356 On the Possihll'ily of alloying 



coal in this experiment beyond tliat in No. 1, revived 34 i^raina 

 of additional produce, and in every respect effected a complete 

 reduction of the ore. 



No. 5. Argillaceous iron ore . . . . . . 500 grs. 



Oxide of manganese .. .. .. 100 



Charcoal 5th of the ore .. .. 125 



A rough metallic button resulted from the fusion of this mix- 

 ture, which with some sniall globules weighed 245 grains, equal 

 to 49 per cent. Ghiss grass green considerably transparent :— 

 four grains of charcoal remained unacted upon. 



No. 6. Argillaceous iron ore .. .. .. 500 grs. 



Oxide of manganese roasted .. .. 100 



Charcoal jth 125 



This mixture was accurately fused, and the result was a per- 

 fect metallic button weighino; 246 grains, which is equal to 

 49-[2g- per cent. Glass light green but clouded, transparent in thin 

 pieces ; not so bright as No. 5, but more transparent than No. 4. 

 The same quantity of charcoal remained over in this as in the 

 last experiment. From the three immediately preceding fusions 

 with aii excess of charcoal, it cannot be decidedly inferred that 

 th.e increased produce beyond that of No. \, 2, 3, was derived 

 from the manganese ; the reverse of this conclusion rather ap- 

 pears on the face of the experiments, seeing more metal was ob- 

 tained without manganese than with it, in the proportion of 

 257 to 245 and 24G. Lest any error should have taken place in 

 the weighing of No. 4, which vicldcd 257 grains of iron, this 

 experiment was twice repeated, and the results were 240 and 

 244, making an average of 245}. — Average of the fusions of 

 No. 5 and G with manganese , . . . 245y. 



On the whole, it appears doubtful whether in these experi- 

 ments there was any real alloy of manganese with iron; the so- 

 litary increase of weight in No. 3 and its peculiar fracture 

 being the only circumstance in favour of this conjecture. On 

 dividing and comparing the fractures of the metallic Ijuttons, 

 No. 4, 1 , and 8 (the same experiment) were exactly alike. In 

 these no manganese was used ; the quality was white cast-iron, 

 with a blueish glance, and a grain indicative of an approach to 

 steei. No. 1 and 2 evidently belonged to the class of steels, so 

 far as a commencement of grain was visible ; but in respect to 

 ductility or softness they were still closely allied to the hardest 

 sorts of cast-iron : the former had experienced an excess of car- 

 bon, the latter had absorbed all that was presented to them in 

 mixture. No. 5 and No. 6 were much alike, and similar to 

 No. 3, in which de-oxidated manganese was used. The frac- 

 tures were of a darker glance than 4, 7, 8, minutely but re- 

 gularly granulated and crystallized. It appeared, therefore, that 



the 



