^30 Mr Miishet on the Dpxtdaiion 
taKen place betweexit llie iron and the eartliy matter. In l)oth 
these experiments 41 pe7' ceni> of oxygen seems I’o have been 
removed from the ore. 
Having thns, as I conceive, rendered it exceedingly probable 
that the iron of the common ironstones of this country, either 
in their native or roasted states, is united with nearly 40 'per 
cent of oxygen, I shall conclude this paper with the result of 
some similar experiments with Lancashire ore, Avhich also seem 
to prove that the iron of these ores is not united to^ a greater 
dose of oxygen than the common ironstone, though their intense 
red colour, in the first instance, leads to a different conclusion. 
Lancashire ore, when reduced to a small size, and heated on 
an iron-plate till it becomes of a black colour, loses only from 
4 to 5 per cent, of its weight; (ironstone in the same time 
would lose from S5 to 85 per cent.^) : as the ore cools the black 
passes off, and the native red colour returns. In this state it has 
acquired a slight, though perceptible, obedience to the magnet. 
If the ore is heated for two hours, it loses in all from 6 to 7 per 
cent., and is then strongly magnetic, rendering it probable that 
the magnetic force in this case is more the result of a new ar- 
rangement of the metallic particles, than of the expulsion of 
so small a quantity of water. 
If the exposure of this ore be still further continued, and 
particularly with an increase of temperature, it will acquire its 
lost weight, assume a vitreous sort of fracture, and lose its obe- 
dience to the magnet ; such increase of weight being undoubt- 
edly owing to a further dose of ox 3 ^gen uniting with the iron, 
which I have never found to exceed that of the water expelled 
in the more moderate temperature ; and it may sometimes 
happen in experiments of this nature, that the ore, after being 
roasted for eight or ten hours, will weigh as much as when first 
introduced into the furnace, though, had it retired at an 
earlier stage of the operation, the deficiency of weight would 
have been as before stated. 
In general, when Lancashire ores (according to the sort,) are 
cemented in contact with carbonaceous matter, a loss of weight 
is sustained of from 25 to 35 per cent, from which is to be de- 
ducted that weight, consisting of water, which the ore would 
have lost by being exposed to a low red heat, leaving the dif- 
