144 Stephen P. Hubbell
program in theoretical and empirical ecology.
Using this approach, profound discoveries reveal-
ing the nearly identical molecular machinery of
all life have been made in evolutionary develop-
mental biology and molecular phylogeny. These
discoveries show a remarkable conservatism and
simplicity in the fundamental regulatory con-
trol mechanisms underlying the vast phenotypic
diversity in eukaryotic organisms in the world
today (Carrollet al. 2001, Cracraft and Donoghue
2004, Carroll 2005, Donoghue 2005). Could it
be that a similar conservatism and simplicity
underlie ecological phenomena?
Ecologists may have a hard time answering
this question unless we fundamentally reinvent
our way of doing science. We need to collec-
tively and routinely ask, what elements of our
theories are absolutely essential to explain this
or that phenomenon? What is the simplest set
of assumptions that is sufficient? What assump-
tions are necessary? Ecology still lingers in the
narrative stage of its development, the stage of
collecting case studies of ecological phenomena.
Some ecologists take the almost post-modernist
view that a collection of unique narratives is
all that ecology can ever achieve and that the
search for generality is pointless and quixotic.
Others who believe generality exists do meta-
analyses on the collected case studies to look for
it. However, there are serious problems with most
current meta-analyses in ecology and evolution-
ary biology (Travis 2006). A big problem is that
many of our studies are confirmatory and are
not designed to reject our favorite hypotheses, but
to support them. A few years ago, I attended a
symposium in Japan on indirect effects in com-
munities (e.g., indirect competition sensu Holt
1977). After listening to many papers confirming
the importance of indirect effects in this or that
system, and several meta-analysis talks assert-
ing the nearly universal prevalence of indirect
effects, I asked the question: So, how many studies
were explicitly done on systems in which indirect
effects were not expected to be important? I would
describe the reaction as stunned silence. All one
can honestly say is – of systems picked for study
because researchers expected indirect effects to be
important in them, in a majority the investiga-
tor’s hunch proved to be correct.The only negative
results would occur in surprise cases when the
investigator’s prior hypothesis was not correct.
Confirmatory results get an added boost from the
reluctance of investigators and journals to pub-
lish negative results.The good-faith effort to falsify
hypotheses thus faces a persistent and systematic
bias, a triple-jeopardy handicap at all stages of our
science, from conception, to execution, to publica-
tion. So it is very difficult to assess the significance
of meta-analyses which almost never discuss the
selection criteria used to pick which systems to
study (e.g., McGillet al. 2006). We are all guilty
of these biases. Do we ever counsel our students
to pick an ecological system in which we think the
process they want to study does not occur?
Some of our most cherished hypotheses have
become sacred cows that are virtually imper-
missible to seriously challenge. I have personal
experience of the difficulties of going against the
grain of conventional wisdom in community ecol-
ogy. Nearly 30 years ago, I asked the question,
what would the patterns of relative tree species
abundance in closed-canopy forests be like if they
were determined purely by demographic stochas-
ticity in birth, death, and dispersal rates (Hubbell
1979)? This paper, my first foray into neutral
theory, was a study of a tropical dry forest in
Costa Rica. It unfortunately appeared during the
height of the wars over null community assembly
rules (Stronget al. 1984), ideas that were a logical
outgrowth of the theory of island biogeography
(MacArthur and Wilson 1967).The sacred cow in
question was then, as now, the hypothesis that
ecological communities are “niche-assembled,”
that is, limited-membership, equilibrium assem-
blies of niche-differentiated species, each the best
competitor in its own niche, coexisting with the
other species in competitive equipoise. These wars
were sufficiently off-putting that they delayed any
further serious discussion of neutral theory in
ecology for nearly 20 years. In my own case, I
revisited neutral theory only in the mid-1990s,
when a student in my Princeton biogeography
class asked, why doesn’t the theory of island bio-
geography include a process of speciation, and
what would happen if it did? I did not know, and I
set about finding out.
Neutral theory, as it appeared in my book
(Hubbell 2001) and in two earlier papers,