488 
Proceedings of the Royal Society 
reach the bottom first, and then the lighter and smaller ones, 
descending more slowly, are deposited upon the larger and heavier 
fragments and crystals from the same eruption. We have a fine 
example of this stratification of submarine tufa in the centre of the 
South Pacific, lat. 22° 2P S., long. 150° 17' W. This specimen is 
entirely covered with peroxide of manganese, and at the base of the 
fragment we see the large crystals of hornblende and particles of 
magnetite. This lower layer is covered by a deposit in which these 
minerals and coarser grains are observed to pass gradually into a 
layer composed of small crystals of felspar, d4bris of pumice, and 
more or less fine material. 
We do not propose to occupy ourselves here with the mode 
of formation of volcanic ashes, and with those of Krakatoa 
in particular. It will suffice to indicate that in the dust of a 
volcano we find all the characters supporting the interpreta- 
tion which regards volcanic ashes as formed by the pulverisa- 
tion of an igneous fluid mass in which float crystals already 
formed, and from which, when projected by gases, the pulverised 
vitreous particles undergo a rapid cooling and decrepitation during 
their passage through the atmosphere. It is not only the micro- 
scopic examination of these volcanic matters that leads us to this 
conclusion, but the prodigious quantity of ashes formed during the 
eruption of this volcano, which do not agree with the interpretation 
that regards these ashes as the result of a pulverisation of a rock 
already solidified in the crater. Indeed one cannot understand how, 
in two or three days, the immense quantity of ashes ejected from 
Krakatoa could be formed by this process, as, for instance, on the 
26th August 1883 and in the May eruption, which was the prelude 
to that catastrophe. 
Second Part. 
The recent brilliant sunsets have been attributed to the 
presence in the atmosphere of minute particles of an extra- 
terrestrial origin, as well as to volcanic dust. This induces us 
to conclude this brief abstract of our observations by a descrip- 
tion of the cosmic particles which we have found, along with 
volcanic ashes and pumice, in those regions of the deep sea far 
from land, where the sediment accumulates with extreme slowness. 
