952 Dr. J. S. G. Thomas on Mutual Action of Convection 



between 1'5 and 2'0 cm. The occurrence of a maximum 

 deflexion on rotating the plane of the wires from the hori- 

 zontal to the vertical position, as shown by the points P' in 

 fig. 2, corresponding to a distance apart of the wires equal 

 to very approximately 0*170 cm., is readily interpreted by 

 means of the four curves given in fig. 3. Thus for values of 

 the distance apart of the wires less than 0"5 cm., it is seen 

 that whether the plane- of the wires be vertical or horizontal, 

 a decrease in the distance apart of the wires is accompanied 

 by a continuous increase in the resistance of each wire. 

 Now the magnitude of the deflexion occurring on rotation 

 through 90°, when the arms are separated by a given 

 distance, may be taken as approximately proportional to the 

 amount by which the difference of ordinates corresponding 

 to that distance for the curves C and D is greater than the 

 corresponding difference of ordinates for the curves A and B. 

 It is sufficient for the purpose of comparison of these differ- 

 ences to disregard the difference of ordinates of the curves 

 A and C in the region considered, as those curves are there 

 practically parallel. The maximum deflexion may therefore 

 be anticipated to occur at approximately that distance apart 

 of the wires for which the difference of ordinates of the 

 curves D and B is a maximum. It is clear from the shapes 

 of the cnrves D and B in the region considered, that owing 

 to the initial, very much steeper gradient of the curve I), as 

 the distance between the wires diminishes below 0*5 cm., 

 such a maximum difference of ordinates occurs corresponding 

 to a distance apart of the wires of about 0*2 cm. The 

 occurrence of a distance apart of the wires such that the 

 deflexion on rotation is a maximum, interpreted physically, 

 signifies that on closer approach than this distance of the 

 wires to one another, the mutual thermal effect on the wires 

 due to such approach becomes increasingly greater compaied 

 with the thermal convection effect experienced by the wires 

 on rotation*. In like manner there is a disposition of the 

 wires of a directional hot wire anemometer affording the 

 maximum sensitivity, 



In fig. 4 are given the forms of the deflexion-inclination 

 calibration curves obtained, when the wires heated by a 



* Professor L. V. King lias recently drawn the author's attention to a 

 paper by B. Ray of the University of Calcutta (Proc. Indian Assoon. for 

 the Cultivation of Science, vol. vi. p. 95), in which, by an optical method, 

 the existence of a stagnant region on the down-stream side of a cylinder 

 12 mm. in diameter was established in the case of both free and forced 

 convection. Such a stagnant region, if present in the case of fine wires, 

 affords a satisfactory explanation of the diminished deflexion on rotation 

 when the wires are very close together. 



