widespread species tend to be better colonists, or
possibly a combination of the two.
Anthropogenic experiments in island assembly: evidence of competitive effects?
It is difficult to assess the importance of competi-
tion in the process of community assembly on
islands. One reason for this is the difficulty of
observing immigration events. Human interven-
tion in island systems has, however, inadvertently
created opportunities to test for island assembly
structure, and in cases provides evidence indicative
of competition as a structuring force.
One such example is a study by Badano et al.
(2005), who use the chequerboard index or C-score
to analyse ant community structure in islands of the
Cabra Corral dam, Argentina. They found that
newly created islands showed a random pattern of
species co-occurrence, but that on older (forest rem-
nant) islands species showed a significant tendency
to co-occur less often than expect by chance. This,
they argued, is a consequence of competitive sort-
ing on the older islands. Contrasting results were
obtained in another study of newly created (forest
remnant) islands from within a South American
reservoir, this time an analysis of forest-interior
birds nesting on islands within Lake Guri,
Venezuela by Feeley (2003). He tested several dif-
ferent assembly rule models, finding strong evi-
dence of nestedness (below) related to habitat
specificity, such that specialist species tend to be
absent from small, species-poor islands. This indi-
cates that the communities have been shaped by
structured losses of species rather than through the
effects of interspecific or interguild competition,
although there was some limited evidence of com-
petitive limitations in the degree of size similarity
permissible in co-occurring species.
The purposeful introduction of land birds to
remote oceanic islands provides another form of
inadvertent experiment on the role of competitive
effects in structuring insular avifaunas. Such intro-
ductions have been for various purposes, including
as game birds, as song birds, and to ‘beautify’ the
native avifauna. Very often the native avifauna has
been depleted concurrent with or before the intro-
ductions, with the exotic species to varying degrees
replacing the lost species (Sax et al. 2002). Although
the natural immigration rate of land birds to such
islands in the absence of human transformation of
the landscape can be assumed to be negligible, high
rates of introductions of species by people have
been documented in several archipelagosISLAND ASSEMBLY THEORY 125compensatory pattern) only in the agricultural
landscape
1 species (brown thornbill, Azanthizapusilla)
showed an isolation threshold in the agricultural
landscape and area thresholds in the other two
landscapes
None of the species was found to show a block
area/isolation pattern
The variability in area/isolation is illustrated
graphically for eastern yellow robin (Figure 1).
Even where the species showed a common type
of response between landscapes, the minimum
area thresholds were found to vary significantly
(Figure 2). These results show that species
incidence functions can vary substantially, even
within the same ecoclimatic region. As the
woodland remnants themselves were similar in
character across the study system, Watson et al.
(2005) concluded that the differences in incidencefunctions between the landscapes were being
determined by the nature of the matrix of non-
woodland habitats in which the woodland
fragments were embedded.
There are many ways in which matrix habitats
might impact on the incidence of woodland
birds within a patchy landscape, and it is not
possible to know which were operative in this
system. Nor is it possible to claim that the
patterns shown, e.g. by the eastern yellow robin,
would be replicated in another suite of urban,
peri-urban, and agricultural catchments
elsewhere within the range of the species. These
findings do, however, imply that although
species incidence functions may show some
emergent tendencies (e.g. as reported for body
size by Bierdermann 2003), they are not
consistent properties within the range of a
species.