on the Action of Water upon Glass. S3 5 
case, more highly alkaline than any modern glass containing 
lead, that has hitherto been examined. 
The specific gravity of common flint-glass was taken by way 
•of comparison with the ancient fragments above mentioned, the 
result of which is here given. Flint-glass, S. G., 3.208. Ancient 
glass, 2.875. It may here be remarked, that the latter acted 
powerfully upon the test paper, by merely moistening it, with- 
out reduction to powder. It cannot be surprising, therefore, 
that ancient glass, which may almost be called pure silicate of 
potash, should be occasionally found in states of such rapid de- 
cay, as the specimens in collections often exhibit. 
Another proof of the action of water, aided by other concomi- 
tant circumstances, in producing decomposition upon glass, is 
an account given in vol. i. p. 135, of the Quarterly Journal of 
Science, of some bottles of wine, found in a quantity of black 
mud at the bottom of an old well, full of burned wood, supposed, 
upon good authority, to be of anterior date to the fire of Lon- 
don (1666). The siliceous earth, in this instance, separated in 
films on the surface of the bottle, in consequence of the abstrac- 
tion of alkaline matter, probably by the action of water, aided 
perhaps originally by a certain degree of heat, and afterwards 
by the long period of their continuance in a situation favourable 
to the decomposing agency. 
In contact with ammoniacal, or decomposing animal matter, 
the disintegration of glass takes place more rapidly. Stable- 
windows, and bottles kept in such situations, often present a 
very beautiful iridescent appearance, in consequence of the sili- 
ceous matter being developed in thin plates on its surface, often 
amounting to a pearly, and sometimes almost metallic, appear- 
ance ; an effect which, it is believed, has not been hitherto in- 
vestigated. 
Solution of potash acts very rapidly upon glass, as the chemist, 
often inconveniently, learns by the effect produced upon the bulb 
of a thermometer, employed to determine its boiling point, and 
which is always found corroded to a considerable extent after the 
experiment. 
It may also here be remarked (although not perhaps imme- 
diately connected with the subject), that from frequent observa- 
tions by a person in the habit of using solid carbonate of am- 
y 2 
