﻿936 Mr. A. H. Davis on Natural 



the limits of our experiments the complete equation for the 

 cooling of wires in fluids is given by 



E/kac(c 2 gd s adlk 2 )%cvjk)-y, . . . (18) 



where x varies from 1/10 to 1/6 



and y varies from 1/10 to 1/8. 



It is noticed that x and y have nearly equal values, and if 

 we put x=y, equation (6) becomes 



H/&OC {d 3 0gac/kp) x , (19) 



which is in agreement with the theoretical equation (16). 

 There appears to be no published evidence to show whether 

 this simplified grouping is of more general applicability, but 

 in this connexion it is interesting to note that Lorenz, in 

 calculating the convective cooling of a vertical plate in a 

 viscous gas, obtained a result reducible to this form. The 

 assumptions he made led in his special case to the result 

 represented by x—y — lj^. This index is rather higher than 

 those given above, but it is known from an extended curve 

 for gases that for larger cylinders x tends to increase to a 

 value of 1/3. 



(b) Representation of results in the form ~K/k—~F(d 3 0g!ac/kv). 



We now proceed to show graphically the extent to which 

 the present experimental results agree with this theoretical 

 equation. 



If satisfactory, results plotted with H/k as ordinate and 

 (d h Qgac\kv) as abscissa should yield one line independent of 

 the nature of the liquid used. Strictly it is necessary to 

 make allowance for the temperature change in the plrysical 

 properties of the liquids. Unfortunately, reliable data do 

 not seem to be available for some of the properties, but it 

 appears that the viscosity provides the most rapid change, 

 and its temperature coefficient is usually known. 



Consequently, while in this paper temperature variations 

 in the other variables are neglected, an attempt will be made 

 to allow for them in the case of viscosity. As a first 

 approximation, the value taken for the viscosity in any 

 experiment is that appropriate to a temperature intermediate 

 between the temperature of the hot wire and that of the 

 main volume of cold liquid in which it is immersed. If the 

 cooling wire is 50° C. hotter than the cold liquid, the viscosity 

 taken is that for a temperature 25° C. in excess of that of 



