Water is changed to ice at a certain temperature under the same law of crystallization by which any inorganic bodies in a fluid state may assume a solid condition, taking the shape of perfectly regular crystals, which combine at certain angles with mathematical precision.  The frost does not form a solid, continuous sheet of ice over an expanse of water, but produces crystals, little ice-blades, as it were, which shoot into each other at angles of thirty or sixty degrees, forming the closest net-work.  Of course, under the process of alternate freezing and thawing, these crystals lose their regularity, and soon become merged in each other.  But even then a mass of ice is not continuous or compact throughout, for it is rendered completely porous by air-bubbles, the presence of which is easily explained.  Ice being in a measure transparent to heat, the water below any frozen surface is nearly as susceptible to the elevation of the temperature without as if it were in immediate contact with it.  Such changes of temperature produce air-bubbles, which float upward against the lower surface of the ice and are stranded there.  At night there may come a severe frost; new ice is then formed below the air-bubbles, and they are thus caught and imprisoned, a layer of air-bubbles between two layers of ice, and this process may be continued until we have a succession of such parallel layers, forming a body of ice more or less permeated with air.  These air-bubbles have the power also of extending their own area, and thus rendering the whole mass still more porous; for, since the ice offers little or no obstacle to the passage of heat, such an air-bubble may easily become heated during the day; the moment it reaches a temperature above thirty-two degrees, it melts the ice around it, thus clearing a little space for itself, and rises through the water produced by the action of its own warmth.  The spaces so formed are so many vertical tubes in the ice, filled with water, and having an air-bubble at the upper extremity.

Ice of this kind, resulting from the direct congelation of water, is easily recognized under all circumstances by its regular stratification, the alternate beds varying in thickness according to the intensity of the cold, and its continuance below the freezing-point during a longer or shorter period.  Singly, these layers consist of irregular crystals confusedly blended together, as in large masses of crystalline rocks in which a crystalline structure prevails, though regular crystals occur but rarely.  The appearance of stratification is the result of the circumstances under which the water congeals.  The temperature varies much more rapidly in the atmosphere around the earth than in the waters upon its surface.  When the atmosphere above any sheet of water sinks below the freezing-point, there stretches over its surface a stratum of cold air, determining by its intensity and duration the formation of the first stratum of ice.  According to the alternations of temperature,