injured, but was released from its anchorage by the breaking of the chain. The Roraima, still nearer the center of the force, was almost overturned, and deprived of its upper works and rigging. The Grappler, still nearer, and lying off the mouth of the Rivière des Pères, was completely annihilated. The decrease of the force from without inward is testified by all the objects. The center of this force seems to have been at the north end of the city, where the destruction was greatest, and every trace of culture, vegetation, buildings, and humanity was almost instantaneously annihilated.

Not only was there a tremendous outward force, but this was followed by a return movement, as if a vacuum had been created. This return movement was gentler than the outward force, and brought life and vitality to many who had been exhausted for the lack of air to breathe. Another marked phenomenon of the explosion was this exhaustion of air, as testified by all witnesses who were within the radius of the blast, indicating that the oxygen had been completely consumed. Then another peculiar thing occurred: from the heavens there fell a rain of thick liquid mud, accompanied by small stones of pumice, which lasted for a considerable interval, some say as long as thirty minutes.

This apparently was not rain from the sky, but moisture which had developed within the cloud, and, mingling with the ash-like lapilli, formed mud within the sky. Wherever this liquid touched, it plastered the surfaces as though it had been placed on by the hands of man; the landscape was bedaubed with a continuous smooth coatingships and trees, stones, and even the corpses in the streets. Trees, smoke-stacks, and rigging were all plastered alike; even the bodies of the survivors on the ships were so coated that the cement-like mass had to be literally cracked off the heads of the negroes.

Besides these phenomena seen, there was another substance, which, though invisible, was perhaps the chief agent in the resulting disaster. I refer to the gases which accompanied it. Yet there is no statement so hard to demonstrate as this. That gases were present there can be no doubt whatever, as testified by the phenomena themselves and the traces which they left upon the wreck. What these gases were is another question, for although the gaseous phenomena of volcanoes are least understood, it is a wellknown fact that the vapors of all the Caribbean soufrières are sulphurous, and it is highly probable that either sulphureted hydrogen, H2S, or sulphur dioxid was crimi

nally implicated in the disaster at St. Pierre. The chief surgeon of the Dixie expedition alleges that some of the people at St. Vincent were killed by the inhalation of sulphur dioxid. Up-to-date chemistry has not been of much service in detecting the nature of these gases, but we must remember that full and specific collections with chemical determinations cannot be made until Pelée ceases erupting. It is a significant fact, however, that twigs collected from the trees by me in St. Pierre showed a sulphurous coating, and that all the silverware was blackened by this substance. It is even stated that the captain of the Suchet picked up pieces of pure sulphur in the streets of St. Pierre. Still more significant is the fact that for weeks before the catastrophe the city was filled with sulphurous smells-"so strong," wrote Mrs. Thomas T. Prentis, wife of the United States consul, "that horses stopped and snorted, and some of them dropped in their harness and died of suffocation."

The problem of exactly how St. Pierre and its inhabitants were destroyed is still before us. There are two theories:

1. The heat-blast theory. This assumes that the lapilli, gases, and steam of the cloud were ejected with sufficient initial force to destroy buildings from two to five miles distant, and were sufficiently hot to inflame the city and destroy the people by singeing, suffocation, and asphyxiation.

2. The aërial gas explosion theory. This postulates that the weight of the cloud, causing it to descend, the exhaustion of air, the flame, and the great aërial force developed, were the products of an explosion caused by the union of the gases of the cloud with the oxygen of the air, which took place in the air, but near the surface of the ground.

From whatever point of view the subject is approached, all the evidence focuses upon a single deduction: that there was a terrific aërial explosion within the cloud after it erupted from the mountain, which developed tremendous destructive forces, and that the situation of St. Pierre adjacent to the bluff

1 Professor Hill informs us that since writing the foregoing he has received from Dr. Emil Deckert of Berlin, who ascended Pelée in 1898, interesting confirmation of the view that the deadly eruption was from the lower lateral crater on the west side of the mountain. Dr. Deckert, in an address delivered in Berlin on the 12th of May, asserted this opinion very emphatically on the ground of his own observations in 1898 and those made by M. Léon Sully on the 25th and 26th of April of the present year, and of the newspaper reports. Professor Hill finds further confirmation of his theory

behind it was such that reverberation caused therefrom assisted in its destruction.

The position of the cliffs inclosing the area of devastation would have had much to do with the destruction of St. Pierre, inasmuch as places outside of the cliff-bound amphitheater and upon the edge of the plateau above its depth were spared from destruction.

The aërial explosion, if it occurred, involves the presence within the cloud of a combustible gas, but science is still unable to state its nature. The distinguishing of explosive gases involves a faculty of scientific specialization which the writer does not possess; but as sudden and mysterious as was the great secret, it has left its traces and clues, which the detectives of science will follow up. Metal surfaces of objects in the ruins will be examined and analyzed for traces of sulphur and chlorids. The deposits from the numerous steaming fumaroles are already within the chemical laboratory. Even the ash and rocks of the island will be submitted to minute investigation.

Haunting my mind is a hypothesis which may be but a dream, and which cannot be proved, yet is one which would explain all the phenomena.

On that fateful morning two clouds erupted almost simultaneously from the mountain, one following the other at a slight interval. The first of these came from the open flue of the summit chimney and floated southward horizontally toward Mont Vert. In this cloud developed a violent storm of electric discharges. The second came from lower down the mountain at the spot probably known as La Soufrière or L'Étang Sec. The ejecta from the latter vent were not from an open flue, as the summit cloud, but were from the initial explosion of an ancient vent long clogged up by an accumulation of old material and sulphurous gases. Being heavy with gases, and while rolling toward the city, what would have happened had that host of electric sparks from the upper cloud flashed into the lower aërial mass of superheated gases, even though they had otherwise refrained from already uniting with the oxygen?1

in the following extract from an article by G. Mazet published in "L'Opinion" of Fort-de-France of May 7, but written before the destruction of the Guérin factory on May 5: "The eruption was accompanied by no serious trepidations of soil, but a rain of ashes, broken through by burning flashes of gas, alone signalized this volcanic eruption. In the neighborhood commotions were produced in the air, the effect of returning shocks, and created the impression of frequent seismic phenomena. However, the soil has only felt a few slight undulations of short duration."-EDITOR.