landing than two more figures appeared and we had to put back for them. Meantime the mountain was sending up clouds of steam from all over its slopes as though it were rifting in a hundred places preparatory to a titanic outburst. But it did not do anything, and I now strongly suspect that what we saw was the product of a smart shower over the mountain from the clouds lowering on its summit. The cool water rills running down the slopes come into contact with beds of hot dry gravel previously thrown out. Wherever such a contact is made a jet of steam at once is formed; we had opportunity to see much of this action later on at St. Vincent, and it is this process that has given rise to many stories of small craters forming.
We spent another day among the ruins, exploring St. Pierre to the mountain slope at its northern end and to the steep roadways that climb the southern cliff. The city was in a cul de sac, hemmed in by a cliff south, a higher cliff east and the ocean west; its northern end was on the actual foot slope of the volcano. The present crater was blown clear of clouds as we steamed past and we saw a cup-on the summit open to the west, walled in on the east, with a huge pile of scaly looking hot bowlders in its midst steaming violently. At the distance from which we saw it, it was estimated roughly that the cup would measure perhaps 2,000 feet across by 800 feet deep. This crater ends in a deep gulch west that extends down to the sea; old photographs show that this gulch was there before the present eruptions. Apparently it was down this gulch that the mud flood came winch overwhelmed the Guerin factory. All along the foot of the mountain are steaming fan-shaped deltas of débris, and far up the slopes these are matched by leaf-shaped arroyos or deep trenches hollowed out of the old earthy volcanic beds of which much of the island is composed. This trenching has been accomplished by the cloud-burst torrents of water laden with grinding sand that fell during and after the eruptions—in part condensed steam and in part heavy rains that recently have been exceptionally abundant. Much of the material which fell in the first outbursts of probably both Pelée and Soufrière was dry and red hot; it was relatively fine, the largest fragments falling near the vent. The grades of the hillsides were already steep and it is probable that this material flowed somewhat like dry sand. This, if seen at night, would account for the reports current of glowing molten lava on Pelée. When I was in Martinique there was no sign of molten rock, nor have I seen any in St. Vincent, and the greater part of the evidence at present makes these eruptions purely steam explosions which have blasted out and comminuted large quantities of the old country rock, or bedrock of the islands, itself an ancient volcanic product.