(Asteraceae), initiated the colonization of the inte-
rior. As emphasized above, colonization of the
strandlines was rapid during this phase, gradually
diminishing thereafter.
Phase 2 represents the period during which the
extensive grasslands waxed and waned, as, increas-
ingly, animal-dispersed trees and shrubs spread out
from their initial clumps to form woodlands. The
interiors have been filled almost exclusively by
species which are primarily either wind- or animal-
dispersed. The relative balance between these two
groups shifts dramatically between phase 1 and 2.
Few ferns colonized during phase 2 despite the effi-
ciency of their dispersal system (microscopic
spores). The interpretation offered for this is that
there are actually relatively few ferns in the regional
species pool which typify such extreme, seasonally
droughted pioneer sites, and they arrived very
quickly during phase 1. Wind-dispersed early suc-
cessional flowering plants also continued to accrue
during phase 2, including Asteraceae and terrestrial
orchids, but very few trees and shrubs. The vast
majority of arboreal species are animal-dispersed,
and after a very slow start, in which only two
species of animal-dispersed species were found by
1897, their rate of arrival took off in the early 1900s.
Phase 3 runs from the point when forest closure
became extensive, to the 1989 datum. During this
phase, the colonization rate of sea-dispersed
species slackens, which reflects the relatively
limited source pool, but also the limited array of
coastal habitats available. In the interiors, it was
only when forest habitats became available that the
second wave of ferns, those of forest interiors,
could build in numbers. Hence their rate of
discovery increased rapidly in the period of forest
formation, peaking around 1920. The ease of dis-
persal of ferns, with their microscopic propagules,
can be demonstrated by comparison of the size of
the Krakatau flora (all islands 1883–1994) with the
native flora of West Java. The ratio for spermato-
phytes is 1:10.1 and for ferns 1:4.2, indicating
Krakatau to have a remarkably rich fern flora
(Whittakeret al. 1997). Wind-dispersed flowering
plants colonizing the newly available forests of
phase 3 were predominantly orchids, many of
which are epiphytic, together with a mix of other
epiphytic and climbing herbaceous species. The
single largest group recorded for phase 3, however,
is the animal-dispersed set, mostly being trees and
shrubs.
More detailed analyses of dispersal mechanisms
suggest further structural features in the data. For
instance, wind-dispersed species can be split into
those with dust-like propagules, plumed seeds, and
winged seeds or fruits. The larger, winged propag-
ules have limited dispersive ability and, although
significant in the regional pool of forest canopy
trees, only one or two species of this category have
colonized Krakatau (Whittaker et al. 1997): they are
thus noted in Fig. 5.6 as improbable colonists.
Colonization of the animal-dispersed species of
Krakatau, with the exception of a few human-
introduced cases, can be attributed to the actions of
frugivorous birds and bats. Whittaker and Jones’s
(1994b) best estimates of their relative roles are
given in Table 5.3. Fruit bats swallow only the
smallest seeds (1 or 2 mm) and it was assumed
that they would be unlikely to carry larger seeds in
their mouths or claws over the many kilometres ofKRAKATAU—SUCCESSION, DISPERSAL STRUCTURE, AND HIERARCHIES 135Table 5.3Estimates of the numbers of plant species found on
Krakatau between 1886 and 1992 for which birds and bats have a
dispersal role. Under the heading of animal dispersal two modes of
transport may be distinguished: those seeds which are eaten,
swallowed and which pass through the gut or which are eventually
spat out, and those transported by external attachment to the
animal. The data set include all four Krakatau islands and all records
(i.e. including some species which have not maintained a presence)
(source: Whittaker and Jones 1994b.)Dispersal mode Number of
speciesAnimal, gut passage (bird and/or bat)* 124
Animal, attached externally (bird) 10
Human introduction, but spread by
feeding birds and/or bats 15
Sea-colonist, then spread by animals 24
Total animal introduction and/or spread 173* Of these 124 species, 50 were considered to be bird-dispersed, 31
could potentially have colonized either by bird or bat transport, and
31 were considered to have arrived by bird but then to have the
potential to be bat-spread within the islands. These estimates should
be regarded as first approximations (e.g. see further work on
bat-dispersal on Krakatau, by Shilton 1999).