430 
PROF. LOUIS VESSOT KING ON THE CONVECTION OF 
Description of Diagram III.—Details and Specifications in the Construction of Hot-Wire Anemometers. 
The most suitable wire for use in hot-wire anemometry was found to be platinum thermometer wire 
drawn down to a diameter of about 3 mils. It is advisable to age the wire before mounting by heating it 
to a red heat for some hours by means of an electrical current. The wire should then be carefully 
examined under a high-power microscope, and a portion of sufficient length selected free from flaws and 
pittings and of as uniform a diameter as possible. The wire may at once be mounted in position in the 
fork described below, heated to a bright red heat, and the potential terminals of 1-mil wire fused in place. 
Care should be taken to avoid allowing the heated wire to come into contact with easily fusible metals or 
organic matter. No connections should be fused in a blow-pipe flame, and no solder should be employed 
except in those connections which remain cool. The Kelvin-bridge connections obtain the condition that 
the reading of the instrument is independent of the contact resistances at Ci and C 2 , which need only be 
tolerably good. The potential leads may be fused to the anemometer at any convenient distance apart, and 
the instrument will give correct readings of air-velocity if the constants are determined experimentally in 
such a manner that the potential terminals are in the same position relatively to the flow of air as in the 
distribution of flow to be measured. If, however, the calibration curve of the instrument is to be 
determined from the dimensions and electrical constants of the wire, making use of the convection 
constants obtained in Part II. of the present paper, the position of the potential terminals is subject to 
the limitations imposed by keeping within narrow limits the errors introduced by the cooling effect of 
the end-connections and potential wires. The maximum values of these errors are investigated in 
Section 12 (ii.); the values of P and p occurring in formulge (60) and (61) are calculated for anemometer- 
wires of 6, 3, and 1 mils diameter, fitted with potential terminals of 1-mil wire. The thermal conductivity 
of platinum is taken to be K = 0 • 7 watts per sq. cm. per sec. per degree C. The temperature of the 
wire is that corresponding to the ratio R/R 0 = 4, adopted as the most suitable for hot-wire anemometry 
and representing a value 9 - d 0 = 1165° C. The heat-loss W employed in the table given beloAV is that 
obtained from experiment at the lowest velocity V = 81 cm./sec.:— 
Table (i .).—Values of p and P. 
Wire. 
6-mil. 
3-mil. 
1-mil. 
Diameter 
Cross-section 
Koj (6 - 9 q ) 
Heat-loss W watts at 1165° C., \ 
for V = 81 cm./sec. J 
p = J(R 0 /E). n /[W/Koj (9 - 0 O )] 
P = V(R/R 0 ). (a'/a) s /< 
•0153 cm. 
1-84 x10 -4 sq. cm. 
•150 
2-00 
1-83 
•213 
•00775 cm. 
' 472 x 10 -4 sq. cm. 
•0385 
1-60 
3-22 
•51 
•00283 cm. 
■ 0628 x 10~ 4 sq. cm. 
•00513 
(1 • 10) (calculated) 
7-31 
2-0 
We now take the case of an anemometer-wire of length 2(1 + Id). The resistance between potential 
terminals at a temperature 9° C. is denoted by 2R/, l being the length between potential terminals. If 
calculated from the resistance per unit length for the same temperature the result would be 2R l. It is 
shown that, owing to the cooling effect of the leads and potential terminals, these differ by a small 
correction-factor given by R = R(1 -e), e being given by formula (62). In designing an anemometer for 
which the constants are to be determined by calculation the length and diameter of the wire and the 
position of the potential terminals must be so disposed that the quantity e remain small. In the following 
table, based on the preceding Table (i.), values of e are given for various diameters and positions of the 
potential terminals and represent overestimates of this correction :—• 
