Iceland. 6. The temperature of the water in the basin was found to be usually 170° to 180°, and that in the tube to increase rapidly, though not regularly, with depth. Moreover, the temperature, both at the surface and at all depths, increased regularly as the time of eruption approached. Just before the eruption it was, at the depth of about forty-five feet, very near the boiling-point for that depth.
1. It is well known that the boiling-point of water rises as the pressure increases. This is shown in the adjoining table. 2. It follows 
from the above that if water be under strong pressure, and at high temperature, though below its boiling-point for that pressure, and the pressure be diminished sufficiently, it will immediately flash into steam. 3. Water heated beneath, if the circulation be unimpeded, is very nearly the same temperature throughout. That it is never the same temperature precisely is shown by the circulation itself, which is caused by difference of temperature, producing difference in density. The phenomenon of simmering is also a well-known evidence of this difference of temperature, since it is produced by the collapse of steam-bubbles rising into the cooler water above. 4. But if the circulation be impeded, as when the water is contained in long, narrow, irregular tubes, and heated with great rapidity, the temperature may be greater below than above to any extent, and the boiling-point may be reached in the lower part of the tube, while it is far from this point in the upper part.
We will suppose a geyser to have a simple but irregular tube, without a cave, heated below by volcanic fires, or by still hot volcanic ejections. Now, we have already seen that the temperature of the water in the tube increases rapidly with the depth, but is, at every depth to which observation extends, short of the boiling-point for that 
Fig. 8. depth. Let absciss a d (Fig. 8) represent depth in the tube, and also pressures; and the corresponding temperature be measured on the ordinate a n. If, then, a b, b c, c d, represent equal depths of thirty-three or more feet, which is equal to one atmospheric pressure, the curve e f passing through 210°, 250°, 275°, and 293°, at the horizontal lines, representing one atmosphere, two atmospheres, three atmospheres, etc., would correctly represent the increasing boiling-points as we pass downward. We shall call this line, e f, the curve of boiling-point. The line a g commencing at the surface at 180°, and gradually ap-
