In the third place, there can now be no doubt that the greater part of the sheet of tuff in the Roman Campagna was accumulated under the sea. This subject was for many years one on which various contradictory opinions were held. Some writers, from the general stratified structure, correctly maintained the marine origin of the tuff. Again, on the evidence of enclosed land-plants and animals, some observers have regarded it as a freshwater deposit, while others have looked upon it as a terrestrial formation. It is true, as I shall point out a little further on, that here and there, especially in its upper parts, the tuff includes intercalated bands of strata containing land and freshwater shells as well as bones of terrestrial mammals, and indicating that the floor of the sea had been converted into low land with brackish lagoons and lakes of fresh water. But as regards the main mass of the tuff of the Campagna, the question of its marine origin may now be considered as definitely settled by the researches of Professor Portis, of the University of Rome. In specimens of different varieties of the rock from all parts of the district, and previously supposed to be entirely unfossiliferous, this careful observer has found that foraminifera are often abundant and well preserved. These organisms are unequivocally marine, swimming freely in the upper waters and sinking when dead to mingle with the silt or to form of themselves an ooze on the bottom. We can thus understand how they might be borne along above a seafloor on which molluscan life was hardly possible. If, then, cones of loose ashes and scoriæ were thrown up on the bottom of the sea, they would obviously be apt to be rapidly lowered by the agitation of currents and ground swell, while those which rose above the surface of the water, as Lipari, Volcano and Stromboli do now, would be subject also to continual erosion by rain and to unceasing attack along their shores by wind-waves. They would thus tend to be ultimately planed down, their materials being strewn over the surrounding sea-bottom, so as to add to the general accumulating sheet of tuff. The rapidity with which this kind of demolition may be completed was impressively exemplified in this very area of the Mediterranean by the history of Graham Island, which in the summer of 1831 was thrown up by a submarine eruption off the southwest coast of Sicily. In the course of less than a month, a cone of loose cinders, scoriæ and pumice was piled up to a height, it is said, of more than 200 feet above sea-level, with a circumference of three miles and a large crater inside. In about three months, this volcano was levelled with the surface of the sea.