different Earths for Carton. %f$ 



This mixture was perfectly fused, and equally destitute 

 t)f revived iron as the former. In Experiment VI, the same' 

 proportion of mixture yielded 47 per cent, of iron. I after- 

 wards found that, when this quantity of oxide was used^ 

 the first symptoms of separation took place with l-5th its 

 weight of charcoal. . . t . 



I know of no satisfactory reason which can be given to 1 

 solve the material difference of the results in these experi- 

 ments, simply arising frofn quantity, unless the additional 

 exposure- requisite to reduce the large quantity destroys a 

 greater portion of the charcoal uselessly than when the small 

 portion of matter is operated upon. The process of sepa- 

 ration divides itself into branches. The first consists in the 

 action of a well known affinity,—- the combination of the 

 tarbon with the oxygen 6'f the oxide, which leaves the par- 

 ticles of metal highly disposed to become the subject of the* 

 second affinity, viz. the ,v combiriation of the carbonaceous 

 matter with the iron. Now, if we suppose that in both 

 Series of experiments there existed in each mixture a particle 

 of carbon for" every particle of metal 5 and one for every par- 

 ticle of oxygen (if I am allowed the expression),- then we 

 'cannot see that in either operation a difference of result 

 Should take place, provided the experiments are alike accu- 

 rately performed, and the time of exposure similar.- The 

 fact, however, turns out very differently, and, I make nor 

 doubt, is guided by some regular and well established cause- 

 May not the last portions of oxygen be more difficult to re- 

 move from the large than fro'rii the small quantity? or, iri 

 other words, May they not require a greater dose of carbon 

 to saturate them under the double circumstance of increased 

 quantity and approximation to fusion ? 



Although I could not reduce the mixture operated upon iri 

 Exp. X, I found this easi-ly effected by the addition of chalk. 



XV. Oxide of iron - - 200 grs; 



Charcoal - * - lOO' 

 Chalk - - - 100 



But in place of finding the revived iron carbufated as iii 

 fhat experiment, I found it quite the reverse. 



A review of these experiments performed with pure iron/ 

 Oxygen, and carbon ^ will convey a pretty accurate idea of 

 the real quantity of the latter necessary to revive certain 

 portions of iron. It is however most difficult to decide 

 what portion of the carbon unites to the oxygen of the ox- . 

 ide, and what to the metallic part. I have uniformly found 

 that all iron ores and oxides take up a portion of carbon 

 before any of the metal is separated. This combination-, 



U 9 bring s 



t 



