radiation is distinguished by niche differentiation
amongst the members of the lineage. As we have
seen, selection and adaptation does not necessarily
lead to radiation. Similarly, as we will discuss later,
radiation within monophyletic lineages does not
have to be adaptive. The term non-adaptive radia-
tioncan be applied to cases where allopatric specia-
tion via founder effect and genetic drift mechanisms
has been dominant in lineage development,
although it is best to think in terms of a gradient
between entirely adaptive and non-adaptive end
points, with most systems occupying a space some-
where in the middle of such a conceptual gradient.
As with the taxon cycle, adaptive radiations are the
result of a combination of driving forces and mech-
anisms; and they are drawn from a common tool
kit. Adaptive radiation scenarios differ from the
taxon cycle model in being concerned solely with
the expansion phase of a monophyletic lineage,
whereas taxon cycle theories attempt to describe
the expansion and subsequent contraction of par-
ticular taxa within an entire evolutionary assem-
blage (e.g. ponerine ants, or land birds).
The availability of vacant niche space is a key fea-
ture in adaptive radiations, as it allows a form of
ecological release, allowing the diversification that
sometimes leads to speciation within a lineage. As
suggested by its name, Metrosideros polymorphais
very diverse and multiform. It occupies habitats
ranging from bare lowlands to high bogs, and
occurs as a small shrub on young lava flows and as
a good sized tree in the canopy of mature forest. It
is the principal tree of the Hawaiian wet forests
and, as an aside, at least some individuals can be
found in flower at any time of the year, a pattern
supported by enough of the nectar-producing flora
to have allowed the evolution of the nectar-feeding
lineages among the honeycreepers. Despite its
radiation of forms, all members of this complex
were originally allocated to M. polymorpha, i.e.
Meterosideroshas radiated a lot but has speciated to
a lesser degree. Some highly distinct populations
have now been recognized as segregate species, but
active hybrid swarms also exist (Carlquist 1995),
indicating continuing gene flow.
We now consider in a little more detail what we
mean by the term adaptive. First, according to the
above quote from Begon et al. (1986), adjustment to
the abiotic and biotic environment works by natu-
ral selection. This is effected by the successful forms
contributing more progeny to the following gener-
ations, a form of filtering that tells us about the past
and which may be of quite varying fit to con-
temporary circumstances (e.g. of altered climate,
depleted native biotas, altered habitats and numer-
ous exotic species). Gittenberger (1991) has argued
that there should be no automatic labelling of
radiations as adaptive, but that a degree of special-
ization of species in to different niches should be
involved before the term is used.
Once these points are registered, there is little to
the core theory of adaptive radiation that we have
not already covered. Adaptive radiation invokes
not so much a particular process of evolutionary
change as the emergent pattern of the most spec-
tacular cases, the crowning glories of island evolu-
tion. For this very reason the best cases are liable to
involve a wide range of the evolutionary processes
previously introduced.
The data in Table 9.2 provide recent estimates for
the degree of endemism and an idea of the degree
of radiation involved in the biota of Hawaii. These
figures are not cast in stone, as they depend on the
taxonomic resolution and assumptions involved in
their calculation. For example, Paulay (1994) cites
an estimate that there may be as many as 10 000
Hawaiian insect species, and that they may have
evolved from 350–400 successful colonists. Not-
withstanding such uncertainties, Hawaii clearly
provides spectacular examples of radiations in taxa
as diverse as flowering plants, insects, molluscs,
and birds. It is estimated that the 980–1000 flower-
ing plant species arose from between just 270 and
280 original colonists (Fosberg 1948, Wagner and
Funk 1995), the pattern of radiation being indicated
by the following statistics: there are 88 families and
211 genera, the 16 largest genera account for nearly
50% of the native species, about 91% of which are
endemic (Sohmer and Gustafson 1993).
The biogeographical circumstances in which
radiations take place are reasonably distinctive.
Radiations are especially prevalent on large, high,
and remote islands lying close to the edge of a
group’s dispersal range (Paulay 1994). MacArthur218 EMERGENT MODELS OF ISLAND EVOLUTION