As shown by the map in the Report of the Royal Society, the group of islands which existed previous to 1883 were but the unsubmerged portions of one vast volcanic crater, built up of a remarkable variety of lava allied to the andesite of the Java volcanoes, but having a larger percentage of silica, and hence falling under the head of "enstatite-dacite."[5] That these volcanic rocks are of very recent origin is shown by the fact, ascertained by Verbeek, that beneath them occur deposits of Post-Tertiary age, and that these in turn rest on the Tertiary strata which are widely distributed through Sumatra, Java, and the adjoining islands. According to the reasoning of Professor Judd, the Krakatoa group at an early period of its history presented the form of a magnificent crater-cone, several miles in circumference at the base, which subsequent eruptions shattered into fragments or blew into the air in the form of dust, ashes, and blocks of lava, while the central part collapsed and fell in, leaving a vast circular ring like the ancient crater of Somma (see Fig. 6, p. 43), and he supposes the former eruptions to have been on a scale exceeding in magnificence those which have caused such world-wide interest within the last few years. (_d._) _Eruption of 26th to 28th of August._--It was, as we have seen, in the month of May that, in the language of Chev. Verbeek, "the volcano of Krakatoa chose to announce in a high voice to the inhabitants of the Archipelago that, although almost nothing amongst the many colossal volcanic mountains of the Indies, it yielded to none of them in regard to its power." These eruptions were, however, only premonitory of the tremendous and terrible explosion which was to commence on Sunday, the 26th of August, and which continued for several days subsequently. A little after noon of that day, a rumbling noise accompanied by short and feeble explosions was heard at Buitenzorg, coming from the direction of Krakatoa; and similar sounds were heard at Anjer and Batavia a little later. Soon these detonations augmented in intensity, especially about five o'clock in the evening; and news was afterwards received that the sounds had been heard in the isle of Java. These sounds increased still more during the night, so that few persons living on the west side of the isle of Java were able to sleep. At seven in the morning there came a crash so formidable, that those who had hoped for a little sleep at Buitenzorg leaped from their beds. Meanwhile the sky, which had up to this time been clear, became overcast, so that by ten o'clock it became necessary to have recourse to lamps, and the air became charged with vapour. Occasional shocks of earthquake were now felt. Darkness became general all over the straits and the bordering coasts. Showers of ashes began to fall. The repeated shocks of earthquake, and the rapid discharges of subterranean artillery, all combined to show that an