388 
PROF. LOUIS YESSOT KINO ON TIIE CONVECTION OF 
For most purposes a specially constructed wind-tunnel is employed in the production 
of wind-velocity. In order to eliminate the effect of the walls and to produce a uni¬ 
form velocity over a considerable cross-section, a large size of tunnel is necessary, 
requiring at high velocities a very large delivery of air obtainable only from massive 
and expensive equipment. Moreover, it is necessary for the actual measurement of 
wind-velocity to depend on a Pitot-tube or other form of anemometer ; in addition it 
is difficult to secure stream-line motion at high velocities. 
(iv.) On the Determination of the Correction Factor in the Measurement 
of Velocity. 
For the purposes of the present experiment the measurement of velocity by means 
of a rotating arm recommends itself as the simplest and the most direct. With the 
wire held parallel to the axis of rotation, the objection to circular motion does not 
hold, while the true velocity at any point on the arm relative to the air can be deter¬ 
mined as follows :— 
We denote by V r the velocity of a point at radius r on the rotating arm relative to 
the room (apparent velocity), by V the true velocity relative to the air, and by v the 
velocity of the vortex at radius r relative to the room. Observation then shows that 
at the same radius the true velocity is proportional to the apparent velocity consistently 
with an expression of the form 
V = V r - v = (!-$) V r .(37) 
The constant s may conveniently be called the swirl and expressed in percentages 
of the apparent velocity. The quantity v = ,s j Y r represents the velocity at radius r of 
the vortex set up by the rotating arm. In order to measure this velocity one of the 
wires is placed in the fork of the rotating arm, and a series of currents required to 
heat up the wire to a determinate resistance for various velocities measured. The 
relation between the current and the apparent velocity is found to be of the form 
i 2 = i 0 2 + K r V r % 
(38) 
and the constants i 0 2 and K r were determined by observation. The same wire was then 
mounted in another fork fixed relatively to the room and placed in such a position 
that the rotating fork passed within a centimetre or less from the fixed wire. The 
whirling table was then set into motion and the current required to bring the 
resistance to the same value for each of a series of velocities measured. In this way 
the apparent velocity of the vortex v r was calculated from (38). It was found that 
the apparent velocity of the arm V r was connected with v r by a linear relation v r = s r V r , 
s r now denoting the apparent swirl. Making use of (37) we have the relations 
v r = v/(l—s) = sV r /(l—s) = s r V r giving s r = s/(l—s) and finally 1— s = l/(l+sv) so 
that 
Y = VJ(l + s r ), 
(3«) 
