Page:Popular Science Monthly Volume 11.djvu/90

80 brought in succession into the action of the spark until the lines of the substance were clearly visible," If a skeptical person refuse to believe the results of Cappel, who tells us that of a milligramme ( of a grain) of nickel will just write the signature of that metal, what will he say when, glancing at Table B, appended hereto, he finds the statement that  of a milligramme ( of a grain) will sign its name in brilliant characters! And yet the writer does not hesitate to say that even a smaller amount of this metal will show a spectrum, for in these experiments a much stronger spark was used than was necessary to show a visible spectrum. When reduced to a minimum, by means of a miniature Leyden jar, improvised out of a test-tube, which still gave a distinct spectrum, the loss in weight, after passing 3,000 sparks, was absolutely inappreciable on the balance. The tables show another curious and unexpected result, viz., that the loss in weight of the volatile metals very slightly exceeds, and in some cases does not equal, the loss of the less volatile metals. Thus, in three different experiments of 3,000 sparks each, copper loses but milligramme, while gold loses  milligramme."

In one experiment the number of sparks was increased to 10,000, and the loss in weight was nearly proportioned to the increased number. In this case the sparks were passed at the rate of about 250 per minute.

The first column shows weight of metallic electrodes (in milligrammes) before passing the sparks.

The second column shows weight after passing 3,000 sparks.

The third column shows total weight of metal volatilized (in fractions of a milligramme).

The fourth column shows the amount of metal volatilized by each spark (in fractions of a milligramme).

The fifth column shows the amount of metal volatilized by each spark (in fractions of a grain troy).