TRANSACTIONS OF SECTION G. 669 



and other hydraulic machines on which experimental investigations will be carried 

 out. The water discharged from these machines passes into one or other of three 

 rectangular channels formed in the floor, and the quantity is measured by allowing 

 the water to pass over weirs. The water then flows into one or other of two 

 large rectangular tanks, each 11 feet square by 5 feet deep, sunk below the lower 

 floor-level of the laboratory, where it is measured again by floats, with rods 

 moving in front of carefully graduated vertical scales. From these lower 

 measuring tanks the water is lifted by an electrically driven 20 horse-power 

 centrifugal pump back to the storage tank in the water-tower. The hydraulic 

 equipment includes a Venturi meter and other forms of meters and a considerable 

 amount of other apparatus for experimental work. 



The paper was illustrated by a considerable number of lantern views. 



2. Glow-Lamps and the Grading of Voltages.^ 

 By Sir W. H. Preece, K.C.B., F.B.S. 



Twenty-five years have passed without much change in the character of 

 carbon filament glow-lamps, but rivals are now springing up which threaten to 

 lower its predominancy. The British Association System of electrical units has 

 become an international language. Their use is universal. The connection 

 between energy, temperature, light, and the properties of matter are now so well 

 known that the production of light is the subject of well-defined laws. The 

 efficiency is indicated by the watts per candle. In regard to the life of the lamp 

 the higher the efficiency the shorter the life. The relative useful life is then easily 

 tested. The useful life is the time the lamp takes to lose 20 per cent, of its 

 specified candle-power. A new graphite filament has recently been produced in 

 the United States which gives greatly improved results. The resistance increases 

 with temperature. It does uot fall as at present. Its efficiency is 2-5 watts per 

 candle, and for street lighting when diffused it gives 1*8 watts per candle. 



The Nernst lamp with an efficiency of 1-325 watts per candle, the Osmion 

 lamp with an efficiency of 1-5 watts per candle, the Tantalum lamp with a mean 

 efficiency of 2'1 watts per candle, the Zirconium lamp with an initial efficiency 

 of 1-07 watts per candle, and the Mercury vapour lamp were all referred to. The 

 rare metals have the great advantage that like the new graphite their resistance 

 increases with their temperature, which makes them self-regulating. Their light 

 is whiter. The life of carbon lamps is affected very much by variation of voltai^e, 

 especially when of poor efficiency like all untreated 220-volt lamps. The use of 

 these lamps is extravagant. The author is using 110-volt lamps in series in pre- 

 ference. A 220-volt lamp wastes its own value in every 500 hours when the 

 mean price is 3^d. per unit. The higher the price the shorter the time of waste. 

 All lamps blacken with age and with such regularity that they can be classified 

 according to their usage by comparison with a scale of standard lamps artiBcially 

 blackened. Thus the useful life is easily determined. 



The standard lamps in candle-power are 8, 16, and 32. It is proposed to add 

 12 candle-power and 25 candle-power. Lamps should comply with a standard 

 specification (which is under consideration), and they should be tested bv 

 authorised distributors or by a public institution. The National Physical 

 Laboratory is fully authorised to do this. 



Clause 18 of the Electric Lighting Act of 1882, still in force, has an injurious 

 and deterrent influence on the progress of electric lighting, and it needs repealing'. 



The most serious evils extant are the departures from standards and tbe great 

 variations of systems and of voltages. Individual faddism seems to have been 

 exercised to see how far it was possible to depart from rule. Of 472 systems 

 only eleven adopt 110 and eighty- two 220 volts. 



The rating of lamps sold in the United Kingdom and their branding is in a 

 very unsatisfactorj' condition. Reliance cannot be placed on the present system. 



' I'ublished in the Electrician, August 10, 1906. 



