For a brief period in the fall of 1904 Vesuvius had been in almost absolute repose, but soon thereafter the Strombolian crater explosions were resumed. On May 25, 1905, a small stream of lava began to issue from a fissure high up upon the central cone, and from this time on the lava continued to flow down to the valley or _atrio_, separating the inner cone from the caldera remnant of Monte Somma. Seen in the night, this stream of lava appeared from Naples like a red hot wire laid against the mountain’s side (Fig. 130). With gradual augmentation of Strombolian explosions and increase in volume of the flowing lava stream, the same condition continued until the first days of April in 1906. The flowing lava had then overrun the tracks of the mountain railway and accumulated in considerable quantity within the _atrio_ (Fig. 131). [Illustration: FIG. 130.—Night view of Vesuvius from Naples before the outbreak of 1906. A small lava stream is seen descending from a high point upon the central cone (after Mercalli).] On the morning of April 4, a preliminary stage of the eruption was inaugurated by the opening of a new radial fissure about 500 feet below the summit of the cone (Fig. 132 _a_), and by early afternoon the cone-destroying stage began with the rise of a dark “cauliflower cloud” or _pino_ to replace the lighter colored steam cloud. The cone was beginning to fall into the crater, and old lava débris was mingled in the ejections with the lava clots blown from the still fluid material within the chimney. From now on short and snappy lightning flashes played about the black cloud, giving out a sharp staccato “tack-a-tack.” The volume and density of the cloud and the intensity of the crater explosions continued to increase until the culmination on April 7. On April 5 at midnight a new lava mouth appeared upon the same fissure which had opened near the summit, but now some 300 feet lower (Fig. 132 _b_). The lava now welled out in larger volume corresponding to its greater head, and the stream which for ten months had been flowing from the highest outlet upon the cone now ceased to flow. The next morning, April 6, at about 8 o’clock, lava broke out at several points some distance east of the opening _b_, and evidently upon another fissure transverse to the first (Fig. 132 _c_). The lava surface within the chimney must still have remained near its old level,—effective draining had not yet begun,—since early upon the following morning a small outflow began nearly at the top of the cone upon the opposite side and at least a thousand feet higher. [Illustration: FIG. 131.—Scoriaceous lava encroaching upon the tracks of the Vesuvian railway (after a photograph by Sommer).] [Illustration: FIG. 132.—Map of Vesuvius, showing the position and order of formation of the lava mouths upon its flanks during the eruption of 1906 (after Johnston-Lavis).]