit any one to recognise 
