52 



NATURE 



[September 20, 191 7 



made in magnesia crucibles heated in an Arsem 

 vacuum furnace by melting- pure iron and afterwards 

 dropping in aluminium, the latter being suspended 

 in the form of wire or rod from a very fine wire 

 extended between insulating posts which passed 

 through the cover of the furnace. At the required 

 moment tliis was fused. In this way alloys containing 

 up to 13 per cent, of aluminium were obtained, and 

 are stated by the authors to be less contaminated with 

 impurities than any previously made. The alloys 

 classed as uncontaminated contain from ooi to 002 

 per cent, of carbon. Other alloys containing more 

 carbon are classed as contaminated and are used to 

 show the effects of carbon. Aluminium is a more 

 powerful deoxidiser than silicon and does not com- 

 mence to combine with iron until all the oxides present 

 have been reduced. It forms solid solutions with iron 

 throughout the range studied. The alloys have been 

 studied chiefly from the point of view of their mag- 

 netic properties. Aluminium, like silicon, has a bene- 

 ficial effect when added in small quantities. The best 

 alloy obtained, containing 04 per cent, of aluminium, 

 has a maximum permeabifity above 35,000 when 

 annealed at iioo°C. The hysteresis loss for B. max.= 

 10,000 and 15,000 is 450 and 1000 ergs per c.c. per 

 cycle respectively. The specific electrical resistance 

 increases, about 12 microms for each per cent, of 

 aluminium added up to 3 per cent. Above this the 

 rate of increase falls off gradually. 



Owing to the shortage of superphosphate prepared 

 in the ordinary way with sulphuric acid various sug- 

 gestions have been made to supply the deficiency (such 

 as the action of nitre-cake on rock phosphate), or to 

 use what ordinary superphosphate there is available 

 to the greatest advantage. For the latter purpose ad- 

 mixture with insoluble phosphate has been advocated. 

 Concerning this proposal, the results obtained by Mr. 

 G. Scott Robertson (Journal of the Society of Chemical 

 Industry, June 30) are of considerable importance. Mr. 

 Robertson finds that when ordinary superphosphate is 

 mixed with basic slag the greater part of the water- 

 soluble calcium phosphate contained in the former 

 reverts to the insoluble form almost at once by the 

 action of the free lime in the basic slag, whilst on 

 keeping the mixture a slower change in the same direc- 

 tion is observed Thus when a superphosphate con- 

 taining 26 per cent, of water-soluble calcium phosphate 

 was mixed with an equal quantity of basic slag (con- 

 taining 17 per cent, of free caustic lime), the mixture, 

 instead of containing 13 per cent, of water-soluble 

 phosphate, contained only 5 per cent, immediately after 

 mixing, and only 2-8 per cent, after keeping for four- 

 teen days, after which period the composition altered 

 but slightly. When the basic slag is replaced by an 

 equal proportion of natural rock phosphate there is a 

 much smaller reversion, so that there is no serious 

 objection to such a mixture being used by farmers. 

 A mixture of Gafsa rock phosphate with an equal 

 weight of superphosphate contained 12-7 per cent, of 

 water-soluble phosphate (instead of 13 per cent.) imme- 

 diately after mixing, and 10-4 per cent, at the end of 

 fourteen days. Gafsa phosphate (which rarely contains 

 more than 075 per cent, of calcium oxide in the form 

 of free carbonate) is probably the most suitable for 

 mixing with superphosphate. Then come Egyptian 

 phosphate, Florida pebble phosphate, Makatea Island 

 phosphate, and Tunisian and Algerian phosphates, in 

 order of suitability. 



The Scottish Motor Traction Company, which runs 

 a number of motor omnibuses in Edinburgh and the 

 district, has taken up the use of coal-gas as a substi- 

 tute for petrol, and a photograph of one of its 

 NO. 2499, VOL. 100] 



vehicles appears in the Engineer for September 14. 



'■ The gas-holder consists of a large flexible container, 



which covers the entire roof of the vehicle, the arrange- 



I ment being similar to that adopted in other vehicles 



' of this type. The simplicity and low cost of the 



I flexible receiver charged with gas at low pressure war- 



i rant careful consideration of the system as a war 



measure. Although the quantity of gas which can 



thus be carried is limited, this disadvantage is to some 



extent compensated for by the ease with which the 



I gas-holder can be recharged. Further, high-pressure 



receivers are not easy to procure at the present time. 



Owing to the low pressure — about 0-2 in. of water — 



a meter delivering into the receiver races unless some 



i form of throttle is interposed. A short length of small- 



[ bore piping, or a diaphragm with a suitable aperture, 



I is recommended so as to retain the pressure drop in 



i the meter within reasonable limits. 



I 



I Writing in the Tohdku Mathematical Journal, vol. 

 xi., 3, Mr. Hastime Tanate discusses the logical foun- 

 dations for negative and imaginary quantities and ex- 

 presses the view that the existence of these quantities 

 may be explained independently of any geometrical 

 considerations. 



OUR ASTRONOMICAL COLUMN. 



Distribution of Spiral Nebula. — The distribution 

 of spiral nebulae has been further investigated by Dr. 

 R. F. Sanford, with the aid of photographs obtained 

 with tjie Crossley reflector (Lick Observatory Bulletin. 

 297). The photographs were taken with exposures of 

 the order of twelve hours, the object being to find out 

 whether new nebulae could be detected in regions of the 

 Milky Way which have hitherto seemed barren of 

 them. They afford no evidence of undiscovered faint 

 nebulae in the regions where they have not previously 

 been found with shorter exposures. It is shown that 

 there is greater average brightness for the extra-galac- 

 tic than for the galactic spirals, and that the nebulae 

 which lie nearest to the Milky Way are on the average 

 of larger angular size than those away from it. F. G. 

 Brown has shown that the larger nebulae in general 

 are the brighter, but this is not true of spiral nebulae 

 near the Milky Way, which are large and faint. Thus, 

 if angular size be taken as a criterion of distance, it 

 follows that something cuts off the light from the 

 galactic spirals, thereby letting only the nearer ones be 

 perceptible, and then only with diminished brightness. 

 An arbitrary and general distribution of the spiral 

 nebulae can be best harmonised with the observed 

 features of the distribution by assuming the existence 

 of an obstructing medium, which is irregularly scat- 

 tered throughout the galaxy. It is considered prob- 

 able that the spirals are not only outside our own 

 system, but that they can have no intimate connection 

 with it dynamically. 



The Great Solar Prominence of 19 16, May 26. — 

 A detailed account of the great eruptive prominence of 

 1916, May 26, which reached a maximum height of 

 half a million miles, and in some parts attained a 

 velocity of 457 km. per second, has been given by Mr. 

 Evershed (Kodaikanal Bulletin, No. 55). One of the 

 most striking results of the measurements of the photo- 

 graphs is to show that all parts of the prominence 

 were moving radially outwards from a point in the 

 chromosphere at the base of the main column. It is 

 considered probable that an eruptive prominence begins 

 as an unusually dense low-lying mass of gas which 

 may persist without much change for several days, and 

 then suddenly become unstable, becoming subjected to 

 a force which tears it to shreds and sends the frag- 



