provide strong selection pressure towards more
agile, smaller forms. Some of the best evidence for
nanism seems to be for fossils of the Cretan form of
dwarf elephant, which show a wide size range,
interpreted as representing stages in the develop-
ment of the dwarf form. ‘The flatland, lumbering
ancestral form, with good swimming capabilities,
had been irrevocably transformed into a clever
climber of rocky slopes’ (Reyment 1983, p. 302). The
prize for the greatest degree of dwarfing goes to the
MaltanElephas falconeri, which was barely more
than a metre in height (Lister 1993) and only a quar-
ter the size of the putative ancestor, E. namadicus.
These remarkable decreases in size may have
occurred quite quickly. Vartanyan et al. (1993)
reported remains of the woolly mammoth, which
appear to have persisted on Wrangel island, 200 km
off the coast of Siberia, when the island became
separated from the mainland c.12 000BP. The mam-
moth became extinct from its continental range by
around 9500BP(either because of climate change, or
humans, or a combination of both), but it persisted
on Wrangel until some point between 7000 and
4000 BP. Not only did the population thus attain
relictual status, but between 12 000 and 7000BPit
appeared (on the basis of fossil teeth) to shrink in
body size by at least 30%. However, size changes
need not be so rapid, nor follow so soon after colo-
nization. Sicilian elephants probably took between
250 000 and 600 000 years to shrink to about half
the dimensions of the ancestral Elephas antiquusand
about 1 million years to reach one quarter of the
ancestral size.
All Pleistocene ungulates, other mammals of any
considerable size, flightless birds, and giant tor-
toises on the Mediterranean islands were almost
certainly wiped out by the first humans to arrive.
This is, however, only recognizable in the archaeo-
logical record on Cyprus and Sardinia. It appears
that the extant wild ungulates on the
Mediterranean islands were all introduced by
humans.
Although the island rule now appears to be
clearly established, there is evidence of Holocene
size reductions in lizards from small islands in the
Caribbean (Pregill 1986). In particular cases, the
evidence is strongly supportive of some form of
co-evolutionary competitive pressure brought
about by formerly allopatric taxa coming into sym-
patry (cf. Roughgarden and Pacala 1989). In other
cases, the activities of humans, such as habitat
alteration and the introduction of terrestrial preda-
tors, may be responsible not only for the extinction
of larger lizards from within island assemblages
but also for size reductions (Pregill 1986). In illus-
tration, in the Virgin Islands, populations of the St
Croix ground lizard (Amereiva polyps) have declined
steeply over the past century, and body size of
mature adults has also decreased. The ecology of
these islands has been greatly affected by humans,
and in particular, by the introduction of the mon-
goose (Pregill 1986). Associations between human
colonization and the extinctions of larger species
from within insular lizard assemblages are also
known from the fossil records of the Canaries,
Mascarenes, and Galápagos archipelagos. Indeed,
Pregill (1986) concludes that insular lizards that
draw our attention because of their exceptional size
may have been rather ordinary in prehistory.
Fossils of extinct large frogs and snakes are also
known from the Caribbean. Humans have also
selectively eliminated larger vertebrates across the
globe during the late Quaternary: so this form of fil-
tering is far from being just an island pattern.
Food chain links may be important in selecting
for both smaller and larger forms. An interesting
example is Forsman’s (1991) study of variation in
head size in adder Vipera beruspopulations on
islands in the Baltic Sea. In accord with the general
trend for smaller vertebrates noted above, it was
found that relative head length of adders was
smallest in the mainland population. An interesting
further feature of this study was that, within the
island set, relative head length increased on the
islands with increasing body size of the main prey,
the field vole, Microtus agrestis. This was inter-
preted as the outcome of stabilizing selection for
head size within each population, i.e. an evolution-
ary response to the variation in body size of the
main prey species and the small number of alterna-
tive prey species available on islands. Schwaner
and Sarre (1988) also linked large body size in a
snake species to food resources. In this case, the
Australian tiger snakes (Notechis ater serventyi) onNICHE SHIFTS AND SYNDROMES 189