from the RIVER DEE. 257 



of the 6 grains in muriatic acid, a portion of black matter fepa- 

 rated. It weighed 0.2 grains, and was totally diflipated before 

 the blow-pipe in a white fmoke. Hence, it muft have been ar- 

 fenic Thefe 1.5 gr. are equivalent to rather more than 1 

 grain of metallic arfenic. Thus, it appears, that the 6 grains 

 contained 1 grain of arfenic, which explains the whitenefs of 

 their colour. The reft was iron. It can fcarcely be doubted, 

 that the proportion of arfenic prefent was originally greater. 

 Some of it muft have been driven off when the iron oxide was 

 heated with oil. 



10. The infoluble refidue, (No. 4.), was with great difficul- 

 ty diflblved in fulphuric acid. When the folution was mixed 

 with ammonia, a white powder fell, which weighed 0.8 grains. 

 It was accidentally loft, before I examined its properties. But 

 I have no doubt, from its appearance, that it was oxide of tita- 

 nium. 



11. Thus, from the 100 grains of iron-fand, trie following 

 conftituents have been extracted by analyfis : 



Black oxide of iron, - 98.70 



White oxide of titanium, 12.65 



Arfenic, - - 1.00 



Silica and alumina, - 1.50 



Total, 113.85 



Here there is an excefs of nearly 14 grains, owing, without 

 doubt, to the combination of oxygen with the iron and the 

 titanium during the analyfis. 



Had the iron in the ore been in the metallic ftate, the ex- 

 cefs of weight, inftead of 14, could not have been lefs than 30. 

 For the black oxide is known to be a compound of 100 metal 



and 



