HEAT FROM SMALL CYLINDERS IN A STREAM OF FLUID. 
417 
Description of Tables III. and IV.—Specimen Observation Sheets. 
In Tables III. and IY. are set out the observations and calculations of the heat-losses in the case of the 
specimens Nos. 1 and 10 of the wires tested at different velocities. Although observations of relative 
humidity and of barometric pressures were made at the time of the experiment, no effect due to the 
variation of these factors could be detected in the final results. The ratio R/R^ refers to the entire length 
of wire between the potential terminals, while the ratio r/rn refers to unit length at the temperature 
under consideration and is corrected for thermal expansion by the formula of Section 10. The time of 
a single revolution of the rotating arm, denoted by T, was obtained from the measurement of 
a chronograph sheet. The fork containing the wire was clamped in one of three positions so that the 
radius L was one of the values 53‘7 cm., 131 • 0 cm., and 261 • 3 cm.; the apparent velocity of the wire 
through the air was given by Y r = 27tL/T. Owing to the “swirl ” set up by the rotating arm, the true 
velocity of the wire through the air was obtained from the formula Y = (1 — s) Y r , the correction factor 
(1 - s) having been determined for each value of the radius in the manner described under Diagram II. 
The values printed in italics in the Tables III. and IY. are the heat-losses expressed in watts per unit 
length, as calculated from the current required to increase the resistance of the wire to the value R, and 
entered in the same compartment. These values when plotted against the square root of the velocity give 
rise to a family of straight lines, the temperature being the variable parameter. These are illustrated in 
figs. 1 and 5, and indicate how closely these curves correspond to the theoretical curve shown in Diagram I. 
It was found impossible to work at very low velocities owing to the disturbing effect of the free convection 
current set up by the heated wire. It was thus impossible to follow out the experimental curves to small 
velocities with a view to comparison with theory. It will be seen from the graph of the theoretical curve 
that the straight-line asymptote lies extremely close to the curve over the interval covered by experiment. 
In the reduction of the observations the line of closest fit through the experimentally determined points 
was taken to represent the equation of the asymptote given by formula (33). 
YOL. CCXIY. 
A. 
3 H 
