Currents arising from two fine heated Platinum Wires. 94? 



of the chamber, as the distance between the wires is 

 diminished, and (b) the increased cooling effect owing to 

 the diminished influence o£ the wall of the chamber in 

 retarding the velocity of the free convection current arising 

 from the wire, as the wires approach nearer to one another. 

 The iuitial increase in the resistance of tha movable wire is 

 clearly attributable to the fact that initially the tendency 

 (a) more than counterbalances that due to (b). Ultimately 

 the movable wire is so far removed from the wall of the 

 chamber that the variation in the effect (a) due thereto becomes 

 negligible compared with the increased cooling effect due to 

 free convection, and the resistance of the wire accordingly 

 falls. It is clear that in the case of the results shown in fig. 5, 

 the effect due to free convection is greater relative to the 

 effect due to proximity of the wall of the chamber, than is 

 the case in the results shown in fig. 3. Accordingly it is 

 to be anticipated that the decrease of resistance of the 

 movable wire would commence at a greater separation of the 

 wires in the former than in the latter case (cf. point R. 

 figs. 3 and 5). Some increased heating effect on each wire 

 is to be anticipated due to radiation and conduction from the 

 other wire of the pair, but this, except when the wires are 

 exceedingly close to one another, may be safely neglected. 

 The initial fall in temperature of the central wire (see curve 

 A, figs. 3 and 5), as the wires approach one another is to be 

 explained as follows : — From a comparison of curves A 

 and C, in both figures it is seen that when the wires are 

 widely separated, the initial resistance of the central wire is 

 greater when the plane of the wires is horizontal than when 

 the plane of the wires is vertical, and the central wire below 

 the movable wire (cf. points S & C in fig. 5). Now in 

 each case the central wire is similarly disposed with regard 

 to the wall of the chamber, and as the movable wire is at 

 the same distance in each case from the central wire, it may 

 be legitimately assumed that the radiation and conduction 

 loss from the central wire, which in any case is very small, 

 remains constant. The initial decrease in the resistance of 

 the central wire can therefore be attributable only to a con- 

 vection effect originating in the movable wire. Moreover, the 

 initial resistance of the central wire when heated at some 

 distance from the heated movable wire, the plane of the 

 wires being horizontal, was found to be greater than when 

 the movable wire was not heated. The respective resistances 

 of the central wire in the two cases, when heated by a 

 current of 1'200 amp., were in the ratio 463 : 152, which 

 ratio is very nearly that of the galvanometer deflexions 



