the Indian Ocean tsunami of 26 December 2004, the
eruption of Montserrat, in the 1990s, and the recent
increase in hurricane activity in the Caribbean
suggest otherwise. In addition, less dramatic envi-
ronmental variation may also perpetuate non-
equilibrium conditions. For example, Lynch and
Johnson (1974), who may have coined the term
‘successional turnover’ in its island ecological
sense, argued that where turnover is successional in
nature, such as produced after a fire, then the
turnover is determined not primarily by
area–distance effects, but by the timing, extent, and
nature of changes occurring in the habitats. They
regard use of an equilibrium model in such
situations as inappropriate. They also argue that
interpretation of faunal change is especially
difficult if both equilibrium and non-equilibrium
(e.g. successional) turnover are involved.
Extreme climatic events, such as freezes, have
been recognized as important determinants of
island avifaunas. An illustration is provided by
Paine’s (1985) observations of the extinction of the
winter wren (Troglodytes troglodytes) from Tatoosh, a
5–6 ha island, 0.7 km from the north-western tip of
the Olympic Peninsula (Washington, USA). The
loss of the species followed an extremely cold
period in December 1978, and it took 6 years to
re-establish a breeding population, which was pre-
sumed to be because of the difficulty of a relatively
sedentary species invading across a water gap.
Abbott and Grant (1976) found evidence of non-
equilibrium in land-bird faunas on islands around
Australia and New Zealand, a result that did not
appear to be attributable to sampling error or to
humans. They suggested in explanation that
islands at high latitudes (such as these) are gener-
ally subject to irregular climatic fluctuations and
may not therefore have fixed faunal equilibria
(cf. Russell et al. 1995). Abbott and Grant (1976)
argued that the extent to which climate fluctuates
about a long-term average value determines the
extent to which species number does the same. ‘In
this sense, species number ‘chases’ and perhaps
never reaches a periodically moving equilibrium
value, hindered or helped by stochastic processes’
(Abbott and Grant 1976, p. 525).
Another study that appears to demonstrate that
equilibrium values change in response to modest,
non-catastrophic changes in island environments
on timescales of years and decades is that of Russell
et al. (1995). They found that a non-equilibrium
model provided improved predictions of observed
turnover of birds for 13 small islands off the coast of
Britain and Ireland. They suggested that turnover
could be viewed as operating on three scales: first,
year to year ‘floaters’ (trivial turnover); secondly,
on a timescale varying between 10 and 60 years, an
intrinsic component equivalent to that envisaged in
the EMIB was observable; thirdly, most islands show
a change in numbers over time due to so-called
extrinsic factors, such as habitat alteration.Long term non-equilibrium systems
Extending the time scale, the glacial/interglacial
cycles of the Quaternary have involved high-
magnitude effects, the boundaries of biomes being
shifted by hundreds of kilometres, sea levels falling
and rising by scores of metres. These changes have
resulted in at least two types of relictual pattern
studied by island biogeographers. The first is that
of land-bridge islands, those formerly connected to
larger land masses (their mainlands), which gained
at least some of their biota by overland dispersal
before becoming islands (e.g. Crowell 1986). The
second well-studied context is where habitat
‘islands’ have become isolated in a similar fashion
within continents (e.g. Brown 1971). In both con-
texts, data often support a non-equilibrium inter-
pretation.
The EMIB would have it that slopes of
species–area curves should be steeper (higher z
values) for distant (oceanic) island faunas than for
nearer (often land-bridge) islands in the same archi-
pelago. This is due to the combined effects of
extinction and low colonization rates on very
isolated islands, and because the impact of both is
greater on small islands (Fig. 6.3a). However, as
Lawlor (1986) noted in a review of data for
terrestrial mammals on islands, large remote
islands tend not to attain the species richness
predicted for them on the basis of their area. This is158 SCALE AND ISLAND ECOLOGICAL THEORY: TOWARDS A NEW SYNTHESIS