occurred as a result of climatic changes and the
associated sea-level changes and habitat alterations
at the end of the Pleistocene, and could be consid-
ered a ‘natural’ part of the biogeographical pattern-
ing of the region; others were doubtless due to
human activities, including hunting, habitat alter-
ation and the introduction of exotics. Interpretation
of the present-day biogeography of the region thus
requires knowledge not only of past faunas but also
of the extent to which humans have been involved
in their alteration.
The difficulty of this task is illustrated by a study
from Henderson Island. In his investigation of
42 213 bird bones, 31% of which were identifiable,
Wragg (1995) found that 12% of the fossil bird
species recorded were accounted for by just 0.05%
of the total number of identifiable bones, indicating
them to be of uncertain status, quite possibly
vagrants. Consequently, biogeographical studies
that rely simply on a list of fossil birds might,
exceptionally, assign resident status to temporary
inhabitants. A different form of bias probably
occurs in other studies because of the large mesh
sizes (6 or 13 mm) commonly used in sieving soil
samples. Larger-boned species such as seabirds,
pigeons, and rails are usually found, but small
passerines and hummingbirds are much less likely
to be recovered (Milberg and Tyrberg 1993).
As yet, we have only a fragmentary picture of
past losses, biased towards groups for which the
fossil record has been revealing, notably verte-
brates and especially birds. Much is still unknown.
What, for instance, has been the impact of the
losses from New Zealand of the giant moas on the
dynamics of forests they formerly browsed? How
might the loss of hundreds of populations and a
few entire species of Pacific seabirds have influ-
enced marine food webs, in which seabirds are top
consumers? (Steadman 1997a). Many island fruit
bats and land birds (including many now extinct)
have undoubtedly had crucial roles as pollinators
and dispersers of plants, and thus their extirpation
must be anticipated to have important repercus-
sions for other taxa, many of which may be ongo-
ing. That is, unless there is more functional
redundancy within oceanic island biotic assem-
blages than most authors currently recognize72 THE BIOGEOGRAPHY OF ISLAND LIFE
understanding of the relative timing of human
arrival and faunal turnover across the islands of the
Mediterranean. Thus, on the one hand, the loss of
Myotragus balearicuson the Balearic islands appears
attributable to human settlement, which predates
the extinction event, while the extinction of hippo
from Cyprus, previously attributed to human hunt-
ing, has recently been queried because clear evi-
dence of temporal overlap of humans with the
hippo is lacking.
Estimates of the numbers of Hawaiian birds that
were lost following Polynesian colonization
c.1500 years ago, but before European contact, now
stands at a minimum of 40 species (Olson and
James 1982). Each of the 16 Polynesian islands that
has yielded in excess of 300 prehistoric bird bones
approaches or exceeds 20 extirpated species (i.e.
lost from those particular islands, but not necessar-
ily globally extinct), pointing to the likelihood of far
more extinct species yet to be catalogued. About
85% of bird extinctions during historical times have
occurred on islands (Steadman 1997a), a rate of loss
40 times greater than for continental species
(Johnson and Stattersfield 1990).
The cataloguing of insular losses is clearly far
from complete, and although we have some basis
for estimating how many species of birds may be
involved, for most taxa such calculations remain to
be made (Milberg and Tyrberg 1993). Morgan and
Woods (1986) appraise the problem for West Indian
mammals. They calculate an extinction rate for the
period since human colonization of one species
every 122 years, and they emphasize that this is a
minimum estimate. The mammals of the West
Indian islands today are thus a highly impoverished
subset of species that escaped the general decima-
tion of the fauna. Morgan and Woods caution that
biogeographical analyses that fail to incorporate the
extinct forms are unlikely to prove reliable.
It is unsurprising, but significant biogeographi-
cally, that in addition to species disappearing alto-
gether, a number of species have gone extinct from
particular islands during the past 20 000 years,
while still surviving on others. This is particularly
evident in the bats, which supply the majority of
persisting species of mammal, but is also a pattern
found among lizards and birds. Many such losses