HEAT FROM SMALL CYLINDERS IN A STREAM OF FLUID. 
387 
(17° C.) was obtained, care being taken to test for the absence of any appreciable 
heating effect of the measuring current since this may become quite noticeable in the 
case of the very fine wires. Points were then determined on the bridge-wire corre¬ 
sponding to a series of predetermined ratios ft/ft 17 , each of these representing when 
balanced against the heated platinum wire a certain temperature as calculated from 
the constants of the wire. The series of temperatures employed was as far as possible 
kept the same for all the wires. In order to measure the heat-loss, the standard 
resistance and the Kelvin-Varley slide were cut out of the circuit, and the terminals 
T 1 T 2 connected to the 110-volt mains through rheostats allowing of the continuous 
adjustment of current from zero to about 5 amperes. The rotating arm was then 
adjusted to a fixed speed, and the current set to such a value that a balance was 
obtained on the galvanometer (shunted to two of its sensitivity) when the adjustable 
contact B, was set on each of the points of the bridge-wire previously determined to 
represent certain temperatures. Headings were also taken with the apparatus at a 
standstill, and the wire in three positions, horizontal, vertical, and inclined at an angle 
of 45 degrees to the vertical. At each point the current was read off on the ammeter, and 
from the resistance of the wire the heat-loss in watts per unit length could be calcu¬ 
lated. This series of readings was repeated for various velocities as high as 25 miles 
an hour. In order to eliminate the end correction the wires tested were for the most 
part of considerable length (about 23 cm.); it was found impossible to go to higher 
speeds owing to the sagging and vibration of the wires under wind-pressure and 
centrifugal force with consequent risk of breaking at high temperatures and the loss 
of a set of observations. 
(iii.) On the Measurement of Wind-velocity. 
The simplest laboratory method of obtaining a stream-line wind-velocity, whose 
value is known directly without reference to the calibration of Pitot-tubes or anemo¬ 
meters, is realized by the use of a whirling table to which is attached a light arm at 
the extremity of which the object to be experimented upon is attached. The advan¬ 
tage of this method is, however, more apparent than real. It is well-known that in 
such a disposition of apparatus a vortex is created in the neighbourhood of the 
rotating arm, so that the velocity of a point at any radius relative to the room does not 
represent its velocity relative to the air. In addition to the difficulty mentioned 
results obtained from air-velocities measured from motion in a circle cannot in some 
classes of work be applied with safety to linear motion. In fact so serious have these 
objections proved themselves to be that the method of obtaining velocity by means of 
a rotating arm has been abandoned in meteorological work and in aeronautical 
problems. ( 26 ) 
( 26 ) These difficulties are described in detail by Fry and Tyndall in a paper “ On the Value of the 
Pitot Constant,” ‘Phil. Mag.,’ 21, p. 352, 1911; also in the ‘Report of the Advisory Committee for 
Aeronautics,’ 1909-10, p. 15; 1910-11, ‘Report,’ No. 34, p. 50, by Messrs. Bramwell and Sillick. 
3 D 2 
VOL. CCXIV.—A. 
