1. A thin slip of platinum rolled in a spiral is placed in a small
crucible of retort carbon closed by a turned cover of the same material.
This is placed in a second larger crucible of refractory clay, and the
intervening space filled with lampblack tightly packed. The whole is
then heated to white heat for an hour and a half in a good wind furnace.
After cooling, the platinum is generally found to have been fused into a
button, with a marked increase in weight due to taking up silicon, which
has penetrated in the form of vapor through the walls of the interior
crucible.

2. If, in the preceding experiment, the lampblack be replaced by a
mixture of lampblack and rutile in fine powder, the slip of platinum
remains absolutely intact, and does not change in weight. Thus the
titaniferous packing recommended by Sainte-Claire Deville for preventing
the access of nitrogen in experiments at high temperatures also prevents
the passage of silicon. A mixture of carbon and finely divided iron is,
on the contrary, ineffectual. These facts seem to indicate that nitrogen
plays a part in the transportation of the silicon, as this is only
prevented by the same means made use of in order to prevent the passage
of nitrogen.

3. The volatility of free silicon at a high temperature is too slight to
account for the alteration of the platinum at a distance. This can be
shown by placing several decigrammes of crystallized silicon on the
bottom of a small crucible of retort carbon, covering the silicon with a
small flat disk of retort carbon upon which is placed the slip of
platinum. The crucible, closed by its turned cover, is then enveloped in
a titaniferous packing and kept at a brilliant white heat for an hour
and a half. The metal is found to have only very slightly increased in
weight, and its properties remain unaltered. This experiment was