Currents arising from two' fine heated Platinum Wires. 951 



of the impressed convection current from the lower wire is 

 exactly compensated by the diminished free convection loss 

 owing to the stream being at a temperature above the 

 normal. Thereafter, with increase in the distance apart of 

 the wires, the temperature of the movable upper wire will 

 fall below its normal value, as represented by the portion QW 

 of the curve D in fig. 3. The subsequent increase in the resis- 

 tance of the upper wire, as shown in fig. 3, is due partly to 

 the retarding influence exerted by the upper part of the wall 

 of the chamber upon the free convection current rising from 

 the upper wire, and in part to the wire when disposed in a 

 vertical plane at a distance from the central wire greater 

 than that corresponding to Wj entering a zone of heated air 

 accumulated in the upper portion of the chamber. A corre- 

 sponding increase in the resistance of the upper wire is not 

 shown by the curve D in fig. 5. As the convection effect in 

 this latter case is much greater than in the case represented 

 in fig. 3, it may be concluded that the increase shown in the 

 latter case is to be attributed mainly to the effect of the 

 warm air collected in the upper part of the chamber. The 

 point is of some consequence in the design of a hot-wire 

 inclinometer in that if the disposition of the wires within the 

 enclosure is such that one or both can enter the heated zone 

 in the upper part of the enclosure, a symmetrical deflexion- 

 inclination curve will be obtained for the upper and lower 

 pairs of quadrants only if the wires are symmetrically 

 disposed about the axis of the chamber. In cases of such 

 disposition of the wires, the observed effects are not to be 

 attributed even mainly to the existence of the free convection 

 current, and deductions with regard to the temperature 

 gradient in the free convection current rising from the wire 

 may be subject to error*. 



The variation in the sensitivity of the hot-wire inclino- 

 meter with increase of distance apart of the wires as shown 

 in fig. 2 can be readily interpreted in the light of the 

 remarks referring to the curves in fig. 3. Thus the point 

 of zero deflexion represented by Q x in fig. 2, corresponds to 

 a distance apart of the wires slightly less than 0*65 cm. 

 (point Q in fig. 3), such that the distances between the 

 curves A and B corresponding to that distance is very 

 approximately equal to that between the curves and D 

 thereat. For values of the distance apart of the wires 

 greater than this, it is clear that the galvanometer deflexion 

 will bo negative and increasino- numerically until a maximum 

 is obtained corresponding to a distance apart of the wires 

 * See Humphreys, he. cit. p. 192. 



