scenarios by such means. Studies of non-volant
mammals on isolated mountain tops in western
North America by J. H. Brown and colleagues pro-
vide an illustration (e.g. Brown 1971; Brown and
Gibson 1983). The mountains of the Great Basin
have been isolated from comparable habitat since
the onset of the Holocene (c.8000–10 000 years).
Brown assumed that immigration across the inter-
vening desert habitats did not occur. It follows that
the system dynamics must be dominated by extinc-
tion events, i.e. they are ‘relaxing’ towards lower
diversity, ultimately to zero species. Small islands
would have had higher extinction rates and have
already lost most of their species, whereas large
islands still retain most of theirs (Brown 1971). This
explained the much steeper zvalues found for the
species–area curve for mammals than for birds. For
birds, the isolation involved was not sufficient to
prevent saturation of the habitats. Thus, the mam-
mals provided a non-equilibrium pattern and the
birds an equilibrium pattern, although not quite to
the EMIB model, as it appeared that on the rare
occasion that a species went extinct, it tended to be
replaced not by a random draw from the mainland
pool, but by a new population of the same species
(Brown and Gibson 1983).
It is possible to test such models by independent
lines of evidence, if for example, subfossil remains
can be found of species once found on the moun-
tains but which have become extinct since isolation.
Some such evidence was found in the Great Basin.
The approach has been applied to a number of
other North American mountain data sets, for
instance by Lomolino et al. (1989), again showing
how palaeoecological data can be combined with
reasoning based on species–area relationships to
derive plausible models of the dynamics of the
insular systems. In this case, however, the montane
isolates in the American South-West (Arizona, New
Mexico, and southern Utah and Colorado) were
argued to be influenced by post-Pleistocene immi-
grations as well as extinctions.
Connor and McCoy (1979) caution against such
procedures, arguing that the slope and intercept
parameters of ISARs should be viewed simply
as fitted constants, devoid of specific biological
meaning, and that it is important to examine the
assumptions involved in such studies critically. For
example, Grayson and Livingston (1993) observed
one of the mountain species, the Nuttall’s cottontail
(Sylvilagus nuttallii) within the desert ‘sea’, suggest-
ing that immigration across these desert oceans
may occasionally be possible. Lomolino and Davis
(1997) provide a re-analysis of data for both the
Great Basin and the American South-west archipel-
agos. Their approach is predominantly statistical.
They conclude that although the most extreme
desert oceans (southern Great Basin) represent an
effective barrier to dispersal, in less extreme cases
(the American South-West) they represent immi-
gration filters, allowing intermountain movements
of at least some forest mammals on ecological
timescales. The scenarios constructed are thus more
complex than the original, invoking both equilib-
rium and non-equilibrium systems and varying
contributions for immigration and extinction.
Although plausible, these deductions ultimately
demand verification by independent lines of
evidence.Species–area relationships in remote archipelagos
It was recognized in the EMIB that species number
could increase through two means, immigration
(I) and evolution (V). In ecological time, only
immigration occurs at measurable rates, but in
remote islands the evolution of new forms may be
more rapid than the immigration rate from main-
lands; the inclusion of Vwas thus necessary for
the theory to extend to such systems. Wilson
(1969) subsequently published a modified ‘colo-
nization’ curve, showing how species number
might rise above the value first achieved via a
process of ecological assortment (e.g. successional
effects) and, over long timespans, rise again via
evolutionary additions. No means of quantifying
the timescale or amplitude of the adjustment was
made, making the adjusted theory difficult to
falsify.
Several authors have applied equilibrium think-
ing to evolutionary biotas on remote islands (e.g.
MacArthur and Wilson 1967; Juvig and Austring
1979). If the islands within a remote archipelago92 SPECIES NUMBERS GAMES: THE MACROECOLOGY OF ISLAND BIOTAS