284 DR. TEMPEST ANDERSON ON THE ERUPTIONS OF THE SOUFRIERE IN 
another still less evident on the east. Neither is sufficiently conspicuous to appear on the 
photographs. Thus all the deeper parts of the crater bear evidence of the severity of 
the eruption of 1902, since it appears that any loose deposits previously existing there 
have been entirely blown out, and that the small bench or talus mentioned above is 
of subsequent formation. The upper parts of the walls of the crater are on the whole 
less precipitous than those lower down, and are in places, especially on the south and 
south-west, covered with a deposit of new ash apparently only a few feet thick 
(Plates 16 and 18). It dips towards the crater at the angle of repose between 30° and 
40° and has been much denuded by the rain. There are only a few places, however, 
where it is sufficiently continuous to obscure the solid beds, and where these are 
visible ash and tuff predominate over lava, the reverse of which is the case lower 
down. On the north the wall is considerably higher than on the south, and the 
precipitous portion extends higher up. The beds of tuff, which are here particularly 
well developed, contain many large ejected blocks (Plate 17). This abundance of 
ejected blocks to the north-east of the crater agrees with the observation of 
Mr. T. M. Macdonald during the eruption of 1902, that most of the stones thrown out 
went to windward, so that the direction appears to have remained unchanged from 
an earlier period. The figure to the right is standing on a lower portion of the rim 
(Gap A in several plates) near the point where the Carib track begins to descend in 
a south-easterly direction to the windward side of the island, and through this gap, no 
doubt, the black cloud and avalanche descended which devastated the Carib 
country. 
This lip of the crater is usually quite narrow, generally only a few feet, occasionally 
a few yards, wide, as shown in Plate 16. It is mostly composed of a bed of 
new ashes a few feet thick, almost everywhere consolidated on the surface into a 
crust, generally less than an inch thick, as in the case of the ridges lower down. 
Where this crust is entire it has preserved the rest of the bed from erosion, but 
whenever it has been broken through, the whole of the deposit has generally been 
washed away, and this is particularly noticeable on the outer slopes, where the 
beds of new ash lie conformably on the old ones and weather off in successive layers. 
The lip is by no means regular or uniform in height. It is highest to the north. 
Besides the gap just mentioned (Gap A), it presents three well-marked gaps, B, 
C, and D, of which B is shown in Plate 16, and D, the more westerly, in greater 
detail in Plate 18. B and C are on the south of the crater; of these the more 
westerly, C, is somewhat the lower, and both are lower than A. It was doubtless 
through them, and presumably through the lower one especially, that the water of 
the crater lake and the incandescent avalanche descended into the Wallibu and 
Rabaka districts. 
The other gap, D, Plate 18, occurs more to the west, where the lip of the crater 
joined the Somma ridge ; as the whole of this part of the rim of the crater is much 
higher than the southern portion, the bottom of the gap, deep as it is, still 
