Tropical Forest Community Ecology

232 Walter P. Carsonetal.

Table 13.3 Regions and forest types for studies that
evaluated the Janzen–Connell hypothesis.

Habitat No. of

studies

%

Lowland tropical forest 38 71.7

Tropical forests in

mature floodplain or

swampy areas

2 3.8

Montane tropical forest 1 1.9

Tropical dry forest 2 3.8

Subtropical forest 2 3.8

Temperate deciduous

forest

7 13.2

Temperate arid forest 1 1.9

Table 13.4 Number and percent of species in
different life-history classes that were investigated for
Janzen–Connell effects.

Life-form No. of

species

%

Herbaceous plant 1 0.6

Liana (woody vine) 1 0.6

Emergent tree 2 1.2

Canopy tree 81 47.7

Pioneer tree 2 1.2

Shrub 15 8.8

Species listed as tree

(no indication of

canopy position)

8 4.7

Understory palm 3 1.8

Understory tree 57 33.5

Total 170

suffer stron gpressure from pests (see below).
Furthermore, if enemies are the cause of these
effects then the degree that these enemies are
specialists or facultative specialists needs to be
quantified. In addition, far greater attention needs
to be given to quantifying the relative abundance
of focal species and more importantly to aspects
of plant species life history, particularly seed size,
dispersal mode, degree of shade tolerance, and
overall habitat breadth. It is important to deter-
mine whether species that are kept in check by

their enemies are superior competitors, that is

they have life-history traits that allow them to

establish, form dense stands, and persist for long

periods in the understory. Most importantly, we

argue that the focus needs to be shiftedawayfrom

whether these effects occur and towards the very

difficult task of evaluatin gtheir impact on local

and regional patterns of diversity (see below).

CHALLENGES AND ISSUES

ASSOCIATED WITH TESTING,

EVALUATING, AND FALSIFYING

THE JANZEN–CONNELL

HYPOTHESIS

The Janzen–Connell hypothesis is

ultimately a community-level

hypothesis

As Janzen (1970) concluded, host-specific or

facultatively host-specific seed and seedlin gpre-

dators will decrease tree population density of

a given tree species and/or increase distances

between new adults. Either of these consequences

of predation will lead “to more space in the

habitat for other species of trees and therefore

higher total number of tree species.” Thus, the

most unequivocal tests of Janzen–Connell, as with

Paine’s original test of keystone predation (Paine

1966), will come from studies that experimen-

tally remove enemies or subsets of enemies over

lon gperiods of time and quantify the chan ge

in species diversity. This task, of course, is not

a trivial undertaking. Nonetheless, other major

challenges have been overcome in studies of trop-

ical forests (e.g., establishing and maintaining

50 ha plots).

Janzen (1970, p. 517) proposed five field

experiments or observational studies that would

test predictions of his model. However, none of

these experiments focused on the key predic-

tion, that the exclusion of host-specific predators

would cause a decrease in diversity as tree species

with greater establishment or competitive ability

formed low-diversity seedlin gand saplin gcom-

munities where dominance was concentrated in

a few species. Connell (1971) did propose such an

experiment: “if all enemies of trees were removed