dried and exposed to a strong heat; a minute quantity of white matter
remained, which, as far as my examinations could go, had the properties
of carbonate of soda. I compared it with similar minute portions of
the pure carbonates of potash, and similar minute portions of the pure
carbonates of potash and soda. It was not so deliquescent as the former
of these bodies, and it formed a salt with nitric acid, which, like
nitrate of soda, soon attracted moisture from a damp atmosphere and
became fluid.

"This result was unexpected, but it was far from convincing me that the
substances which were obtained were generated. In a similar process with
glass tubes, carried on under exactly the same circumstances and for
the same time, I obtained a quantity of alkali which must have been more
than twenty times greater, but no traces of muriatic acid. There was
much probability that the agate contained some minute portion of saline
matter, not easily detected by chemical analysis, either in combination
or intimate cohesion in its pores. To determine this, I repeated this a
second, a third, and a fourth time. In the second experiment turbidness
was still produced by a solution of nitrate of silver in the tube
containing the acid, but it was less distinct; in the third process
it was barely perceptible; and in the fourth process the two fluids
remained perfectly clear after the mixture. The quantity of alkaline
matter diminished in every operation; and in the last process, though
the battery had been kept in great activity for three days, the fluid
possessed, in a very slight degree, only the power of acting on paper
tinged with turmeric; but its alkaline property was very sensible to
litmus paper slightly reddened, which is a much more delicate test;
and after evaporation and the process by carbonate of ammonia, a barely
perceptible quantity of fixed alkali was still left. The acid matter in
the other tube was abundant; its taste was sour; it smelled like water
over which large quantities of nitrous gas have been long kept; it did
not effect solution of muriate of barytes; and a drop of it placed
upon a polished plate of silver left, after evaporation, a black stain,
precisely similar to that produced by extremely diluted nitrous acid.

"After these results I could no longer doubt that some saline matter
existing in the agate tubes had been the source of the acid matter
capable of precipitating nitrate of silver and much of the alka