Testing and Falsifying the Janzen–Connell Hypothesis 237explanations originally pointed out by Connell
(1971, e.g., allelopathy and intraspecific competi-
tion; see also Wright 2002). Additionally, Connell
(1971) argued that predation was more critical
durin gthe seedlin gand early saplin gsta ge and
not durin gthe seed sta ge (but see Harmset al.
2000). Because there is little evidence for com-
petition amon gseedlin gs of tropical trees even
at high densities (Paineet al. in review), experi-
ments may need to be conducted for many years
up to and through the sapling stage as large and
dense understory layers begin to thin (Chazdon
Chapter 23, this volume). Furthermore, a Janzen–
Connell effect may be driven by periodic outbreaks
of specialist insects in the understory or overstory
(Wonget al. 1990, Carsonet al. 2004). If these
outbreaks drive the Janzen–Connell effect, then
studies will need to run through a typical outbreak
cycle, which for the vast majority of species will
almost certainly be lengthy.
There are additional serious logistical and con-
ceptualchallengesthatimpedetesting theJanzen–
Connell hypothesis. These difficulties include:
(1) identifyin gthe key enemies to exclude in
experimental tests; (2) the ability to remove or
substantively reduce entire or even partial enemy
trophic levels for lon gperiods of time (patho gens
are a particular hurdle); (3) directly linkin gdis-
tance and/or density effects to their putative
causes (i.e., enemies); and (4) demonstratin gthe
entire chain of events necessary, specifically that
enemies respondin gin a distance- or density-
dependent fashion reduce the abundance of puta-
tive superior competitors and thereby cause an
increase in alpha-diversity.
Overall, rejectin gthe Janzen–Connell model is
extremely difficult. The key experiments testing
the prediction of higher diversity remain to be
done and will require large-scale, logistically dif-
ficult, long-term studies. An important but more
modest step for testin gJanzen–Connell is to link
repelled recruitment syndromes to life-history
strategies or traits that lead to high survivorship
in low light in the absence of enemies. Clearly,
theoretical approaches that incorporate results
from field studies will also have to play a major
role in evaluatin gthe relative role of various
diversity promotin gmechanisms includin gthe
Janzen–Connell hypothesis.
For the Janzen–Connell hypothesis, if only a
relatively small number of tree species are kept
in check by their enemies (5–20%), then this
may explain why these tree species do not spread
to exclude others, but it still may not explain
the coexistence of many additional species in
these species-rich communities (see, e.g., Hubbell
1980, Penfold and Lamb 1999).Thus the Janzen–
Connell hypothesis would be a necessary but
insufficient explanation of tree diversity. Still,
Janzen–Connell may readily promote coexistence
in combination with other processes necessary to
explain hyper-diverse communities (e.g., Wright
2002, Barot and Gignoux 2004, Leighet al.
2004). We agree completely with Barot and
Gignoux (2004) who concluded the critical task is
to “assess the respective influence of each mecha-
nism [of coexistence] for different communities...
and determine the main processes that shape their
biodiversity.”ACKNOWLEDGMENTS
We thank Tim Nuttle and Liza Comita for com-
ments on earlier drafts. This work was supported
by NSF grants to Walter Carson.REFERENCES
Alvarez-Buylla, E.R. (1994) Density dependence and
patch dynamics in tropical rain forests: matrix models
and applications to a tree species.American Naturalist
143, 155–191.
Andow, D.A. (1991) Vegetational diversity and arthro-
pod population response.AnnualReviewof Entomology
36, 561–586.
Augspurger, C.K. (1983) Seed dispersal of the tropi-
cal tree,Platypodium elegans, and the escape of its
seedlings from fungal pathogens.Journal of Ecology
71, 759–771.
Augspurger, C.K. (1984) Seedling survival of tropical
tree species – interactions of dispersal distance, light-
gaps, and pathogens.Ecology65, 1705–1712.
Augspurger, C.K. and Kelly, C.K. (1984) Pathogen mor-
tality of tropical tree seedlings – experimental studies
of the effects of dispersal distance, seedlin gdensity,
and light conditions.Oecologia61, 211–217.