Thus, bottlenecks can variously diminish het-
erozygosity (a general indicator of fitness), or in
contrast can serve to ‘shake things up’, generating a
sort of mini-genetic revolution (sensuCarson and
Templeton 1984), contributing to rapid phases of
evolutionary change in island populations.
Secondly, although the number of studies is grow-
ing, providing some support for the operation of
these mechanisms, their relative contribution to
island evolutionary change remains uncertain
(Grant 2002). Thirdly, notwithstanding recent
developments in molecular biology and genetics, it
is hard to be sure how well the specific phenotypic
properties upon which the success of a lineage
depends are measured by tools such as allozyme
electrophoresis (see papers in Clarke and Grant
1996). It is surely notable that in comparison to
related continental lineages, island endemic forms
often display far greater morphological differences
than are reflected in estimates of average heterozy-
gosity (e.g. Crawford et al. 1998), which implies that
something about these island environments
encourages rapid lineage diversification through
throwing specific switches in the genetic make-up.
That ‘something’ likely reflects the general impov-
erishment, disharmony, and environmental/biotic
novelty of remote islands. While founder events
and other largely stochastic bottlenecks are part of
the story, the emergent patterns encourage evolu-
tionary scientists to invoke significant roles for
deterministic causes alongside or in preference to
stochastic founding effect processes, as will be dis-
cussed later.
7.2 After the founding event: ecological responses to empty niche space
Do islands possess vacant niche space? The rela-
tionship between isolation and impoverishment
may not necessarily take a simple linear or log–
linear form (Chapter 4), but it is accepted that
oceanic islands possess fewer species for their area
than equivalent areas of continent. Yet, if this repre-
sents an equilibrium condition (whether dynamic
or static), this could be interpreted as indicating
that the islands involved are actually ‘full’ and that
the ‘missing’ ecological niches are lacking because
there is no space for them. Although this argument
might apply to older oceanic islands, surely there
must be plenty of vacant niche space in the species
poor and disharmonic early stages of the lifespan of
an island? This is generally assumed to be the case,
but the concept of empty niche space turns out to
be another slippery one to pin down and quantify.
This is partly because the term ecological nichehas
itself been variously defined. When first introduced
it reflected the figurative usage of ‘a place in the
community’, but developed into the description of
the types of opportunities that occur for species to
use resources and avoid predators (Schoener 1989).
Hutchinson’s (1957) usage was of a different form:
then-dimensional hypervolume of environmental
axes within which a species occurs. A plant or ani-
mal has a certain range of an environmental
resource, such as temperature, at which growth is
optimal. Outside this optimal range it may survive
but not thrive, until a point is reached, the limits of
its species tolerance, beyond which the species can-
not persist. Each resource gradient forms a part or
dimension of the hypervolume.
In the island context, the term empty nicherefers
to the lack of representation of a particular main-
land ecological guild or niche, providing evolution-
ary opportunities for colonizing taxa to exploit the
vacancy: the idea being that a fundamental niche
exists and it is just a matter of filling it. As the
assembly rules debate shows, it can be quite diffi-
cult to determine if such a fixed form of architec-
tural blueprint exists and so while the concept of
the empty niche is of intuitive value, it can prove
extremely difficult to apply.
The literature discussed in the following sections
mostly refers to a form of niche theory that derives
from the resource gradient idea. Typically, it does
not encompass a multidimensional characterization
of the niche, but is limited to a single or very lim-
ited set of characters, such as mobility, body size, or
jaw size, each of which provides a surrogate meas-
ure for one dimension of the n-dimensional hyper-
volume of the organism. The ambiguities of the
niche concept are avoided in many such studies by
describing the distributions of species populations
along resource spectra. As Schoener (1989) points172 ARRIVAL AND CHANGE