comparable to closing the safety valve on a steam boiler
while the boiler continued to build up pressure inside.
Eventually, the “boiler” in this case, the volcano had to
give way and explode. The explosion began with a series
of powerful blasts on August 26–27.
Loud explosions were heard at Batavia and Butzenbourg,
about 100 miles (161 km) from Krakatoa, early on the after-
noon of August 26. A few minutes later, the captain of the
ship Medea about 75 miles (121 km) from the island, saw
smoke and vapor rise from the volcano to an estimated alti-
tude of more than 15 miles (24 km). At about 5 p.m., some
four hours after the final series of explosions began, the noise
of Krakatoa’s destruction was audible all over the island of
Java. Around sunset, the captain of the ship Charles Bal,
some 10 miles (16 km) south of Krakatoa, was so intimi-
dated by the eruption that he ordered the ship to retreat east-
ward. About this time, large chunks of pumice began landing
on the decks of the Charles Bal. Meanwhile, the captain of
the ship Sir Robert Sale, also in the vicinity of Krakatoa, wit-
nessed spectacular displays of lightning in the eruption cloud
and detected an odor of sulfur in the air. The sulfur odor
became intense during the night, as bright flashes of light
from the volcano lit up the surrounding area in the manner
of a giant strobe light. A thick cloud from Krakatoa cov-
ered Batavia soon after dawn on August 27, and a muddy
rain began to fall, followed by a fall of small clumps of dust
held together by moisture. This strange precipitation ended
in midafternoon. Although atmospheric effects hid Krakatoa
from view during this period, the sound of its final destruc-
tion was apparent. Four great explosions, starting at 5:30
a.m. on August 27 and ending shortly before 11:00 a.m., sent
pieces of rock showering down on ships in the vicinity and
were heard as far away as Rodrigues, an island in the Indian
Ocean some 3,000 miles (4,800 km) away from Krakatoa.
Heat and hot gases emitted from ejecta allowed them to
glide on a cushion of air across the surface of the water like a
hovercraft for as many as 60 miles (97 km) where they came
ashore on islands and started fires. The explosions destroyed
the north half of Krakatoa and cut Rakata in two vertically
from base to summit. The shock waves were registered by
every barometer on Earth.
The explosive destruction of Krakatoa generated tsuna-
mis that caused extensive destruction and loss of life along
shores in the region. In some places, the waves are estimated
to have come ashore at heights of 120 feet (37 m). The
men of the ship Loudon turned their ship’s bow toward the
approaching tsunami and rode out the wave’s passage and
then watched as the tsunami engulfed and destroyed a town.
Similar scenes occurred along the coasts of Sumatra and Java
near Krakatoa. In one location, a town some 10 miles (16
km) from the water was submerged by the wave. Some 3,000
residents of Karang Antoe were killed by the tsunami, as were
more than 2,000 at Anjer and Batavia and 1,500 at Bantan.
The total loss of life from the tsunami that followed Kraka-
toa’s explosion is uncertain, but estimates range in the tens of
thousands (most recently 36,000). The tsunami from Kraka-
toa was strong enough to twirl ships around their anchors
at the harbor in Colombo, Ceylon (now Sri Lanka). Some
deaths from this eruption were unrelated to the tsunami.
In 1990, more than a century after the eruption, an
expedition to Krakatoa, including a team from the Univer-
sity of Rhode Island Graduate School of Oceanography,
examined the role of pyroclastic flows in generating the
tsunamis that originated from Krakatoa during the 1883
eruption. This expedition took samples of clastic material
from the ocean floor around Krakatoa. Sediment cores taken
in this way showed evidence of a large, poorly sorted main
deposit, overlain by another deposit of volcanic gravel and
sand. Pieces of pumice large enough to impede the action of
the sediment-coring mechanism were reported, and the inves-
tigators noted, in addition to the chunks of pumice, pieces
of rock more than 12 inches (31 cm) across in some cases.
These large fragments of rock are contained in a matrix with
a silty and sandy character, the same as deposits on land.
The similarity between the submarine and terrestrial depos-
its indicated that the underwater deposits originated as pyro-
clastic flows. The expedition also determined that submarine
mounds near the Steers and Calmeyer island platforms just
north of Krakatoa were made up of pyroclastics from the
eruption in 1883. On the Calmeyer island platform, the
investigators noticed signs of turbidity currents that may
have resulted when pyroclastic material at high temperature
came in contact with seawater, causing steam explosions
that gave rise to ash-filled turbidity currents. At one point on
the seafloor west of Krakatoa, the expedition found a pyro-
clastic flow deposit some 300 feet (91 m) thick laid down by
the eruption, including fragments of pumice several feet in
diameter.
The 1990 expedition found evidence that one or more
pyroclastic surges preceded the tsunami from Krakatoa, and
two or more additional pyroclastic surges came after the tsu-
nami. Deposits included pieces of coral that appear to have
been torn from the seafloor and deposited with the pyroclas-
tic material by the tsunami. The investigators concluded that
deposits on land from Krakatoa’s 1883 eruption are made up
largely of material from pyroclastic flows as are underwater
deposits from that eruption. The pattern of these deposits
shows they were laid down in all directions around the island,
not principally to the north, as had been presumed before.
Accumulations are greatest to the west of Krakatoa. Deposi-
tion of pyroclastic material on the west side of the volcano is
thought to have generated powerful tsunamis.
That the meeting between ocean and pyroclastic flows
was not gentle is indicated by the small amount of mixing
between seawater and pyroclastics. The pyroclastic materi-
als show poor sorting, evidence that there was little mixing
between them and the water. Had mixing occurred on a
significant scale, water action would have resulted in more
effective sorting of material. This pattern of poor sorting
persuaded the scientists on the 1990 expedition that masses
of pyroclastic material from the volcano in 1883 displaced
water in such amounts as to cause the destructive tsunamis
associated with the eruption of Krakatoa. Apparently, five
or more units of pyroclastic flow occurred in the eruption.
The results of the 1990 expedition indicate that one particu-
lar tsunami, at approximately 10 a.m. on August 27, 1883,
was linked with the formation of the caldera at Krakatoa
during an especially violent portion of the eruption, possi-
140 Krakatoa