SOUFEIERE, AND ON A VISIT TO MONTAONE PELFE, IN 1902. 
437 
enters a shallow open broad channel, which is partly filled with terraced masses of ash. 
This is known as “ the lava bed,” and the accumulation is said to be the product of 
the eruption of 1812. Before that time the river is believed to have flowed more or 
less continuously, but since then it often dries up for weeks together, and its flow is 
considerable only after rains. That at Wallibu so much more water should be found 
than in the larger stream at Babaka is certainly a curious fact, and one for which it is 
not possible to offer an explanation without making a thorough study of tlie geology 
and physical features of its drainage area—an investigation which could not be 
undertaken in the present condition of affairs. 
In the upper part of the Babaka Dry Biver enormous accunuflations of hot sand 
have been piled up exactly like tho.se already described in the Wallibu Valley 
(see Plate 32). We were told by Mr. Spence, who is well acquainted with 
the district, and visited it along with us, that in some places this deposit could 
not be less than 200 feet thick. Certainly it had almost completely olfliterated 
the old valley of the river at more than one point, and where formerly there 
had been a deep trough, there was now a rolling plain of black sand, through 
which, after rains, a .sluggish stream of mud wound its way, but on a dry day the 
steaming channel was streamless. Here erosion had produced much smaller effects 
than on the leeward side, and in many places the asli had been practically unattacked, 
and the original nature of the surface could be well seen (see Plate 34, fig. I). It 
was flattened 021 the whole, but rolling, hummocky, and uneven, so tliat pools of water 
had gathered in the hollows, and in these little deposits of alluvium had formed from 
the fine materials washed into tliem by the luin. The.se pools were now mostly diy, 
and the dark surface of the a.sh was freely steaming. After a couple of hours of dry 
weather, spots of grey would ap})ear where the heat had been sufficient to dry the 
uppei- layers. Then our porters coifld with difficiilty cross it witli bare feet ; when 
wet it was quite cool to walk upon. The deeper parts of this deposit must have been 
at a very high temperature. Where the hot, grey ash was exposed by landslips in 
the banks of the stream, it had a stifling stuffy smell. A good deal of charred wood 
was mingled with the sand, and there were many spongy, vitreous bombs, but angular 
broken pieces of the old rocks of the hill made up far the greater portion of the coar.se 
material in the mass. As a whole, however, it was an accumulation of sand and dust; 
bombs and ejected blocks were more numerous than on the Wallibu, but formed only 
a small percentage of the mass. 
In the main valley of the Babaka Dry Biver vast quantities of this material had 
been heaped, but, strange to say, the tributary valleys contained far less of it, and 
the great pile of ash in the main channel often formed dams across their mouths, 
behind which lakes of water and of liquid mud had formed (see Plate 34, figs. 
1 and 2). On several occasions great and .sudden floods of mud had rushed down the 
stream, and it was clear that they might have originated by the water eroding 
these dams till they were so weakened that they gave way before the pressure of 
