Testing and Falsifying the Janzen–Connell Hypothesis 233from an entire forest, each species would prob-
ably form small groves and the more rapidly
growing species would gradually spread over the
habitat.... The final result would be a lower pat-
tern diversity and as a consequence fewer species
in any local area of forest.” We suggest that
new studies should now be designed to test the
diversity prediction and thus build upon species-
specific studies that have demonstrated patterns
consistent with the Janzen–Connell model.
There have been a small number of community-
level evaluations of Janzen–Connell effects in the
tropics (Connellet al. 1984, Conditet al. 1992,
Harmset al. 2000, Hubbellet al. 2001, Peters
2003,Willsetal.2006).However,thesestudiesdid
notdirectlytest the diversity prediction, were not
experimental, and did not determine the causes
(e.g., pest pressure versus intraspecific competi-
tion) for patterns found to be consistent with
Janzen–Connell effects. As Wright (2002) pointed
out, “field measurements only demonstrate that
niche differences, Janzen–Connell effects, and
negative density dependence occur. Implications
for species coexistence and plant diversity remain
conjectural.”Is the Janzen–Connell hypothesis a
special case of keystone predation?We suggest that the Janzen–Connell hypothesis
is a type of, or special case of, keystone preda-
tion. To some degree Janzen acknowledged this
in his original paper (Janzen 1970, pp. 502,
522). Janzen wondered how you pack so many
species in a tropical forest. His answer was that his
research was an extension of Paine’s (1966) find-
ings that “local animal species diversity is related
to the number of predators in the system and
their efficiency in preventin gsin gle species from
monopolizing” space or resources. Thus, we sug-
gest the Janzen–Connell hypothesis can be viewed
in this very general context (see also Connell
1971,p.307).Underthekeystonespeciesconcept,
natural enemies limit the abundance of superior
competitors that would otherwise displace subor-
dinate species, thereby enhancin galpha-diversity.
Thus, the suppression of superior competitors
always has the capacity to maintain diversity.
The number of keystone predator species influ-
encin gthe abundance of potentially dominant
prey species (woody species that are superior com-
petitors) does not change the nature of keystone
predation though we acknowledge that classic
ideas about keystone predation did not focus on
rare species advantage. Regardless, what remains
unresolved is: How many plant species would
increase in abundance and depress diversity if
their natural enemies were eliminated?Complex trade-offs underlie the
Janzen–Connell hypothesis: to what
degree are tolerance to pest pressure
and tolerance to low light correlated?Implicit in the Janzen–Connell hypothesis is a
trade-off between establishment or competitive
ability and vulnerability to seed and seedlin gpre-
dation. Janzen (1970, p. 512) pointed out this
trade-off (Janzen 1970, pp. 509, 516, 521) where
large-seeded species are typically more vulnerable
to seed predators or less likely to be produced in
sufficient quantity to satiate predators but have
a greater likelihood of establishing relative to
small-seeded species, particularly in deeply shaded
microsites. Janzen (1970) concluded that “a tree
may persist in the face of very heavy predation if
theoccasionalsurvivingseedlingisaverysuperior
competitor, and a tree with very light predation
may be a very poor competitor yet survive by
repeated trials at establishment.” Connell (1971)
suggested that the trade-off was between vulner-
ability to predation and rapid growth. Regardless,
both suggested that trade-offs likely play a cen-
tral role in how Janzen–Connell effects operate in
tropical forests, an idea that has been recognized
by others (Wright 2002, Leighet al.2004).
In both temperate and tropical forests, there
appears to be a continuum of species from pio-
neers that have rapidly growing saplings in high
light to mature forest species that have saplings
that persist for years in the shaded understory
(e.g.,Wright2002,Pacalaetal.2003,Wrightetal.
2003, Leighet al.2004). Typically, species that
are classified as shade tolerant have seedlings and
saplings with a suite of correlated traits (estab-
lishment within shaded understories, dense wood,