.SYA' WILLIAM SIEMENS, F.R.S. 



203 



of hrat by radiation should be made independent of accidental 

 s such as currents of air or rapid variations of the external 

 temjxTntmv, for which purpose the strip is put under a glass 

 shade, and the instrument itself should be placed in a room where 

 a tolerably uniform temperature of say 15 C. is maintained. 

 I'lidtT these circumstances, the rate of dissipation by radiation 

 and conduction (considering that we have to deal with low degrees 

 of heat) increases in arithmetical ratio with the temperature of the 

 strip ; the expansion of the strip, which affects the position of the 

 lever (L), is proportionate to the temperature which is itself 

 proportionate to the square of the current a circumstance highly 

 favourable to the sensitive action of the instrument. 



Suppose that the current intended to be passed through the 

 instrument is capable of maintaining the sensitive strip at a 

 temperature of say 60 C., and that a sudden increase of current 

 take< place in consequence either of an augmentation of the 

 supply of electricity or of a change in the extraneous resistance to 

 be overcome, the result will be an augmentation of temperature, 

 which will continue until a new equilibrium between the heat 

 supplied and that lost by radiation is effected. If the strip is 

 made of metal of high conductivity, such as copper or silver, and 

 is rolled down to a thickness not exceeding 0'0f> milim., its 

 capacity for heat is exceedingly small, and its surface being 

 relatively very great, the new equilibrium between the supply of 

 heat and its loss by radiation is effected almost instantaneously. 

 But, with the increase of temperature, the position of the regula- 

 ting lever (L) is simultaneously affected, causing one or more 

 contacts to be liberated, and as many additional resistance coils to 

 be thrown into circuit : the result being that the temperature of 

 the strip varies only between very narrow limits, and that the 

 current itself is rendered very uniform, notwithstanding consider- 

 able variation in its force, or in the resistance of the lamp, or 

 other extraneous resistance which it is intended to regulate. 



It might appear at first sight that, in dealing with powerful 

 currents, the breaking of contacts would cause serious incon- 

 venience in consequence of the discharge of extra current between 

 the points of contact. But no such discharges of any importance 

 actually take place, because the metallic continuity of the circuit 

 is never broken, and each contact serves only to diminish to some 



