of Edinhurgli, Session 1883 - 84 . 
477 
of vitreous splinters derived from the trituration of pumice stones. 
The description of the phenomena connected with the Krakatoa 
eruption gives us a complete explanation of these observations. 
The specimens of pumice from Krakatoa, which have been collected 
floating on the sea and which we have examined, are in like manner 
rounded. The angular surfaces are all worn away just as in 
pebbles ; the only asperities to be observed consist of crystals and 
fragments of crystals, which project beyond the general surface of 
the vitreous matter, which last, on account of its structure, presents 
less resistance to wear and tear than the minerals which are 
embedded in it. 
We may recall the fact that the Bay of Lampoung, in the Straits 
of Sunda, was blocked by the vast accumulation of pumice, formed 
in a few hours by the eruption of Krakatoa, which completely filled 
the bay. This floating bar of pumice stones was about 30 kilometres 
long, 1 kilometre broad, and 3 to 4 metres in depth, 2 or 3 metres of 
which were below the surface of the water, and 1 metre above. 
These numbers give about 150 millions of cubic metres of ejected 
matter. This moving elastic wall rose and fell with the waves and 
tide,* and was carried by currents thousands of miles from the point 
of eruption over the surface of the ocean. The rounded form of 
blocks of pumice met with everywhere floating on the surface of 
the sea, as well as of those samples which, after having floated some 
time, became waterlogged and sank to the bottom, may be perfectly 
explained if we remember the friability of this rock, and, at the same 
time, the agitation to which it is submitted by the waves, through 
which the pieces are continually being knocked against each other. 
We understand also how this wear and tear gives rise to an immense 
quantity of pulverulent pumice fragments, which contribute in a great 
measure to the formation of oceanic deposits. As a matter of fact, 
rounded fragments of pumice have been met with floating on the 
surface of every ocean, and during the last few years many samples 
have been sent to us by captains of ships and missionaries. As has 
been already pointed out, they are universally distributed in oceanic 
deposits, although frequently highly altered. 
If it be easy to pronounce upon the volcanic nature of these larger 
fragments, it becomes, on the other hand, exceedingly difficult when 
we have to deal with particles reduced to powder, and when 
* Comptcs rendus d& V Academie des Sciences, 19 Nov. 1883, p. 1101. 
