Let us try an experiment. Take a small glass flask, terminating in a long neck, say of four to six inches, and of small diameter. Suppose the water in the glass to be at 50 degrees Fahrenheit. Fill the flask with water until it stands halfway up the neck at 50 degrees temperature. Now immerse the flask gradually in hot water, and observe the effect. For a moment the water will lower in the neck of the tube, but this is due to the fact that the glass expands before the heat is communicated to the water and enlarges its capacity. But immediately the water will begin to rise as the heat is communicated to it, and will continue to expand up to the boiling point. Now take the flask out of the hot water and gradually introduce it into a freezing mixture made of broken ice and salt. Immediately the water will begin to fall in the tube, showing that it is contracting under the cold, and it will continue to contract until it reaches a temperature of 39 degrees Fahrenheit, when it will come to a standstill and then proceed to expand as the temperature of the water lowers. When it reaches the freezing point the fluid can no longer rise in the neck of the flask, which is broken by the sudden expansion that takes place at this point. To show what an irresistible power resides in the atoms of which the body is made, let us take an iron flask with walls one-half inch or more in thickness; fill it with water and seal it up by screwing on the neck an iron cap; now plunge it into the freezing mixture, and the first effect will be to contract the water unless it is already below 39 degrees Fahrenheit, but when it reaches that point expansion sets in, which continues to the freezing point, when a greatly increased expansion takes place suddenly. The walls of the iron flask, although a half-inch in thickness, are no longer able to resist the combined efforts of the billions upon billions of the atoms of which the water is made up, in their individual clamor for more room, hence the flask is shivered into pieces. There are one or two other substances which are exceptions to the general rule, but we will mention only one, which is the metal bismuth. If we should melt a sufficient amount to fill an iron flask, such as we have described, and subject it to the same freezing process, the flask will be broken the same as in the experiment made with the water. A query arises, Why this phenomenon? Why does water follow a different law in cooling from that of nearly all other substances?