Pierre. In the same fashion Richmond is buried, 45 feet deep at its northern end, 3 feet at the southern. In the same way the one masonry building in Richmond village was. thrown down, and five-foot blocks of masonry blown forty feet away from the mountain. The very odor over the ruin is the same. Just south of the Soufrière there is a group of high mountains. These blocked the passage to leeward southwest, and the heavier material thrown out of the crater's throat accumulated in the basin between these mountains and the volcano. Great drifts of fiery hot sand and gravel fell here and remained hot for weeks. The local torrential streams at the beginning of the rainy season, working on these beds, are converted into steam and make spectacular explosions, which alarm the natives, but are really quite harmless. At the sea-front west of the Soufrière there have been submarine landslips on a considerable scale, leaving in some instances vertical earthen walls 50 feet high where before there was a peaceful little village of thatched huts. The sea laps the foot of these unbalanced precipices, and a twelve-foot oar three feet from the strand-line finds no bottom. To a geologist one of the most remarkable effects of such an eruption is the rapid wearing away of the land that succeeds it. The protecting matter of tropical jungle has been burned and buried under two to ten feet of angular sand-grains; heavy rains cut torrent trenches in this material and all the old slopes are suddenly steepened by the falling away of the seashore and the piling of volcanic débris on all the crests of the land. The mountain is soaked in steam and water, local showers are at work all the time, and the result is like playing the hose on a steep mud pie; trenched and rill-marked in every direction, landslides are the rule in such topography and no slopes are safe. I landed for only a few moments to collect a specimen in one of the deep canyons northwest of Soufrière; a few blows of the hammer started a twenty-foot block of solid rock out of the gulch-slope a hundred yards away, and immediately afterwards a great bank of earth some twenty-five feet high came crashing down only a few feet from me. Needless to say I beat a hasty retreat to the boat from such an unstable land. We learned from this experience that the only safe place on new volcanoes is on the crest—unless one includes the Irishman's position, on 'leave of absence.'
There are a number of questions constantly asked of a geologist in this field, and some of these I must try to answer briefly here, without going further into a description of geological details. Was there any forewarning of the eruptions? Clearly there was; the prediction quoted at the head of this article was based on well-authenticated data. At Pelée, the lake in the crater was warm, and the smell of sulphuretted hydrogen perceived as far back as January. In April
