exploit the opportunities available, through combi-
nations of adapative, non-adaptive and archipel-
agic radiation.
Yet, still some taxa appear to radiate more than
others, with some lineages persisting for millions
of years on islands without radiating. This sug-
gests that some taxa have a greater inherent degree
of genetic plasticity than others (e.g. Werner and
Sherry 1987) or that they possess traits related to
their breeding systems that favour rapid evolu-
tionary change on islands (see discussion in
Chapter 7).
9.7 Biogeographical hierarchies and island evolutionary models
Williamson (1981) has suggested redefining island
types as oceanic where evolution is ‘faster’ than
immigration, and as continental if immigration is
faster than evolution. He also offered the observa-
tion that it is useful for any particular island context
to distinguish between those groups for which the
island is continental, readily reached by dispersal,
and those for which the island is oceanic, leading to
little dispersal. For instance, the Azores could be
regarded as oceanic for beetles, intermediate for
birds, and almost continental for ferns. The gradual
diminution in the range and numbers of taxa that
have been able to cross ever-increasing breadths of
ocean means that, in general, the more remote the
island the more disharmonic the assemblage,
and the greater the amount of vacant niche space.
Thus, the greatest opportunities for adaptive radia-
tion are towards the edge of the range of particular
taxa, i.e. in the radiation zoneof MacArthur and
Wilson (1967), reaching their synergistic peak in the
most remote archipelagos. Thus, in the absence of
ants (amongst other things), there have been great
radiations of carabid beetles and spiders in Hawaii
and south-east Polynesia (Paulay 1994).
Although diversification in certain lineages has
undoubtedly been enabled by the disharmony of
very remote islands, the absence of particular inter-
acting organisms (e.g. pollinators for particular
plant species) may have restricted the colonization,
spread, and evolution of other lineages. That is to
say, hierarchical relationships within the island
biota, partly predictable and partly the chance
effects of landfall (e.g. the highly improbable
arrival of the first drosophilids on Hawaii at a par-
ticular point in time), must have a strong influence
on the patterns that have unfolded.
Once on an archipelago, the relative importance
of interisland and intraisland speciation may vary
(Paulay 1994). Only a handful of isolated archipela-
gos show large endemic radiations of birds or
plants. The occurrence of a number of islands in
proximity to one another appears to be particularly
important to birds, suggesting that interisland
effects must be important (cf. Grant and Grant
1996 b). Intraisland speciation is restricted to islands
large enough to allow effective segregation of pop-
ulations within the island. This is taxon dependent,
so that for land snails and flightless insects an
island of a few square kilometres is sufficient.
As several authors emphasize, different island
groups have their own special circumstances and
histories, and few theories can span them all. A
corollary of this is that each theory may have its
own constituency. The solution, as elsewhere, must
be to set up multiple working hypotheses and seek
evidence to distinguish between them. Yet, freque-
ntly, a particular lineage reflects not the operation
of a single process, but of several. Therefore it may
not be possible to explain the overall biogeography
of an island lineage by means of a single model,
such as character displacement, a double invasion,
or the taxon cycle. Moreover, alongside the evolu-
tionary hypotheses we should also be on the look
out for anthropogenic effects. The historical impact
of humans in eliminating species from many
islands, and thus selectively pruning phylogenies,
turning multiple-island endemics into single-island
endemics, and altering distributions of extant
species (e.g. Steadman et al. 1999, 2002), requires
careful consideration in evaluating island evolu-
tionary models and phylogenies (e.g. compare
Pregill and Olson 1981; Pregill 1986 with Ricklefs
and Bermingham 2002).
Having established the relevance of factors such
as the regional biogeographic setting, environmen-
tal change, dispersal differences between taxa,
island area, island habitat diversity, and island iso-
lation and configuration, we should consider howBIOGEOGRAPHICAL HIERARCHIES AND ISLAND EVOLUTIONARY MODELS 245