(compare Rainey et al. 1995; Vitousek et al. 1995;
Olesenet al. 2002).
It has been observed that the rapidity of human-
induced extinctions of island birds has been influ-
enced by the nature of the island. The species of
high, rugged islands, and/or those with small or
impermanent human populations, have probably
fared better than those of low relief and dense, per-
manent settlement. This is exemplified by the find-
ing that a larger percentage of the indigenous
avifauna survives on large Melanesian islands than
on small Melanesian islands, or on Polynesian or
Micronesian islands. This may be due to the buffer-
ing effects that steep terrain, cold and wet montane
climates and human diseases have had on human
impact (Steadman 1997a; Diamond 2005).
Steadman (1997a) has pointed out that the
Kingdom of Tonga did not qualify in a recent
attempt to identify the Endemic Bird Areas of the
world because of its depauperate modern avifauna,
yet bones from just one of the islands of Tonga,
‘Eua, indicate that at least 27 species of land birds
lived there before humans arrived, around 3000BP.
Forest frugivores/granivores have declined from
12 to 4, insectivores from 6 to 3, nectarivores from 4
to 1, omnivores from 3 to none, and predators from
2 to 1. As is consistent with predation by humans,
rats, dogs, and pigs, the losses have been more
complete for ground-dwelling species. It is highly
likely that the means of pollination and seed dis-
persal, not just of the plants of ‘Eua but of islands
across the Pacific, have over the past few thousand
years been diminished greatly as a consequence of
this pattern of loss, replicated, as it has been, across
the Pacific. That the attrition of endemic species
from oceanic islands may have so altered distribu-
tional patterns, compositional structure, and eco-
logical processes serves as a caution on our
attempts in the following chapters to generalize
biogeographic and evolutionary patterns across
islands.
3.6 Summary
This chapter set out to establish the biogeographi-
cal and biodiversity significance of island biotas,
particularly remote island biotas, and to provide an
indication of the ways in which island assemblages
are distinctive. In global terms, for a variety of taxa,
islands make a contribution to biodiversity out of
proportion to their land area, and in this sense col-
lectively they can be thought of as ‘hotspots’. Some
of course, are considerably hotter in this regard
than others. Their high biodiversity value is widely
acknowledged, as is the threat to island biodiver-
sity, and island thus typically feature prominently
in global and regional conservation prioritization
schemes.
Accompanying their high contribution to global
diversity (through the possession of high propor-
tions of island endemic forms), islands are, how-
ever, typically species poor for their area in
comparison to areas of mainland. The more isolated
the island, and the less the topographic relief, the
greater the impoverishment. Remote island biotas
are typically disharmonic (filtered) assemblages in
relation to source areas. This can take the form of a
climatic filter and thus a sort of biome shift on the
island, but more interestingly involves the
sampling of only a biased subset of mainland taxa.
Such biases may take the form of the absence of
non-volant mammals, the relative lack of represen-
tation within a particular family of plants, and so
forth, and are interpretable in terms of the relative
dispersal powers of the lineages and taxa under
consideration. An island may be a remote isolate
(oceanic in character) for a taxon with poor ocean-
going powers, but be linked by fairly frequent pop-
ulation flows (continental in character) to a source
area for a highly dispersive taxon.
Traditional zoogeographic and phytogeographic
analyses have enabled biogeographers to identify
unidirectional dispersal filters (so-called sweep-
stake dispersal routes), filter effects between differ-
ent biogeographic regions, and cases where islands
have been subject to the influence of multiple source
areas. We appraise the issues involved in placing
oceanic islands into traditional phytogeographical
and zoogeographical regionalism schemes, showing
that remote islands typically receive colonists from
multiple sources and via multiple routes.
Historical biogeography has been for some time
embroiled in debates between those advocating a
prominent role for long distance transoceanicSUMMARY 73