Testing for Janzen–Connell Effects in a West African Montane Forest
, Hazel Chapman
,and Dave Kelly
School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand
The Janzen–Connell hypothesis proposes that density dependent seed and seedling mortality, combined with increasing seed and seedling survival away from the
parent tree, together promote regular spacing of species and thus adiversity. This hypothesis has rarely been tested in tropical Africa, and rarely in montane forests
anywhere. We tested this hypothesis using a combination of ﬁeld experiments and observations in the most ﬂoristically diverse dry submontane forest in Nigeria. We
investigated distance effects on seedling herbivory, seedling survival and seedling height growth. We found a signiﬁcant decrease in herbivory with distance from
conspeciﬁc adult trees for all three species of experimentally planted seedlings (Entandrophragma angolense,Deinbollia pinnata and Sterculia setigera), and also for
naturally occurring seedlings of Pouteria altissima but not of Newtonia buchananii or Isolona pleurocarpa. The relative density of large seedlings/saplings of P. altissima,
N. buchananii and I. pleurocarpa increased signiﬁcantly at greater distance from conspeciﬁc adult trees; however, we found no signiﬁcant distance effect on survival or
height growth over 3 mo for all three experimentally planted species. Taken together, our results are some of the ﬁrst to show that Janzen–Connell effects occur on the
African continent and in particular montane tropical forest and suggest that such effects may be pantropical.
Key words: Afromontane forest; density dependence; insect herbivory; species diversity.
ONE OF THE MOST FASCINATING QUESTIONS IN ECOLOGY has been what
drives the high species diversity of tropical forests. The processes
responsible remain largely unresolved but one of the main con-
tenders is a hypothesis based on negative density dependence (Jan-
zen 1970, Connell 1971, Freckleton & Lewis 2006, Comita &
Hubbell 2009). The most tested and compelling explanations focus
on the role of host-specific, natural plant enemies which, by de-
stroying seeds and seedlings, create space for the establishment of
other species and so contribute to the diversity of the forest. First
articulated by Janzen (1970) and Connell (1971), this model
(called the Janzen–Connell hypothesis) proposes that density de-
pendent seed and seedling mortality, combined with increasing
seed and seedling survival away from the parent tree, together pro-
mote regular spacing of species. Host-specific seed and seedling
predation (Schupp 1992), herbivory (Basset 1992, Novotny &
Basset 2005) and/or pathogen attack (Freckleton & Lewis 2006,
Augspurger & Wilkinson 2007) drive these mechanisms.
The complex set of assumptions, predictions and hypotheses
generated by the Janzen–Connell hypothesis has been tested more
than 50 times with variable results (Carson et al. 2008). There is,
however, compelling support for the density and distance predic-
tions of the hypothesis from both lowland tropical forests (Clark &
Clark 1984, Burkey 1994, Hammond & Brown 1998, Romo et al.
2004) and temperate forests (Packer & Clay 2000, Lambers et al.
2002). Other studies, mainly from Central and South America, do
not support the hypothesis (e.g., Hubbell 1979, 1980; Schupp
Givnish (1999) stressed the need to evaluate whether Jan-
zen–Connell effects are important in different forest types and Ba-
rot and Gignoux (2004) suggested that Janzen–Connell effects may
vary among different communities. In a review of tests of the Jan-
zen–Connell hypothesis, Carson et al. (2008) found only 1 (Wenny
2000) of 53 studies had been carried out in a tropical montane for-
est and the vast majority of studies had been conducted in the neo-
tropics within just a small number of forest types and locations.
Here, for the ﬁrst time, we test for the presence of Janzen–Connell
effects in a dry, relatively species-rich, and relatively undisturbed
African submontane forest. We focused on the role of insect herb-
ivory because there is good evidence for host-speciﬁc invertebrate
herbivores in tropical forests (Janzen 1980, Augspurger 1984, Clark
& Clark 1987, Howe 1990, Gilbert et al. 1994, Norghauer et al.
Speciﬁcally, we sought to test two predictions of the Jan-
zen–Connell hypothesis using six relatively common forest tree
species, with a mixture of manipulative and observational methods:
(1) seedling survival increases with distance from conspeciﬁc adult
trees and (2) seedling leaf herbivory decreases with distance from
conspeciﬁc adult trees.
STUDY SITE AND SPECIES.—Ngel Nyaki Forest Reserve includes ca
7200 ha of dry, submontane forest along the escarpment of the
Mambilla Plateau, located on the Nigerian/Cameroon border in
Taraba State, Nigeria. The forest lies on a steep southwest facing
slope at 1400–1600 m asl. The mean annual rainfall of ca 1800 mm
occurs mainly between mid-April and mid-October (Nigerian
Montane Forest Project Rainfall data). Mean monthly maximum
and minimum temperature for the wet and dry season are 261C and
131C, and 231C and 161C, respectively (Upper Benue River Basin
Sixty-nine tree species have been recorded from Ngel Nyaki
forest, which is the most ﬂoristically diverse forest of its type in
Received 20 September 2009; revision accepted 8 April 2010.
Corresponding author; e-mail: firstname.lastname@example.org
Current address: Bioprotection Division, Department of Primary Industries,
PO Box 48, Frankston, Victoria, Australia
BIOTROPICA 43(1): 77–83 2011 10.1111/j.1744-7429.2010.00664.x
r2010 The Author(s) 77
Journal compilation r2010 by The Association for Tropical Biology and Conservation
Nigeria (Chapman & Chapman 2001). The three main emergent
species are Pouteria altissima (Sapotaceae), Entandrophragma ango-
lense (Meliaceae) and Newtonia buchananii (Leguminosae:Mimoso-
idae; Dowsett-Lemaire 1989) and all three were included as study
species in this research. The three other tree species investigated
were Isolona pleurocarpa (Annonaceae), Deinbollia pinnata (Sap-
indaceae) and Sterculia setigera (Sterculiaceae).
Entandrophragma angolense,D. pinnata and S. setigera were
used to test survival and leaf herbivory of artiﬁcially raised and
planted seedlings at increasing distance from parent/conspeciﬁc
adult trees, because their seeds were readily available at the begin-
ning of the study period (end of February 2005) and because they
are relatively common in Ngel Nyaki forest. Pouteria altissima,N.
buchananii and I. pleurocarpa were used to test the survival and leaf
herbivory of naturally occurring seedlings at increasing distance
from parent/conspeciﬁc adult trees because their seedlings are mor-
phologically distinct and thus can be identiﬁed with conﬁdence.
Moreover, the seedlings of all three species are common in the forest,
especially P. altissima that produces dense understories of seedlings.
SURVIVAL,GROWTH AND LEAF HERBIVORY OF ARTIFICIALLY RAISED AND
PLANTED SEEDLINGS.—The seedling survival and herbivory hypoth-
eses were both tested using experimental plantings of a minimum of
540 seedlings of each of the three species E. angolense,D. pinnata
and S. setigera. Seeds were collected from at least ﬁve individual
trees and then sown in polypots in forest soil and grown in a forest
nursery to an average height of 15 cm. The seedlings were then
planted out in the forest in rows of ten individuals 20 cm apart, at
1 m, 12.5m and 25 m from conspeciﬁc adult trees. Before planting,
each seedling was scored for leaf loss by visually estimating the per-
cent of the total leaf area of a seedling that was missing in 5 percent
intervals. The loss of whole leaves was included in the estimate of
whole leaf tissue loss. Loss of whole leaves was evident from leaﬂess
petioles (Stotz et al. 2000). The number of replicates for each spe-
cies was limited by the number of surviving potted seedlings, so our
sample sizes were N= 18, 19 and 20 conspeciﬁc adult trees for S.
setigera,D. pinnata and E. angolense (540, 570 and 600 seedlings),
respectively. The individual conspeciﬁc adults were chosen so that
they were at least 50 m from any conspeciﬁcs to avoid any overlap-
ping of seed shadows and the potential sharing of seedling predators
among them, as predicted by the Janzen–Connell model (Janzen
After planting in the forest, seedling survival and heights were
measured and leaf herbivory was estimated (as described above)
once a month for 3 mo. Seedling survival to month 3 and seedling
height growth from month 1 to 3 were used to measure plant per-
formance following the method of Howe (1990), while amount of
herbivory at month 3 measured herbivore impacts. Remains of pet-
ioles, which indicated that whole leaves had been lost, were in-
cluded in the herbivory estimate. Seedlings were measured in the
same order they were planted in to keep time intervals between
measurements as consistent as possible.
HEIGHT AND LEAF HERBIVORY OF NATURALLY OCCURRING SEEDLINGS.—
The ﬁrst hypothesis (seedling height as an estimate of survival
increases with distance from conspeciﬁc adult trees) was tested us-
ing naturally occurring seedlings around 20 adult trees of each of
the three tree species P. altissima,N. buchananii and I. pleurocarpa.
As in the previous experiment, all of the conspeciﬁc adult trees un-
der which seedlings were measured were at least 50 m apart. None
of the adult trees were located near light gaps.
All naturally occurring seedlings were measured within plots at
ﬁve distance intervals away from the adult trees (0–5, 5–10, 10–15,
15–20 and 20–25 m). We used a maximum distance of 25 m away
from the adult tree because at greater distances it became difﬁcult to
ensure that plots were not closer to a different conspeciﬁc. At the
ﬁrst three distance intervals (0–5, 5–10 and 10–15 m), 5 5-m
plots were used. At 15–20 m distance, plot size was increased to
510 m, and at 20–25 m distance, further increased to 515 m,
to obtain a better representation of seedling height distributions
where conspeciﬁc seedling densities are lower. Within the
plots, seedling height classes were used as a proxy for seedling age
(Webb & Peart 1999). Seedling heights were measured to the near-
est centimeter and later grouped into height categories, depending
on the relative abundance of seedlings within each species (see
The second hypothesis (seedling leaf herbivory decreases with
distance from conspeciﬁc adult trees) was tested in these same plots
by visually estimating seedling leaf herbivory as described above
(Stotz et al. 2000) on up to 20 seedlings in each plot. Although this
method has limitations (Clark & Clark 1985), it has been shown to
be effective in ﬁeld situations (Sessions & Kelly 2001).
SURVIVAL AND HERBIVORY OF EXPERIMENTALLY PLANTED SEED-
LINGS.—Seedling survival was analyzed by calculating the propor-
tion of the ten seedlings in a row that survived to month 3 at each
distance (1, 12.5 and 25 m). Any differences between distances in
seedling mortality were expected to be greatest after 3 mo when
they had been exposed to mortality factors for the maximum time
available in the experiment. We used one generalized linear mixed
model (GLMM) with a binomial distribution and a logit link
function to test for effects of species, distance from conspeciﬁc
adult and the species distance interaction. The R statistics pro-
gram (v. 2.10.1; R: A Language and Environment for Statistical
Computing 2009) with library lme4 was used for all analyses.
Parent tree was entered as a random effect. If Janzen–Connell
effects are consistent, we would predict a signiﬁcant positive dis-
tance effect on survival and no signiﬁcant species distance in-
Seedling height growth in surviving seedlings was used as a
second estimate of seedling performance (Howe 1990). A second
GLMM was used to detect whether a signiﬁcant proportion of the
variance in height increase from month 1 to 3 was explained by
species, distance and the interaction species distance, again using
parent tree as a random effect. The GLMM was run with a Gauss-
ian distribution and an identity link function. Probability values for
the parameters from the Gaussian GLMM were estimated using
Monte Carlo Markov chain (MCMC) methods implemented in
the ‘pvals.fnc()’ command in the ‘languageR’ library for R.
78 Matthesius, Chapman, and Kelly
Similarly, the mean percentage leaf area per plot lost to herb-
ivory at month 3 was also investigated using a Gaussian GLMM.
We tested the proportion of leaf area lost to herbivores (arcsin
square root transformed) against species, distance and species dis-
tance with parent tree as a random effect, and MCMC probability
values from pvals.fnc().
HEIGHT AND LEAF HERBIVORY OF NATURALLY OCCURRING SEED-
LINGS.—Tests for changes in the relative distributions of various-
sized seedlings with distance used linear mixed models to test the
proportion of seedlings in each plot which were in the largest size
categories. The three species varied in the numbers of seedlings in
different size classes, so the size classes were chosen for each species
to ensure that each size class contained reasonable numbers of seed-
lings. We initially used four size classes for P. altissima (0–10,
11–20, 21–40, 41–200cm) and I. pleurocarpa (0–20, 21–40,
41–70, 71–200 cm), and three for N. buchananii (0–10, 11–20,
21–200cm). For analysis, in all species, this was subsequently sim-
pliﬁed to the proportion of total seedlings that were in the largest
size class. In each case, the same ﬁve distance categories were retained
(0–5, 5–10, 10–15, 15–20 and 20–25m). For each species, a bi-
nomial GLMM run using lme4 in R tested for an effect of distance
on the proportion of seedlings per plot which were in the largest size
class. As before, parent tree was entered as a random effect.
A general linear model (GLM) was used to investigate the
effect of distance from the parent tree on the average percent leaf
herbivory of each species of seedlings. The response variable was the
mean proportion lost to herbivory for all the seedlings in each plot,
and the data were arcsin square root transformed before analysis.
Parent tree was included as a block effect in the GLMs. We did also
run these analyses as linear mixed models using lme4 (treating par-
ent as a random effect rather than a block effect) and the conclu-
sions were identical.
GROWTH AND HERBIVORY OF EXPERIMENTALLY PLANTED SEEDLINGS.—
The proportion of the planted seedlings of all three species that
survived decreased over the 3-mo period. By the end of the 3 mo
across all three species, there were signiﬁcant differences among
species (Table S1) but there was no overall trend in the linear mixed
model for an increase in survival with increasing distance from the
conspeciﬁc adult trees (distance was nonsigniﬁcant, Table S1).
There was one modestly signiﬁcant species distance interaction,
for S. setigera (P= 0.038) but with a small negative slope (Fig. 1A),
contrary to predictions. Similarly, seedling height growth showed
no support for Janzen–Connell effects, with signiﬁcant variation
among species but no variation with distance, and no species dis-
tance interactions (Table S1; Fig. 1B). Herbivory decreased signiﬁ-
cantly for all three species (D. pinnata,S. setigera and E. angolense)
with distance from conspeciﬁc adults. There was no signiﬁcant spe-
cies distance interaction (Table S1; Fig. 1C).
HEIGHT AND HERBIVORY OF NATURALLY OCCURRING SEEDLINGS.—A
total of 4031 P. altissima, 4037 N. buchananii and 383 I. pleuro-
carpa seedlings were measured within all of the plots. As predicted
by the Janzen–Connell hypothesis, the proportion of taller seed-
lings increased signiﬁcantly with distance from conspeciﬁc adults
FIGURE 1. The performance of experimentally planted seedlings (mean
SEM) at three distances away from conspeciﬁc adult trees (1, 12.5 and 25 m)
for Deinbollia pinnata (Dein, circles), Entandrophragma angolense (Enta, trian-
gles) and Sterculia setigera (Ster, squares). (A) Proportion of seedlings surviving
to 3mo, (B) seedling height growth from month 1 to 3, and (C) mean seedling
herbivory (% leaf area loss) at month 3.
Testing Janzen–Connell Model in West Africa 79
for all three species (Fig. 2). The larger seedling size classes were
much more common at greater distances from the parent, and after
allowing for parent tree in the linear mixed models, there were
highly signiﬁcant effects of distance in all three species (Table 1). In
each case, the sign was positive, i.e., there was an increase in the
relative abundance of larger seedling classes with increasing dis-
tance, consistent with higher survival of seedlings further away.
The average percent leaf area lost to herbivory in P. altissima
seedlings varied signiﬁcantly among parent trees and decreased sig-
niﬁcantly with distance from parent trees (Table 2). The mean
herbivory reduced from a ﬁtted value of 34.9 percent of leaf area at
2.5 m to 29.0 percent of leaf area at 22.5m distance from the parent
tree (Fig. 3). For N. buchananii, there was a signiﬁcant parent effect
but no distance effect, and in I. pleurocarpa neither the parent or
distance effects were signiﬁcant (Table 2).
EVIDENCE FOR JANZEN–CONNELL EFFECTS.—Our results are the ﬁrst
tests of the Janzen–Connell hypothesis from an Afromontane forest
that we are aware of. As such, they make a valuable contribution to
the understanding of latitudinal gradients in species richness.
As in many studies from lowland tropical and, to a lesser ex-
tent, temperate forests (Carson et al. 2008), our results were mixed
in terms of support for the Janzen–Connell predictions we tested.
Our strongest supporting evidence came from the natural popula-
tions of P. altissima in which both predictions: (1) increased pro-
portion of larger seedlings and (2) decreased herbivory, with
distance from parent tree were fulﬁlled. In the case of naturally oc-
curring I. pleurocarpa and N. buchananii while prediction (1) was
supported, no signiﬁcant distance effects on levels of seedling herb-
ivory were found. Similar mixed results from studies using compa-
rable methodologies in a range of studies around the world are
summarized in Carson et al. (2008). Several tropical studies (Augs-
purger 1984, Clark & Clark 1984, Webb & Pert 1999, Harms
et al. 2000) found increasing numbers of large seedlings (and/or
seedling survival) with distance from the parent tree was consistent
across species. Other tropical studies, however, have found such
patterns to be species (Chapman & Chapman 1996, Cintra 1997)
or scale (Queenborough et al. 2007) dependent. They are far less
common in temperate forests (Hyatt et al. 2003). Decreased host-
speciﬁc herbivory with distance from the parent tree has been re-
corded many times before in the tropics (Janzen 1980, Augspurger
1984, Clark & Clark 1987, Howe 1990, Gilbert et al. 1994, Nor-
ghauer et al. 2006) and it was surprising to us that this was not the
case for all three of our naturally occurring seedling species, partic-
ularly in view of the fact that survival of larger seedlings was appar-
ently higher with increasing distance away from the parent tree.
Decreasing herbivory with distance, however, was found in all three
species of the experimentally planted seedlings (D. pinnata,E. an-
golense and S. setigera). Possible explanations for why we did not
ﬁnd evidence of decreased herbivory with distance for I. pleurocarpa
and N. buchananii are discussed below.
In the case of the planted seedlings, there was no signiﬁcant
distance effect on seedling survival or height growth with distance
from conspeciﬁc adults over the 3 mo of the experiment, but as
mentioned above, there was a signiﬁcant decrease in herbivory with
FIGURE 2. The percentage (mean SEM) of all naturally occurring seedlings
of (A) Isolona pleurocarpa, (B) Pouteria altissima and (C) Newtonia buchananii
which were in the largest height class (71–200, 41–200 and 21–200cm, respec-
tively, see text) at various distances (0–5, 5–10, 10–15, 15–20 and 20–25 m)
from the parent tree. There was a signiﬁcant increase in the percentage of large
seedlings at greater distances for all three species (see Table 1).
80 Matthesius, Chapman, and Kelly
increasing distance from the parent tree for all three species. The
apparent contradiction between lower herbivory with distance, yet
no distance effect on height growth or seedling survival, might be
due to the short duration of the planting experiment (3 mo) not
leaving time for herbivory to affect seedling vital rates. The fact that
the three species included in the experiment (D. pinnata,E. ango-
lense and S. setigera) are so different in terms of ecology and yet all
failed to show demographic effects of herbivory makes this a pos-
sible explanation. Alternatively, it is possible that microhabitat
preferences could negate the beneﬁts of reduced herbivory away
from the parent tree. For example, farther-away seedlings may be
outside their optimal habitat, or light gaps could obscure Jan-
zen–Connell effects (Clark & Clark 1984, 1987; Cintra 1997;
Cintra & Horna 1997). Light gaps have been shown to locally re-
duce pathogen loads (Augspurger 1984) and may allow for com-
pensatory growth in response to herbivory relative to seedlings
outside of gaps (Blundell & Peart 2001). While light gaps were not
measured in this study, there was environmental heterogeneity
within the experimental plots. However, it may be that leaf herb-
ivory does not affect survival and growth in these species (Blundell
& Peart 2001); Chapman and Chapman (1996) found a distance-
survival response in experimentally planted seedlings in Uganda to
be species speciﬁc.
The lack of distance effect of herbivory on seedlings of N. bu-
chananii and I. Pleurocarpa suggests that their apparently higher
survival at greater distances may have been driven by other density/
distance factors such as pathogens (Augspurger 1984, Coley & Bar-
one 1996, Bell et al. 2006, Norghauer et al. 2008) or intraspeciﬁc
competition (Comita & Hubbell 2009). Possible nondensity de-
pendent explanations include, for example, allelopathic affects be-
tween conspeciﬁc parent trees and seedlings (Janzen 1970, Connell
1971, Wright 2002) and leaf litter under a conspeciﬁc adult (Mo-
lofsky & Augspurger 1992) can reduce survival rates in conspeciﬁc
seedlings. In this study, litter may have contributed to seedling
death under parent trees of P. altissima. Seedlings of this species
were often absent from areas with heavy leaf litter (A. Matthesius,
pers. obs.) and P. altissima does periodically shed all of its leaves
(Chapman & Chapman 2001).
Alternatively, the possibility exists that we simply did not de-
tect density dependent host-speciﬁc leaf herbivory in N. buchananii
and I. Pleurocarpa because we measured herbivory at a single point
in time. This approach could hide the pattern predicted by Jan-
zen–Connell effects for some species simply because host-speciﬁc
seedling predators were not abundant at the time of the experiment
(Connell 1971) or because their host seedlings were in short supply
TABLE 1. Generalized linear mixed model analyses of the effect of distance away
from the parent tree (0–25 m) on the proportion of plants in each plot
which were in the largest size class for Isolona pleurocarpa,Pouteria
altissima and Newtonia buchananii. In each case the analysis was a
linear mixed model with parent plant as a random effect and distance as
a ﬁxed effect, using a binomial error distribution with a logit link.
Fixed effect Estimate SE Z-value P
Intercept –2.7189 0.4250 –6.397 o0.001
Distance 0.0810 0.0234 3.456 o0.001
Intercept –4.712 0.5285 –8.916 o0.001
Distance 0.0910 0.0155 5.860 o0.001
Intercept –5.0588 0.6215 –8.140 o0.001
Distance 0.1075 0.0210 5.126 o0.001
TABLE 2. GLM analyses of effects of distance from the parent tree on herbivory in
naturally occurring seedlings, in Isolona pleurocarpa,Pouteria altissi-
ma and Newtonia buchananii. The GLMs used parent plant as a block
effect and distance (0–25 m) as a variate, in Gaussian GLMs with mean
percentage leaf loss per plot (arcsin square root transformed) as the re-
Predictor Deviance df F P
Parent tree 0.419 19 1.22 0.272
Distance 0.0032 1 0.179 0.674
Residual 1.07 59
Parent tree 1.33 19 8.54 o0.001
Distance 0.0495 1 6.00 0.0165
Residual 0.646 79
Parent tree 0.565 20 5.15 o0.001
Distance 0.0007 1 0.131 0.718
Residual 0.368 67
FIGURE 3. The change in mean percent leaf area lost to herbivory per plot
against distance from conspeciﬁc adult (0–5, 5–10, 10–15, 15–20 and 20–25 m)
for Pouteria altissima. Shown are residuals from a GLM ﬁtting a parent plant
effect. The distance effect was signiﬁcant in a GLM with parent plant ﬁtted ﬁrst
(see Table 2).
Testing Janzen–Connell Model in West Africa 81
at some point in time before making the measurements (Janzen
1970). Another difﬁculty in determining herbivory rates at one
point in time in the observational datasets is that there is no infor-
mation on leaf turnover, which makes accurate determination of
the time scale over which damage has accumulated very difﬁcult
(Coley & Barone 1996) (this was less of a problem in our 3-mo-old
planted seedlings). In addition, not all insects leave obvious evi-
dence of leaf consumption (e.g., phloem feeders; Coley & Barone
1996) and we may have sometimes underestimated herbivory
through loss of whole leaves in cases where the petiole was also lost.
In summary, while our results were mixed, we did ﬁnd some
Janzen–Connell effects in Ngel Nyaki forest, which suggests that
such effects cannot be discounted in our understanding of species
diversity in Afromontane forests more generally. As pointed
by Carson et al. (2008), Janzen–Connell effects on just a few of
the common species are sufﬁcient to lead to increased species
FUTURE WORK.—Considering the paucity of tests of the Jan-
zen–Connell hypothesis in African forests, this is an important ﬁrst
step. More African studies need to be undertaken so that more di-
rect comparisons between latitudes can be made. Such investiga-
tions will need to be longer term and on a larger scale than this
present study. They should include more life stages such as seeds as
well as seedlings.
Research to determine whether natural enemies (both herbi-
vore and pathogen) inﬂict density dependent damage and mortality
are required. Manipulative experiments, along the lines of Freckle-
ton and Lewis (2006) and Augspurger and Wilkinson (2007) will
be important. Such manipulations should involve the addition and
removal of pests to communities over sufﬁcient time periods for
their absence or increased abundance to have an effect (Carson et al.
2008). Only once such studies are available will it be really possible
to include Africa in latitudinal comparisons of distance/density
effects on species diversity.
We thank U. Usuf and M. Zubairu and other staff of the Nigerian
Montane Forest Project for invaluable ﬁeld assistance. Very helpful
comments on this manuscript were gratefully received from W. P.
Carson, the editor, and two anonymous reviewers. We are grateful
to the Taraba State Forest Service for their invitation and permis-
sion to conduct research. Financial support came from The North
of England Zoological Society, Nexen Inc. and the A. P. Leventis
Additional Supporting Information may be found in the online
version of this article:
TABLE S1. GLMM mixed model analyses of means of experimen-
tally planted seedling survival, height growth and seedling herbivory at
three distances away from conspeciﬁc adult trees for Deinbollia pin-
nata, Entandrophragma angolense and Sterculia setigera.
Please note: Wiley-Blackwell is not responsible for the content
or functionality of any supporting materials supplied by the au-
thors. Any queries (other than missing material) should be directed
to the corresponding author for the article.
AUGSPURGER, C. K. 1984. Seedling survival of tropical tree species: Interactions
of dispersal distance, light gaps and pathogens. Ecology 65: 1705–1712.
AUGSPURGER,C.K.,AND H. T. WILKINSON. 2007. Host speciﬁcity of pathogenic
Pythium species: Implications for tree species diversity. Biotropica 39:
BAROT, S., AND J. GIGNOUX. 2004. Mechanisms promoting plant coexistence:
can all the proposed processes be reconciled? Oikos 106: 185–192.
BASSET, Y. 1992. Host speciﬁcity of arboreal and free-living insect herbivores in
rain forests. Biol. J. Linn. Soc. 47: 115–133.
BELL, T., R. P. FRECKLETON,AND O. T. LEWIS. 2006. Plant pathogens drive den-
sity-dependent seedling mortality in a tropical tree. Ecol. Lett. 9:
BLUNDELL, A. G., AND D. R. PEART. 2001. Growth strategies of a shade-tolerant
tropical tree: The interactive effects of canopy gaps and simulated herb-
ivory. J. Ecol. 89: 608–615.
BURKEY, T. V. 1994. Tropical tree species diversity: A test of the Janzen–Connell
model. Oecologia 97: 533–540.
CARSON, W. P., J. ANDERSON,E.LEIGH,AND S. A. SCHNITZER. 2008. Challenges
associated with testing and falsifying the Janzen–Connell hypothesis: A
review and critique. In W. P. Carson and S. A. Schnitzer (Eds.). Tropical
forest community ecology, pp. 210–241. Blackwell Publishing, Oxford,
CHAPMAN, C. A., AND L. J. CHAPMAN. 1996. Frugivory and the fate of dispersed
and non-dispersed seeds in six African tree species. J. Trop. Ecol. 12:
CHAPMAN, J. D., AND H. M. CHAPMAN. 2001. The forests of Taraba and Ada-
mawa states, Nigeria: An ecological account and plant species checklist.
University of Canterbury, Christchurch, New Zealand.
CINTRA, R. 1997. A test of the Janzen–Connell model with two common tree
species in Amazonian forest. J. Trop. Ecol. 13: 641–658.
CINTRA, R., AND V. HORNA. 1997. Seed and seedling survival of the palm As-
trocaryum murumuru and the legume tree Dipteryx micrantha in gaps in
Amazonian forest. J. Trop. Ecol. 13: 257–277.
CLARK, D. A., AND D. B. CLARK. 1984. Spacing dynamics of a tropical rain forest
tree: Evaluation of the Janzen–Connell model. Am. Nat. 124: 769–788.
CLARK, D. A., AND D. B. CLARK. 1985. Seedling dynamics of a tropical rain for-
est tree: Impacts of herbivory and apical meristem damage. Ecology 66:
CLARK, D. A., AND D. B. CLARK. 1987. Population ecology and microhabitat
distribution of Dipteryx panamensis, a neotropical rain forest emergent
tree. Biotropica 19: 236–244.
COLEY, P. D., AND J. A. BARONE. 1996. Herbivory and plant defences in tropical
forests. Annu. Rev. Ecol. Syst. 27: 305–335.
COMITA, L. S., AND S. P. HUBBELL. 2009. Local neighborhood and species’ shade
tolerance inﬂuence survival in a diverse seedling bank. Ecology 90:
CONNELL, J. H. 1971. On the role of natural enemies in preventing competitive
exclusion in some marine mammals and in rain forest trees. In P. J. Boer
and G. Gradwell (Eds.), pp. 298–310. Dynamics of populations.
PUDOC, The Netherlands.
DOWSETT-LEMAIRE, F. 1989. Physiography and vegetation of the highland for-
ests of eastern Nigeria. In R. J. Dowsett (Ed.). A preliminary natural
history survey of Mambilla Plateau and some lowland forests of eastern
Nigeria. Tauraco research report 1, pp. 6–12. Tauraco Press, Ely, UK.
82 Matthesius, Chapman, and Kelly
FRECKLETON, R. P., AND O. T. LEWIS. 2006. Pathogens, density dependence and
the coexistence of tropical trees. Proc. R. Soc. B 273: 2909–2916.
GILBERT, G. S., S. P. HUBBELL,AND R. B. FOSTER. 1994. Density and distance-to-
adult effects of a canker disease of trees in a moist tropical forest.
Oecologia 98: 100–108.
GIVNISH, T. J. 1999. On the causes of gradients in tropical tree diversity. J. Ecol.
HAMMOND, D. S., AND V. K. BROWN. 1998. Disturbance and phenology and life
history characteristics: Factors inﬂuencing distance/density-dependant
attack on tropical seeds and seedlings, pp. 51–78. In D. M. Newbury, H.
H. T. Prins, and N. D. Brown (Eds.). Dynamics of tropical communi-
ties: 37th symposium of the British ecological society, pp. 51–78.
Cambridge University Press, Cambridge, UK.
HARMS, K. E., S. J. WRIGHT,O.CALDERON,A.HERNANDEZ,AND E. A. HERRE.
2000. Pervasive density dependant recruitment enhances seedling diver-
sity in a tropical forest. Nature 404: 493–495.
HOWE, H. F. 1990. Survival and growth of juvenile Virola surinamensis in Pan-
ama: Effects of herbivory and canopy closure. J. Trop. Ecol. 6: 259–280.
HUBBELL, S. P. 1979. Tree dispersion, abundance, and diversity in a tropical dry
forest. Science 203: 1299–1309.
HUBBELL, S. P. 1980. Seed predation and the coexistence of tree species in trop-
ical forests. Oikos 35: 214–229.
HYATT, L. A., M. S. ROSENBERG,T.G.HOWARD,G.BOLE,W.FANG,J.ANAST-
ASIA,K.BROWN,R.GRELLA,K.HINMAN,J.P.KURDZIEL,AND J. GUREV-
ITCH. 2003. The distance dependence prediction of the Janzen–Connell
hypothesis: A meta analysis. Oikos 103: 590–602.
JANZEN, D. H. 1970. Herbivores and the number of tree species in tropical for-
ests. Am. Nat. 104: 501–528.
JANZEN, D. H. 1980. Speciﬁcity of seed attacking beetles in a Costa Rican de-
ciduous forest. J. Ecol. 68: 929–952.
LAMBERS,J.H.R.,J.S.CLARK,AND B. BECKAGE. 2002. Density-dependant mortal-
ity and the latitudinal gradient in species diversity. Nature 417: 732–734.
MOLOFSKY, J., AND C. K. AUGSPURGER. 1992. The effect of leaf litter on early
seedling establishment in tropical forest. Ecology 73: 68–77.
NORGHAUER, J. M., J. R. MALCOLM,AND B. L. ZIMMERMAN. 2006. Juvenile mor-
tality and attacks by a specialist herbivore increase with conspeciﬁc adult
basal area of Amazonian Swietenia macrophylla (Meliaceae). J. Trop.
Ecol. 22: 451–460.
NORGHAUER, J. M., J. R. MALCOLM,AND B. L. ZIMMERMAN. 2008. Experimental
establishment of big-leaf mahogany (Swietenia macrophylla King) seed-
lings on two soil types in native forest of Para
´, Brazil. For. Ecol. Manage.
NOVOTNY, V., AND Y. BASSET. 2005. Host speciﬁcity of insect herbivores in
tropical forests. Proc. R. Soc. B 272: 1083–1090.
PACKER, A., AND K. CLAY. 2000. Soil pathogens and spatial patterns of seedling
mortality in a temperate tree. Nature 404: 478–481.
QUEENBOROUGH, S. A., F. R. DAVID,P.BURSLEM,N.C.GARWOOD,AND
R. VALENCIA. 2007. Neighborhood and community interactions deter-
mine the spatial pattern of tropical tree seedling survival. Ecology 88:
R: A Language and Environment for Statistical Computing. 2009. R Founda-
tion for Statistical Computing, Vienna, Austria. Available from http://
www.r-project.org (accessed 3 March 2010).
ROMO, M., H. TUOMISTO,AND B. A. LOISELLE. 2004. On the density-depen-
dence of seed predation in Dipteryx micrantha, a bat dispersed rainforest
tree. Oecologia 140: 76–85.
SCHUPP, E. W. 1992. The Janzen–Connell model for tropical tree diversity:
Population implications and the importance of spatial scale. Am. Nat.
SESSIONS, L.A, AND D. KELLY. 2001. Methods for monitoring herbivory and
plant growth on New Zealand mistletoes (Loranthaceae). N.Z. J. Ecol.
STOTZ, H. U., B. R. PITTENDRIGH,J.KROYMANN,K.WENIGER,J.FRITSCHE,A.
BAUKE,AND T. MITCHELL-OLDS. 2000. Induced plant defense responses
against chewing insects: Ethelene signalling reduces resistance of Arab-
idopsis against Egyptian cotton worm but not diamondback moth. Plant
Physiol. 124: 1007–1019.
WEBB, C., AND D. R. PEART. 1999. Seedling density dependence pro-
motes co-existence of Bornean rainforest trees. Ecology 80: 2006–
WENNY, D. G. 2000. Seed dispersal, seed predation, and seedling recruitment of
a neotropical montane tree. Ecol. Monogr. 70: 331–351.
WRIGHT, J. S. 2002. Plant diversity in tropical forests: A review of mechanisms
of species coexistence. Oecologia 130: 1–14.
Testing Janzen–Connell Model in West Africa 83