This idea, of ‘adaptive landscapes’, was originally
introduced by Sewall Wright (1932) and has been
incorporated into a number of different evolution-
ary models (see e.g. Templeton 1981, Futuyma
1986, Otte and Endler 1989).
Intentionally, the account given here is a simple
rendition of ideas culled from a much larger body
of literature. Based on laboratory evidence, compet-
itive speciation may occur in as few as 10–100
generations, not as fast as polyploidy (almost
instantaneous), but faster than geographical specia-
tion, which seems to require thousands or even
hundreds of thousands of years (Rosenzweig 1995).
Grant and Grant (1989) recorded an early phase
in the process of sympatric speciation in a study of
a temporary reproductive subdivision within
Geospiza conirostris, one of the Galápagos (Darwin’s)
finches occurring on Isla Genovesa. The population
was partly subdivided ecologically for a brief
period, but this division then collapsed again
through random mating. All of the known sym-
patric species of ground finches differ by at least
15% in at least one bill dimension, suggesting that
this may indicate a threshold in niche separation
needed to sustain separate species. During the
temporary subdivision, the two groups of males
differed by a maximum of 6% in bill dimensions,
and this, occurring as it did over but a brief period,
was insufficient to foster any discrimination in a
mating context. One of the dry-season food niches
sustaining the division declined catastrophically
and the division collapsed. Nonetheless, this study
shows how sympatric division may be fostered in
taxa in which polyploidy does not occur. Grant and
Grant would have had to be extraordinarily lucky
to come along just at the right time to record a sym-
patric split that ‘stuck’. They therefore did not
regard the failure of this event as evidence against
the idea, which they concluded to have some rele-
vance for vertebrates. Rather, they took it to indi-
cate that in order to foster sympatric divergence
leading to speciation, the niches or habitats to
which different members of a population adapt
must be markedly different and display a long-
term persistence. This is more likely to occur on
larger, higher islands.
How important is competitive speciation in the
island context? This is difficult to judge as it cannot
be decided merely by investigating the degree of
geographical overlap between sibling species in
ecological time. This is because, given the proba-
bilistic nature of dispersal and the magnitude of
past environmental change, populations currently
isolated from one another may once have been
sympatric and vice versa. Indeed, it may commonly
be the case that archipelago speciation—diversifi-
cation within archipelagos—involves a mix of
allopatric and sympatric phases of subtly varying
combination from one branch of a lineage to the
next (cf. Clarke and Grant 1996). Over the lifespan
of a large island, the same may also apply within a
single island. The relative roles of allopatric and
competitive speciation within the island context
cannot therefore be quantified, but one thing
remains clear; isolation is crucial, either directly or
indirectly (respectively).Polyploidy
One important class of sympatric speciation is
through polyploidy, a condition comparatively
common in plants and many invertebrates, but not
in higher animals: it is unknown in mammals and
birds, for instance (Grant and Grant 1989;
Rosenzweig 1995). Polyploid species are those that
have arisen by an increase in chromosome number.
Two main forms of polyploid can be distinguished.
If the new species has twice the chromosome com-
plement of a single parent species it is termed an
autopolyploid. If it has the chromosomes of both of
two parent species (i.e. a crossing of lineages) it is
called an allopolyploid. According to Rosenzweig
(1995), autopolyploids are rare, but at a conserva-
tive estimate 25% or more of plant species may be
allopolyploids. Barrett (1989) states that current
estimates for the angiosperm subset are as high as
70–80% of species being of polyploid origin.
Rosenzweig (1995) presents an intriguing analysis
of the proportionate importance of polyploids in
different floras, showing that there is a general
decrease in the relative contribution of polyploids in
floras with decreasing latitude. The best trend wasMECHANISMS OF SPECIATION 205