370 
DR. P. E. SHAW ON THE NEWTONIAN CONSTANT OF 
it must be borne in mind that temperature should be uniform radially rather than 
longitudinally in the long tube ; for radiometric pressure, not convection, is likely to 
be the main gas trouble. The plan adopted is to have water tubes wound over the 
whole vacuum tube up to the level of the top of the torsion fibre. Similar metal water 
tubes are wound over the tube covering the window. In fig. 8, the covering tube, but 
not the water tubes, are shown. There is a space of about 6 cm. near the window on 
the long tube which cannot be wrapped with metal tube, so rubber tube is used, with 
a specially large amount of closely-packed cotton-wool, in this region. Thus we have 
several layers of cotton wool both within and outside the water helix ; while the hot 
spheres, M, M, are themselves covered doubly in cotton-wool and then in tin-foil. 
Water from the tap passes to the window region and branches into four parts, viz. : 
(a) the upper tube ; ( b) the lower tube ; (c) the window cover ; and ( d ) the window 
region. Each section thus receives water at one and the same temperature. 
After leaving the water-jacket, the water passes into vessels where its temperature 
can be watched. A Page thermostat was attached, but proved unnecessary. 
Next, the whole system of M, M, with the turn-table above (see figs. 9 and 10) is set 
co-axial with the torsion fibre, i.e., until masses, M, M, clear the wrapped tubes equally 
all round ; M, M being so far apart as to just clear the water-jacket everywhere. The 
heating coils of nichrome inserted in insulating tubes in M, M (see fig. 9) are 
connected in series with a 200-volt circuit, the leads being taken right up the copper 
suspending wires to the turn-table to avoid any hampering of the free motion of the 
heavy mass system. 
The last adjustment is to bring the beam to rest. The period of the beam may be 
anything from 2 mins, to 10 mins. The angle of rotation after the above violent 
adjustments is sure to be considerable and in high vacuum the swings are practically 
undamped. 
There are three ways of reducing the swing :— 
(1) By moving the vacuum tube to one side and so bringing it against the beam 
system. Sometimes this succeeds, but great judgment and timing are required or the 
amplitude will be increased instead of decreased by the contact. 
(2) Unclamp the torsion head and rotate the magnet, E, and so time the movement 
of the head that the torsion of the fibre is always acting against the motion of the 
beam. The head must be left clamped finally. 
(3) Gravitational damping by moving large masses, M, M, to always oppose the 
motion of the beam system. This method is too weak to be effective, except as a 
finishing touch to reduce the swing when it is already small, say, 5 degrees of arc. 
The process of damping the swing may take many hours, but it must be done so 
that finally the mirror when at rest would face the centre of the scale (of 500 mm.) 
which is placed about 5 m. away and immediately over the large reading telescope. 
The apparatus is now ready, but it is found that no accurate work can be performed 
