October 23, 1919] 



NATURE 



j6i 



utilised their available bauxite, including the small 

 deposits of poor quality in Germany ; but the neces- 

 sity of finding a more widespread source was felt, 

 and a process was discovered and successfully put 

 into operation by Dr. Buchner, of Heidelberg, for 

 winning the metal from kaolin and kaolin-rich clays. 

 (Jlay is extracted with sulphuric acid, and, after 

 removal of the iron, the alumina is precipitated from 

 the solution with ammonia, four tons of a clay with 

 an alumina content of 30 per cent, yielding one ton 

 of the oxide. This process seems to have a con- 

 siderable future before it, and plans are laid for its 

 introduction on a large scale into Sweden. 



There is no bauxite in Norway, and it was first 

 proposed to work the clays after the Buchner and 

 other suggested methods. This, however, has for 

 the most part proved impracticable owing to the un- 

 favourable character of the clays, which are relatively 

 unweathered glacial deposits, not only poor in 

 alumina, 16-20 per cent., but with part of it bound 

 in alkali felspar, and therefore unavailable. In 1917 

 Prof. V. M. Goldschmidt, of the Mineralogical 

 Institute, Kristiania, conceived the idea of using 

 labrador-stone as a source of aluminium.' Labrador- 

 stone is a white rock extensively developed in south- 

 western Norway, and especially in the inner Sogne- 

 fjord district, where it builds the huge laccolitic 

 mountain masses so familiar to tourists, by 

 whom it is commonly mistaken for marble. The 

 main constituent of this rock is a plagioclase felspar 

 of the labrador group, the more felspathic varieties 

 containing only a very small amount of iron-bearing 

 . pyroxenes, and with an alumina content of 30 per 

 cent. Prof. Goldschmidt has found that the felspar 

 is surprisingly soluble in dilute acids, so that it can 

 be dissolved out, leaving a residue of insoluble ferro- 

 magnesian minerals and silica. The calcium and 

 sodium oxides of the felspar, 13 per cent, and s per 

 cent, in amount respectively, go into solution with 

 the alumina, and, using sulphuric acid, there is thus 

 a considerable loss in the form of a useless by- 

 product. This method is rejected for the manufac- 

 ture of alumina, although it is used in the prepara- 

 tion of sufficient quantities of aluminium sulphate to 

 satisfy Norwegian needs. By using nitric acid as a 

 solvent, not only is a waste of acid avoided, but the 

 precipitation with ammonia can be dispensed with— -a 

 valuable consideration in Norway, where ammonia 

 cannot be obtained cheaply in quantity. _ 



The main features of the process, which has been 

 worked out by Prof. H. Goldschmidt, are as follows : 

 — The labrador-stone is extracted with dilute nitric 

 acid, the 30 per cent, acid, first raw product of the 

 electrical air-industry, serving for this purpose. The 

 silica and greater proportion of the iron minerals 

 remain insoluble, aluminium, calcium, and sodium 

 going into solution together with a little iron. .'Vfter 

 removal of this, iron the solution is evaporated down 

 and the residue heated to a certain temperature at 

 which the aluminium salt alone is decomposed, the 

 nitric acid driven off being collected as a valuable 

 concentrate. Bv washing with water the nitrates cf 

 calcium and sodium are removed, to be recovered 

 and used in agriculture, the alumina remaining. 



This process seems to be full of promise for Norway 

 —a country with such abundant water-power, a 

 flourishing nitric acid industry, and an unlimited 

 quantitv of a raw material which few other countries 

 possess'; and hopes are entertained that a product will 

 be obtained which will not only suffice forlocal needs, 

 but also win a footing in the world's rapidly expand- 

 infj aluminium market. L- Hawkes. 



i " Om Aluminiumfrem«liilinK av Norske RaastofTer." By V. M. Gold- 

 schmidt. Siertryk av Tidsskrlft. for Kemi, No, 2, 1919. 



NO. 2608, VOL. 104] 



A' 



H YDRO-ELECTRIC DE VELOPMENT 

 WORKS.i 



N extremely able and informative paper has 

 recently been contributed to the Institution of 

 Electrical Engineers by Mr. J. W. Meares, chief en- 

 gineer of the Hydroelectric Service of India, dealing 

 with the general principles of the development and 

 storage of water for electrical purposes — a subject 

 which is of the greatest interest at the present time 

 from an industrial and economic point of view. Mr. 

 Meares 's paper is a general survey of the various 

 problems connected with the inception of hvdro- 

 electric installations; it outlines the conditions 

 I essential to the satisfactory development of any 

 i scheme of water-power, for it must, of course, be 

 I borne in mind that it is quite possible for a country 

 j to have considerable water resources, say, in the 

 I form of rivers, which are incapable of economical 

 development. The paper treats of all the preliminarv 

 , considerations relating to the gathering of supplies, 

 ! flow and storage, the lay-out and efficiency of hvdro- 

 ! electric plant, and the principles underlying the 

 I design of headworks, canals, and delivery mains. 

 i Supplies of water at different heads entail distinct 

 ' methods of treatment. The heads may be broadly 

 grouped as high, medium, and low, in which, with- 

 ' out too rigidly defining the boundary lines, high heads 

 are taken at from 300 ft. or 400 ft. to a possible 

 5000 ft., low heads from 3 ft. as a minimum to, say, 

 80 ft. or 100 ft., with medium heads between these 

 limits. A high head is associated with small volu- 

 metric flow, and a low head with a large flow; the 

 former is adapted to jet-impulse wheels of the Pelton 

 type, and the latter to pressure, or reaction, turbines. 

 In areas dependent on the collection of rainfall 

 for supplies, the amount of fall and the run-off are 

 important considerations. The following empirical 

 tfible devised by Mr. G. T. Barlow, formerly Chief 

 Engineer of the United Provinces, India, gives a 

 working hypothesis for preliminary calculations which, 

 while perhaps inapplicable to many parts of Europe 

 or -America, affords a clear indication of the nature 

 of the variations to be met with in a particular 

 locality : — 



Perrentaee run-off 

 A B C D E 

 Light falls, say under ij" in 



24 hours 135 10 15 



Medium falls, say from ij" 



to 3" in 24 hours ... 

 Heavy falls, say above 3" 



A is flat, cultivated and black cotton soil catchment. 



B is flat, partly cultivated and stiff soils. 



C is average catchment. 



D is hills and plains, with little cultivation. 



E is very hilly, steep, and rocky, with very little 

 cultivation. 



The paper also contains a table giving the over-all 

 commercial efficiency of hydro-electric plant as 

 follows : — - 



10 15 20 25 33 

 20 33 40 SS 70 



For 500 kilowatts 

 ,, 1000 ,, 

 ,, 1500 ,, 

 ,, 2000 ,, 

 ,, 3000 ,, and over 



74 per cent. 

 76 ,, 



78 .. .. 



80 „ ,, 



82 „ ,, 



.As a rough approximation, the capacity of plant in 

 kilowatts may be obtained by dividing by is the 

 product of the quantity of water in cubic feet per 



1 "The General Principles of the Development and Storage of Water for 

 Electrical Purposes." By J. W. Meares. 



