The loss of dispersal powers
Given that oceanic islands are so hard to reach, it
may appear paradoxical that so many flightless
forms occur on them. For a distant island popping
out of the sea for the first time, there will inevitably
be a strong bias towards groups that are highly dis-
persive, as these will, by definition, have a greater
chance of colonizing. This was reflected in the bio-
geographical patterns discussed in Chapter 3, of fil-
ter effects, the gradual loss of taxa with increasing
distance from continental source areas, and of
disharmony. The relevance of these effects is broad,
and wide ranging, from the evolutionary opportu-
nities thus presented on islands like Hawaii, to the
ecological structuring in less remote islands such as
Krakatau (Chapter 5). The island biogeography of
remote islands thus bears the imprint of dispersal
filtering of faunas and floras as its first organiza-
tional principle.
Once on an isolated island, however, there may
be a strong selective force against dispersal ability,
particularly in insects and birds. Examples given by
Williamson (1981) of flightless forms include the 20
endemic species of beetle on Tristan de Cuhna, all
but two of which have reduced wings. A number of
ideas have emerged in explaining such tendencies.
One put forward by Darwin (letter to J. D. Hooker,
7 March 1855; Roff 1991) is that highly dispersive
propagules are more liable to be lost from the gene
pool, for instance by a highly dispersive plant or
insect being blown off to sea, so selecting for less-
dispersive forms among the island population. A
second idea, applicable to flightless birds such as
the famously extinct dodos, is that the lack of pred-
ators removes the selective advantage of flight in
ground-feeding species. Indeed, flight might be dis-
advantageous in that the energy used in maintain-
ing and using flight muscles is effectively wasted.
However, flightless species within generally volant
groups can also be found on continents, so before
attempting to test ideas concerning loss of disper-
sal, it might be prudent to confirm that the inci-
dence of flightlessness on islands is higher than
expected by chance alone.
Roff (1991, 1994) has undertaken just such analy-
ses for insects and for birds. Darwin’s original
observations on insects were based, of course, on
relatively few data, and in his analysis of the much
larger data sets now available Roff (1991) found
that statistically there is no evidence of an associa-
tion between oceanic islands and insect flightless-
ness. He argues that the large size of most oceanic
islands makes flightlessness of little positive selec-
tive value: only a small fraction of the population is
likely to be lost from such sizeable land masses (see
also Ashmole and Ashmole 1988). It should be
stressed, however, that Roff’s analyses are based on
a dichotomy between volant and non-volant forms.
He does not take into account island forms that
have evolved a reduced flight capability or behav-
iour without actually losing the ability altogether.
In his second article, on birds, Roff (1994) points
out two important factors that complicate the
analysis. First, in the case of most bird groups, the
small sample size limits the potential for formal sta-
tistical analysis. Secondly, the phylogenetic con-
straints on flightlessness must be considered. The
condition may have evolved once, followed by sub-
sequent speciation within a flightless lineage, or it
may have evolved several times, in extreme cases
once for each flightless species. For example,
among important flightless bird groups, the 18
known penguin species are considered to be mono-
phyletic, having speciated from a common flight-
less ancestor, whereas the 41 ratite species
(ostriches, etc.) represent between one and five evo-
lutionary transitions to flightless forms, depending
on whose phylogeny one accepts. In contrast,
flightlessness in rails has most probably evolved
separately for each of the 17 (or more) flightless
species, among the 122 (or more) recognized rail
species. Truly flightless rails are only found on
islands and they represent about half of the island
rail species (Roff 1994). More ratites occur on
islands than on continents, but the taxonomic dis-
tribution does not indicate that flightlessness is
more likely to evolve on islands in this group.
Examples of flightlessness from other groups of
birds include the following island forms: the New
Caledonian kagu (Rhynochetos jubatus), the two
extinct dodos (Raphus cucullatusin Mauritius and
Pezophas solitariain Rodrigues), New Zealand’s
kakapo (Strigops habroptilus, the only flightless182 ARRIVAL AND CHANGE