HEAT FROM SMALL CYLINDERS IN A STREAM OF FLUID. 
389 
expressing the true velocity in terms of the apparent velocity and the apparent swirl, 
both easily measurable quantities. Details of measurements by means of which the 
swirl was obtained for the three positions on the rotating arm employed in the present 
experiment are given briefly under the description of Diagram II. We note in passing 
that the value of the swirl will depend on the construction and disposition of the 
apparatus as well as on the dimensions and shape of the room in which the rotating 
arm is set up ; in the present case the swirl amounted to about 5 per cent, and could 
be determined to within 5 per cent., so that the true velocity of the wires relatively to 
the air of the room could be obtained to an accuracy well within ^ of a per cent. In 
this manner the simplicity and ease of working of the rotating arm can be employed 
without loss of accuracy. ( 27 ) 
(v.) Note on the Construction of a Constant Low Resistance Set of Mercury 
Contact Slip-rings. 
In order that the resistance of the wire between potential terminals may be measured 
with accuracy at the extremity of the rotating arm, it is necessary to provide some 
form of slip-ring; since the contacts enter into the connections of the bridge, it is 
important that the contact resistances be as small as possible and remain constant 
while the apparatus is in rotation. A specially designed set of four slip-rings which 
fulfils these requirements is shown in (h) Diagram II., while the specifications are 
given under the accompanying description. The insertion of this set of slip-rings in 
the circuit P^B^ of fig. 2, increased the resistance by the amount 0'0077 ohm. 
only and was found to remain extremely constant. No change in the balance of the 
bridge when connected to measure the resistance of the wire at room temperature 
could be detected when the rotating arm was set into motion at various speeds. 
(vi.) Note on the Fork Carrying the Wire to he Tested and on the Whirling Table. 
The rotating arm consisted of a light strip of pine (6'3 x 1'3 cm.) about 2'6 m. 
long fastened to a whirling table in the manner shown in Diagram II. and illustrated 
by Plate 8. At three determinate positions on this arm could be clamped the fork 
for carrying the wires. This fork was designed to disturb as little as possible the flow 
of air past the wire, and is illustrated in detail in (c) Diagram II. In order to 
eliminate the cooling effect of the terminals and their disturbance on the flow of air 
past the ends of the wires, potential leads of extremely fine platinum wire (l mil 
diameter) were fastened at the points PrP 2 at distances of about 2 cm. from the ends 
CjC 2 . These potential leads were most easily fixed in position by heating the wire to 
( 27 ) qh e results here described and the method of correction for the swirl agree with the observations of 
Bramwell and Sillick (footnote ( 12 )), who found nearly the same value for the ratio s = vjY r , I he 
correction is now regularly made in tests carried out on a rotating arm. 
