December 23, 1915J 



NATURE 



467 



[iinch. The visual observations made by von Guth- 

 ick during- the summer of 191 1 {Astronomische Nach- 

 ichten, 4818) show that the average amplitude of the 

 jht curve is 025 mag. in each case, and the periods 

 re just about one day. With magnitudes about 58 

 md spectra of the usual advanced helium type (B8-B9) 

 ither would supply the Ottawa observers with just the 

 titf kind of spectroscopic problem they seem to revel 



The light curve of a new Algol star, 3, 1915 Coronse 

 TD + 3o°, 2688, 9-1 mag.) has been published by von 

 'uno Hoffmeister (Astronomische Nachrichten, 4821) 

 ;^ho first detected its variability during August, 1914. 



series of observations, including fifteen minima, gives 

 %e following elements: — Period, i7h. 26m. 4-185.; 



)rmal magnitude, 9-28; minima, 9-76; duration of 

 jase, 4-3 hours. 



A TUNGSTEN ARC LAMP. 



A NEW and extremely interesting development of 

 ■^*- the nitrogen- or argon-filled electric lamp is 

 described in a paper by Messrs. E. A. Gimingham 

 and S. R. Mullard in the Journal of the Institution of 

 Electrical Engineers for December i. The lamp is the 

 result of experiments started in 1913 in the Edison 

 and Swan United Electric Light Co.'s laboratory. 

 Instead of the light being emitted by an incandescent 

 spiralised filament of tungsten, it is given out by an 

 actual arc between tungsten electrodes. 



The first problem to be solved was, of course, to 

 strike the arc. In the first forms of lamp, two tung- 

 sten electrodes were normally in contact, and an ex- 

 pansion strip consisting of a strip of molybdenum, to 

 one side of which was welded a thin strip of copper, 

 was fixed to one of the electrodes. This strip was 

 heated by a spiral filament of tungsten, in series with 

 the electrodes. It was found, however, that the elec- 

 trodes were frequently partially fused together, so that 

 the expansion strip failed to separate them, in addition 

 to which a certain amount of spluttering occurred, 

 which shortened the life of the lamp. To overcome 

 this defect, an entirely difi'erent method was employed, 

 enabling the arc to be struck between fixed electrodes 

 consisting of two small globules of tungsten. A tung- 

 sten filament was made to glow close to the electrodes, 

 and ionised the gas between them. This made the 

 gas conducting. The ionising filament was connected 

 in parallel with the arc circuit, and was connected up 

 for a few seconds only by means of a switch, and 

 then disconnected as soon as the arc was properly 

 struck. 



In continuous-current lamps, however, difficulty was 

 experienced in inducing the arc to leave the filament 

 and pass to the negative electrode. The heat of the 

 arc, moreover, destroyed the ionising properties of the 

 filament. To cure this, two expedients were adopted. 

 First of all, instead of a simple tungsten filament, it 

 was found that better ionising properties could be 

 obtained from a mixture of tungsten with zirconia, 

 yttria, thoria, and other refractory oxides, and that, 

 in addition, such a filament had a longer life. 

 Secondly, the use of the expansion strip was reverted 

 to for the purpose of moving the positive electrode 

 along, after the arc had been struck, to another 

 part of the filament, which acts as the negative 

 electrode, and saves the central part of the latter 

 (opposite to which the positive electrode returns when 

 the lamp is switched off) from being rendered inactive. 



Such a lamp is shown diagrammaticallv in Fig. i. 



The current first passes through the circuit A and the 



filament BB', while at the same time there is the full 



potential difference of the mains between the tungsten 



NO. 2408, VOL. 96] 



globule, E, and the filament. The gas in this gap L' 

 ionised, and the arc strikes. As the main current 

 then passes through the relay C, this is actuated, and 

 breaks the ioniser circuit. Meantime the heat from 

 the arc causes the expansion strip F to warp, and 

 moves E further along the filament. 



Alternate methods, used for lamps of higher candle- 

 power, are to arrange the ionising filament and the 

 electrode, so that, after striking, the arc rises away 

 from the active part of the filament, or to employ 

 two electrodes and a change-over switch. In the 

 latter case, the arc is first struck between the filament 

 (connected to the negative pole) and the smaller elec- 

 trode as anode, and as soon as the latter becomes 

 brightly incandescent, the filament is switched off and 

 the polarity changed, so that the arc is formed between 

 a larger positive electrode and the smaller one, which 

 is now negative. 



The efficiency of the lamp, for a life of a few 

 hundred hours, is in the order of ^ watt per c.p., and 

 the Intrinsic brilliancy of the light source about 

 10,000 c.p. per sq. in. Spectrum analysis shows per- 

 fect continuity and strength over the whole visible 

 spectrum, and at the same time richness In the ultra- 

 violet. 



One of the immediate applications of the new lamp 

 is for projection purposes, and the manufacturers have 

 already placed on the market a practical form of lan- 

 tern lamp, with a resistance arranged so that It may 

 be set for any voltage. This lamp is simply started 

 by means of a push switch, which closes the ioniser 

 circuit, and is to be released as soon as the 

 arc is struck. In Its smaller sizes the lamp 

 will doubtless displace entirely the Nernst fila- 

 ment which. In spite of Its disadvantages owing 

 to the negative temperature coefficient render- 

 ing it very sensitive to pressure variations, has been 

 used very largely for small projectors, surgical exam- 

 ination lamps, etc., and has remained popular until 

 the stock of the German Nernst burners in this 

 country at the outbreak of the war became exhausted. 

 For photographic enlargement lamps It should also 

 have a useful field, and, ultimately, as a substitute 

 for arc lamps for kinema theatres, and the Internal 

 lighting of shops and large buildings. The spectrum 

 shows that It should be Invaluable for colour matching. 



