Spatial Variation in Tree Species Composition 19they live in, a mechanism that would indirectly
lead to habitat specialization (Fineet al. 2004).
In all these cases, species that have not devel-
oped particular adaptations to cope with the
specific environmental conditions will be unable
to establish a population.
The biotic filter operates through interspecific
competition for available resource. In the resource
use theory developed by Tilman (1980, 1982),
individuals of each species immobilize resources,
for instance nutrients. One species may draw these
resources to below a level at which potential com-
petitors would survive (Tilman’sR∗theory). This
process of resource-driven competitive exclusion
is formally equivalent to that predicted by the
Lotka–Volterra competition theory (Levin 1970).
This theory is compatible with the assumption
that soil environments are a major cause of
the landscape-scale variation in plant distribu-
tions (Ashton 1964, 1976). However, according
to Tilman’s theory, the presence of a species in
a landscape is not fully determined by its envi-
ronmental requirements: the failure to encounter
the species locally may be attributable to either
competitive displacement or samplin glimitation.
Thus, environmental variation may determine the
spatial contours of thefundamentalniche for any
species, but theirrealizedniche is a smaller area
that also depends on competitive displacement by
other species. One important limitation of this the-
ory is that it assumes perfect mixin gin resources,
a reasonable assumption in marine environments,
but likely incorrect in terrestrial environments
(Huston and DeAngelis 1994, Loreau 1998). In
the presence of limited resource mixing, the com-
petitive exclusion principle holds only in a small
area around each individual, and the biologi-
cal filter has only a limited influence on the
presence/absence of a species.
What predictions do niche-assembly theories
make about beta-diversity patterns? The capacity
of a species to exclude other species from its niche
depends on its status within a competitive hier-
archy. At a very small scale (micro-sites), there is
one and only one competitively dominant species,
provided it is present in the local community. If
such a clear competitive hierarchy amon gspecies
does exist, two environmentally similar sites will
tend to be floristically similar. More precisely, the
higher the environmental dissimilarity across the
landscape, the higher the beta-diversity index, a
statement that can be tested with standard statis-
tical approaches. However, this theory makes no
prediction regarding the quantitative relationship
between biological and environmental diversities.Dispersal-assembly theoriesDispersal-assembly theories place an emphasis on
demographic and seed dispersal processes. The
presence/absence of a species within a landscape
is due not to its preference for a specific envi-
ronmental condition, but to the ability of the
plant to reach maturity at one site, havin gdis-
persed from another one. Moreover, few of the
produced seeds are dispersed, and few of the dis-
persed seeds will ever produce a mature plant.
The fundamental premise of this theory is there-
fore that the limited ability of species to colonize
remote habitats should be a major explanatory
factor for their heterogeneous distributional pat-
terns. Species should then be relatively insensi-
tive to environmental differences, so lon gas the
resources are not too limiting. Soil and climate
maps show that huge expanses of tropical land
arerelativelyhomogeneous,andcoveredbyforests
growing typically on oxisols, with rainfall of more
than 1500 mm year−^1 , and experiencin ga dry
season of less than 3 months. Under such condi-
tions, most plant species should be able to persist
in tropical forests irrespective of the abiotic envi-
ronment. Won gand Whitmore (1970) made a
similarproposalintheirstudyof treespeciesdistri-
butions in a rainforest in Peninsular Malaysia. To
examine how well a model with no environmen-
tal constraints could fit to biodiversity patterns,
Hubbell (1979) and Hubbell and Foster (1986)
constructed a neutral theory by assumin gthat all
individuals in all species have the same prospects
of reproduction and death, irrespective of the
environmenttheygrowin.Theyalsoassumedthat
seed rain is not homogeneous at the landscape
scale, but clustered around the parent plants.
Hubbell’s neutral theory makes a number
of quantitative predictions about biodiversity
patterns. Chave and Leigh (2002) produced a
spatially explicit version of the neutral theory for