occurs solely through branching (Herben et al.
2005).) Thus, although bottlenecks can serve to
reduce the genetic base of a new colony, in some
instances they may serve to provide enhanced
genetic variability within a population, the very
stuff on which natural selection may then go to
work!
Laboratory studies of Hawaiian Drosophilahave
provided a number of important insights into
founder effects, but recently these phenomena have
been studied in an increasing range of island taxa
(e.g. Clarke and Grant 1996). For example,
Abdelkrimet al. (2005) describe how the use of
complementary genetic techniques of analysis
allows the detection of varying severities of found-
ing events in initial colonization and subsequent
interisland dispersal events for ship rat (Rattus
rattus) in the Guadeloupe archipelago. Another
Hawaiian example is provided by endemic mem-
bers of the plant genus Silene. Populations appear
to be much more polymorphic on the older island
of Maui than on the younger island of Hawaii (i.e.
the main or Big Island), suggesting that genetic
variation had been lost in colonization of the latter
(Westerbergh and Saura 1994). The process of
restricted gene flow followed by genetic drift also
seems to be repeated as younger volcanic terrain is
colonized within the island of Hawaii.
Although some regard founder effects as a key
part of the most likely mode of speciation in most
Hawaiian Drosophilidae (e.g. Kaneshiro 1989,
1995), others dispute their significance. For
instance, Barton (1989) observes that enzyme het-
erozygosity within Hawaiian drosophilids is simi-
lar to that of continental species, and queries
whether, in practice, founder effects cause much
loss in genetic variation. Barton (1989) regards
other elements in the island laboratory as being of
greater significance than founder effects, and he
mentions specifically what he terms faunal drift,
the random sampling of species reaching an
‘empty’ habitat, and thus providing a novel
(disharmonic) biotic environment in which selec-
tion drives divergence into a variety of new niches
(see also Vincek et al. 1997).
Sarre et al. (1990) examined morphological and
genetic variation among seven island and three
mainland populations of the sleepy lizard,
Trachydosaurus rugosus, in South Australia. Their
work supports the significance of random genetic
drift and inbreeding in small isolates. All but two of
their islands were isolated from the mainland
between 6000 and 8500 years ago. Contrary to some
other recent studies of vertebrates on South
Australian islands, their allozyme electrophoresis
analysis showed that heterozygosity levels did not
vary significantly among the populations, i.e. there
has been little erosion of heterozygosity within the
insular populations. Nonetheless, the island popu-
lations exhibited reduced allelic diversity, with rare
mainland alleles tending not to be present on the
islands. This may have been due in part to bottle-
necks occurring at some point or points since isola-
tion of the populations. There was also a greater
degree of genetic divergence shown between island
populations than between mainland populations,
and this they attributed to genetic drift on the
islands.
Allozymesare different forms of an enzyme
specified by allelic genes, i.e. by genes that occupy
a particular locus. Allozyme electrophoresis has
become a conventional method of measuring het-
erozygosity levels. It is thus of interest that two
measures of developmental stability varied signifi-
cantly among the island populations in a manner
distinct from the results of the allozyme study
(Sarre and Dearn 1991). The morphological meas-
ures used were: fluctuating asymmetry—the
degree of random difference between left and right
sides in bilateral structures (see Box 10.1); and per-
centage gross abnormalities, involving features
such as missing ear openings, missing toes, club
feet, and deformed head and labial scales. Three of
the island populations stood out as having dis-
tinctly higher levels of these forms of developmen-
tal instability. This was interpreted as being due to
the expression of deleterious genes (in turn due to
inbreeding depression) or the genetic disruption of
co-adapted genotypes. One important conclusion
drawn from their study is that reduced levels of
heterozygosity per seare unlikely to be the prime
cause of the observed developmental instability. In
short, the allozyme technique was failing to reflect
the genetic basis for the deleterious morphologicalFOUNDER EFFECTS, GENETIC DRIFT, AND BOTTLENECKS 169