Tropical Forest Community Ecology

Symmetric Neutral Theory 153

predict depends on whether the metacommunity
is large and continuous, or is an archipelago of
very isolated islands.
But what about the structure of tropical rain-
forest tree communities on large spatial scales?
If the assumption of a large, continuous for-
est is correct, then neutral theory asserts that
Fisher’s logseries should be the distribution of rel-
ative species abundance in the metacommunity,
and not the lognormal. Is this, in fact, correct?
Over the last two decades, a dataset consisting of
288,973 individual tree records has been assem-
bled from a large number of small plots all over the
Amazon basin (ter Steegeet al.2006). The plots
extend throughout the Brazilian Amazon into
Amazonian Colombia, Ecuador, and Peru to the
west, and into the Guianan shield to the northeast.
According to neutral theory, samples aggregated
from many small scattered samples collected over
a large area reduce the impact of dispersal lim-
itation and will better reflect the distribution of
relative abundances for the entire metacommu-
nity. Although many taxonomic problems remain
at the species level with these data, the generic-
level determinations are much more reliable. This
is fortunate because we can test the fit of the
logseries and the lognormal to the abundances
of Amazonian genera. Neutral theory asserts that
generic- and familial-level clades should also obey

the same metacommunity dynamics as species,

the only difference being that they should have

lower rates of origination and extinction than

species do.

Neutral theory predicts that the abundances

of Amazonian tree genera and families should

be distributed according to Fisher’s logseries, and

not Preston’s lognormal, and this prediction is,

in fact, correct (Figure 9.3; Hubbellet al.2008).

The figure shows a tight fit of the logseries with

a value ofθ(Fisher’sα) of approximately 71.

The inset graph shows a Preston-type plot of

species binned into doubling abundance classes.

The flat to pof the Preston curve over many

doubling abundance classes for rare species is

predicted by the logseries in species-rich assem-

blages on large spatial scales, but does not agree

with the pattern predicted by Preston’s lognor-

mal. Given this result, it is highly unlikely that

the species distribution will be a Preston canonical

lognormal.

Perhaps the most challenging test for neu-

tral theory in community ecology is to predict

community dynamics, a much more stringent

test than simply showing that it can fit snap-

shot static data on relative species abundance

(e.g., Volkovet al.2003). Only one or perhaps

two of the large-plot studies of the Center for

Tropical Forest Science have time series that are

0 100

10

1

0.01

0.001

% Relative genus abundance

0.0001

0.00001

200 300

50

No. of genera

40

30

20

(^01234567891011) Log
2 generic abundance
12 13
10
0
400
Genus rank in abundance
500 600
Figure 9.3 Fit of Fisher’s logseries to the relative abundance data of Amazonian tree genera (data provided by ter
Steege). Fisher’sα(θ) is about 71. The rank abundance curve is shown in the full figure, and the Preston-style
histogram of species binned into doubling classes of abundance is shown as the inset graph (after Hubbellet al.
2008). The flatness of the Preston curve for seven doubling classes of abundance at the rare-species end of the curve
is not predicted by the canonical lognormal of Preston (1962). (After Hubbellet al. 2008).