438 RECORD OF SCIENCE FOR 1887 AND 1888. 



mentioued. Third, by passing the chlorine over a heated mixture of 

 alumina, salt, and carbon, a double cliloride of alnminnm and sodium is 

 produced. Fourth, the aluminum is extracted by heating, in appropri- 

 ate furnaces, 80 pounds of the aforesaid double chloride, 25 pounds of 

 sodium, and 30 pounds of cryolite, the latter serving as a flux. The 

 charge is maintained for two hours at a temperature of 1,000°, and 

 yields about 8 pounds of aluminum containing not over 2 per cent, of 

 impurities. The present cost of the metal is from $10 to $12 a pound, 

 and under this process it is estimated that the price can be reduced to 

 about $1.25. (Chem. Kews, lviii, 04.) 



Some jnoducts of the Gowles electric furnace. — The following observa- 

 tions are due to Prof. C. F. Mabery. 



First, as to the reduction of aluminum from corundum : When that 

 mineral, mixed with charcoal, is submitted to the action of an electric 

 current in the furnace it fuses and reduction proceeds rapidly. The 

 liberated metal, vaporized, partly condenses in tlie upper layer of char- 

 coal, while part escapes into the air and is burned. The mixture of 

 carbon and metal contains from 10 to 50 per cent, of the latter. If sili- 

 con is present in the ore it is also reduced and the silicon alloys with the 

 aluminum in variable proportions up to 10 per cent, of the former. With 

 clay a similar reduction takes place, and when the alloy obtained is 

 treated with hydrochloric acid, (;rystalline or gra])hitoidal silicon is left 

 behind. The slags which are produced usually contain some metallic 

 aluminum, rarely iron, and almost no silicon. When corundum is re- 

 duced in presence of iron, the latter retains aluminum to a greater or 

 less extent. One product, which seems available for foundry purposes 

 and for introducing aluminum into steel, contains from 6 to 10 percent, 

 of aluminum. It resembles in api)earauce a hard, finegrained white 

 iron. 



Secondly, as to the reduction of silicon. Sand is more easily reduced 

 in the furnace than corundum. It first fuses, and reduction immedi- 

 ately follows, yielding crystallized silicon. Copper mixed with the 

 charge dissolves the silicon up to 14 per cent., giving a bronze of re- 

 markable strength. The product obtained when silicon is reduced in 

 the absence of metals contains more or less of an amorphous greenish 

 substance of vitreous luster, which always occurs between the unre- 

 duced sand and the free silicon. In composition it appears to be the 

 hitherto unknown silicon monoxide, SiO. Its specific gravity is 2.893, 

 or somewhat higher than quartz. (Am. Chem. Journ., ix, 11.) 



Alum in haking-powders. — In a series of experiments upon baking- 

 powders used in the United States, Professor Mallet has reached the 

 following conclusions: The greater part of the aluminous powders con- 

 tain alum, acid calcium phosphate, sodium bicarbonate, and starch. 

 They give off very different proportions of carbon dioxide, and there- 

 fore require to be used in different amounts. Generally there is an ex- 

 cess of the alkaline ingredient in these powders, but sometimes the re- 



