432 
PROF. LOUIS YESSOT KING ON THE CONVECTION OF HEAT, ETC. 
instrument and in the event of being able to employ standardized anemometer-wires, the scale of the 
ammeter could be graduated to read velocities directly. 
The galvanometer employed was a Weston portable instrument with jewel bearings and capable of 
detecting a current of about 10 -6 amperes. This degree of sensitivity is more than necessary; in fact 
a millivoltmeter was found to be sufficiently sensitive for most purposes. The constants of damping are 
very important in determining the rapidity with which observations can be made, and it was found that 
equally sensitive galvanometers varied within wide limits in this respect. By employing an alternating 
current and a telephone receiver instead of a galvanometer it was found that the same calibration curve 
was obtained as in the case of direct current measurements. 
It was found possible to test for any suspected change in the constants of the anemometer-wire by 
measuring the current required to bring the wire to the standard temperature in a stagnant atmosphere, 
care being taken to protect the instrument from draughts. 
In analysing a complicated distribution of air-flow, the direction of the current of air may be determined 
from the effect discussed in Section 14; if the anemometer is rotated about an axis perpendicular to the 
wire, it will be at right-angles to the stream when the current required to bring it to the standard 
Diagram III. Details of hot-wire anemometers and connections. 
temperature is a maximum. This effect is sufficiently marked to enable the direction of a stream to be 
fixed with a fair degree of accuracy. 
In figs, (c) and (d) of Diagram III. are illustrated simple methods of making the connections to an 
anemometer-wire so as to. dispense with a galvanometer. In fig. (c) a low resistance storage-cell is 
connected by means of a three-way key in such a way that the anemometer-wire and a manganin 
resistance of four times the resistance may be successively included in the circuit; if the current be 
adjusted so that the ammeter reading is unchanged on successively including these two resistances, they 
must have the same resistance. Fig. (cl) illustrates an adaptation of Kelvin’s method for measuring a 
galvanometer resistance. The ammeter a is in one of the arms in series with a manganin resistance so that 
the total resistance is four times that of the anemometer-wire in the adjacent arm. The ratio resistances 
R are adjusted to equality and have approximately the same value as that of the remaining arms. 
A contact key K replaces the galvanometer. If the current in the bridge be adjusted so that on 
depressing the key there is no change in the ammeter reading, it is easily seen that the condition 
of balance of the Wheatstone-bridge is satisfied and the resistances A and B are equal. 
The two methods just described suffer the disadvantage that the contact resistances of the anemometer- 
wire are included in the resistance of the wire itself, and is likely to be uncertain if the wire is to be 
employed at high temperatures. The potential terminals and the Kelvin-bridge connections are recom¬ 
mended for use as the most satisfactory and should be employed if possible. 
