sea separating Krakatau from source areas. Thus,
small-seeded animal-dispersed species might be
introduced either by birds or bats, but larger-seeded
species can be considered strictly bird-assisted
colonists. Beyond a certain threshold seed size it is
likely that only the largest fruit-pigeons can effect
an introduction. However, once a species has
established on an island, a wider range of dispersal
agents may be involved in local dissemination.
Where plant species have two rather contrasting
dispersal vectors, they are termed diplochores.
Twenty-four sea-colonist, but potentially animal-
spread species occur on Krakatau, of which 14 are
bat-spread, 4 may be bat- or bird-spread, and 6 are
bird-spread. This mixed group of diplochores
includes several of the earliest colonists of the
Krakatau strandlines. They likely had an important
role in kick-starting the succession of animal-
dispersed species. As shown, restrictively animal-
dispersed species were laggardly colonists. There
cannot be much advantage in visiting a barren
island if you happen to be a frugivorous bird or bat.
Once the fringing plant communities of sea-
dispersed species established and began to fruit,
however, the islands would have provided a suit-
able food supply in the form of the fruit of the
diplochorous species. From the early accounts we
know that many of the first true animal-dispersed
plants were concentrated around the coastal
fringes. Admittedly the interiors were less well
explored, and certainly pockets of trees quickly
became dotted throughout the interiors, yet we can
be fairly confident that the first points of landfall,
and the first roosting points for frugivores and the
first food supply, were provided by the coastal veg-
etation.
The first frugivores observed on the islands
were birds, six species being observed in 1908 (the
first bird survey), whereas it was not until the next
zoological survey in 1919 that bats were noted.
Birds may thus have been the more important
group in terms both of simply the number of
animal-dispersed colonists and being earlier in
arrival (although survey data are too poor to be
conclusive). However, bats have had a differing
role to birds. First, they have spread some species
that birds do not transport. For instance, the
important coastal and near-coastal tree Terminalia
catappaonly invaded the island interiors after the
arrival of the bats. Secondly, Whittaker and Jones
(1994b) suggested that they might be more impor-
tant to the early seeding of open habitats because
of differences in their foraging behaviour (in part
mediated by interactions with predators). More
recent detailed work by Shilton (1999; Shilton et al.
1999) supports these ideas, and also crucially
demonstrates the ability of bats to retain viable
seed in the gut for over 12 hours between the final
feed of one night and the start of feeding on a sub-
sequent night, by which time they may easily have
moved from mainland Java or Sumatra to the
Krakatau islands (or vice versa). Although their
role in long-distance seed dispersal seems limited
to only a few small-seeded plant genera with the
right types of fruit (especially Ficusspecies), we
are now convinced that their role in kick-starting
forest succession was more important than earlier
researchers (e.g. Docters van Leeuwen 1936) rec-
ognized.
The point of detailing these patterns and
processes here is to illustrate that on real islands (in
the sea and some kilometres in area) succession is
complex and demonstrates hierarchical interde-
pendency across trophic levels. In order to
understand how compositional patterns develop in
the vegetation and flora, it is necessary to consider
the ecological attributes of the plant species, their
habitat relationships, and their hierarchical ecolog-
ical links with animals (Bush and Whittaker 1991).
Animals effect their dispersal, in cases their polli-
nation, and of course also act as seed predators. In
turn the vegetation provides the habitat and food
resources of the animals; no fruit supply means no
resident frugivores.Colonization and turnover—the dynamics of species lists
Given the basic elements of vegetation succession
and colonization patterns, broken down into plant
dispersal types, it is possible to interpret much of
the pattern in the rates of immigration and
turnover of both plants and animals of the
Krakatau islands. What follows here flows on from136 COMMUNITY ASSEMBLY AND DYNAMICS