950 Dr. J. S. Gr. Thomas on Mutual Action of Convection 



central wire of some heat. The close approach of the wires 

 disposed in a vertical plane — of the order of 2 mm. — before 

 the lower wire is effectively cooied due to the presence of 

 the upper wire, indicates that the free convection current 

 does not extend to any considerable distance below eilher 

 wire, and that the air supply to the free convection current is 

 derived mainly from a lateral How of air. The curves A and 

 C clearly indicate that the sensitivity of the hot wire 

 inclinometer — the wires must be used close together, as w T ill 

 be evident from the sequel — does not depend to any con- 

 siderable extent upon the thermal effect experienced by the 

 lower wire of the pair. Referring to curve D in fig. 3, it is 

 seen that, as the distance between the wires is increased, the 

 movable wire being disposed vertically above the central 

 wire, the resistance of the former falls off at a continually 

 decreasing rate. For values of the distance apart less than 

 about 0'65 cm., the movable wire experiences a heating effect 

 due to the free convection current arising from the central 

 wire. When the distance apart of the wires is increased 

 beyond this value, the movable ware experiences a cooling- 

 effect due to the same cause, the rate of fall of temperature 

 decreasing continually, until a minimum value of the resist- 

 ance is attained corresponding to the point W. Thereafter 

 the resistance of the ware increases slightly as the distance 

 apart of the wires is increased. With the wires disposed in 

 a vertical plane, the thermal effect experienced by the upper 

 wire, owing to the presence of the low 7 er v ire, depends both 

 upon the velocity and the temperature of the free convection 

 current rising from the latter. It is clear that the tempera- 

 ture of this free convection current at any point decreases in 

 the direction vertically above the wire. The initial increase 

 in the temperature of the upper wire, when the wires are 

 very close together (point X, figs. 3 and 5), shows that the 

 heating effect due to the increased temperature of the free 

 convection current in the neighbourhood of the upper wire, 

 more than counterbalances the effect of the increased 

 velocity of the current. At this short distance from the 

 lower wire, the free convection current is being accelerated, 

 and consequently it is to be anticipated that at greater 

 distances apart of the wires, the' effects upon the upper wire 

 due to the temperature, and to the velocity of the convection 

 stream rising from the lower wire, will be both operative in 

 the same direction, tending to reduce the temperature of the 

 upper wire. The temperature of the upper wire will 

 attain its normal value, i. e. its value if the lower value were 

 absent, when the heat convected from it due to the velocity 



