Tropical Forest Community Ecology

Symmetric Neutral Theory 145

published obscurely (Hubbell 1995, 1997), was
the result, also discovered independently by
Graham Bell (2000, 2001). The obscurity of my
first two papers was not by design. I tried to pub-
lish my second paper, which ultimately appeared
inCoral Reefs(Hubbell 1997), in many prominent
places, includingNature,Science,Proceedings of the
National Academy of Sciences of the United States
of America,American Naturalist, andEcology, all
of which rejected the manuscript, three with-
out review (“not of sufficient interest”), and two
with very brief and very chilly reviews. Excerpts:
“Ecology is not ready for yet another null model of
community assembly. Let sleeping dogs lie.” “Any
theory [of community assembly] based on such
an obviously false assumption does not merit pub-
lication here, or anywhere else.” “If this paper
is published, I will never review for this jour-
nal again.” “Has Hubbell lost his mind? I don’t
understand the math, but his conclusions are
anti-intuitive and must be wrong.” It is perhaps
ironic that these and many other leading journals,
collectively, have published a very large number of
papers on neutral theory in ecology since 2001,
including special issues in some journals fully
devoted to the subject. I do not think the edi-
tors ofCoral Reefswanted to publish the paper
either, but I was one of the keynote speakers for
the Eighth International Coral Reef Symposium
in Panama in 1996, and the editors had agreed
in advance to publish the keynote addresses. The
take-home message of this experience: if you
can’t get your iconoclastic paper accepted in a
peer-reviewed journal, expand its scope and write
a book.
In my synopsis of some of the high points
of neutral theory here, I could proceed histor-
ically with how the ideas originally developed,
but I have chosen instead to develo pthe theory
conceptually in steps of increasing complexity,
starting from the simplest possible model. This
approach draws on some major advances in the
theory that occurred after my book was published
(Hubbell 2001). I was fortunate that my book
attracted the attention and interest of a number
of brilliant statistical physicists who have con-
siderably improved and generalized the theory I
originally presented. Because this chapter is not a
review of all the developments in neutral theory

since 2001, I apologize to the many people with

papers on the subject to which I do not refer.

THEORETICAL RECIPE: START

SIMPLY, ADD AS FEW FREE

PARAMETERS AS POSSIBLE, STIR

VIGOROUSLY

Afree parameteris a number in a theory that can-

not be derived from the theory itself. In my view,

the best theories are those which make the largest

number of testable predictions per free param-

eter, a qualitative judgment in the spirit of the

Akaike information criterion (AIC), which penal-

izes models the more free parameters they have.

Neutral theory is attractive because it has very

few free parameters, all of which have ready bio-

logical interpretations. Yet from its small set of

free parameters, neutral theory leverages a large

number of predictions about diverse phenomena

in community ecology and biogeography. Here I

limit discussion to its predictions for the static

and dynamic patterns of relative species abun-

dance, the subject receiving the most attention in

recent years, and I will confront these predictions

with data on tropical tree communities. Before

doing so, however, there is a rather long theo-

retical preamble, which I feel is necessary but for

which I apologize to readers more interested in the

biological punch line.

The simplest possible neutral model of rela-

tive species abundance is to imagine a large,

self-contained and homogeneous, biogeographic

region in which the only processes at work are

speciation and extinction. In neutral theory, this

is called themetacommunity. The metacommu-

nity is the evolutionary–biogeographic unit in

which most member species live out their entire

evolutionary lifespans. In our bare-bones model,

let us for the moment ignore almost all of our

favorite ecological processes such as density- and

frequency dependence, niche differences, disper-

sal limitation, and so on. Let us further imagine

that the metacommunity is in aninterregnum

period of steady-state species richness between

evolutionary punctuational events, so that the

speciation and extinction rates are in balance.

What patterns of relative species abundance do