biological processes and some, at least, of the
characteristics of the environment.
●Thirdly, it is empirically operational, making
robust, qualitative predictions that can be tested
(although, as we saw in chapter 4, the model turns
out be much more difficult to test than once
thought).
Brown notes that the fact that the EMIB has been
repeatedly and unequivocally falsified does not
diminish the contribution it has made to advance
our understanding of diversity. In contrast to the
judgement by Sauer (1969, above), he therefore
characterizes it as a useful beginning:
The model implies that the determination of diversity is a
very stochastic process. Species continually colonize and
go extinct, the biota at equilibrium is constantly changing
species composition, and all of these processes occur
essentially at random. MacArthur and Wilson knew that
this was not so, but ‘... it was a useful simplifying
assumption...‘ (Brown 1981, p. 883).
Perhaps the key virtue of the model as a general
model is that it is dynamic. The distinction between
historicalanddynamic hypothesesis based on
the repeatability or probability of recurrence of a
particular state or form. Whereas, historical, or
time-bound, knowledge refers to the analysis
of complex states having very small probabilities
of being repeated (i.e. states of low recoverability),
physical, or dynamic, or timeless, knowledge
refers to the analysis of states having a high degree
of probability of being repeated (Schumm 1991).
The search for dynamic hypotheses of species
richness can thus be likened to the search for the
general laws of ecology, from which historical
contingency supplies the deviations (Brown 1999;
Whittakeret al. 2001). Hence, the abiding attrac-
tion of the MacArthur and Wilson model and of
the dynamic, equilibrial framework they put in
play.
Many of the problems with the essentially
stochastic EMIB of MacArthur and Wilson have
been discussed in the previous two chapters, as we
have focused on the extent to which equilibrium
models apply; our ability to predict island species
numbers and identities, and on what forces struc-
ture island biotas. Table 6.1 summarizes some of the
key issues. There is no doubting the lasting heuris-
tic value of the dynamic equilibrium model, but
more in question is the continuing widespread use
of the model, its direct derivatives and assumptions
in predicting species losses in relation to habitat
loss and fragmentation: an issue we return to in
Chapter 10.146 SCALE AND ISLAND ECOLOGICAL THEORY: TOWARDS A NEW SYNTHESIS
Table 6.1Some of the limitations identified in MacArthur and Wilson’s equilibrium model of island biogeography (EMIB)Author Limitations of EMIBSauer (1969) Theory ignores autoecology—but species are not interchangeable units (e.g. see Armstrong
1982 on the effects of rabbits introduced onto islands)
Brown and Kodric-Brown (1971) Distance may affect extinction rate (E) (through the rescue effect), whereas Eis treated as a
function only of area in the EMIB
Lynch and Johnson (1974) The data are rarely adequate for testing turnover
Hunt and Hunt (1974) Turnover can be confounded by trophic-level effects
Simberloff (1976) Re: mangrove data set, most turnover involves transients, i.e. is pseudoturnover
Gilbert (1980) Turnover at equilibrium has not been demonstrated
Williamson (1981) Area may affect immigration rate (I), whereas Iis treated as a function only of
isolation in the EMIB
Williamson (1981, 1989a,b) Immigration, extinction, and species pool are each poorly defined. EMIB is imprecise on
reasons for extinction, and most turnover is ecologically trivial (cf. Simberloff 1976)
Pregill and Olsen (1981) Ignores historical data and role of environmental change
Haila (1990) EMIB has a narrower domain of applicability than originally thought (may be operational on
the population scale)
Bush and Whittaker (1991) Ignores successional effects and pace, and the hierarchical links between taxa