In every case in which he had procured much soda, the glass[61] at the point of contact with the wire seemed considerably eroded; when by substituting an agate for a glass cup, no fixed saline matter could be obtained. Its source therefore, in the former case, was evidently the glass. [61] It is perhaps a fact not very generally known, that glass, to a certain extent, is decomposable by water: if some of it in a powdered state be triturated with distilled water, in a short time the turmeric test will indicate a portion of alkali in solution. With respect to Mr. Sylvester's experiment, already noticed, it was sufficient to say that he conducted his process in a vessel of _pipe-clay_, which not only contains lime, but may also include in its composition some of the combinations of a fixed alkali. On resuming the enquiry, it was Davy's first care to remove every possible source of impurity: he accordingly procured cups of agate, which, previously to being filled, were boiled for several hours in distilled water; and a piece of very white and transparent _amianthus_, a substance first proposed for this purpose by Dr. Wollaston, having been similarly purified, was made to connect the vessels together. Thus was every apparent source of fallacy removed; but still, after having been exposed to Voltaic action for forty-eight hours, the water in the positive cup gave indications of muriatic acid, and that in the negative cup, of soda! The result was as embarrassing as it was unexpected; but it was far from convincing him that the bodies thus obtained were _generated_:--but whence arose the saline matter? Did the agate, after every precaution, still contain some very minute portion of saline matter, not easily discoverable by chemical tests? To determine this question, the experiment was repeated a second, a third, and a fourth time: the quantities of saline matter diminished in every successive operation, which sufficiently proved that the agate must at least have been _one_ of the sources sought for; but four additional repetitions of the process convinced the operator that it could not be the only one; that there must exist some other source from which the alkali proceeded, since it continued to appear to the last, in quantities sufficiently distinct, and apparently equal, in every experiment. This was extremely perplexing: every precaution had been taken--the agate cups had even been included in glass vessels, out of the reach of the circulating air--all the acting materials had been repeatedly washed with distilled water; and no part of them in contact with the fluid had ever touched the fingers.