PHYSICS. 495 



influence of magnetization is extremely feeble. {Wied. Awn.,Y, xiv, p. 54; 

 J. Phys., June, II, i, p. 291.) 



The Thomson effect, or the transference of heat by an electrical cur- 

 rent, has been investigated by Trowbridge and Penrose, both in nickel 

 and carbon. The strip of nickel was 45<'™ long, 2.0*^'" wide, and 2™'" 

 thick. The transference of the heat was found to be negative; *'. c, heat 

 is absorbed by a current which passes from hot to cold, and evolved 

 when it passes from cold to hot. On testing the question whether the 

 phenomenon is reversible, the results were quite inconclusive, though 

 tending to prove rather than to disprove it. No efiect upon the Thom- 

 son eflect was produced by a magnetic field. Eepeating with copper the 

 first experiment with nickel, the relative values for the two metals were 

 obtained. Calling copper 2, the effect for nickel is 2.25, but of opposite 

 sign. The carbon employed was that of a Faber pencil, and the direc- 

 tion of the Thomson effect was found to be negative, like nickel. {Am. 

 J. ScL, November, lll,xxiv, 379.) 



Elster and Geitel have investigated the electrical properties of flame, 

 using the Thomson quadrant electrometer. They find that as long as 

 either of the electrodes is outside the flame and the other inside, the 

 outside one is positive and the other negative. The film of hot air 

 outside the flame is always positive and the flame inside relatively neg- 

 ative. The result is the same with gas flames, caudle flames, alcohol 

 flames, and even in air flames burning in coal gas, the difl'erence of poten- 

 tial being from 1^ to 1^ times that of a Daniell cell. If in place of two 

 platinum wires, the upper one is copper, the E. M. ¥. rose to 2 Daniell 

 cells, w'ith aluminum it rose to 3, and with magnesium to 3.2 cells. 

 With a lump of clean sodium it even reached 5 cells. Using 25 spirit 

 lamps in series, a curved piece of platinum wire passing from the base 

 of one flame to the tip of the next, a flame battery was i)roduced. 

 Finally, a difference of potential was observed between an incandescent 

 platinum wire heated by a current and a second wire whose tip entered 

 the hot-air currents from the former, no flame or combustion products 

 being present. Hence a flame in itself is not a source of electrification 

 at all. The authors conclude, (1) that the production of electricity by 

 flames is independent of the size of the flame; (2) dependent on the 

 nature and gtate of the surface of the electrodes; (3) dependent on the 

 nature of the gases that are burning in the flame ; and (4) dependent on 

 the state of ignition of the electrodes. Hence it is a thermo-electric 

 phenomenon and analogous to the counter e. m. f. of the arc. ( Wied. 

 Ann.^Y, XX7I, p. 193; Fhil. Mag., September, V, xiv, p. 161; Nature, 

 August, XXVI, p. 321.) 



3. Electrical Measurements. , 



In his address as president of the British Association at the South- 

 ampton meeting, Dr. C. W. Siemens gave a resume of the progress 

 made in the establishment of electric units, culminating finally in the 



