102 SPECIES NUMBERS GAMES: THE MACROECOLOGY OF ISLAND BIOTAS
publication). A number of analyses of these
phenomena have used formulations in which rates
are expressed as a function of the size of the biota on
an island during the course of the study, and/or as
a function of the size of the mainland pool (e.g.
Thorntonet al. 1990). Where Pis unknown, or
unknowable, then evaluation of the theory must fol-
low one of the other routes provided by MacArthur
and Wilson (1967). Particular formulations may be
optimal for different contexts (e.g. early periods of
build-up, or near-equilibrium condition) and
arguably should be given a different nomenclature
(see the useful discussion in Thornton et al. 1990).The rescue effect and the effect of island area on immigration rate
The EMIB promises a workable, testable model
because extinction rate is influenced by area (or its
correlates) and immigration rate by isolation. If,
however, it can be shown that area also influences
immigration, and isolation influences extinction,
then the model shown in Fig. 4.1 is in some jeop-
ardy (Sauer 1969).
It will be recalled that immigration rate is the
term given to the arrival of species not already pres-
ent on an island. It is logical to assume that a near-
shore island for which there is a high immigration
rate will also continue to receive additional immi-
grants of the species which arepresent. This might
be termed supplementary immigrationto distin-
guish it from immigration proper. An island popu-
lation declining towards extinction might be
rescued by an infusion of new immigrants of the
same species, thereby lowering extinction rates on
near-shore islands compared with distant islands.
This effect of supplementary immigration was
termed the rescue effectby Brown and Kodric-
Brown (1977), who censused the number of indi-
viduals and species of arthropods (mostly insects
and spiders) on individual thistles growing in
desert shrubland in south-east Arizona. Their find-
ings conformed quite well with the predictions of
EMIB, with the exception that turnover rates were
higher on isolated plants than on those in close
proximity to others. This was argued to be because
of increased frequencies of immigration events forspecies present on the thistles at time 1, either pre-
venting extinctions or perhaps effecting recoloniza-
tion before a subsequent survey at time 2. Either
way, each census was likely to record the species as
present on the less-isolated thistles. It has been
pointed out that this study was based on rather a
notional form of island (individual thistles) and
was not based on breeding populations. Since the
phrase was coined, the rescue effect has been
invoked by a number of authors in explanation of
their data (e.g. Hanski 1986; Adler and Wilson
1989); often, as in the original, this seems to be
based on its good sense rather than on any direct
proof (but see Lomolino 1984a, 1986).
Similarly, immigration rate may be affected by
area. A large island presents a bigger target for ran-
dom dispersers. A simple study of sea-dispersed
propagules on 28 reef islands in Australia showed
that the fraction of the dispersing propagule pool
intercepted by each island is proportional to
beach length (Buckley and Knedlhans 1986).
Larger islands also present a greater range of
habitats and attractions to purposeful dispersers,
such as some birds, elephants swimming across a
stretch of sea, or mammals crossing ice-covered
rivers in winter (Johnson 1980; Lomolino 1990).
Williamson (1981) points out that MacArthur
and Wilson recognized the possible influence of
area on immigration in their 1963 paper but
discreetly sidestepped this issue in their 1967
monograph.
At least one study has reported both the above
‘problems’. Toft and Schoener (1983) undertook a
two-year study of 100 very small central Bahamian
islands, in which they recorded numbers of indi-
viduals and species of diurnal orb-weaving spi-
ders. They found that the extinction rate was
related positively to species number and to dis-
tance, and was negatively related to area.
Immigration rate was positively related to area and
negatively related to species number and distance.
In consequence, the turnover rate showed no
strong relationship to any variable, thus the net
result of these interactions is that the effects of area
and distance on turnover tend towards the inde-
terminate. Put simply, the EMIB does not work for
this particular insular system.