5 The taxon cycling hypothesis(Givnish 1998).
Successful island colonizers are usually weedy
species that established themselves easily in
open/marginal habitats, where they can grow rap-
idly. Increased specialization and selection in com-
petitive environments, with scarce resources, may
have precipitated a shift toward persistence through
woodiness (Givnish 1998, Panero et al. 1999). This
hypothesis is an extension of hypothesis (1).
It is too early to conclude which of the above might
provide the best explanation, if indeed there is a
single answer.
Size shifts in island species and the island rule
It presents another apparent paradox that two
conflicting traits are viewed as common in island
forms:gigantism, examples of which include the
Galápagos and Indian Ocean tortoises (Darwin
1845; Arnold 1979), and nanism(dwarf forms),
illustrated by island subspecies of ducks, which
are commonly smaller than mainland species
(Lack 1970). Of course, size variations within taxa
are not restricted to islands, and so there is the
danger of inferring a general rule from a few strik-
ing cases. For instance, the famous Aldabran and
Galápagos giant tortoises (Geochelonespp.) reach
up to 130 cm in length. However, giant tortoise are
readily eliminated by humans, and other giant
Geocheloneoccur, or occurred, in South America,
Madagascar, the Seychelles, the Mascarenes, and
on the African mainland. It is not certain whether
large size evolved after reaching islands or
whether it merely bestowed the capabilities of
dispersing to islands as remote as the Galápagos
(Arnold 1979).
Unusual size has been reported for many island
taxa, including plants and insects, but it is amongst
vertebrates that the most striking general patterns
are found, for there appears to be a graded trend
from gigantism in small species to dwarfism in
larger species. This has been termed the island rule
(van Valen 1973, Lomolino 1985, 2005), and has
been found to hold for mammalian orders, non-
volant mammals, bats, passerine birds, snakes and
turtles (Fig 7.4, Box 7.1). How well particular taxa
fit this island rule depends, as ever, on the assump-
tions involved in the analysis (e.g. compare
Lomolino 2005 and Meiri et al. 2006), but the syn-
thesis provided by Lomolino (2005) provides both a
compelling body of evidence and a persuasive
interpretation of apparent exceptions. Analyses of
data for continental mammal species have sug-
gested that there is an optimum body size for
energy acquisition at about 1 kg body weight
(Damuth 1993). Linking this observation to the
island rule, it would seem that islands favour the
evolution of a body size most advantageous for
exploiting resources in particular dietary cate-
gories, and that this must in turn be linked in some
way to the constrained area of islands, and the
absence of the full array of competitors and preda-
tors found within a continent.
The body size at which equivalence in body size
between island and continental populations is
reached varies between taxa. Lomolino (2005) inter-
prets these values as indicating the optimal size for
a particular body architectural plan (bauplan) and
ecological resource exploitation strategy (Fig. 7.5).
Whereas in species-rich mainland communities,
interspecific interactions (especially predation and
competition) facilitate evolutionary divergence
from the optimal size, on islands selection favours
change towards the optimum body size. For
instance, islands often lack terrestrial vertebrate
predators, but do have birds of prey. This switch in
predation regime might favour larger rodents,
which would then become dominant in the popula-
tion through intraspecific competitive advantage
(Reyment 1983, Adler and Levins 1994). At the
other end of the scale, carnivores (Box 7.1) and
artiodactyls (even-toed ruminants) typically
become smaller, suggesting small size to be advan-
tageous in restricted insular environments of lim-
ited food resources (Lomolino 1985).
There is considerable variability around the trend
line (Fig 7.4), some of which may be explained by
variables such as island isolation and area. For
example, body size of the tricoloured squirrel
(Callosciurus prevosti) decreases with increasing iso-
lation, and increases with island area (Heaney
1978). Some size changes can only be understood186 ARRIVAL AND CHANGE