that conditions of climate and habitat were
appreciably different during the last glacial period
and indeed earlier in the Pleistocene (cf. Buskirk
1985). Lowered sea level combined with climatic
changes to produce connections between islands
presently separated and also produced a general
increase in the extent of xeric habitats, such as arid
savanna, grassland, and xeric scrub forest. They list
several examples from the fossil record of species
characteristic of dry, open country, which once had
a much wider distribution than they have had in
late Holocene times; these include the burrowing
owl (Athene cunucularia), the Bahaman mockingbird
(Mimus gundlachii), thick knees (Burhinus), falcons
and caracaras (PolyborusandMilvago), curly-tailed
lizards (Leicophalus), and rock iguanas (Cyclura).
The significance of such losses of species on parti-
cular islands is exemplified by the figures for New
Providence Island, where 50% of the late Pleistocene
avifauna and 20% of the fossil herpetofauna no
longer occur, many of these species being xerophilic.
Pregill and Olson (1981) provide a number of
other examples of species distributions that they
regard as most parsimoniously explained by refer-
ence to long-term environmental changes. For
instance, they list a number of pairs of Hispaniolan
bird species, such as the todies (Todus subulatusand
T. angustirostris) and the palm-tanagers (Phaeni-
cophilus palmarumandP. poliocephalus), which they
regard as most probably having differentiated at
times during the Quaternary when Hispaniola was
divided into north and south islands. Hence, in
opposition to the taxon cycle model, they offered
the argument that the distributional patterns iden-
tified might be primarily an evolutionary outcome
of climates becoming wetter (and warmer) since the
end of the Pleistocene rather than being interpretable
simply in terms of interspecific interactions.
At the time of this debate the distributional and
fossil data were inadequate for a clear resolution of
the argument. However, recent genetic analyses
have provided a new angle on this controversy.
Within the taxon cycle model it is assumed that a
widespread species moves out in a single colon-
ization phase from a mainland source, and spreads
out according to a stepping-stone process of colon-
ization through the archipelago, producing a
monophyletic group within which birds on islands
close to one another should be more closely related
than those distant from one another. Klein and
Brown (1994) point out that the ornithologist James
Bond (after whom the fictional character was
named) proposed multiple colonizations from
mainland sources for West Indian island popula-
tions of widespread species. Multiple colonization
events would mean that the representatives of a
species on different islands would not form a
monophyletic group relative to samples from the
various mainland source pools. Modern techniques
allow Bond’s idea to be tested.
Klein and Brown (1994) studied mitochondrial
DNA (mtDNA) from specimens of yellow warbler
(Dendroica petechia) collected from North, South,
and Central American sites, as well as from the
West Indies. The most parsimonious tree con-
structed from their phylogenetic analyses indicated
colonization of some islands in the Lesser Antilles
by Venezuelan birds, and colonization of the
Greater Antilles by Central American birds.
Furthermore, they found evidence of multiple colo-
nizations not only of the West Indies as a whole, but
also of individual islands. Such events can involve
the introgression of characters from one lineage to
another. Moreover, the phylogenetic data indicated
that birds of adjacent islands are not always each
other’s closest relatives, against the assumptions of
a stepping-stone model for colonization. The multi-
ple colonizations of the West Indies by yellow war-
bler are at variance with the taxon cycle model as
outlined above. At least some of the differences
between populations are a consequence of what
could be thought of as a series of founder effects,
rather than exclusively because of in situselective
pressures and drift. Klein and Brown (1994) note
that a number of recent studies of bats in the
Caribbean provide similar lines of evidence for
multiple colonizations of individual islands, and of
one group of related haplotypes being widespread
while another was confined to the Lesser Antilles.
Similar analyses based on mtDNA have now
been conducted on many other bird lineages in the
West Indies (papers reviewed in Ricklefs and
Bermingham 2002), allowing a comprehensive
re-evaluation of the debate over the taxon cycle214 EMERGENT MODELS OF ISLAND EVOLUTION