maximum duration of about three centuries, for
which two otherwise allopatric species have
co-occurred. Furthermore, within the Long group,
the two species occur together in all habitats and at
all altitudes, often being found in the same flower-
ing tree. They are thus sympatric even at the finest
intraisland scale. Diamond et al. (1989) found that
the Long populations are significantly more diver-
gent from each other than are the allopatric source
populations. The larger species, M. pammelaena, is
bigger still on Long than in the source populations,
and the smaller M. sclateri is even smaller.
Comparing samples of the two species from
allopatric populations, the weight ratio was found
to lie between 1.24 and 1.43, but for the sympatric
Long island populations the value was 1.52.
The 300-year time span may be unremarkable by
analogy with similar niche shifts in other circum-
stances. Conant (1988) found that Laysan finches
(Telespyza cantans) introduced to Pearl and Hermes
Reef underwent measurable shifts in bill shape
within 18 years, in adjusting to their local food sup-
ply. Similarly, significant change in diaspore size in
Lactuca muralisin Barkley Sound took place within
just five plant generations (see below; Cody and
Overton 1996).
7.4 Sex on islands
There are many ways in which plants and animals
reproduce, the simplest dichotomy being between
sexual and asexual modes. What works best in the
island laboratory? How does the possession of a
particular breeding system affect the chances of col-
onization or speciation on islands? Does hybridiza-
tion create new forms and thus engender
diversification, or does it restrict the tendency to
radiation by pulling back two diverging forms from
the brink of irrevocable parting?
Dioecy and outcrossing
Plants exhibiting sexual reproduction may as a sim-
ple generalization be classified as monoecious
(male and female flowers on the same plant), her-
maphrodite (bisexual flowers), or dioecious (plants
that, generally, bear either male or female flowers
throughout their lifespan). At least theoretically, it
should be more difficult for dioecious species to
colonize a remote island, given the apparent need
to have both sexes present for success (Ehrendorfer
1979; Sohmer and Gustafson 1993). In practice,
early comparison by Bawa (1980, 1982) suggested
that islands might have relatively high proportions
of dioecious species. He noted that 28% of the
Hawaiian plant species are dioecious, compared to
a figure of 9% for a well-specified tropical mainland
flora from Panama. Since then, the Hawaiian figure
has been revised downwards, to 15% for species
that are strictly dioecious, or to 21% if a broader
grouping of sexually dimorphic species is used.
This is still the highest incidence of dioecy of any
well-specified flora (Sakai et al. 1995a). However, it
is by no means clear that Hawaii is representative:
indeed dioecious species are reportedly scarce in
both the Galápagos (McMullen 1987) and the
Canaries, where only 2% of the endemic flora is
dioecious (Francisco-Ortega et al. 2000).
The high proportion of dimorphic species in the
Hawaiian endemic flora is intriguing. Sakai et al.
(1995a) find that dimorphy is frequent in part
because many colonists were already dioecious, but
that autochthonous evolution of dioecy has
occurred in at least 12 lineages, including several
species-rich lineages. Interestingly, one-third of cur-
rently dioecious species derive from monomorphic
colonists. Further study of other island floras will
be needed to establish how representative these
patterns are of other oceanic island floras.
Baker (1955) argued that hermaphroditic plants
with self-compatible reproduction systems, able to
establish (potentially) from a single founder indi-
vidual, would have better chances of island colo-
nization than dimorphic or obligated out-crossers.
From the data available it appears that the floras of
Hawaii (Carr et al. 1986), New Zealand (Webb and
Kelly 1993) and the Galápagos (McMullen 1987) do
have smaller proportions of self-incompatible taxa
than nearby continents, consistent with Baker’s
rule. Moreover, in some species, such as Fragaria
chiloensis, dioecious in America but hermaphroditic
in Hawaii, or Coprosma pumila, dioecious in New
Zealand but monoecious in Macquarie (Ehrendorfer
1979), there is evidence that ‘continental’ allogamySEX ON ISLANDS 177