This is an interesting fact which it will be well to remember when we come to explain glacial action, or rather the law that governs glacial action. If we take two pieces of melting ice and bring them together they immediately congeal at the point of contact. This phenomenon is called "regelation." Ice has some of the properties of a viscous substance. It will yield slowly to pressure, especially when near the melting point, but if put under a tensional strain it will break, as any brittle substance will, so that it has the properties of both viscosity and brittleness. Ordinarily we are in the habit of treating water as a fluid and ice as a solid, but from what has gone before the reader must understand that in a certain sense ice should be treated as having semi-fluidic properties. CHAPTER XXIV. WHY DOES ICE FLOAT? Nature is full of surprises. By a long series of experimental investigations you think you have established a law that is as unalterable as those of the Medes and Persians. But once in a while you stumble upon phenomena that seem to contradict all that has gone before. These, however, may be only the exceptions that prove the rule. It is recognized as a fundamental law that heat expands and cold contracts; that the atom when in a state of intense motion (which is the condition producing the effect that we call "heat") requires more room than when its motions are of a less amplitude. In other words, an increase in the amplitude of atomic motion is heating, while a decrease is cooling. It follows from the above statement that the colder a body becomes the smaller will be its dimensions. There are two or three, and perhaps more, exceptions to this rule, and the most notable one is that of water. Water follows the same law that all other substances do under the action of heat and cold, within certain limits only. If we take water, say, at 50 degrees Fahrenheit and subject it to cold it will gradually contract in bulk until it reaches 39 degrees Fahrenheit. At this point, very curiously, contraction ceases, and here we find the maximum density of water. If the temperature is still lowered we find the bulk is gradually increasing instead of diminishing (as is the rule with other fluids), and when it reaches the freezing point there is a sudden and marked expansion, so much so that a cubic foot of ice, which is solidified water, will not weigh as much as a cubic foot of water before it freezes--hence it floats.