bats, act as important seed vectors for the introduc-
tion and/or spread of a particular subset of
Krakatau plant species (Shilton 1999; Shilton
et al. 1999).
The large fruit-pigeon, Ducula bicolor, is another
important species of frugivore, which for island
ecological purposes is considered as a Krakatau
resident (Thornton et al. 1993), yet which exhibits a
roaming behavioural pattern, in which large flocks
move between offshore islands in the region
(Dammerman 1948). The two species of Krakatau
Ducula(the other is D. aena) are believed to have
played crucial roles in introducing animal-
dispersed (i.e. zoochorous) plants—particularly
those with larger seeds—to Krakatau (Whittaker
and Jones 1994b). The particular identities and
timing of the trees introduced by such means may
have important ramifications for the species of
animals which may subsequently be supported
within the islands.
All these species of frugivores have been highly
significant to the structure inherent in the assembly
process for Krakatau, the interior forests being pre-
dominantly composed of bird- or bat-dispersed
plants. Arguably, the behaviours of these species
exhibit characteristics of connectedness to other
locations which are more akin to metapopulation
models (Chapter 10) than to the notions of resi-
dency demanded by the EMIB. In short, for some
animal taxa and some guilds, there are potentially
complex hierarchical links between their turnover
patterns and floral development and turnover.
Other forms of hierarchical links may also be
found, such as those linking predator–prey groups
in an area. Although such observations are of
explanatory value, it is, as we saw with the debates
about island assembly rules, a considerable chal-
lenge to incorporate such processes into general
predictive models.
Immigration and extinction curves are theorized
in the EMIB to have smooth forms, the former
falling and the latter rising to a point of intersec-
tion: the dynamic equilibrium. The patterns of
arrival and disappearance from the Krakatau lists
do not correspond with this expectation (Chapter 5),
in part because of hierarchical, successional
features evident in the data, which are simply
absent from the model. The big early kinks in the
rates relate to the key switch from open to closed
habitats. It is conceivable, however, that this switch
marks the end of significant autogenically derived
(i.e. biotically driven) switches in trends. In which
case, once beyond the early phases of succession,
the EMIB may become more realistic, i.e. at some
stage, an island must fill up with species and some
form of equilibrium be established. However, given
the complexity of succcessional processes in tropi-
cal lowland forests, Bush and Whittaker (1993)
question whether it is reasonable to suppose that
population processes can bring the system to equi-
librium before the next major disturbance moves
the system away from that condition again.
Forest succession is a slow process: the lifespan
of individual canopy trees can exceed 300 years,
and even early successional species may live for
decades. If a true biotic equilibrium in the flora
were to be established beyond doubt, it must be
after a period of several generations of biotically
mediated interactions, i.e. hundreds of years.
However, such conditions are unlikely to be reached
in very small patches because of varying forms of
environmental variability and disturbance. Events
such as hurricanes, volcanic eruptions, flooding,
landslides, and other high-magnitude phenomena
occur within continental and island landscapes,
and typically have periodicities of less than several
hundred years (Chapter 2).
For this argument to hold force for animal
groups/taxa, there must be a strong pattern of
dependency of animal groups on plants for habitat
and food resources, such that their patterns of colo-
nization, carrying capacity and turnover are tied to
the dynamics of the plant communities. In turn,
animals can have key roles in shaping plant succes-
sion, through roles as seed dispersers and pollina-
tors (Bush and Whittaker, 1991, Elmqvist et al.
1994). If the degree of dependency is weak, it may be
that merely the recovery of forest biomass and pro-
ductivity, which is likely to be more swiftly accom-
plished (Whittaker et al. 1998), may be sufficient to
allow equilibration of particular animal groups. In
practice, some 120 years on from the initiation of the
Krakatau experiment, it remains difficult to provide
a definitive answer to this question.SCALE AND THE DYNAMICS OF ISLAND BIOTAS 149