450 MKSSRS. R. S. HUTTON AND J. E. PETAVEL ON ELECTRIC 



to the seat of reaction by means of a hollow electrode, the stream of gas thus not only 

 effectively diluting the carlxm monoxide, but serving to carry forward the metallic 

 vapour into a zone more favourable for its condensation. 



Even in the absence of carbon monoxide, carburisation can occur by direct union 

 of the metal with solid carbon. 



Some unpublished work of W. H. PATTERSON carried out in this laboratory has, 

 however, shown that in the absence of carbon monoxide this reaction only occurs 

 above a bright red heat, thus explaining the results already quoted in which the 

 metal was obtained, although doubtless it had not altogether escaped contact with 

 carbon. 



We are therefore in the following position : we have proved the facility with which 

 the direct reduction of alumina by carbon can be effected, and have shown that the 

 minimum temperature at which it can occur is already sufficiently high for the metal 

 to be produced in the form of vapour. 



Future work must be directed towards the application of high pressure for reducing 

 the vaporisation of the metal at the temperature of reaction, the rapid removal or 

 dilution of the carbon monoxide by a stream of inert or reducing gas, and a modifica- 

 tion of the regime to facilitate the condensation and prevent the collected metal from 

 flowing into a bed of highly heated carbon. 



Thus the necessary conditions for the successful direct reduction of alumina by 

 carbon seem to be fairly well defined, the outstanding problem being chiefly a matter 

 of the arrangement and construction of the furnace. 



The cost of most of the apparatus, which was specially constructed for the above 

 research, was defrayed by funds awarded by the Government Grant Committee of the 

 Royal Society. We have also been materially assisted, so far as the gas preparation 

 plant is concerned, by Messrs. Brunner, Mond & Co., Ltd., and the Tudor 

 Accumulator Co., Ltd. 



With regard to running expenses, the work has been much facilitated by the kind 

 way in which the ample resources of the Physical Laboratory of the Manchester 

 University have been placed at our disposal. 



In concluding, we desire to express our heart-felt gratitude to Professor ARTHUR 

 SCHUSTER for the never-failing interest and encouragement which he has given us 

 during the several years over which the research has extended. 



