ST. VINCENT IN 1902, AND ON A VISIT TO MONTAGNE PELEE. 
279 
Further up the mountain the remains of the avalanche became more abundant in 
the valley bottoms (Plate 13), and here they were also often better preserved, so that 
traces of the feather-pattern erosion, so noticeable in 1902, # were still visible on the 
surface. This was mainly due to the surface of these ash deposits, like those to be 
presently mentioned on the plateaux and on the ridges, having consolidated into a 
crust almost like a cement pavement which resists the action of the rain. Another 
very curious and, it is believed, novel point was observed with regard to these massive 
ash deposits. Instead of one stream re-establishing itself along the centre of the 
deposit, the tendency is for a new stream to form on each side at or near the 
junction of the new ash with the old valley walls, and as these streams deepen 
themselves, two new valleys are formed where only one previously existed, and the 
walls of each are composed on the one side of the new ash, and on the other of 
older tuff with occasional terraces of new ash. Sometimes the two valleys coalesce 
by the washing away of the central mass, but quite often the two remain distinct, 
as in the case of the Trespe Riverf and Wallibu Dry River to be mentioned 
shortly. The fact of this side formation of streams is clear; it is seen in the 
ravine just noticed, also in the upper Rozeau, and in several of the plates in 
Part 1, where, though not noticed at the time owing to its being then in an early 
stage, it is distinctly visible when looked for, but the cause is not so clear. It 
appears to be due to the fact that the water from the old slopes in running down 
into the original valley meets the soft new ash, and at once turns down along the 
valley and so starts the new stream, and it seems likely that the chief cause of its so 
turning is that the surface of the deposit tends to be higher along the middle of the 
valley than at the sides, the shape of the mass somewhat resembling a glacier, which 
it is well known is usually higher in the middle because of the more rapid motion of 
that part. 
The Wallibu Dry River and Trespe River. —These are two small and short rivers 
to the north of and parallel with the lower part of the Wallibu valley, and both run 
in deep gorges in the floor of a wide valley bounded on the south by the Wallibu 
plateau. In 1902 both these gorges were filled with new ash to the level of the 
main valley floor, and the process of re-excavation of one of them is shown in 
Part I., Plate 30, fig. 2. It was noticed that this main valley was wider and 
more open than the Wallibu valley,| but no explanation was forthcoming at the 
time. Mr. Duncan Macdonald now informs me that before the 1812 eruption 
the Wallibu River flowed down the Wallibu Dry River, and that its course 
was changed after that eruption. Mention has already been made of the Wallibu 
valley making an abrupt turn to the south and turning again to the west round the 
* Part I., Plate 26 and Plate 28, fig. 1. 
t I heard this river called also Cobree. The confusion is due to these gorges being liable to be filled up 
and re-excavated in slightly different positions. 
f Part I., p. 429. 
