307. We now take a common finger-glass and put into it a little pounded ice and salt. On this we place the flask, and then build round it the freezing mixture. The liquid column retreats down the tube, proving the contraction of the liquid by cold. We allow the shrinking to continue for some minutes, noticing that the downward retreat of the liquid becomes gradually slower, and that it finally ceases altogether. 308. Keep your eye upon the liquid column; it remains quiescent for a fraction of a minute, and then moves once more. But its motion is now _upwards_ instead of downwards. _The freezing mixture now acts exactly like the flame._ 309. It would not be difficult to pass a thermometer through the cork into the flask, and it would tell us the exact temperature at which the liquid ceased to contract and began to expand. At that moment we should find the temperature of the liquid a shade over 39° Fahr. 310. At this temperature, then, water attains _its maximum density_. 311. Seven degrees below this temperature, or at 32° Fahr., the liquid begins to turn into solid crystals of ice, which you know swims upon water because it is bulkier for a given weight. In fact, this halt of the approaching molecules at the temperature of 39°, is but the preparation for the subsequent act of crystallisation, in which the expansion by cold culminates. Up to the point of solidification the increase of volume is slow and gradual; while in the act of solidification it is sudden, and of overwhelming strength. 312. By this force of expansion the Florentine Academicians long ago burst a sphere of copper nearly three quarters of an inch in thickness. By the same force the celebrated astronomer Huyghens burst in 1667 iron cannons a finger breadth thick. Such experiments have been frequently made since. Major Williams during a severe Quebec winter filled a mortar with water, and closed it by driving into its muzzle a plug of wood. Exposed to a temperature 50° Fahr. below the freezing point of water, the metal resisted the strain, but the plug gave way, being projected to a distance of 400 feet. At Warsaw howitzer shells have been thus exploded; and you and I have shivered thick bombshells to fragments, by placing them for half an hour in a freezing mixture. 313. The theory of the shafts and pits referred to at the beginning of this section is this: The water at the surface of the shaft is warmed by the sun, say to a temperature of 39° Fahr. The water at the bottom, in contact with the ice, must be at 32° or near it. The heavier water is therefore at the top; it will descend to the bottom, melt the ice there, and thus deepen the shaft.