fine scale of analysis, the ‘island’ assumption is at
least a rational starting point for thinking about the
SLOSS question. However, at coarse scales of analy-
sis, the relevance of island theory is less clear, and so
conservation biogeographers have developed other
approaches, designed to make use of the improved
availability of distributional data in digital form (at
least for some taxa). For instance, computer algo-
rithms have been developed to achieve goals such as
finding the 10% of sites/grid cells that maximize
representation of rare or endemic species (e.g. see
Araújoet al. 2005a; O’Deaet al. 2006; and papers cited
therein). Such analyses provide a far more direct
approach to conservation planning than the original
SLOSS guidelines and effectively supercede them at
coarse scales of analysis.
To sum up, on theoretical grounds, the answer to
the SLOSS debate is equivocal (e.g. Boecklen and
Gotelli 1984; Shafer 1990; Worthen 1996). The answer
depends, as we have argued here, on the ecology of
the species (or species assemblage), on system grain
and on extent. It also depends on a range of other
system properties, which we will now explore.
10.6 Physical changes and the hyperdynamism of fragment systems
Large environmental changes can be involved in
ecosystem fragmentation, particularly where
forests are fragmented (Lovejoy et al. 1986;
Saunderset al. 1991). The immediate and subse-
quent differences in fluxes of radiation, wind,
water, and nutrients across the landscape can have
significant consequences for the ecology of the
remnants, and have been summed up by Laurance
(2002, p. 595) in the idea of hyperdynamism, which
he defines as ‘an increase in the frequency and/or
amplitude of population, community, and land-
scape dynamics in fragmented habitats’. The
following examples are illustrative.
●Radiation fluxes. In south-western Australia,
elevated temperatures in fragmented landscapes
reduced the foraging time available to adult
Carnaby’s cockatoos (Calyptorhyncus funereus
latirostrus) and contributed to their local extinction.
●Wind. When air flows from one vegetation type
to another it is influenced by the change in height
and roughness. At the edge of a newly fragmented
woodland patch, increased desiccation, wind-
damage, and tree-throw can occur (e.g. Kapos
1989). Increased wind turbulence can affect the
breeding success of birds by creating difficulties in
landing due to wind shear and vigorous canopy
movement. Wind-throw of dominant trees can
result in changes in the vegetation structure, and
allow recruitment of earlier successional species.
●Water and nutrient fluxes. Removal of native
vegetation changes the rates of interception and
evapotranspiration, and hence changes soil mois-
ture levels (Kapos 1989). In parts of the wheat belt
of Western Australia, the new agricultural systems
cause rises in the water table, which can bring
stored salts to the surface, and this secondary salin-
ity has caused problems both to agriculture and the
remnant patches. In the fenlands of eastern
England, drainage for agriculture has led to peat
shrinkage and a drop of 4 m in land level in
130 years. Remnant areas of ‘natural’ wetland now
require pumping systems to maintain adequate
water levels (see also Runhaar et al. 1996).
●Fire regimes. Modified landscapes can be a major
source of surface fires, for example from burning of
adjoining pastures. Penetration of such fires into
fragment interiors can increase plant mortality, dis-
turb the fragment boundaries and in time cause an
‘implosion’ of forest fragments (Laurance 2002).
The physical effects of edges thus have important
knock-on effects for the biota, especially soon after
fragmentation (Table 10.2). In time, the system
adjusts to the new physical conditions, and the
woodland edge fills and becomes more stable. Yet,
just as there is a wide range of disturbance regimes
across real islands, so must there be for continental
habitats, and as we create new habitat islands, we
inevitably alter disturbance regimes (Kapos 1989;
Turner 1996; Ross et al. 2002). Furthermore, land-
use of the matrix in which the habitat islands are
embedded typically continues to change (Laurance
2002). It might therefore be anticipated that many
habitat islands will be characteristic of the dynamic
non-equilibrial position identified in Fig. 6.1 (cf.
Hobbs and Huenneke 1992). Indeed, it has been
suggested that smaller habitat islands may be sub-
ject to greater disturbance impacts than larger268 ISLAND THEORY AND CONSERVATION