388 
DES. T. ANDEESON AND J. S. FLEET ON THE EEUPTIONS OF THE 
was there has survived to tell the tale. If then, as we believe, the lake still occupied 
the crater till about 1 o’clock on the Wednesday afternoon, one very curious fact 
requires to be explained. There can be no doubt that the reflection of fire was seen in 
the steam cloud at 5.20 p.m. on Tuesday and repeatedly after that. All witnesses 
agree on this point. It could not have been due to the burning of the sulphuretted 
hydrogen gas, as this gives a colourless flame, and after passing ujd through boihng 
water would require contact with sometliing incandescent to ignite it. On the other 
hand, it may have been, and very probably was in part, the burning of the withered 
bush which clothed the interior walls of the crater. Many describe the fire, seen after 
the great oidbui’st at midnight on Tuesday, as “ like fire running through a cane 
brake.” On Tuesday morning the bush in the crater Avas green ; on AVednesday 
morning it was l)urnt, and it is in every way likely that some of the “ flames ” AA^ere 
due to bush fires. They may have been ignited by flashes of lightning or by incan¬ 
descent stones Avhich had fallen, sufficiently hot to set fire to the timber. In the 
latter case it is necessary to suppose that a cone Avas built up Avithin the crater 
sufficiently high to lise above the surface of the lake, and that in the centre of this 
cone there Avas a minor crater from which the main explosions proceeded, and in which 
a surface of molten rock was exposed. 
The configuration of the crater floor beneath the AA^aters of the lake is not fulh^ 
knoAvn, and it is by no means improbable that such a cone existed before this eruption, 
but concealed from vieAA^ Descriptions of the crater as it Avas before the eriq^tion of 
1812, state that within the lake there Avas a conical hiU seA-eral hundi’ed feet 
high (see p. 462). From the eAudence of the fish .sellers it is clear that Meam AA-as 
rising only from the centre of the lake, and this Avould indicate the presence of 
an orifice there, comparatively near the surface. Moi'eoA’er, they stated that on 
Tuesday .stones Avere being cast up Avith the puffs of steam, and this Avould ineAutably 
tend to build a cone around the outlet. In this regard the eAudence of a little boy 
Avho accompanied these Avonien may be of some importance. He stated that on 
Tuesday morning he looked into the crater and saAv the lake at its usual leA’el, but 
discoloured and boiling at the centre. Where the Avater boiled he “ saAv a stone 
floating,” and the Avater boiled only Avhen it touched the stone. Though stringently 
cross-examined, he insisted that there Avas a floating stone on tlie surface of the 
Avater. This may have been the summit of an interior cone just projecting at that 
time above the water level. No red reflection Avas seen till at any rate six hours 
later, and by that time the lake may have someAvhat diminished, or the cone 
have l)een built up to such a height as to rise freely above its surface. 
According to the soundings pidolished by Mr. Foster Huggins, since Ave left 
St. Vincent, the depth of the de})ression increased on all sides toAvards the centre 
aAvay from the shore. The Admiralty chart shoAvs the lake to haA’e been about half 
a mile in diameter, and as Mr. Huggins gives the greatest distance from the shore on 
the north -east and Avest, sides to Avhich he carried a chain of consecutiA^e soundings 
