evaporate, it would accumulate until it rose above, and therefore closed, the opening at a. The steam, now having no outlet, would condense in the chamber b until its pressure raised the water into the pipe, and caused it to overflow the basin. The pressure still continuing, all the water would be driven out of the cave, and partly up the pipe. Now, the pressure which sustained the whole column a d would not only sustain, but eject with violence, the column c d. The steam would escape, the ejected water would cool, and a period of quiescence would follow. If there were but one geyser in Iceland, this would be rightly considered a very ingenious and probable hypothesis, for without doubt we may conceive of a cave and conduit so constructed as to account for the phenomena. But there are many eruptive springs in 
Fig. 7.—Mackenzie's Theory or Eruption. Iceland, and it is inconceivable that all of them should have caves and conduits so peculiarly constructed. This theory is therefore entirely untenable.
The investigations of Bunsen and his theory of the eruption and the formation of geysers are among the most beautiful illustrations of scientific induction which we have in geology. We therefore give it, perhaps, more fully than its strict geological importance warrants.
Bunsen examined all the phenomena of hot springs in Iceland. 1. He ascertained that geyser-water is meteoric water, containing the soluble matters of the igneous rocks in the vicinity. He formed identical water by digesting Iceland rocks in hot rain-water. 2. He ascertained that there are two kinds of hot springs in Iceland, viz., acid springs and alkaline-carbonate springs, and that only alkaline-carbonate springs contain any silica in solution. The reason is obvious: alkaline waters, especially if hot, are the natural solvents of silica. 3. He ascertained that only the silicated springs form geysers. Here is one important step taken—one condition of geyser-formation discovered. Deposit of silica is necessary to the existence of geysers. The tube of a geyser is not an accidental conduit, but is built up by its own deposit. 4. Of silicated springs, only those with long tubes erupt—another condition. 5. Contrary to previous opinion, the silica in solution does not deposit on cooling, but only by drying. This would make the building-up of a geyser-tube an inconceivably slow process, and the time proportionally long. This, however, is not true, for the Yellowstone geyser-waters, which deposit abundantly by cooling, evidently because they contain much more silica than those of
