56 KENNELLY-SANBORN— INFLUENCE OF [April 24. 



electric resistance, was discovered to vary as the square root of the 

 wind velocity, which was measured by the speed of the moving wire 

 through otherwise tranquil air. In other words, it was discovered 

 experimentally that in order to dissipate double the power from the 

 wire, at constant resistance and temperature-elevation, it was neces- 

 sary to quadruple the speed of the wire through the air. This rela- 

 tion was found to hold, within observation errors, for several differ- 

 ent sizes of thin copper wire, and for various temperature elevations, 

 between wind-velocities of 2 and 20 meters per second. Below 2 

 meters per second, the relation deviated towards the case of free 

 convection from a hot wire at rest. That is, at low wind velocities, 

 empirical corrections became necessary for the free convection which 

 naturally occurs from a wire at rest, or moving at zero speed through 

 the air. The possible application of the square-root law of wind 

 cooling to anemometry was also pointed out. 



After the results were published in 1909, our attention was drawn 

 to papers by Professor Boussinesq in the Comptes Rendus for 1901, 

 Vol. 133, p. 257, and the Journal de Mathcmatiques, 6th Series, Vol. 

 I, 1905, in which is given the theory of the convection of heat by a 

 stream of liquid from the surface of a cylindrical rod, placed at 

 right angles to the stream. The liquid is assumed to be incompres- 

 sible and devoid of viscosity. The formula arrived at by Boussinesq, 

 as given by Russell, is : 



IsakVa 

 H 



= se^' 



where H is the heat carried off convectively per second from unit 

 length of cylinder. 



.s is the specific heat of the liquid. 



o- is the density of the liquid. 



k the thermal conductivity of the liquid. 



V is the velocity of the liquid. 



a the radius of the cylinder. 



6 the temperature elevation of the cylinder. 

 This means that the linear forced convection, or ergs per second per 

 cm. of the cylinder, is proportional to its temperature-elevation above 



