Tropical Forest Community Ecology

Testing Explanations for Species Diversity using FDPs 99

reasonable number of individuals of rare species
in addition to common species. Observing the
changes in the tree spatial arrangement over time
sheds light on the factors that affect a species’
population growth and survival and the outcome
of species interactions (Condit 1995). In order to
study the diversity and ecology of tropical forests
around the world, some researchers have adopted
the approach of using large tropical forest dynam-
ics plots (FDPs), repeatedly censused over the long
term. By comparing the spatio-temporal dynamics
of populations in forest communities observed in
the FDPs with the patterns predicted by different
explanations for community diversity, we may be
able to understand the mechanisms that drive the
community patterns. The research in many tropi-
cal FDPs is coordinated by the Center for Tropical
Forest Science (CTFS) of the Smithsonian Institu-
tion (Losos and Leigh 2004). Most of the FDPs are
50 ha and are usually re-censused every 5 years.
This chapter describes the contribution of these
large tropical FDPs to testing explanations for the
diversity, structure, and composition of tropical
forest tree communities.

ECOLOGICAL EXPLANATIONS

FOR SPECIES DIVERSITY

In this chapter, we focus on three explanations

regarding the interpretation of species diversity,

structure, and community dynamics in tropical

forest that have dominated the literature, and that

have been tested using large FDPs (Chesson 2000,

Wright 2002, Leighet al. 2004). First, we address

neutral theory (Hubbell 2001), an extension of

the theory of island biogeography (MacArthur

and Wilson 1967) that incorporates speciation,

and uses individuals, not species, as the basic

units of propagation. Neutral theory is contro-

versial (Chave 2004, Missa 2005) because, unlike

other theories, it assumes that a tree’s prospects of

death or reproduction are not affected by what

species it is, or what species are in its neigh-

borhood, and thus establishes that differences

among tree species are irrelevant to maintain-

ing tree species diversity. Even if neutral theory

is not an adequate explanation on its own, it does,

however, constitute a null model for community

structure and dynamics, against which other

potential explanations for high species diversity

in ecological communities can be assessed (Chave

2004). We consider two departures from the

neutral theory null model: the first is negative

density dependence (NDD), a tenet of the Janzen–

Connell hypothesis (Janzen 1970, Connell 1971)

and other biological effects that provide for a rare

species advantage. With NDD, unlike in neutral

theory, an individual’s reproductive prospects dif-

fer according to whether it belongs to a species

that is common or rare. The second departure

from neutral theory we consider is the role of

canopy disturbance and gap specialization, and

the degree to which variation in light availabil-

ity allows light-demanding and shade-tolerant

species to coexist via niche partitioning (Connell

1978).

DEVELOPMENT OF THE CTFS

NETWORK OF LARGE FDPS

CTFS plots are located in different biogeographic

regions, and were placed to encompass the

extremes and means of tropical forest envi-

ronments (Ashton 1998, Condit 1998). Rele-

vant environment considerations included total

annual rainfall, its seasonal distribution, soils

and topography, and natural disturbance regimes

(frequency of fires, hurricanes, etc.; Losos 2004).

Most CTFS plots are located at altitudes between

0 and 500 m; they have between 0 and 6 dry

months, and have yellow–red zonal soils rang-

ing from basic to acid. The specific locations of

CTFS plots were dictated by the need to balance

accessibility and available research resources, and

freedom from potential human disturbance. The

CTFS network in Asia has a larger program

because, so far, it has been better funded than

those in Africa and Latin America (Table 7.1).

There are currently 18 FDPs in the CTFS network,

in which scientists are monitoring approximately

three million trees representing 6000 species.

Most plots have completed at least two censuses

(Table 7.1) to allow the study of the static dis-

tribution of trees and also the short-term forest

dynamics.The FDP on Barro Colorado Island (BCI)

in Panama has been censused more times than