.S7A' WILLIAM SIEMENS, F.R.S. 299 



to add alumina in the shape of aluminous iron ore ; manganiferous 

 ir-in ore may also be added with advantage. A charge of say 20 

 <-\vi. of ore is put into the furnace when fully heated, while it is 

 slouly ivvuhin;.r. In about forty minutes this charge of ore and 

 fluxing material will have been heated to bright redness, and at 

 this time from f> cwt. to 6 cwt. of small coal of uniform size (not 

 larger than nuts) are added to the charge, whilst the rotative 

 velocity is increased for a short time in order to accelerate the 

 mixture of coal and ore. A rapid reaction is the result : the 

 peroxide of iron being reduced to magnetic oxide begins to fuse, 

 and at the same time metallic iron is precipitated by each piece of 

 carbon, while the fluxing materials form a fluid slag with the 

 siliceous gangue of the ore. The slow rotative action is again 

 resorted to, whereby the mass is turned over and over, presenting 

 continually new surfaces to the heated lining and to the flame 

 within the rotator. 



During the time of this reaction, carbonic oxide gas is evolved 

 from the mixture of ore and carbon, and heated air only is intro- 

 duced from the regenerator, to effect its combustion within the 

 rotating chamber. The gas from the gas-producers is entirely, or 

 almost entirely, shut off during this portion of the process. When 

 the reduction of the iron ore is thus nearly completed, the rotator 

 is stopped in the proper position for tapping off the fluid cinder ; 

 after this the quick speed is imparted to the rotator, whereby the 

 loose masses of iron contained in it are rapidly collected into two 

 or three metallic bails. These are taken out and shingled in the 

 usual way of consolidating puddled balls ; the furnace is tapped 

 again and is ready to receive another charge of ore. The time 

 occupied in working one charge rarely exceeds two hours ; and 

 supposing that 10 cwt. of metallic iron is got out per charge, the 

 apparatus is capable of turning out at least o tons of puddled bar 

 per 24 hours. If anthracite or hard coke is available for effecting 

 the reduction of the ore, it should be crushed much finer than 

 when coal or brown coal is used, the idea being that each particle 

 of the reducing agent should be fully consumed during the period 

 of chemical reaction. If wood is used, it has to be charged for 

 the same reason in still larger pieces. 



If it is not intended to make iron, but cast steel, the balls may 

 be transferred from the rotator to the bath of a steel-melting 



