containing iron more difficult to be revived than ever, 
owing to the great quantity of oxygen combined with it* 
In cases of this kind, the disengaged hydrogen manifests 
its escape, by changing the colour of the flame from a 
mellow white to a pale sickly blue. 
It must from all this appear obvious, that, when torre- 
faction has been properly conducted, a very considerable 
part by weight of the whole will be dissipated : the ab- 
sence of these volatile substances always leaves the iron- 
stones more or less magnetic. The loss of w eight, how- 
ever, is very different in the various classes of iron-stones, 
even when they are all exposed for the same length of time 
to a degree of heat capable of expelling those mixtures, 
which, under such circumstances, assume the gaseous 
state. 
1. Calcareous iron-stone, when properly exposed to 
torrefaction, loses more of its w eight than either of the 
other two classes. Where the lime is abundant, I have 
found this iron-stone lose 38 per cent, but more common- 
ly 35 and 36 per cent, of water, carbonic acid, and sul- 
phur. 
2. Argillaceous iron-stones, exposed to a similar de- 
gree of heat, and treated in the same manner, commonly 
lose in weight from 32 to 35 per cent. 
3. Siliceous iron-stones always give out less weight 
when exposed to torrefaction under similar circumstan- 
ces with other stones. In common they lose from 
27 to 30 per cent. I have, though rarely, found them to 
give out 25 per cent. only. 
The extreme loss of weight in each class must be con- 
sidered as the utmost point to which torrefaction can with 
safety be urged, without exposing the iron-stone to an ac- 
cumulation of weight by the combination of oxygen. 
From the variously compounded natures of iron-stones 
arise the various calculations of the loss which they are said 
