Though the water may at its surface have the normal boiling temperature and be there in quiet ebullition, the boiling point for all lower levels is raised by the weight of the column of superincumbent liquid, and so for a time the formation of steam within the mass is prevented. In Fig. 202 is shown a cross section of the Icelandic _Geysir_ from which our name for such phenomena has been derived, and to this section have been added the actual observed temperatures of the water at the different levels as well as the temperatures at which boiling can take place at these levels. From this it will be seen that at a depth of 45 feet the water is but 2° Centigrade below its boiling point. A slight increase of temperature at this level, due to the constantly ascending steam, will not only carry this layer above the boiling point, but the expansion of the steam within the mass will elevate the upper layers of the water into zones where the boiling points are lower, and thus bring about a sudden and violent ebullition of all these upper portions. Thus is explained the almost universal observation that just before geysers erupt the hot water rises in the bowls and generally overflows them. [Illustration: FIG. 203.—Apparatus for simulating geyser action in the lecture room (by courtesy of Professor B. W. Snow).] The water ejected from the geyser is considerably cooled in the air; and after its return to the tube must be again heated by the ascending vapors before another eruption can occur. The measure of the cooling, the time necessary to fill the tube, and the supply of rising steam, all play a part in fixing the period which separates consecutive eruptions. If the top of the tube be narrowed from its average caliber, as is commonly observed to be true of the geysers within the Yellowstone National Park, the escape of the steam is further hindered, and frequent geyser eruption promoted. An artificial geyser for demonstration of the phenomenon in the lecture room is represented in Fig. 203. The cut has been prepared from a photograph of an apparatus designed by Professor B. W. Snow of the University of Wisconsin. In this design the tube is contracted so as to have a top diameter one fourth only of what it is at the bottom, where heat is directly applied by multiple Bunsen lamps. The water once sufficiently heated, this artificial geyser erupts at regular intervals of time which are dependent upon the dimensions of the apparatus and the quantity of heat applied.