Testing Explanations for Species Diversity using FDPs 113a particular mechanism, even phenomenologi-
cal, will eventually allow researchers to discount
its importance. Once detected phenomenologi-
cally, however, experimental studies are needed to
identify causative factors.
A further limitation of FDPs is a result of the
census method, in which only stems≥1 cm dbh
are censused throughout the whole plot. This
size limit omits the most dynamic size class, the
seedlings, in which the majority of mortality
occurs. The huge number of seedlings precludes
the monitoring of the total seedling population,
but several FDPs now conduct repeated cen-
suses in numerous small plots to sample seedling
dynamics (e.g., Harmset al. 2000, Uriarteet al.
2005b). Most FDPs exclude lianas from their
samples (one exception is the Korup plot, in
Cameroon), because it is almost impossible to
reliably recognize an individual, and most of the
liana biomass is supported up in the tree canopies.
We recognize that lianas may significantly affect
ga pdynamics in some tro pical forests (Schnitzer
et al. 2000, Schnitzer and Bongers 2002).
Finally, there is the “bigfoot effect,” a corollary
of the Heisenberg uncertainty principle (Malakoff
2004): the trampling caused by field workers may
obscure the real forest dynamics (Phillipset al.
2002, Wright 2005). Most FDPs are re-censused
at 5-year intervals or longer, hopefully minimizing
any severe impact. Our observations at Luquillo,
however, suggest that the overall research activ-
ity at FDPs causes significant human disturbance,
morethansuggestedbyaonce-every-5-yearsvisit.
However, it is reassuring that Goldsmithet al.
(2006) found no significant differences in seedling
density and dispersion, height-class distributions,
species richness, evenness, and overall composi-
tion between plots inside and outside the FDP
at BCI. Because the number of plots sampled
was large, the comparison had high statistical
power.
FUTURE RESEARCH IN FDPS
Research from large tropical FDPs in the CTFS
network, particularly large, multi-plot compar-
isons (Conditet al. 2006, Willset al. 2006),
is beginning to distinguish among explanations
for the diversity of tropical forest shrubs and
trees. What role do large tropical FDPs have in
future research? Willset al. (2006) clearly show
that diversity enhancement through time is a
common feature in the FDPs, but cannot dis-
tinguish between the Janzen–Connell hypothesis,
niche complementarity, or the species herd effect
(facilitation). Understanding the precise mecha-
nisms underlying diversity enhancement should
be a focus of future research. Second, some the-
oretical explanations remain untested in large
FDPs (e.g., the storage effect; Chesson 2000).
Detailed studies of monodominant forests in the
tropics (Connell and Lowman 1989), such as that
provided by the Ituri plots (Makanaet al. 2004),
will provide an important perspective on low-
diversity forests, as will complementary studies
of temperate forests (e.g., HilleRisLamberset al.
2002). As a large network of forest plots measured
consistently over time and around the globe, the
plots will be essential in evaluating the impacts
of climate change on tropical forests (Conditet al.
1996, Phillipset al. 2002, Wright 2005). Finally,
human disturbance has fragmented many tropi-
cal forests, and many others are now in secondary
growth (e.g., Grauet al. 2003, Wright 2005).
Studies are needed to determine if there are dif-
ferent mechanisms that promote species diversity
in old-growth forest communities versus second-
growth forests (Thompsonet al. 2002, Thomas
2004) or if it is only the relative importance of the
different mechanisms that has changed with dis-
turbance. Such investigations will be important
in developing management options for human-
disturbed forests. In sum, large FDPs will continue
to be important in refining our understanding of
tropical forest diversity, be it high versus low or
under the influence of natural or anthropogeni-
cally modified conditions.ACKNOWLEDGMENTS
We acknowledge and express our appreciation
to the many people, too numerous to mention,
who have undertaken the huge task of cen-
susing these forest dynamics plots. We thank
Elizabeth Losos, Kyle Harms, Richard Condit, and
two anonymous reviewers for their comments