exchange species moderately frequently, species–
area relationships may be detectable at the
archipelago level. Moreover, taxa capable of rapid
rates of speciation may perhaps so rapidly saturate
even geologically young islands that a clear positive
correlation could result between species richness
and island area. This appears to apply to the highly
volant Hawaiian Plagithmysusbeetles, but not to the
flightless Hawaiian Rhyncogonusweevils, which
instead show a pattern of increasing species number
with increasing island age, and a decline with
island area (the largest Hawaiian island being the
youngest) (Paulay 1994). Juvig and Austring (1979)
have undertaken an analysis of species–area
relationships within the Hawaiian honeycreeper-
finches, which produced results they claimed
were consistent with equilibrium interpretations.
However, their analysis was undertaken before the
extent of the anthropogenic species extinctions
became clear. The full distribution of this family
before Polynesian colonization is unknown, casting
doubt upon interpretations of such species–area
calculations (but see the attempt by James (1995) to
correct for extinctions).
In addition, the history of introductions to remote
islands suggests that, at least for particular taxa (e.g.
plants), the total number of species on such islands
can rise significantly, suggesting that the islands in
question are effectively unsaturated and therefore
non-equilibrial (e.g. Groombridge 1992, p. 150). This
is perhaps a weak line of argument because carrying
capacities have undoubtedly been changed by
human alterations of habitats and this, in tandem
with the anthropogenic increase in immigration
rate, will set a new equilibrium point for the island
in question (Sax et al. 2002). However, this does not
mean that a new dynamic equilibrium will be
achieved. Gardner (1986) has shown that the lizards
of the Seychelles do not produce a significant ISAR,
and that there are unfilled niches on some islands.
The patterns of distribution of endemic and more
widespread species were found to be best explained
by reference to historical events, such as periods
of lower sea level in the Pleistocene, and to non-
equilibrium hypotheses (see also Lawlor 1986).
The applicability of equilibrium ideas to remote
archipelagos may ultimately depend on the
effective response rate of particular taxa and the
scale at which they interact with the island environ-
ment. It is not sufficient to consider demonstration
of an ISAR as being proof of dynamic equilibrium in
those island groups in which evolutionary phe-
nomena are dominant over ecological phenomena
(cf. Williamson 1981, 1988; Lawlor 1986).Scale effects and the shape of species–area relationships
Scale
As a number of authors have pointed out, it is
important to take account of scale factors in
analysing species–area relationships (e.g. Martin
1981; Lomolino 1986; Rosenzweig 1995;
Sfenthourakis 1996). The details as to which vari-
ables best explain variations in Sin a given limited
data set hold fascination principally in their specific
context. They are difficult to generalize. This is in
large part because each group of islands (real or
habitat) has its own unique spatial configuration
and range of environmental conditions. The rela-
tive significance of such key factors as area and
isolation will depend in part on how great the
range (how many orders of magnitude) the sample
islands encompass for each variable. Studies of
plant diversity where maximum distance is of the
order of 500 m, but where elevational range
encompasses distinct habitats, vertically arranged,
and where area varies by several orders of
magnitude, are unlikely to show a significant
isolation effect, but are likely to show both area and
habitat effects. This is a description of Buckley’s
(1982) Perth islets data set, in which the purpose of
the study required that isolation be effectively
controlled. However, it is not just a matter of how
great the range is but what the upper and lower
limits of the variables are relative to the ecology
of the taxon under consideration. This can be
appreciated in terms of habitat—the addition of a
patch of mangrove habitat will not do the same for
plant diversity on a tropical island as the addition
of an area of montane cloud-forest.
In terms of isolation, the effective mobility
of different taxa (e.g. birds versus terrestrial
mammals) and of different ecological guildsISLAND SPECIES NUMBERS AND ISARS: WHAT HAVE WE LEARNT? 93