proaching the boiling-point line, but everywhere within it, would represent the actual temperature in a state of quiescence. Now, Bunsen found that, as the time of eruption approached, the temperature at every depth approached the boiling-point for that depth, i. e., the line a g moved toward the line e f There is no doubt, therefore, that, at the moment of eruption, at some point below the reach of observation, the line a g actually touches the line e f—the boiling-point for that depth is actually reached. As soon as this occurs, a quantity of water in the lower portion of the tube, or perhaps even in the subterranean channels which lead to the tube, would be changed into steam, and the expanding steam would lift the whole column of water in the tube, and cause the water in the basin to bulge and overflow. As soon as the water overflowed, the pressure would be diminished in every part of the tube, and consequently a large quantity of water before very near the boiling-point would flash into steam and instantly eject the whole of the water in the pipe; and the steam itself would rush out immediately afterward. The premonitory cannonading beneath is evidently produced by the collapse of large steam-bubbles rising through the cooler water of the upper part of the tube; in other words, it is simmering on a huge scale. An eruption is more quickly brought on by throwing stones into the throat of the geyser, because the circulation is thus more effectually impeded. The theory given above is substantially that of Bunsen for the eruption of the Great Geyser, but modified to make it applicable to all geysers. In the Great Geyser, as already stated, Bunsen found a point, forty-five feet deep, where the temperature was nearer the Fig. 9. boiling-point than at any within reach of observation, though doubtless beyond the reach of observation the temperature again approached and touched the boiling-point. This point, forty-five feet deep, plays an important part in Bunsen's theory. To illustrate: if e f (Fig. 9) represent again the curve of boiling-point, then the curve of actual temperature in the Great Geyser tube would be the irregular line a g h. At the moment of eruption, this line touched boiling-point at some depth, h, beyond the reach of observation. Then followed the lifting of the column, the overflow of the basin, the relief of pressure by which the point g was brought to the boiling-point, the instantaneous formation of steam at g, and the phenomena of an eruption. But it is extremely unlikely that this