outnumber the historical losses. Table 11.3 provides
a figure of 158 extinctions of island endemic birds
in prehistoric times, but a slightly higher figure, of
just over 200 species (recorded as subfossils, and
whose disappearance is probably due to prehistoric
man), is given by Milberg and Tyrberg (1993). There
is no doubt that further archaeological and taxo-
nomic work will add to the number of extinctions
for the prehistoric period. It should therefore be
understood that the year AD1600 did not mark a
key point in island history globally, and that its use
by those compiling the statistics is essentially a
matter of convenience. The primary place at which
we ought to divide the analysis is the point at
which islands have been colonized by humans.
Steadman (1997a) estimates that rates of natural,
pre-human extinction in island birds may be at
least two orders of magnitude lower than post-
human rates. The world’s avifauna would have
been about 20% richer in species had the islands of
the Pacific remained unoccupied by humans. As
many as 2000 bird species may have become
extinct following human colonization but before
European contact; the list being dominated by
flightless rails, but including moas, petrels, prions,
pelicans, ibises, herons, swans, geese, ducks,
hawks, eagles, megapodes, kagus, aptornthids,
sandpipers, gulls, pigeons, doves, parrots, owls,
owlet-nightjars, and many types of passerines. The
emphasis on rails stems from observations that vir-
tually every Pacific island that has been examined
palaeontologically has been found to have had one
or more rails of such reduced flight capabilities
that they must have been endemic species. By
extrapolation, it is conceivable that more than 800
Pacific islands were once graced with one or more
rail species in one of the most dramatic examples
of island radiation to have persisted into the
Holocene (cf. Trewick 1997). Only a handful of
these species remain today.
Island species continue to be disproportionately
at risk. For instance, approximately one in six plant294 ANTHROPOGENIC LOSSES AND THREATS TO ISLAND ECOSYSTEMS
Table 11.3Estimates of actual and threatened losses of island endemic birds (data from Johnson and Stattersfield 1990; Steadman
1997 a). The total of 97 extinctions since AD1600 is slightly at variance with the estimate of 92 species losses given in Table 11.2,
and contrasts with the value of 21 species given in that table for the number lost from continental areas over this periodRegion Number of extinctions Endemism
Prehistorica AD1600–1899 AD1900–94 Approximate % endemics
number of threatened
endemicsPacific Ocean 90 28 23 290 38
Indonesia and Borneo – 0 2 390 22
Indian Ocean 11 30 1 200 33
Philippines – 0 1 180 19
Caribbean Sea 34 2 1 140 22
New Guinea and Melanesia 10 2 3 500 10
Atlantic Ocean 3 3 1 50 50
Mediterranean Sea 10 0 0 – –
Total 158 65 32 1750 23
aData are lacking for prehistoric losses from Indonesia and the Philippines. The prehistory category consists of species from
prehistoric cultural contexts, and includes only species already described. Sources of possible overestimation include: (1) the so-called
Lazarus effect, i.e. where species thought to be extinct are eventually rediscovered, (2) identification of species that did not exist, e.g.
there were once thought to have been up to 11 species of flightless giant Moa on New Zealand, but recent genetic analyses of
bones indicates that there were only 2 species which, however, exhibited exceptional morphological dimorphism (Bunce et al. 2003).
However, overall, there are doubtless many more extinct species yet to be discovered by the archaeologists; we can thus regard these
estimates as minimum figures.