Mr. Mallet, who is one of the best authorities on the subject, considers that submarine eruptions must account for them. A volcanic upheaval of the sea bottom would produce crevices, by which the sea is brought directly in contact with subterranean fires. An explosion is the result, like those that have occasionally occurred at foundries from dumping masses of fiery slag into a snowbank. So what began with a gradual upheaval ends with a sudden concussion, the vibration of which passes along the sea bottom to the mainland. Every one who has lived in the city is familiar with the fact that the vibration produced by a carriage may be felt at the top of a very tall building. But the idea that the explosion always occurs at the sea bottom leaves no way to account for the fact that a volcanic eruption acts as a safety-valve. Mr. Mallet’s conclusions are largely based on personal observations of earthquakes in England, where no active volcano exists. That earthquakes are more violent and volcanoes more numerous on islands or near the sea coast is well-known. It is also well established that shocks frequently occur at sea, which are not perceptible on the land. The shock is similar to that produced by striking on a reef. Often have sailors been mystified, on receiving such shocks and hastily heaving the lead, to find the ocean unfathomable. Again, shocks which are most violent on land are not perceptible at sea, unless a great sea wave be produced; but such a wave in the open sea, as often experienced, produces no shock but passes under a vessel like a heavy swell. And a shock at sea is sometimes severe enough to snap a spar, or wrench loose bolts like the blow of a reef, yet no trace is perceptible on shore. Lastly, earthquakes often happen in inland regions, and affect but a small area. Clearly it will not do to attribute effects so different to explosions at the sea-bottom. Those who attribute all earthquakes to subterranean heat and gases, whether local or general, find it easy to account for the occurrence of violent earthquakes in regions remote from active volcanoes. In case of the gradual decline of volcanic action, such as we know from the great numbers of extinct volcanoes, old trap-dykes, and ancient lava beds, to be continually taking place in one region or another, the old vents or safety-valves would cool and close. The pent up power would in consequence gradually accumulate, till finding no outlet, it would burst the crust over a wide area, and so relieve the pressure.