addition to such non-selective drift there may also
be distinctive selective features of the novel island
environment. The outcome of these phenomena
could, in theory, be a rapid shift to a new,
co-adapted combination of alleles, hence contributing
to reproductive isolation from the original parent
population.
Although it is possible to demonstrate the occur-
rence of genetic drift, as Berry et al. (1992) have
done in their study of a small population of house
mice on Faray in the Orkneys, the contribution of
founding effects and drift to island speciation
remains the subject of keen debate (Barton 1989;
Clarke and Grant 1996). It is particularly problem-
atic that founding events(i.e. colonizations) have
been theorized to produce a variety of rather differ-
ent founding effects, and that the latter are difficult
to isolate empirically from other microevolutionary
processes. In fact, there are at least three distin-
guishable models of how founder effects may lead
to speciation:
●Mayr’s peripatric speciation
●Carson’s founder-flush speciation
●Templeton’s genetic transilience.
These theories place greater degrees of stress,
respectively, on:
●the impact of the general loss of heterozygosity
combined with random drift
●the role of variations in selection pressure (partic-
ularly its relaxation immediately after colonization)
●changes in a few major loci within the genome
(see discussions in Templeton 1981; Barton and
Charlesworth 1984; Clarke and Grant 1996).
The founding event represents a form of bottle-
neck, a term implying a sharp reduction in popula-
tion size. Bottlenecks may occur at other points in
the lifespan of a lineage, e.g. following catastrophic
disturbance of a habitat, or in the context of species
pushed to the brink of extinction by human action.
Studies of the Hawaiian Drosophilaindicate that
founder events have occurred and at three levels:
in the colonization of the archipelago, of its con-
stituent islands, and of patches of forest habitat
(kipukas) within islands (Carson 1983, 1992; Carson
et al. 1990). Significantly, this work indicates that
bottlenecks produce a marked loss of heterozygosity
only if they are sustained or are repeated at short
intervals. In contrast, if a bottleneck lasts only one
generation and recovery to a large population size
is rapid, then increased genetic variance for quanti-
tative traits can occur.
According to theories of founder-effect specia-
tion, the narrowing of the gene pool during a bot-
tleneck may serve to break up the genetic basis of
some of the older adaptive complexes, liberating
additive genetic variance in the new population
(Carson 1992). Novel phenotypes might subse-
quently result from natural selection operating on
the recombinational variability thus provided in
the generations immediately following the bottle-
neck. According to a model proposed by Kaneshiro
(e.g. 1989, 1995), an important element in this sce-
nario is differential sexual selection. He found that
a minority of male drosophilids perform the major-
ity of matings, and that females vary greatly in their
discrimination, and acceptance, of mates. In nor-
mal, large populations the most likely matings are
between males that have a high mating ability (the
studs), and females that are less discriminating in
mate choice. This provides differential selection for
opposing ends of the mating distribution and a sta-
bilizing force maintaining a balanced polymor-
phism. In a bottleneck situation, however, there
would be even stronger selection for reduced dis-
crimination in females, and if this was maintained
for several generations, it could result in a destabi-
lization of the previous co-adapted genotypes.
In such bottleneck situations, with reduction in
mate discrimination by females, the likelihood of a
female accepting males of a related species
increases, thus allowing hybridization to occur. The
term given to the incorporation in this way of genes
of one species into the gene pool of another via an
interspecific hybrid is introgressive hybridization.
With the development of molecular tools,
hybridization between animal species (especially
insects) has been discovered to occur naturally
much more often than had previously been sus-
pected (Grant and Grant 1994; Hanley et al. 1994;
Kaneshiro 1995; Clarke and Grant 1996). (This pres-
ents a small but perhaps too often overlooked chal-
lenge to the assumption that lineage development168 ARRIVAL AND CHANGE