206 Proceedings of the Royal Society 
and gave much annoyance by preventing the valves of the apparatus 
from closing. To remedy this inconvenience I enclosed the glass 
tube in a tube of stout brass, closed at the bottom only, but was 
surprised to find that it was crushed almost flat on the first trial. 
This was evidently due to the fact that water is compressible, and 
therefore the relaxation of pressure (produced by the breaking of 
the glass tube) takes time to travel from the inside to the outside 
of the brass tube ; so that for about a second that tube 
was exposed to a pressure of four or five tons weight per square 
inch on its outer surface, and no pressure on the inner. The im- 
pulsive pressure on the bottom of the tube projected it upwards, so 
that it stuck in the tallow which fills the hollow of the steel-plug. 
Even a piece of gun-barrel, which I substituted for the brass tube, 
was cracked, and an iron disc, tightly screwed into the bottom of 
it to close it, was blown in. I have since used a portion of a thicker 
gun-barrel, and have had the end welded in. But I feel sure that 
an impulsive pressure of ten or twelve tons weight would seriously 
damage even this. These remarks seem to be of some interest on 
several grounds, for they not only explain the crushing of the open 
copper cases of those of the “ Challenger” thermometers which gave 
way at the bottom of the sea, but they also give a hint explanatory 
of the very remarkable effects of dynamite and other explosives 
when fired in the open air. [It is easy to see that, ceteris paribus, 
the effects of this impulsive pressure will be greater in a large 
apparatus than in a small one.] 
4. On the Cause of the Sounds produced in the Microphone 
Receiver. By Professor James Blyth. 
In the microphone, or loose contact receiver, we have two or 
more small blocks of carbon resting lightly against each other, 
through which the interrupted current from the transmitter is 
passed. At the points of contact, a strong heating and cooling 
effect must take place every time the current is interrupted or 
otherwise varied. This heating and cooling will cause corres- 
ponding expansions and contractions of the air or other medium 
surrounding the points. These produce undulations in the medium, 
which, striking against the sides of the containing vessel, set them 
