240 EMERGENT MODELS OF ISLAND EVOLUTION
exclusive tendency to single-island endemism. Most
species thus appear to evolve on an island early in its his-
tory and thereafter remain confined to that island.
Colonists arriving at newer emerging islands tend to form
new species, a finding that has led to the serious consid-
eration that speciation may be somehow related to
founder events.
Studies from the Juan Fernández islands by
Crawford et al. (1992) also support the idea of high
initialrates of radiation. Electrophoretic data for the
endemic genus Robinsonia(Asteraceae) suggest that
the founding population arrived early in the
4.0 million year history of Masatierra island, radiat-
ing and speciating rapidly after colonization
(Crawford et al. 1992).
Evolutionary change can be very slow paced but
it can also occur with great speed, so the notion that
speciation rates may vary during the life history of
an island is not unreasonable. In illustration of
rapid change, Losos et al. (1997) were able to
demonstrate significant adaptive differentiation
over a 10–14 year period in populations of Anolis
sagreiintroduced experimentally into small islands
in response to varying vegetation of the recipient
islands. Diamond’s study of character displace-
ment in the myzomelid honeyeaters is another
example of the sort of context in which a fairly pre-
cise time frame, a maximum of c. 300 years, could
be put to a given evolutionary change (Chapter 7;
Diamondet al. 1989). The limitation of such studies
is that they are concerned with microevolutionary
changes rather than full speciation: for the latter we
depend upon geological and biological dating tech-
niques. Repeated speciation is demonstrated by the
70 species of Mecyclothoraxbeetles endemic to 1 mil-
lion-year-old Tahiti, and the picture-wing Drosophila,
in which 25 of the 26 species on the island of Hawaii
are restricted to that approximately 0.6–0.7 million-
year-old island (Paulay 1994), while the cichlid
fishes of Lake Victoria provide the most spectacular
example of explosive radiation (see Box 9.2).
Although such large islands as Hawaii were once
assumed to be fairly well buffered from climatic
variations, the large-scale global climatic fluctua-
tions of the Pleistocene are now known to have
affected them (Chapter 2; Nunn 1994) and must
have some bearing on the biogeography of such
island systems (above). On ecological timescales,
present-day weather conditions can, after all, be
highly significant determinants of breeding success
in bird species, including, it has been found, in one
of the Hawaiian honeycreepers, the Laysan finch
(Morin 1992). Some of the most interesting insights
in this field come from the remarkable studies of
hybridization in Darwin’s finches by Grant and
Grant (1996a), showing how influential the climatic
anomaly of the 1982–83 El Niño was in affecting
gene flow through hybridization. Thus, although
speciation rates may indeed be higher at early
stages in the life of an oceanic island, it remains to
be established if and how such patterns have been
influenced by global climate change during the
Pleistocene (cf. Kim et al. 1996).9.6 Variation in insular endemism between taxa
The majority of island lineages have not radiated
spectacularly at the species level, but some types of
organism have done so repeatedly, in widely
separated archipelagos. Examples include genera
within the Asteraceae in islands as remote from each
other as Hawaii, Galápagos, and Macaronesia, and
the finch lineages of Hawaii and the Galápagos.
What traits might determine the extent to which
particular taxa undergo speciation on islands?
The most obvious trait is dispersal ability. This
trait varies hugely between different taxa, and
largely determines which fail to colonize, which
colonize very occasionally, and which may colonize
repeatedly. The first group obviously cannot form
island endemics, whilst very frequent gene flow
between islands and source regions may also limit
speciation. We can illustrate this by looking at
endemism in ferns and seed plants on two archi-
pelagos. As already mentioned, the Juan Fernández
archipelago contains two main islands, about
150 km apart. Within the angiosperm genera con-
taining endemic species, 31% have at least one
species on both islands; the corresponding figure
for pteridophytes (ferns) is 71% (Stuessy et al. 1990).
Endemism within the Galápagos flora is distrib-
uted across major taxa as follows: 8 of the 107
pteridophyte species are endemic, 18 of the 85