580 Dr. Schaf haeuti on the Different Species of 
little. By actual analysis, a considerable quantity of carbon 
was found in it, as well as arsenic, but no trace of silicon. In 
the retort, treated with hydrochloric acid as usual, the evolu- 
tion of gas lasted upwards of three weeks : the evolved gas 
had no smelly which proves that carbon, at least alone^ cannot 
be the cause of the bad smell of this description of hydrogen. 
The residuum of the solution was black, smoking very much 
during ignition, and leaving a small quantity of dirty red re- 
siduum, which was entirely soluble in hydrochloric acid, 
leaving only a few small black scales behind. The solution 
contained iron, but no traces of silica. 
I must here mention a fact but little known, that all piles of 
iron which are to be welded in a reverberating-furnace, must 
rest on a bottom which contains a large quantity of free silica. 
When the pile of iron is heated in such a reverberating-furnace, 
the silicon and the iron on the surface of the pile become 
oxidized, forming a very tough half-melted slag, which does 
not at all prevent the access of air, and the iron would burn 
into cinders did not the silica of the bottom combine with the 
slag next to it, forming a liquid silicate and giving an equal 
quantity of silica to the uppermost bar of iron in the pile, un- 
til this liquid slag is spread over the whole pile and its inter- 
stices. Iron piles heated upon a slag bottom will not weld 
but burn, a circumstance which I found always overlooked. 
When iron is heated on a bottom composed of siliceous 
matter, for a long time and at the highest degree of heat, the 
silicon of the bottom is reduced by the carbon of the iron as 
well as of that of the flames, and combining with the iron ren- 
ders its texture loose, makes it finally melt, and produces that 
which is usually termed burnt iron. 
The silicon is, in fact, the cause of the welding property of 
the iron. Thus silica is sometimes the cause of malleable 
iron melting in our common fires. The general idea, that 
malleable iron can be melted even in Sefstroem’s or Knight’s 
blast-furnace, is quite erroneous. An accurate analysis of the 
iron before and after fusion will soon convince us of the truth 
of this assertion, and we find invariably that the iron during 
fusion had combined either with carbon, or with silicon, or 
with both. We have seen above, that iron, even though it 
contains a large quantity of carbon, sometimes developes a 
perfectly inodorous hydrogen, and an inodorous hydrogen is 
therefore no proof of chemically pure iron. The process of 
welding iron consists in heating the skeleton grains of iron, 
contained in the mass, in order to excite all their attractive 
forces, but at the same time to prevent their combining with 
any other body, especially carbon, in which case only the 
