58 : Mr. G. JOHNSTONE STONEY, 
the bottom, where the pressure of the water is greatest, thickest 
above. So long as there is any communication between the 
polarized layer and the atmosphere, the lateral stresses within 
the layer will -be equal to P, while those in the direction in which 
the heat penetrates will be P+; but both of these will suffer 
an increase if the ball is plunged deeper after the communication 
with the atmosphere has been cut off. No one can see this splendid 
experiment for the first time without a feeling of astonishment. 
- A Crookes’s layer formed in the same way, but without the 
exquisite beauty which it has in this experiment, may be seen 
any day ina smith’s forge, whenever the smith has occasion to 
quench white-hot iron in water. 
A phenomenon closely resembling the experiment with the 
glowing ball was witnessed lately by my brother and two 
other friends while out walking. There was a shower when they 
reached some rather deep water. The afternoon had become 
chilly, and the phenomenon that presented itself shows that the 
water must have retained a temperature higher than that of the 
air. As the rain-drops fell into the water, some of them 
(estimated at one in twenty) became spheroidal drops floating on 
the water, and of these some (estimated at one in six) were 
visibly submerged before floating about as spheroidal drops. 
They sank, perhaps about half a centimetre before they rose to 
the surface, and while under water looked like silvered pills, 
owing to the total reflection from the boundary between the 
water and the film of polarized air which enveloped each drop. 
Several times, in the course of this communication, I have had 
occasion to speak of the feebleness of conduction or penetration, 
compared with the rapid outpour of heat which takes place on 
direct contact between a very hot and a cold body. This is well 
illustrated by an experiment of M. Boutigny, in which a 
spheroidal drop of water is formed inside a hot copper bottle, and 
the neck of the bottle partially stopped by a cork through which 
a thin tube passes. So long as the drop continues in the 
spheroidal state, a mixture of air’and vapour slowly escapes 
through the tube in the cork, but the instant the spheroidal state 
ceases, and the water comes into contact with the copper, a suffi- 
cient portion of the water flashes off so suddenly into steam that 
the cork is driven out with explosive violence. 
