478 Proceedings of the Royal Society 
recourse must be bad to tbe microscope. Let us see what are tbe 
microscopic characters by which we recognise the particles of this 
dust. 
We may here point out that it is not so much the presence of 
volcanic minerals which enables us in a marine sediment, as well as 
in an atmospheric dust like the ashes of Krakatoa, to recognise that 
the small fragments have an eruptive origin, as the microscopic 
structure of the small vitreous particles. It is well known that 
minerals reduced to small dimensions and irregularly fractured, as 
in the case of volcanic ashes, often lose their distinctive characters. 
Their size does not allow us to judge of their optical properties ; 
their form, irregular and fragmentary, renders it difficult to de- 
termine the characteristic extinction of the species ; the phenomena 
of coloration, of pleochroism, and the tint peculiar to the mineral, 
all lose so much of their intensity, that they no longer serve for the 
identification of isolated minerals like those of the volcanic ashes 
which we have to study. As a result of our observations, we believe 
that in most cases where a mineral, under the conditions we have 
just described, reaches dimensions less than 0'05 mm., its deter- 
mination with certainty is no longer possible, and consequently its 
origin can no longer be established ; whilst a vitreous fragment, like 
those of volcanic ashes or triturated pumice, continues to be dis- 
cernible when its dimensions are less than 0’005 mm. A reason 
for showing that the absence or rarity of crystals, or of fragments 
of volcanic crystals, ought not to be taken as a proof that a 
sedimentary matter, either from the atmosphere or from the deep 
sea, is not of volcanic origin, is the sorting process to which these 
matters are subjected in the air and in the water, a phenomenon to 
which we shall presently recur. 
The most reliable distinctive character is always found in the 
structure of the small vitreous particles which are derived from the 
trituration of pumice or have an analogous origin, inasmuch as they 
have been ejected from the volcano in the state of ash. The struc- 
ture peculiar to these materials is seen in their fracture, which leaves 
its impress upon the smallest fragments of debris, in which the 
microscope can decipher no characteristic properties except such as 
have relation to form. In order to assure ourselves that these 
characters of pumice remain constant to the extreme limits of pul- 
verisation, such as are employed in tlie preparation of silicates for 
