SOUFRIERE, AND ON A VISIT TO MONTAGNE PELEE, IN 1902. 
431 
widening of the gorge by the action of the gently Howiiig stream of thick mud is 
resumed, though now at a somewhat lower level. A new terrace is in this Avay 
formed inside of the old one which has been partly destroyed. 
In the upper part of the Wallibu Valley, at a distance of about a mile from its 
mouth, it widens out to form a comparatively broad expanse, where the principal 
tributaries join to form the main stream above Petit Wallibu. Here was a wide 
plain of new volcanic ash (see Plate 28, fig. 2), which at the time of oui’ visit had as 
yet' been ■ comparatively little eroded, and pi'eserved in some measure its original 
surface configuration. The streams wliich cross this flat had each cut its gorge in 
the deposit, but between these there still lay wide stretches of the recent sand, w'hich 
had hardly been attacked by erosive agents. The surtace, .though on the whole 
level, w’as rolling and irregular, and bore several pools and one lake of greenish 
turbid w’ater. As seen from a distance of less than a mile this lake was perhaps 
200 yards long, and steamed strongly, the water being apparently quite hot. It 
was impossible to get near to it, as the gorges cut in the hot smoking ash could not 
be crossed in safety. The swelling, hummocky, gently-rounded features wdiich had 
originally characterised the ash-deposits, had in this area not yet been masked by 
the effects of erosion as in the stream valley lower down. 
The ash which had gathered in the valleys of the Wallibu and Wallibu Dry Ptivers 
was a fine sand, nearly black when wet, but, when diy and hot, brovuiish-grey or 
buff in colour. In the month of June the coating of fine grey dust which had 
at first formed the surface layer had been washed away, and the mass of the material 
was about as coarse as an ordinary sea sand or a little coarser. In it lay many lapilli, 
scoriaceous, yellowdsh, with white shining crystals of plagioclase and dark pyroxenes. 
Large rounded bombs were less frequent, but many could be found varying from a 
few inches up to 3 feet in diameter. They were rounded or flattened, witli rough 
tuberculate or nodular outer surfaces, and when broken up were black in the interior, 
vitreous and vesicular, with scattered crystals of pyroxene, plagioclase, and olivine. 
Ejected blocks of the w^eathered andesites and reddened decomposed andesitic tuffs, 
which constitute the walls of the crater, were quite as numerous as the fresh 
spongy vitreous bomljs, and entirely distinct from them in character and appearance. 
About half a mile north-west of Wallibu works were some ejected blocks, roughly 
cubical and measuring 5 feet ’oy 4 feet by 4 feet. As they lay upon a layer of 
the new ash they could not have fallen from the cliffs behind, Ifom which also 
they were several hundred yards distant. In the same place were bombs over 2 feet 
across. 
Many of these ejected blocks when struck with the hammer broke with great ease, 
and large, flat pieces commonly flaked oft’ their surfaces. They had been intensely 
heated before being projected from the crater, and rapidly cooled in their passage 
through the air, and it seemed as if concentric cracks or j^lanes of weakness had been 
produced by the contraction of the chilled surface on the hot interior, recalling in 
