milar
loose cones of scoriae, probably of as high antiquity as the whole mass
of Etna, stand uninjured at inferior elevations above the level of the
sea.

CHAPTER XXVI.

Volcanic eruption in Iceland in 1783--New island thrown up--Lava
currents of Skaptar Jokul, in same year--their immense
volume--Eruption of Jorullo in Mexico--Humboldt's theory of the
convexity of the plain of Malpais--Eruption of Galongoon in
Java--Submarine volcanoes--Graham island, formed in 1831--Volcanic
archipelagoes--Submarine eruptions in mid-Atlantic--The
Canaries--Teneriffe--Cones thrown up in Lancerote, 1730-36--Santorin
and its contiguous isles--Barren island in the Bay of Bengal--Mud
volcanoes--Mineral composition of volcanic products.

_Volcanic eruptions in Iceland._--With the exception of Etna and
Vesuvius, the most complete chronological records of a series of
eruptions are those of Iceland, for their history reaches as far back as
the ninth century of our era; and, from the beginning of the twelfth
century, there is clear evidence that, during the whole period, there
has never been an interval of more than forty, and very rarely one of
twenty years, without either an eruption or a great earthquake. So
intense is the energy of the volcanic action in this region, that some
eruptions of Hecla have lasted six years without ceasing. Earthquakes
have often shaken the whole island at once, causing great changes in the
interior, such as the sinking down of hills, the rending of mountains,
the desertion by rivers of their channels, and the appearance of new
lakes.[583] New islands have often been thrown up near the coast, some
of which still exist; while others have disappeared, either by
subsidence or the action of the waves.

In the interval between eruptions, innumerable hot springs afford vent
to subterranean heat, and solfataras discharge copious streams of
inflammable matter. The volcanoes in different parts of this island are
observed, like those of the Phlegraean Fields, to be in activity by
turns, one vent often serving for a time as a safety-valve to the rest.
Many cones are often thrown up in one eruption, and in this case they
take a linear direction, running generally from northeast to southwest,
from the northeastern part of the island, where the volcano Krabla lies,
to the promontory Reykianas.

_New island thrown up in 1783._--The convulsions of the year 1783 appear
to have been