0198566123.pdf

1990). The position adopted in this book, however,
is that the EMIB has a place in a larger framework
of island ecological theory. It is not that the theory
is either true or false, but that the effects it models
may be either strong or weak. If the effects are very
weak for the habitat island system being consid-
ered, as often is the case, then the answer to the
SLOSS question will not be supplied by the equili-
brium theory.
The next obvious criticism of the attempts to
solve SLOSS using the EMIB is that the theory is
insufficiently precise on the amount of composi-
tional overlap that will occur across a series of
isolates of varying area. Deployment of the EMIB
rationale often comes down simply to the use of
species–area equations. Such an approach may give
a rough idea of numbers of species on habitat
islands, but not which habitats contribute most to
richness, nor which species are most likely to be lost
from the remnants (Saunders et al. 1991; Simberloff
1992; Worthen 1996). At its simplest, a strongly non-
nested series of small reserves may be anticipated
to hold more species than a single large reserve, but
where the system is perfectly nested, the largest
reserve will hold most species (discussed more
fully below).
Interpretations of species–area relations based on
equilibrium theory have it that the regression line
provides the equilibrium number of species. The

statisticalinterpretation of the scatter is simply that

the regression line represents the average number

of species for a given area and a point above it

represents a positive error (Boecklen and Gotelli

1984; Boecklen 1986). If a newly created habitat

island lies above the line it doesn’t necessarily fol-

low that it is supersaturated and destined to lose

species. It could just be that habitat quality or het-

erogeneity allows it to maintain more than the

average number of species. In his analyses of a

large US bird census data set, Boecklen (1986)

found that habitat heterogeneity is a significant

predictor of species number even after area has

been factored out. The significance of habitat effects

in the data set points to the possibility that in

particular cases several small reserves can incorpo-

rate a wider (or better) array of habitat types and

thus support more species than the single large

option.

The discussion of SLOSS issues that follows will

mostly be limited to biological and physical param-

eters. However, it should be recognized that there

are often overriding practical considerations. For

instance, the costs involved in establishing and

maintaining conservation and game parks in devel-

oping countries depend on questions such as the

following (Ayers et al. 1991). Is perimeter fencing

required? Is it necessary to patrol to prevent poaching

or encroachment into the reserve? How many

THE ‘SINGLE LARGE OF SEVERAL SMALL’ (SLOSS) DEBATE 265

D

E

F

A

B

C

Better Worse Better Worse

Figure 10.6The suggested geometric principles for the design of nature reserves which were supposedly derived from island biogeographic
studies, and which were at the centre of the so-called SLOSS, or Single Large Or Several Small debate (redrawn after Diamond 1975b). These
‘principles’ have been challenged on both theoretical and practical grounds (see text).