features. The relative contribution of environmental
and genetic factors to developmental stability in the
sleepy lizard cannot at this stage be determined.
In contrast to the emphasis given in this study to
drift, Thorpe and Malhotra (1996) stress the impor-
tance of natural selection for current ecological con-
ditions in their studies of Canarian and Lesser
Antillean lizards. It may be significant that Thorpe
and Malhotra’s islands, unlike Sarre’s, are large and
contain a considerable range of habitats, i.e. the rel-
ative importance of stochastic processes versus
directional selection may turn out to be strongly
dependent on the environmental context.
Although opinions differ as to the overall signif-
icance of founding effects for the rapidity of evolu-
tionary change and ultimately speciation on
islands, it has increasingly become evident that
founding effects can be detected by several differ-
ent techniques, and for a range of plant and animal
taxa (e.g. Clarke and Grant 1996; Drotz 2003; Hille
et al. 2003; King et al. 2003). Moreover, in several
well-specified case studies, it is possible to see a
nested sequence of events, from the initial coloniza-
tion of an archipelago, through interisland move-
ments, down even to intraisland colonization
events (Westerbergh and Saura 1994; Abdelkrim
et al. 2005).
Implications of repeated founding events
What are the implications of such scenarios? First,
island lineages are likely to contain less genetic
diversity than the mainland source population.
Second, because there is a general trend towards
loss of dispersability once established on a remote
island (discussed below), and because of the often
significant interisland distances, the repetition of
such founding events is likely to lead to further
differentiation between populations within an
archipelago, especially if there is significant envi-
ronmental variation amongst the islands to add a
selective component to lineage development.
Available data tend to confirm these expectations
(e.g. DeJoode and Wendel 1992; Frankham 1997;
Hilleet al. 2003). DeJoode and Wendel (1992) com-
piled data in the form of allozyme variability for
approximately 60 Pacific island endemic plant taxa
(a mix of species and varieties), finding that they
exhibit roughly half the variability previously
reported for continental plant species, and two
thirds of that shown by continental endemics of
restricted range. Recent analyses of populations of
silvereyes (Zosterops lateralis) that colonized islands
in the southwest Pacific from mainland Australia
afterAD1830 allow an unusually well-specified
quantification of founder effects. The analyses indi-
cated that single founder events had little impact
on genetic diversity, partly because founding popu-
lations have apparently been quite large (24–200 or
so), but that repeated island hopping was accom-
panied by a gradual decline in allelic diversity
(Clegget al. 2002; Grant 2002).
In general, differences in genetic variability
among island and mainland populations have been
found to be larger: (1) the smaller the size of the
founder event, (2) the larger the differences in
population sizes, (3) the smaller the immigration
and dispersion rates, (4) the smaller the island
size, and (5) the greater the distance from the conti-
nent (Frankham 1997). Indeed, analyses of a few
tree species colonizing the Krakatau islands, which
are isolated from mainland Java and Sumatra by
only about 40 km, indicate no loss of genetic
variation in the island populations, and suggest
that for some species the island and mainland
populations can be considered panmictic, i.e. as
if involving random mating within a single
extensive population (Parrish 2002). Moreover, sim-
ilar results were found for two species of Krakatau
fig wasps (Zavodna et al. 2005); despite the small
size of these insects, such a distance may be
scarcely a barrier to colonization and gene flow.
As Grant (2002, p. 7819) commented about the
silvereye study... Perhaps it is more parsimonious to invoke single-
immigration events for each island, but I think this is
likely to be wrong. If it is wrong, it carries an important
implication: except for the most isolated islands, repeated
immigration may obscure or obliterate any founder-effect
changes that take place following the initial colonization.
Reviewing allozyme variation in 69 Canarian
endemic plant species, Francisco-Ortega et al. (2000)
report an average species-level genetic diversity at170 ARRIVAL AND CHANGE