the point we develop later, that the EMIB holds rel-
evance to far more restricted geographical circum-
stances than many of its advocates have envisaged
(Haila 1990).
4.3 Island species numbers and ISARs: what have we learnt?
Area and habitat diversity
Two facets of the relationship between area and
species number have commonly been recognized:
first, a direct area or area per seeffect, and second,
a habitat diversity effect. Although some authors
have argued to the contrary, MacArthur and
Wilson’s (1967) theory is not really an ‘area per se’
argument. They wrote: ‘Our ultimate theory of
species diversity may not mention area because it
seldom exerts a direct effect on a species’ pres-
ence. More often area allows a large enough
sample of habitats, which in turn control species
occurrence.’ (p. 8).
Theareaper seeffectrefers to the idea that given
a uniform habitat, very small amounts of space can
support only a limited number of individuals, and
thus (through the properties of the SAD), will not
be able to sustain all the species capable of occupy-
ing the habitat. As the area increases, more individ-
uals can accumulate, sufficient for breeding
populations of more species, and richness thus
increases. The range in area over which richness
will continue to rise will vary depending on the
ecological characteristics of the system. It may be
very small for lichens growing on rock outcrops, or
very large for top carnivores. Gibson’s (1986) data
for habitat islands of grassland plants within an
English woodland indicated that an island biogeo-
graphical effect of this sort, if present in the herb
flora, was detectable only in patches of less than
0.1 ha. The habitat diversity effectis that as area
accumulates, so typically does the variability of
habitat, such that species specializing in e.g. fresh
water, boggy, upland, or low pH habitats are added
as the size and complexity of the island increases.
Numerous studies have since been published
which quantify, via regression, values for z, and the
varying explanatory values of area, isolation, alti-
tude, island age, and various other measures of
habitat diversity. Comparative studies of different
taxonomic groups from the same islands typically
return different explanatory models. For example,
multiple regression analyses of Lesser Antillean
data selected habitat diversity and maximum
elevation for herptiles, isolation for bats, and
isolation, area, and habitat diversity for birds
(Morand 2000). Area is very often the most impor-
tant explanatory variable because (via area per se
and habitat diversity mechanisms) area is a prime
determinant of the resource base available to
biological communities. With larger islands or
reserves, it becomes increasingly difficult to sepa-
rate the area per seeffect from the habitat diversity
effect (but see Rafe et al. 1985; Martin et al. 1995).
Depending on the size of the species pool, we can
nonetheless appreciate that with increasing area of
a uniform habitat there may come a point where
there is sufficient area to accommodate breeding
populations of a large array of species, but where
the lack of particular microhabitats prevents all
species from colonizing, e.g. the Krakatau islands
lack mangrove habitats and thus mangrove species,
although propagules are frequently cast up on the
beaches.
Very often, habitat diversity may be expected to
accrue in fairly simple fashion with increasing
island area. However, as shown in many regression
studies, particular habitat properties can account
for significant amounts of variance in ISARs.
Moreover, some studies have found evidence of
thresholds whereby a degree of non-linearity is
introduced into the ISAR over particular ranges of
area. An early example is Whitehead and Jones’
(1969) study of plant data from Kapingamarangi
atoll (a Polynesian outlier in the west Pacific), in
which they argued for threshold effects as follows.
The smallest islands, up to 100 m^2 area, are effec-
tively all strandline habitats: their flora is therefore
restricted to the limited subset of strandline species.
A little larger, and a lens of fresh water can be
maintained, and the next set of plants can colonize,
those intolerant of salt water. A little larger still and
the island would be liable to human occupancy
and would be topped up with introduced species.
Such modifications to the form of ISARs indicate a
greater complexity than captured in the simplestISLAND SPECIES NUMBERS AND ISARS: WHAT HAVE WE LEARNT? 89