522 
MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
disks will hang about of an inch below the centre of the bulb ; that much having to 
be allowed for the contraction of the silk when the air is exhausted. 
85. The bulb-tube is firmly clamped in a vertical position, so that the index hangs 
freely, and the pump is set to work, the bulb being surrounded with a vessel of water 
which is kept boiling all the time exhaustion goes on. The gauge soon rises to the 
barometric height ; but the operation must be continued for several hours beyond this 
point, in order to get the best effects. If the bulb is not heated during the exhaustion, 
the index loses sensitiveness after it has been sealed up for a few days, probably owing 
to the evolution of vapour from the pith ; when, however, the precaution is taken of 
heating the pith, the apparatus preserves its sensitiveness. On this account it is 
necessary to tie the silk on to the loop in the centre of the glass stem, instead of 
adopting the easier plan of cementing it with shellac. During the latter stages of 
the exhaustion, oil of vitriol (which has been boiled and cooled in vacuo) should 
gently leak into the pump through the funnel-stopper at the top of the fall-tube (44). 
This covers each globule of mercury, as it falls, with sulphuric acid, and stops mercury 
vapour from getting into the apparatus*. I cannot find that any vapour is evolved from 
oil of vitriol. 
When the exhaustion is carried to the desired degree, a spirit-flame is applied to the 
contracted part of the tube at a (fig. 1), and it is sealed off. The apparatus is then 
unclamped and the tube is again sealed off at b. This double operation is necessary to 
secure strength at the final sealing, which can only be got by holding the tube hori- 
zontally and rotating it in the flame, watching the glass to prevent it softening too 
suddenly. 
86. The best material of which to form the index in these bulb-tubes is pith, either 
in the form of a needle or bar, or as disks at the end of a glass stem. On 
December 11th, 1873, and again on April 22nd, 1874, I exhibited before the Royal 
Society a glass bulb 4 inches in diameter, having suspended in it a bar of pith 3|-X-^ 
inches. It had been exhausted in the manner above described ; and so sensitive was it 
to heat, that a touch with the finger on a part of the globe near one extremity of the 
pith would drive' the bar round 90°, whilst it followed a piece of ice as a needle 
follows a magnet. 
To get the greatest delicacy in these apparatus there is required large surface with a 
minimum of weight (75, 76). Thin disks of pith answer these requirements very satis- 
factorily ; but I have also used disks cut from the wings of butterflies and dragonflies, 
dried and pressed rose-leaves, very thin split mica and selenite, iridescent films of blown 
glass, as well as the substances mentioned in my former paper (25). Quantitative 
experiments to prove this law were attempted ; but the bulb-apparatus was found too 
imperfect for accurate measurements, so another form was devised which will be described 
further on (102), together with the experiments tried with it. 
* By adopting this precaution it is not difficult to raise the mercury in the gauge higher than that in the 
very perfect barometer hy its side, the latter being somewhat depressed by the tension of mercury vapour. 
