cases where the early stages of succession on a
series of islands allow researchers to examine nest-
edness during the assembly process, and thus
establish more directly the predictability of the
colonization process. A good example is Kadmon’s
(1995) study of seven islands created by the filling
of the Clarks Hill Reservoir, Georgia, USA. The
islands were logged and cleared of woody plants
before their separation from the mainland. Thirty-
four years later, species number was found to be
inversely related to island isolation. The woody
plant floras were found to be nested when ranked
by richness, or isolation, but not by area. These
results point to the significance of differential colo-
nization in structuring the data set. Only a small
proportion of the species included in the analysis
contributed to the observed nestedness, principally
being those that lacked adaptations for long-range
dispersal. Wind-dispersed species showed no
evidence for nestedness. This study thus pointed to
the significance of plant dispersal attributes in
structuring the rebuilding of insular assemblages, a
feature also of the recolonization data from
Krakatau, as we shall shortly see.
●Nestedness can also result from differential area
requirements across species, again as may berevealed in the individual species’ area-incidence
functions of the component species. Where nested-
ness within a former land-bridge island data set is
strongly area-related but is unrelated or only
weakly related to isolation within a data set, this is
taken to indicate a strong extinction signal in the
structuring of assemblages. This rationale was
invoked in classic early studies of nestedness of
mammals on land-bridge islands (e.g. Patterson
and Atmar 1986). This is on the grounds that before
they became islands, the land-bridge islands were
part of the mainland and can be assumed to have
had a full complement of the mainland pool, with
extinction then the dominant process in the sorting
of the island biotas. Similarly, extinction-led sort-
ing of remnant mountain-top ‘island’ has been
invoked for the nestedness of mammal assem-
blages in the south-western USA (Patterson and
Atmar 1986).
Given that nestedness appears to be so prevalent, it
is worth considering factors that may constrain it:
●First, consider the implications of supertramp
distributions: species that tend to occur preferen-
tially on species poor islands, and to disappear in
species rich assemblages. Such patterns, along with
certain other of Diamond’s assembly rule patterns
(above) will form exceptions to any general system-
wide tendency to nestedness.
●Second, nestedness must perforce be diminished
as the geographic extent of the study system is
expanded to incorporate different species pools, i.e.
biogeographical turnover in species composition.
●Third, Wright et al. (1998) suggest that where the
sample system involves mostly very small patches,
the degree of nestedness may be diminished by
high habitat heterogeneity between patches (i.e. the
opposite of habitat nestedness).
It is now well established that nestedness is a
common but not universal feature of many insular
biotas. In an applied context, it has also been shown
that within a given landscape, different taxa can
respond in vastly different ways to fragmentation.
For instance, Fischer and Lindenmayer (2005),
examining data from fragmented forest habitats in
south-eastern Australia, found that whereas birds128 COMMUNITY ASSEMBLY AND DYNAMICS
IslandsSpeciesDispersal zonesB C DAMainlandFigure 5.4Hypothetical scenario for the production of nested
subsets via differential colonization. From Patterson and Atmar
(1986, Fig. 4), who contended, nonetheless, that extinction is a more
potent structural determinant in the majority of insular systems.