170 
POPULAR SCIENCE REVIEW. 
charcoal cools, it condenses the very small residue of gas there 
may be present. They test the vacuum by the passage of an 
electric spark. 
Professor Dewar then noticed that the effects of light and 
heat had been tried by many experimenters, and that in the 
u Edinburgh New Philosophical Journal” for 1828 there is an 
account of experiments by Mark Watt on the subject, made 
with apparatus differing little in appearance from that used by 
Mr. Crookes. A general account of Mr. Crookes’s results was 
given, and the statement was made that this observer considered 
the results “ inexplicable.” 
A substance opaque to heat-rays was then interposed between 
the candle and the balance, a layer of water \ inch thick 
diminishing the deflection to J of the original amount. A 
smoked piece of rock salt or a solution of iodine in bisulphide 
of carbon were then interposed, and the deflection was found to 
be only diminished by the obstruction due to the glass walls of 
the tank. 
Two equal discs, one of rock salt the other of glass, were 
attached to the glass fibre. The rock salt was inactive to the 
beam, the glass was active. The reason was given that the 
former, being transparent to heat, was not heated, whereas the 
glass was. The back of the rock-salt disc was then coated with 
lamp-black. Still there would be attraction. The heat and 
light, passing through the rock salt, were absorbed by the lamp- 
black at the surface of contact. The lamp-black is heated, but 
is so bad a conductor that before the heat can be conducted 
through it, it is conducted through the rock salt, heating it, and 
causing repulsion. The next modification was to substitute 
clear and ordinary sulphur for the salt. Clear sulphur, when 
acted on by light, resumes the appearance of ordinary sulphur, 
with disengagement of heat. A beam being thrown on this, 
the effect was attraction, the back being heated, and repulsion, 
there being attraction on the other side. 
When the action takes place at ordinary pressure, it is pro- 
bably due to convection currents. The air in front of the disc 
is heated and ascends, leaving a vacuum, and hence the disc 
advances. To understand the action when the exhaustion is 
more perfect, the quantity of gas actually present must be con- 
sidered. A vessel of a like capacity would hold a bubble 
inch in diameter at ordinary pressure. It has been shown that 
the average path of molecules between two collisions is about 
j, 00 millimetre. If the pressure be reduced to 4 0 o o ooo ^e 
mean path will be 400 millimetres, or about a foot and a half. 
When, therefore, the gas is rare, the particles may get a long 
way off before they meet others, and so the action becomes per- 
ceptible. It was shown that the total mechanical action on a 
