364 
Cast Steel* 
upper surface of this glass was frequently of a pure pear* 
ly white colour. In one experiment, where -^th of char- 
coal was used, I obtained a large quantity of glass cellu- 
lar throughout. Each cell was surmounted upon the top 
with concentric circles of pearly lines, forming a curious 
and pleasing effect. 
Having fully satisfied myself that operations performed 
in close vessels thus prepared were subject to uncertainty, 
arising as well from the formation of glass as from some 
unknown affinity exerted upon the charcoal ; and having 
performed several experiments with well filled open cruci- 
bles, with charcoal alone, wherein I found little compara- 
tive loss, I performed a very accurate series of experi- 
ments, viz. 
Having selected a parcel of well-prepared Stourbridge 
clay crucibles, with covers exactly fitting, I proceeded 
to make the following experiments upon the quantity of 
charcoal which forms crude iron and steel ; first premis- 
ing that both crucible and cover were brought to a bright 
red heat before the substances acted upon were introduc- 
ed. This was done with the greatest possible caution, to 
avoid volatilizing any part of the charcoal, and rendering 
the result inaccurate. From the approximation of these 
results in repeating most of these experiments, I found 
that no material difference had occurred* Grains. 
Exp . I. Swedish bar iron - - 885 
Charcoal 2 , or - grs. 442 
This mixture was exposed for half an hour, and 
a perfect button of supercarbonated crude iron 
was obtained. Along with the metal was found 
of intensely black charcoal not taken up - 290 
Charcoal disappeared, equal to 34*4 per cent. 1 52 
The metallic button now obtained was found to weigh 928 
