moreover, those tests that have been conducted
have proved equivocal. At the outset, MacArthur
and Wilson (1963, 1967) produced a way of quanti-
fying the effect for Pacific birds. First, they showed
that there was a fairly tight, straight-line log–log
ISAR for the closely grouped Sunda Islands, but
that an equivalent analysis for 26 islands and arch-
ipelagos from Melanesia, Micronesia, and
Polynesia had both a much greater scatter and a
tendency away from a straight-line relationship.
Secondly, they reasoned that a line drawn through
the points for near islands could be taken to repre-
sent a ‘saturation curve’, and the degree of impov-
erishment of islands below the line could then be
estimated as a function of area by comparison with
the saturation curve. In general, the degree of
impoverishment, when corrected for area, was
found to increase from islands ‘near’ to New
Guinea (less than 805 km) to intermediate, to far
islands (greater than 3200 km).
Williamson (1988) suggests that this demonstra-
tion is invalid because the example used archipela-
gos rather than single islands, and included
archipelagos of different climate or deriving their
avifaunas from different sources. He reports that,
in fact, the line tends to be flatter the more distant
the archipelago, such that (with some caveats)
slopes of 0.28, 0.22, 0.18, 0.09, and 0.05 can be
derived for birds in the progressively more isolated
archipelagos of the East Indies, New Guinea, New
Britain, Solomons, and New Hebrides (Vanuatu)
archipelagos, respectively. It appears that the
assumptions made in analysing species–area
curves in relation to isolation, e.g. whether to lump
separate islands into archipelagos, whether to cal-
culate isolation within or between archipelagos,
whether to include islands of rather different cli-
mates, etc., are critical to the outcome of the analy-
sis (compare Itow 1988; Williamson 1988;
Rosenzweig 1995). In any case, the discovery since
the 1960s of numerous ‘new’ bird species from sub-
fossil remains on many of the islands of the Pacific
(Chapter 3; Steadman 1997a) renders this particular
part of the MacArthur and Wilson analysis prob-
lematic in hindsight, as several points on their
graph are below their ‘natural’ level. Again, we
re-evaluate the distance effect below.
Turnover, the core model (EMIB), and its immediate derivatives
It was these observations—of ISARs, SADs and dis-
tance—which were combined in the dynamic equi-
librium theory, through the mechanism of the
turnover of specieson islands. The theory postu-
lates that there are two ways in which islands gain
species, by immigration or by evolution of new
forms, and that these means of increasing species
number will be balanced in the equilibrium condi-
tion by processes leading to the local loss of species
from the island in question. Before immersing
ourselves in the detail of the model, it is important
to take note of the working definitions of the key
terms involved (Box 4.3).
It is a reasonable generalization that immigration
rate diminishes as a simple function of increasing
distance, hence lowering the equilibrium point for
more remote islands. This is only partially coun-
tered by increased speciation rates on remote
islands, because of the relatively slow pace of phy-
logenesis. For the moment we will set aside the role
ofin situspeciation and concentrate solely on immi-
gration. The local loss of a species population may
be accomplished either by out-migration or by the
death of the last representatives on an island, in
either case leading to the local extinction (extirpa-
tion) of the species. The greater resource base of
larger islands should mean that extinction rates for
a given species richness are lower for larger islands
than for smaller. Thus the EMIB postulates that the
number of species found on an island represents a
dynamic balance between immigration (I) and
extinction (E), with immigration varying with
distance from source pool and extinction with
island area. This was presented in simple, accessi-
ble, diagrammatic form (Fig. 4.1)—possibly one of
the keys to the broad uptake of the theory (Brown
and Lomolino 1989).
MacArthur and Wilson (1967) also considered
the effects of slightly more complicated configura-
tions, for example where chains of islands
(‘stepping stones’) are found strung out from a
mainland, producing alterations in the expected
immigration functions and thus in equilibrium
values. However, the basic ideas may be most84 SPECIES NUMBERS GAMES: THE MACROECOLOGY OF ISLAND BIOTAS