Vol. 66.] 
VOLCANO OF MATAVANTJ IN SAVAII. 
6 
oo 
')') 
watched the formation of ordinary corded lava on Vesuvius by a 
similar process of budding'. When the hot lava of the interior has 
got vent and formed a pool, a soft scum forms on the surface, 
which is pushed forward by the fluid lava moving beneath, and 
is raised into a wrinkle or cord, then others are formed in the same 
Avay, until the surface is entirely solidified into a succession of 
cords (PL XLVIII, fig. 1). The same sort of process can often be 
watched on the surface of a slowly moving river when rubbish 
floating on the surface is stopped behind a log of wood. Here it 
was different: there was no trace of the formation of wrinkles and 
cords. The whole surface seemed to be chilled at once, as the 
waves rolled on and off, and examination of cooled specimens 
between high and low water-mark at the edge of the lava in the 
lagoon confirmed this. The surface at and below the water-level 
was roughly granular, like that of air-chilled bombs such as I have 
seen on Stromboli and Haleakala, while higher up the ordinary 
corded or pahoehoe structure was seen (PI. LII). I was vividly 
reminded of a section that I have seen at Aci Castello in Sicily. 
before the eruption a small river came down from the mountains, 
passed through the villages of Patamea and Samalaeulu, and fell 
into the sea at Sataputu. Both the higher and the lower reaches 
of the river have been covered over and filled up with lava, leaving 
the channel at Patamea and Samalaeulu unobstructed, but usually 
empty. The rain, which falls on the former gathering-ground, is 
soaked up by the lava and possibly percolates elsewhere, so that 
scarcely any water now comes down the channel except in very 
heavy rain-storms, which have only occurred once or twice since 
the eruption. On one such occasion the water could not get away 
owdng to the obstruction near the old mouth, and the country 
round was flooded. The empty channel is a good example of a 
water-eroded gorge cut in geologically recent beds of lava, and 
shows the usual water-holes, as also small falls in its bed. At one 
place where I saw it, it was nearly 30 feet deep and not much 
wider. At others it was perhaps 3 or 4 times as wide. 
The Destruction of Buildings. 
The native houses in the villages destroyed consisted merely of 
thatched roofs supported on wrnoden poles. The churches of 
Baleaula and Sataputu were, however, substantial stone bu ildings 
Clements, J. M., Monogr. TJ.S. Greol. Surv. xxxvi (1899) p. 124, thinks 
that piLow-lava is a form of aa, not pahoehoe. The aa referred to is 
not the common scoriaceous form, but one described by Dana as above. 
Having now seen the actual formation of all three—aa, pahoehoe, and 
pillow-lava, I am satisfied that the scoriaceous form of aa is produced by 
watery vapour expanding from inside the pores of the lava; pahoehoe from 
the slow cooling of lava containing lit tie vapour; and pillow-lava (at any rate 
one form) from the chilling action of water on the surface before it has time 
to assume the corded structure of pahoehoe. I have not seen the formation 
of aa as described by Dana. 
