637 
1907-8.] Preparation of a Glass to Conduct Electricity. 
having proved unsuitable on account of this variable P.D. produced in the 
glass itself, it was found best to momentarily close the battery key and 
look for a kick of the galvanometer needle at the moment the normal 
P.D. between the inner and outer faces of the bulb was zero. By reversing 
the battery current and waiting till the bulb when joined direct to the 
galvanometer gave no deflection, reliable readings of the resistance were 
obtained by balancing the bridge until a momentary current gave no sudden 
kick of the needle. It was easy, in this way, to distinguish between the 
deflection due to want of balance and that produced by the rapid rise of 
a back P.D. brought about by polarisation. 
To ensure that this momentary contact be made with regularity, a heavy 
brass cylinder (amalgamated) was rolled down two inclined ebonite rails, 
portions of which were removed and filled in with strips of brass connected 
in series with the battery and bridge. The roller thus completed the circuit 
between the brass strips, and the time of contact could be adjusted by tilting 
the rails. The smoothness of the two curves shown in fig. 4 is good evidence 
that the readings may be relied upon. The sets of observations plotted 
there were obtained with two different bulbs, and show the variation of 
conductivity for a rise of temperature. It is seen that the resistivity 
rapidly falls at first, but that at about 90° C. a change occurs, and beyond 
that a further rise of temperature produces comparatively little alteration 
in the conductivity. From measurements of the bulb the specific resistivity 
was calculated and found to be 5 x 10 s ohms at 20° C. Although this value 
is high, it is about a thousand times less than that of ordinary glass at 100° C. 
Index of Refraction. 
The value of n for this glass was calculated by means of Fraunhofers 
formula, from data obtained by measurements upon a 60° prism. At 17° C. 
a = 1*510. 
Coefficient of Expansion. 
The coefficient of expansion is large enough to admit of a direct 
measurement. It is approximately equal to that of brass. The mean of 
two measurements upon a rod 292 mms. long was *00016. The remarkable 
expansive property is due to the large proportion of sodium present in the 
glass. 
Its relative contraction on cooling may be well shown by allowing two 
beads, one of conducting and the other of, say, ordinary soda glass, to slightly 
weld together in a flame before drawing them out into a thread. The curl 
