Diamond (1975a) himself suggested that the parallel
was quite a close one, although he noted that while
in the taxon cycle ants colonizing a species-poor
island initially occupy lowland non-forest habitats
and subsequently undergo niche shifts to occupy the
interior, many colonizing birds already occupy low-
land forests. This suggests that the superior average
dispersability of the birds may be important to the
differences in the patterns observed.
There can of course be great differences in dis-
persal abilities withintaxa. This can be deduced for
New Guinea birds by reference to those species
absent from non-land-bridge islands. On the
assumption that a dispersive species should some-
where have found one or more suitable islands in
the non-land-bridge set, only 191 of the 325 low-
land bird species have shown a capacity to cross
water gaps greater than 8 km (5 miles) (Diamond
1975 a). An example of an apparently poor disperser
is the flycatcher Monarcha telescophthalmus, which
occurs on all the large land-bridge islands around
New Guinea, but no others. More direct evidence of
dispersal abilities can be derived from the recolo-
nization of islands such as Long, Ritter, and
Krakatau, which are known to have been defau-
nated, or by direct observations of birds moving
between islands (e.g. Diamond 1975a; Thornton
1996; Thornton et al. 2001). This observational infor-
mation also played a part in the formulation of
island assembly theory.
Chequerboard distributions
In developing island assembly theory, the incidence
functions provided the first step in the reasoning.
The second step relied upon a different form of
distributional patterning, chequerboard distri-
butions, found for pairs of congeneric bird species
in the Bismarcks. A chequerboard (from the game
chequers (US checkers), or draughts) is where two
(or more) species have mutually exclusive but
interdigitating distributions across a series of iso-
lates, such that each island supports only one
species. Several examples were given by Diamond
(1975a). One example is provided by two species of
flycatcher. Pachycephala melanura dahliis found on
18 islands, and its congener P. pectoralison 11, but
they do not occur together on any island. Similar
patterns were found for two species of cuckoo-
doves of the genus Macropygia(Fig. 5.2). Macropygia
mackinlayihas a weight range of 73–110 g and mean
of 87 g, M. nigrirostrishas a range of 73–97 g and
mean of 86 g. Given these values, they might rea-
sonably be expected to have similar resource
requirements. The interpretation offered for these
distributional patterns was that either chance deter-
mined first arrival, or that slight ecological advan-
tages favoured one species over the other on a
particular island. Once established, the resident
would then be able to exclude the congener.
Chequerboards and incidence functions are
different facets of the same distributional data sets.
For instance, Macropygia mackinlayiwas one of the
supertramps identified by the analysis of incidence
functions. Its congener occurs on all the larger
land masses and only a few of the small islands.
Again, Diamond invoked competitive exclusion to
explain the failure to find both species resident on a
single island. It will doubtless have occurred to the
reader that as some islands lack both these species
of cuckoo-dove, and as the identification of
M. mackinlayias a supertramp involves analysis of
entire assemblages, not just other cuckoo-doves,
there must be more to explaining the distributions
of such congeners than simply their relationships to
one another. The data are certainly consistent with
a role for competitive interactions, but other factors
and other biogeographic scenarios (e.g. Steadman
1997 b) cannot be ruled out on the basis of Fig. 5.2.
One means of testing the competition hypothesis
would be provided in the event of one of the pair of
congeners colonizing an island ‘held’ by the other.
According to the interpretation of the incidence
functions, there ought to be an asymmetry in their
competitive interactions, and M. mackinlayishould
generally be the loser. Such observations, unfortu-
nately, are hard to come by.Combination and compatibility—assembly rules for cuckoo-doves
Macropygia mackinlayi, although not overlapping
withM. nigrirostris, does occur on islands with one
of two larger species of cuckoo-dove: namelyISLAND ASSEMBLY THEORY 111