island populations, but there appears to be good
evidence for their operation for many island line-
ages, providing selective environments that
encourage genetic adjustments in the island popu-
lations. From first principles, we can anticipate that
they will be particularly prevalent on remote
islands at early stages of faunal and floral coloniza-
tion. We go on to consider other niche shift trends
in a later section.
7.3 Character displacement
Character displacement can, and frequently does,
refer to size changes among island forms. What dis-
tinguishes this form of niche shift is the causal
interpretation: that it is competition between two
fairly similar varieties or species that brings about
selection, in one or both, away from the region of
resource overlap (Brown and Wilson 1956).
Diamondet al. (1989, p. 675) define ecological char-
acter displacement as ‘the effect of competition in
causing two initially allopatric species to diverge
from each other in some character upon attaining
sympatry’. Darwin (1859) termed this ‘divergence
of character’, and it has also been termed ‘character
coevolution’ or ‘coevolutionary divergence’,
although there is actually not one, but a suite of
related theoretical ideas centred around this notion
(Otte 1989).
Caribbean anoline lizards provide examples of
character displacement, Schoener (1975) finding
shifts in perch height and diameter consistent with
competitive effects. Species of similar size were
found to affect one another more than dissimilarly
sized species, while larger species affected smaller
ones more than the reverse (cf. Roughgarden and
Pacala 1989). Furthermore, according to a distribu-
tional simulation using Monte Carlo analysis, the
distribution of lizard species over microhabitat
types on satellite islands of the Greater Antilles was
also found to be consistent with competitive effects
(Schoener 1988). However, the results of this analy-
sis depended on the assumptions built into the null
model––a general problem with null models in
ecology and biogeography (Chapter 5).
There is also some evidence for competitive
displacement in Hawaiian crickets. Otte (1989)
presents an array of observations, including data
on song variation in Laupala, a genus of swordtail
crickets. The songs were found to be hyperdis-
persed, the divergence in signals being greater
between coexisting populations than between
allopatric populations, suggesting that competitive
interactions are important. Otte concluded that
character displacement is probably extremely com-
mon in the Hawaiian crickets and, by extension,
elsewhere, although it is difficult to demonstrate
unequivocally. While the most convincing demon-
strations of character displacement tend to be from
systems involving few interacting species (cf. the
anoline lizard studies), such effects may also be
detectable, and certainly are relevant, where a
somewhat larger network of interacting species is
involved.
Given the difficulty in validating ideas of charac-
ter displacement (Connell 1980), it would be ideal if
an experiment were to be devised that controlled
for other complicating factors. The principal prob-
lem is that character changes with time are gener-
ally not observed, but are merely inferred by
comparison of the characteristics of populations in
allopatry with those in sympatry. Such studies pro-
vide corroboration but not validation.
Diamondet al. (1989) have reported a natural
experiment which comes close to solving this prob-
lem, in demonstrating divergence within popula-
tions in sympatry for less than three centuries. In
the mid-seventeenth century, Long Island, off New
Guinea, erupted violently and destructively, initiat-
ing a primary succession on both Long itself and
the nearby islands of Tolokiwa and Crown. By
analogy with the recolonization of Krakatau, it is
reasonable to assume a delay of a few decades for
re-vegetation to proceed to the point at which a for-
est bird assemblage would have developed. Thus
the bird populations of the three islands can be
taken to have been founded less than 300 years ago.
The avifauna contains two different-sized species
of honeyeaters of the genus Myzomela—M. pamme-
laenaandM. sclateri. They are the only islands on
which the two coexist. Allopatric populations on
other islands can be identified as the probable
sources for the founding populations of the Long
group. Thus, the serendipitous ‘experiment’ has a176 ARRIVAL AND CHANGE