PHYSICS. (!.'^)1 



trification thus produced is always positive, and tberelbie llie discharge 

 is favored at the cathode and rendered more diificult at the anode. 

 Indeed, the latter is, in a certain sense, prolonged by the layer of gas, 

 positively electrified, which surrounds it, and which produces the dark 

 -space around this electrode. In this space the discharge is convective 

 and continuous, becoming luminous and discontinuous only beyond it 

 at a variable distance depending upon the pressure of the gas and 

 upou its temperature. ( Wied. Ann.^ xxii, 305 ; J. Phys., December, 

 1885, II, IV, 570.) 



Jfaccari and Guglielmo have continued their investigations on the 

 heating of the electrodes jjroduced by the induction spark in rarefied 

 air. They had previously shown that for pressures of air above 10'"'" 

 the negative electrode heated more than the positive in the ratio of 1 

 to between 2 and 4. For lower pressures they now show that this ratio 

 increases slowly up to a pressure of 5™'», more rapidly from 5 to 1-2'"'", and 

 very rapidly up to a pressure of 0-27™™, when it reaches a maximum 

 value of 28. It then decreases, and for the lowest pressures obtainable 

 is reversed and becomes less than 1. (J. Phys., December, 1885, II, iv, 

 561.) 



Ayrton and Perry have communicated to the Physical Society of 

 London a paper on the most economical potential difference to employ 

 with Edison incandescent lamps. They point out the fact that it is not 

 sufficient to know that when a lamp is giving out a certain number of 

 candles it absorbs so much power per candle, and when giving out a 

 much larger number of caudles it absorbs so much less power i)er can 

 die. What must be known in addition is the life of the lamp ar each 

 of these two candle-powers before we can decide upon the most economi- 

 cal temperature for it ; since if the efficiency is low at low temperatures, 

 the hfe is great, and at high temperatures the larger efficiency will be 

 balanced somewhat by its short life. From a curve given by calcula- 

 tion, and assuming the cost of the lamp 5,S'., the number of hours of burn- 

 ing per year 560, and the cost of one electric horse-power for this lime 

 as £5, the authors show that the minimum cost per candle per year is 

 lid., and is obtained with a potential value of 101-4 volts, the cost ris- 

 ing to Is. per year if the potential falls to 98-7 or rises to 104 volts. 

 {Pliil Mag., April, 1885, V, xix, 304 5 Nature, March, 1885, xxxi, 450.) 



Fleming has made extended investigations into the phenomena of 

 incandescent lamps with special reference to their efficiency. From 

 statistics concerning the life, resistance, efficiency, and potential differ- 

 ence of such lamps he has constructed empirical equations showing the 

 mutual relations of these variables. A curve showing the relation of 

 any one of these to any other is called a characteristic curve for that 

 lamp. His results confirmed the law formulated by Ayrton and Perry, 

 that for a certain class of lamps {i. e. the Edison) the potential difierence, 

 minus a constant, varies as the cube root of the efficieucj', the latter 

 quantity being measured in candles per horse-power. The constant 



