Content uploaded by Marc-Antoine Dubois
Author content
All content in this area was uploaded by Marc-Antoine Dubois on Dec 03, 2017
Content may be subject to copyright.
RESEARCH PAPER
© 2003 Blackwell Publishing Ltd. http://www.blackwellpublishing.com/journals /geb
Global Ecology & Biogeography
(2003)
12
, 237–248
Blackwell Science, Ltd
Colonization front of the understorey palm
Astrocaryum
sciophilum
in a pristine rain forest of French Guiana
PIERRE CHARLES-DOMINIQUE,
1
*, JEROME CHAVE
2,3
, MARC.-A. DUBOIS
4
, JEAN-JACQUES DE
GRANVILLE
5
, BERNARD RIERA
1
and CECILE VEZZOLI
1
1
Laboratoire d’Ecologie Générale, Muséum National d’Histoire Naturelle and CNRS UMR 8571, 4 Avenue du Petit Château, F-91800 Brunoy,
France,
2
Laboratoire d’Ecologie Terrestre, CNRS UMR 5552, 13 Avenue du colonel Roche, F-31029 Toulouse cedex 4, France,
3
Department
of Ecology and Evolutionary Biology, Princeton University, Princeton NJ 08540, U.S.A.,
4
Service de Physique de l’Etat Condensé DRECAM,
CEN-Saclay l’Orme des Merisiers, F-91191 Gif-sur-Yvette, France and
5
Herbier de Guyane, IRD, BP 165, 97323 Cayenne Cedex, France
ABSTRACT
Aims
Astrocaryum sciophilum
(Miq.) Pulle (Arecaceae) is an
understorey palm, endemic to north-eastern South America
with a patchy distribution. We tested the hypothesis that the
spatial distribution of this palm species is not in equilibrium
but is slowly colonizing the forest understorey.
Location
Inventories and seed dispersal studies were con-
ducted in the undisturbed tropical forest close to the Nour-
agues research station, French Guiana. Additional data were
collected in the entire territory of French Guiana.
Methods
We studied the demography of
A. sciophilum
on a
20-ha plot located at the edge of its distribution. The age of the
palms was estimated by postulating an exponentially decreasing
abundance by age class. Direct seed dispersal experiments were
also conducted, to estimate dispersal parameters. The seeds of
A. sciophilum
were dispersed only by rodents. This informa-
tion was used to parameterize a forest growth simulator, to
study the spatial spread of this species.
Results
Within the sampling plot, the density of
A. sci-
ophilum
dropped sharply from about 500 individuals per
hectare to zero. The maturation age was estimated to be
170
±
70 years, and over 55 years with 95% confidence.
Seed-dispersal experiments yielded an average seed dispersal
distance of 11 m and a maximum estimated dispersal dis-
tance of 125 m across a generational span of 55 years to
maturity. Therefore, the maximal estimated colonization
speed is 2.3 m/y.
Conclusions
Empirical results and numerical simulations
suggest that the boundary of the
A. sciophilum
population
is a colonization front, and that the range of this species
is slowly expanding. The implications of this result in re-
spect of palaeoenvironmental changes in this region are
discussed.
Key words
Astrocaryum sciophilum
, French Guiana, neo-
tropical palaeoecology, palm ecology, population edge, ref-
uges, scatter-hoarding, seed dispersal, spatial pattern.
INTRODUCTION
Small-scale disjunctions are frequent for plants, but several
mechanisms may cause these patterns. Spatial environmental
heterogeneity is a plausible hypothesis (Ashton, 1969;
Gartland
et al
., 1986), especially of small-sized plants such
as pteridophytes (Poulsen & Balslev, 1991; Tuomisto, 1998),
epiphytes (Gentry & Dodson, 1987), or some species restricted
to swamps (Svenning, 1999). This ‘equilibrium’ hypothesis
postulates situations where species quickly reach a local
demographic equilibrium and co-exist by occupying niches
that partition the environmental axes (Grubb, 1977). How-
ever, the evidence for habitat specialization is weak or absent
for many tree species at the scale of 50 hectares in a forest of
Panama (Harms
et al
., 2001). Indeed, a species that is associ-
ated strongly with a given habitat could also be present,
although less abundant, in other habitats. A recent survey in
the Peruvian Amazon, along the Rio Manu, found that only
25% of the 381 studied species are habitat specialists (Pitman
et al
., 1999).
Historical factors provide another plausible explanation
for the disjunction of species ranges. A species could have
gone extinct in certain areas because of sudden environmental
changes (drought, fire, outbreak of pathogens) and be in the
process of recolonizing these areas from relict populations.
* Corresponding author: e-mail: pierre.charles-dominique@wanadoo.fr
238
P. Charles-Dominique
et al.
© 2003 Blackwell Publishing Ltd,
Global Ecology & Biogeography
,
12
, 237–248
Such a picture is consistent with non-equilibrium theories for
the maintenance of diversity (Hubbell, 1979; Hubbell &
Foster, 1986; Hubbell, 2001), suggesting that forest communities
are maintained by drifts in species abundance.
In this context, it is interesting to study species that are
known to be locally abundant in some areas while totally
absent in others. Such a situation could be stable, if the spe-
cies is a habitat specialist, or it could be related to the spatial
dynamics of the population. We have selected the species
Astrocaryum sciophilum
(Miq.) Pulle (Arecaceae), an under-
storey palm endemic to the Guiana Shield. This species is very
common in many areas of the Guianas, but completely absent
in others (Kahn & de Granville, 1992). Preliminary observa-
tions by Sist (1989a, 1989b) have demonstrated that this
palm spends a long time in the immature stage and that its
crop is dispersed mainly over short distances.
A demographic study of a population of
A. sciophilum
was
conducted on the edge of its distribution. A complete inven-
tory was carried out in 1998 and 2001, in a 20-ha permanent
plot located in the pristine lowland rain forest of the Nour-
agues Research Station, central French Guiana. Additional
surveys were performed all around the country to test for
habitat specialization. We also present quantitative results on
its growth pattern and seed shadow in the core of the popu-
lation, 1.3 km from the edge. An individual-based forest
growth model was used to simulate the spatial dynamics of
the population. Our results are discussed in the light of pur-
ported palaeo-environmental changes in this region.
METHODS
Study species
We study the palm species
Astrocaryum sciophilum
(Miq.)
Pulle. The neotropical genus
Astrocaryum
(Arecaceae) is con-
spicuous in the Amazonian forest, with over 35 species falling
into four taxonomic sections (Kahn & Millán, 1992), which
are supported by a recent molecular phylogeny (Pintaud
et al
., in press).
A. sciophilum
is the only species of section
Munbaca in French Guiana, and is morphologically and
genetically differentiated from the closely related species
A.
sociale
in Brazil (Kahn, 2000; Pintaud
et al
., 2003).
This species is identified easily in the field (Henderson
et al
., 1995). The different growth stages we use here are
based on the morphological criteria selected by Sist (1989a),
with slight modifications. We distinguish four juvenile stages,
depending upon the life form of the palm, and one adult stage
(Charles-Dominique
et al
., 2001a). Stage 0 begins at germina-
tion and includes all individuals with entire leaves. The first
leaf reaches 15–20 cm in length, and the most developed
individuals in this category can bear up to 11 leaves, 100–
120 cm long (mean 66.6 cm, max. 225 cm). Stage 1 includes
individuals with leaves bearing one to eight pinnae, still asym-
metrically arranged on either side of the rachis. According to
Sist (1989a,b), individuals in this stage bear seven leaves on
average (2–12), between 25 and 390 cm in length (mean
156 cm, SD 49 cm). When individuals reach stage 2, the pin-
nae are arranged symmetrically on either side of the rachis.
Individuals display eight leaves on average (mean 8.1, SD 2)
that bear between eight and 58 symmetrical pinnae. At this
stage, individuals still have at least one pinna larger than
the others. Individuals at stage 3 still bear relatively delicate
leaves (5–17 leaves, mean 9.7): the rachis is only 26 mm in
diameter on average (SD 6.9 mm), while this value reaches
34.1 mm (SD 2.5 mm) for adults.
For adults, the mean length of the leaves reaches 633 cm
(SD 154 cm) and the rachis is much sturdier than in Juv 3
individuals. Young adults lack an above-ground stem but
none the less show traces of fructification (one or several
recent peduncular bracts, or 1–3-year-old remains of infruct-
escence). All older adults have a visible above-ground stem.
The tallest individuals had a stem 8.6 m in height.
Site location
We collected most of our data at the Nouragues Research Sta-
tion in French Guiana (4
°
05
′
N, 52
°
41
′
W), in an old-growth
tropical rain forest (Poncy
et al
., 1998; Bongers
et al
., 2001).
Rainfall is around 3000 mm/y (14-year average) with a
2-month dry season from September to November (< 100 mm/
month), and a short dry season in March. The monthly mean
maximum temperature varies between 32.1 and 35.8
°
C, and
the monthly mean minimum between 19.5 and 20.8
°
C.
About 95% of French Guiana is lowland wet rain forest. The
remaining 5% corresponds to a narrow 3–10 km wide
coastal strip of savanna, mangroves and marshes (Boyé
et al
.,
1979). Additional data were collected at the Piste de Saint
Elie Research Station and at the Paracou Research Station
(Fig. 1).
A presence/absence map was constructed for
A. sciophilum
in French Guiana, using botanical inventories (Fig. 1), and
by looking for the species
c
. 10 km around the Nouragues
research station. Available data are scarce due to the difficult
access of the interior; however, we believe that the map pro-
vides at least a first approximation. We use this distributional
information to help us evaluate our theory that
A. sciophilum
is not limited by edaphic factors or other environmental
factors.
Study of the population edge
We first located the distributional border of the population of
A. sciophilum
4 km to the South-west of the Nouragues Station.
On this border, we established a 20-ha permanent plot (500 m
×
400 m) perpendicular to the edge of the palm population
(Figs 2 and 3). This plot was located in the middle of a
Spatial patterns of
Astrocaryum sciophilum 239
© 2003 Blackwell Publishing Ltd,
Global Ecology & Biogeography
,
12
, 237–248
plateau in order to avoid hydrographic or topographic
effects.
A first census was carried out in October 1998 and a sec-
ond in October 2001. All adults were recorded over the entire
20 ha, and juveniles were censused in a subsample of 14 ha.
Each 1-ha quadrat was subdivided by trails into 20
×
20 m
subplots to facilitate the mapping and tagging of each indi-
vidual. Leaves were counted and measured and the number
Fig. 1 Map of French Guiana showing the different study areas (1: Nouragues; 2: Piste de St Elie; and 3: Paracou), the distribution of the various
habitat types (forest, coastal savanna and swamps and mangroves) and the distribution of A. sciophilum: black circle = present, open circle = absent.
240
P. Charles-Dominique
et al.
© 2003 Blackwell Publishing Ltd,
Global Ecology & Biogeography
,
12
, 237–248
Fig. 2 Map of the Nouragues Station indicating the location of the study plot (P) on the edge of the A. sciophilum population and the camp (C).
The population edge is drawn as a solid line, black circle = presence of A. sciophilum, white circle = absence of A. sciophilum. After IGN NB22
II Ib Paris 1983.
Spatial patterns of
Astrocaryum sciophilum 241
© 2003 Blackwell Publishing Ltd,
Global Ecology & Biogeography
,
12
, 237–248
of pinnae noted. The stem height and phenological state of
adults were recorded (inflorescences, infructescences, remains
of old infructescences).
Several demographic parameters were collected less than
1000 m from this plot. Every 3–6 months, young unfolded leaves
were recorded and measured for juvenile (
n
= 325) and adult
(
n
= 47) individuals alike in order to evaluate the phenology
of leaf production. A population of 189 adults was examined
in the core population (about 1.3 km from the plot) and stem
height was recorded.
Age estimation
Different approaches were used to estimate the time needed
by a palm to reach the adult stage (Charles-Dominique
et al
.,
2001a). First, if the population is at local equilibrium in age
distribution, assuming a constant mortality rate, we expect
the number of individuals to decrease exponentially with age
class. We then fit the abundance in the juvenile stages to an
exponential, which allows us to estimate the time spent in
each stage. This analysis was performed on the data available
in the core of the palm population, 1.3 km from the edge.
Secondly, the count of leaf scars left on the subterraneous
stem of uprooted palms provides a lower bound estimate of
the age of the palm. Indeed, the oldest (and thinnest) part,
which corresponds to the earliest phase of the palm’s life, is
usually decayed or desiccated. We analysed the leaf scars of
35 unearthed stems overall.
A. sciophilum
spends its entire life cycle in the understorey.
Treefall events are important for this species. The mechanisms
Fig. 3 Distribution of A. sciophilum on the study plot (400 × 500 m). Adults were mapped individually (in relation to height). Because juveniles
were too numerous to be mapped individually, only the edge of the distribution is shown (solid line between the trails M and N). Numbers
(39–43) and capital letters (I–M) indicate the trails.
242
P. Charles-Dominique
et al.
© 2003 Blackwell Publishing Ltd,
Global Ecology & Biogeography
,
12
, 237–248
that maintain the terminal bud underground during the
entire juvenile phase (Sist, 1989a; Tomlinson, 1990) greatly
reduce the risk of complete destruction, which occurs only if
a large tree trunk falls directly atop a young individual. To
provide the third age estimate, we assume that juvenile
mortality is caused only by treefalls. From the fraction of
individuals in the juvenile stage we obtain an upper bound of
the maturation age, as all the other possible causes of mortal-
ity (such as predation) are discarded.
Adult palm age was correlated to adult palm stem height
(SE 5%), measured from ground level (at the base of the heap
of decomposed leaf-litter accumulated at the foot of each
individual) to the base of the oldest leaf still in function.
Seed removal and germination experiments
The seeds of
A. sciophilum
are animal-dispersed. Other means
of dispersion can be eliminated easily. In particular, the seed
has a specific density greater than 1, and thus it cannot be
transported by runoff waters. Its fruits are predated mainly
by the squirrel
Sciurus aestuans
, the agouti
Dasyprocta lepo-
rina
and the acouchi
Myoprocta exilis
, and to a lesser extent
by the spiny rat
Proechimys
spp. Peeling the seed takes a long
time, because the animals must gnaw through the endo-
carp with their incisors, for 10–15 min for the agouti, and
nearly 50 min for the squirrel, before reaching the lipid-rich
endosperm (personal observation, C.V.). True frugivorous
vertebrates such as monkeys ( Julliot, 1997), bats and birds
(personal observation, P.C.D.) have never been observed to
exploit this pulpless species. Similarly, tapirs, known to carry
seeds of the palm
Attalea maripa
kilometres away from the
site of production (Fragoso, 1997) do not consume this very
tough seed (personal observation P.C.D.).
Fruit production is irregular so rodents hide seeds for later
use, a scenario already described for other plant–rodent inter-
actions (Kiltie, 1981; Sist, 1989a; Forget, 1991, 1997).
Agoutis and acouchis bury the unpredated seeds in small
holes 3–9 cm deep (Henry, 1999; Jansen & Forget, 2001),
dug carefully and sealed afterwards. Only the buried seeds
that have not been dug up and eaten will germinate (Sist,
1989a).
A 2-month seed dispersal experiment was undertaken
twice with
A. sciophilum
seeds, during the rainy season
(March–June 1998, experiment 1) and during the dry season
(September–November 1998, experiment 2). A yellow thread
with a coloured flag was affixed to each seed (
n
= 600 for
experiment 1;
n
= 290 for experiment 2). Replicates were
scattered within the experimental plot. We searched for the
seeds that were removed up to 30 m away from each removal
experiment plus along four perpendicular 100 m trails
radiating from the removal sites.
To analyse the seed dispersal curve we pooled the distance
data into bins of 5 m, and fitted the curve by the maximum
likelihood technique using several empirical dispersal kernels
(Portnoy & Willson, 1993; Willson, 1993; Turchin, 1998;
Clark
et al
., 1999). We used a decreasing exponential func-
tion, a Weibull function (type VII of Turchin, 1998) and a fat-
tailed 2Dt function (Clark
et al
., 1999).
To assess the seed germination potential, we monitored
72 seeds that we buried in the forest, and 19 seeds buried by
rodents. This experiment has been followed for 36 months.
Statistical analyses and simulations
We performed correlation analyses to ensure that the population
edge was a population front and not a stable limit of the
population range. We applied two statistical tests. We tested
whether the population increase from the edge of the popula-
tion was significant (linear regression method). We also tested
for the increase in length of juvenile leaves away from the edge
using the non-parametric Kruskal–Wallis test (Zar, 1996),
applied to palms grouped in 50-m distance classes (
n
= 8 groups).
Analyses of variance were not possible, as the distribution
of leaf length is markedly leptokurtic. Pairwise comparisons
across groups were finally performed using Tukey’s test.
The individual-based model
troll
(Chave, 1999, 2001)
was used to test the hypothesis of the advance of the
A. sci-
ophilum
population in a primary forest. In this model, each
tree is described in three dimensions. Competition for light is
also described in three dimensions on the square metre scale,
and gap dynamics are modelled. We had previously defined
12 plant functional types for the French Guiana rain forest
(Chave, 1999).
A. sciophilum
was introduced as an addi-
tional functional type, having the biological characteristics
measured in the field.
RESULTS
Edge of the population
Our plot was set out on an edge of the palm population. No
palm was encountered further south, except a ‘micropopula-
tion’ spreading over 100 m east–west and 1000 m north–
south and confined to a small wet valley 1200 m to the south.
We tested the hypothesis that the gradient corresponds to a
true front of colonization. We recorded the individuals as a
function of the edge of the population, discarding the last 100
m, which would have induced a sampling bias. Distance
classes of 50 m have been used.
The total number of juveniles, subadults and adults increased
steadily and significantly from the population edge (Table 1).
The total population density varied between 27 and 648
individuals per hectare along the south–north gradient. In
the first 100 m, the ratio of juvenile to adults (RJA) was 23,
and decreased to 14 from 300 to 400 m of the edge. In the
core of the population the RJA was between 10 and 12. The
Spatial patterns of
Astrocaryum sciophilum 243
© 2003 Blackwell Publishing Ltd,
Global Ecology & Biogeography
,
12
, 237–248
Kruskal–Wallis test on the leaf length of juveniles showed a
significant difference between 50-m distance classes (
n
= 8,
H
c
= 1919,
P
< 0.001). We analysed further the pairwise dif-
ferences between classes. We found that the six classes from
0 to 300 m were not significantly different (Tukey test with
unequal sample size, Q < 1.22,
Q
0.05,7
= 3.03), while palms
from distance 300–400 m had significantly
smaller
leaves
(
Q
= 3.30). Pooling subadults with juveniles led to the same
result. The differences were marginally significant for stem
height (0.05 <
P
< 0.10).
Finally, we tested for the stability of the population edge. If
the edge was stable, adult palms located on the population
edge should display poorer reproductive success. Of these
adult palms, 44% (
n = 34, 10 ha) were fertile during one year.
In the core population, the same experiment gave percentage
of fertility of 34% (n = 161, 10 ha). Therefore the fertility of
the population was slightly higher at the edge of the population
than in the core (χ2 = 13.6, P < 0.01). This observation runs
against the hypothesis of a stable population edge.
Surveys completed along the coastal savanna/forest boundary
near the Piste de Saint Elie Research Station and the Paracou
Research Station (Fig. 1) revealed the presence of a population
front about 3 km behind the edge of the forest. Counts under-
taken along two transects along another population edge at
Piste de Saint Elie showed the same general trends as at the
Nouragues station: the abundance of adults increased from
25 to 190 individuals per ha in 200 m.
Life history
The time-span between leaf production was 16 months (8–
22 months, SD = 1, n = 129) both for juveniles and adults.
However, Van der Steege (1983) found a time-span of
1.1 years for adults in Surinam. In 3 ha (1211 individuals),
an average of 49 palms/ ha had apparently been damaged
by treefalls or branchfalls (12.5% of the total population),
50 palms/ha showed traces of predation (13% of the total)
and 18 palms /ha of hydric stress (4.5%). Overall the mortality
rate was very low. A small area monitored during 3 years
gave the following results: three juveniles of 120 (2.5%) and
one adult of 61 (1.6%) died. No lethal predation was recorded
on juveniles in this area, although herbivory of this palm by
caterpillars of Brassolis sophoroe (or the closely allied
species B. astyra) has been observed at the Piste de Saint-Elie
station (D. Sabatier & J. Olivier, personal communication).
The methods applied to estimate the maturation age
gave the following results. Method 1 (demographic ana-
lysis): 180 ± 40 years; method 2 (subterraneous stem scars count):
162 ± 88 years; and method 3 (treefall disturbance rate):
300 ± 100 years (about 1% of the forest floor is affected
yearly by direct treefalls; Riéra & Alexandre, 1988). Method
3 provides only a very rough estimation of the maturation
age. Using only Methods 1 and 2, the time duration of the
immature phase is estimated reasonably at about 170 ± 70
years. The error bars should be taken with caution, for they
reflect only an average history in one sampled population.
The one-sided 95% confidence interval is 55 years, i.e. there
is a 95% chance that a palm that reaches maturity is older
than 55 years. In the rest of this paper, we retain this value of
55 years as a conservative estimate for the age of maturation.
Concerning the adult phase, stems grew by 27 (± 2) mm in
height with each new leaf (every 16 month). A typical 2-m tall
palm is therefore c. 270 years old. The tallest recorded indi-
vidual had a stem 8.6 m in height.
Ta b le 1 Num
b
er o
f
in
d
ivi
d
ua
l
s in t
h
ree
d
i
ff
erent stage c
l
asses (juveni
l
es, su
b
a
d
u
l
ts an
d
a
d
u
l
ts) as a
f
unction o
f
t
h
e
d
istance
f
rom t
h
e popu
l
ation
edge. All subadults and adults were sampled in the 400-m-wide strip from the population edge 400 m deep inside the population. Juveniles were
censused in a 100-m-wide strip to 400 m inside the population edge. The total number of sampled individuals, their local density (in individuals/
ha), the average length and the maximal length of the longest leaf are reported here. For adults, we report the average height and maximal stem
height. All dimensions are in cm
Distance class (m) 0–50 50–100 100–150 150 –200 200–250 250–300 300 –350 350–400
No. of juveniles 7 14 47 66 132 156 225 288
No. per ha 14 28 94 132 264 332 450 576
Average leaf length 136 141 171 140 147 152 123 122
Max. leaf length 240 275 570 430 530 600 560 550
No. of subadults 3 4 19 29 31 46 67 59
No. per ha 1.5 2 9.5 14.5 15.5 23 33.5 29.5
Average leaf length 473 527 579 573 587 573 540 550
Max. leaf length 530 645 790 909 800 810 870 770
No. of adults 1 3 13 19 17 54 75 84
No. per ha 0.5 1.5 6.5 9.5 8.5 27 37.5 42
Average stem height 0 132 84 131 170 141 177 181
Max. stem height 0 240 280 400 650 500 860 680
244 P. Charles-Dominique et al.
© 2003 Blackwell Publishing Ltd, Global Ecology & Biogeography, 12, 237– 248
Fruit production and seed dispersal
About 40% of the adults of A. sciophilum produced at least one
infructescence (mean number 2.2, 1–7), each bearing an aver-
age of 65 ripe fruits (SD 177). Because fructification is irregu-
lar and depends also on the size of the infructescences, the
mean production of ripe fruit can be estimated at 33 fruits/
individual/year (see also Sist, 1989a). The dry pulpless fruit is
covered by a thin dry and spiny pericarp that can easily be
discarded. The size of the endocarp — pear-shaped and ros-
trate at the apex — has been estimated for n = 325 seeds. It
measured 44.6 mm in length (28–60 mm, SD = 14.3), 27 mm
in diameter (14–39 mm, SD = 6.2), and weighed 14.5 g (4.5–
33 g, SD = 4.3). The entire fruit (endocarp plus exocarp)
weighed 23 g (8–36 g, SD = 8.5, n = 300). Although seeds of
this size could, in principle, be eaten by many animal species
(van Roosmalen, 1985), the endosperm is protected by an
extremely tough and thick endocarp.
Germination studies show that less than one seed per adult
palm germinates each year. Seed dormancy is long. After
36 months, 25% of the seeds buried in the experimental plot
had germinated (n = 72 seeds in total), as had 26.3% of the
seeds that had been buried — and not uncovered subsequently
— by acouchis (n = 19 seeds in total). These percentages are
not statistically different.
The 2-month seed dispersal experiments during the rainy
season (March–June 1998, experiment 1, n = 600 seeds) and
during the dry season (September–November 1998, experi-
ment 2, n = 290 seeds) gave the following results. We found
247 seeds of 595 removed by rodents during the first experi-
ment: 95.5% (n = 236) were scatter-hoarded by acouchis and
agoutis; 1.62% (n = 4) were eaten by squirrels and 2.83%
(n = 7) were simply removed and left on the ground. In experi-
ment 2, 80% (n = 232) of the seeds we recovered were eaten
by rodents, 20% (n = 58) were hidden, and none were left in
situ. These figures reflect the higher predation rate and the
lower storage rate during the dry season, with a low food
availability. The largest dispersal distance recorded was 40 m
(experiment 1) and 21 m (experiment 2). Of the 374 lost
seeds, most of them were probably eaten by rodents. How-
ever, it is also possible that some seeds were dispersed further
away than 40 m. Therefore, we need to model the likelihood
of such long-distance dispersal events.
The seed shadow was best fit by an exponential distribution
N(r) = N(0)exp(−r/l) (Fig. 4). The likelihood corresponding to
the Weibull function (parametric form N(r) = exp(a
−
brc)) was
not significantly higher. We found an average dispersal distance
of l = 11 m (experiment 1) and l = 7.5 m (experiment 2).
However, to estimate the colonization speed of the population,
the mean dispersal is less relevant than the extreme dispersal
events. The distance beyond which 5% of the seeds have been
dispersed is r95% = 33 m. An even more conservative defini-
tion of the extreme dispersal distance is the distance beyond
which one seed at most has been dispersed during a genera-
tion. Suppose a colonization edge comparable to the one at
the Nouragues station, where about 50 mature palms con-
tribute to the crop production. The average seed production
per mature individual is 33 seeds/individual/year, this gives a
total production of about n = 160 000 seeds dispersed during
55 years. We seek the distance such that N(rmax) = 1, that is
rmax = ln(N)l = 125 m. This value is consistent with the
maximum dispersal distance observed for the seeds of Carapa
procera scatter-hoarded by acouchis at Nouragues (124 m,
Jansen et al., 2002).
To model the dynamics of the population edge, we used
the troll simulator (Chave, 1999, 2001). We took a typical
maturation age of 170 years (SD 70 years), and a negative
exponential seed shadow, with a 95% dispersal distance of 50
metres. We then conducted two types of computer simulations.
We first started with a ‘forest’ made up only of seedlings
(including A. sciophilum seeds), and we monitored the forest
growth to see how long it took the A. sciophilum population
to reach demographic equilibrium. The population took almost
1000 years to reach its equilibrium, characterized by around
600 individuals per hectare, in good agreement with the field
data. We then performed a second experiment, introducing palm
seeds from the edge of a strip of forest (size 2000 × 400 m).
During the simulation, we monitored the speed of advance of
the population front. Figure 5 shows the patterns observed in
this simulation. The colonization front advanced at an average
speed of 2 m/year and with a heterogeneous pattern.
Environmental constraints
Our explorations have underlined the omnipresence of
A. sciophilum along a 100-km-wide strip on the northern coastal
edge forest, from Saint-Georges de l’Oyapock, eastwards, to
Fig. 4 Seed shadow for A. sciophilum; 516 dispersed seeds which
were recovered within a distance of 40 m from the removal site. An
exponential function fitted the data very well up to 25 m (r2 > 0.99).
Spatial patterns of Astrocaryum sciophilum 245
© 2003 Blackwell Publishing Ltd, Global Ecology & Biogeography, 12, 237–248
near St Laurent du Maroni, north-westwards (Fig. 1). The
species is absent, however, from the first 3 km of the coastal
forest. The Nouragues population (100 km from the coast)
and the ‘coastal’ population, present throughout the coastal
region, are both part of a single continuous population.
Incidentally, the palm Geonoma oldemanii exhibits the same
spatial distribution (de Granville, 1989). The continental
edge of the A. sciophilum population lies a few kilometres to
the south of the Nouragues station, following roughly an east–
west direction (Figs 1 and 2). No A. sciophilum populations
were found to the south of this borderline for several dozen
kilometres, excepted one isolated population in the area of the
Mont Chauve, about 25 km south of the edge (C. Sarthou,
personal communication). Other localized populations were
found southwards, in the central and southern regions of
French Guiana, on a rugged landscape of granite outcrops as
well as on tabletop mountains on basic bedrock (Sabatier &
Prévost, 1987; Kahn & de Granville, 1992; personal observa-
tion, J.J.dG. and P.C.D.). North-eastwards, the population
extends towards the Venezuelan Guiana (Galeano, 1992). As
mentioned above, the populations encountered in the region
of Manaus probably belong to another species, A. sociale
(Kahn, 2000; Pintaud et al., 2003).
No obvious change in the physiognomy of the forest can
explain the heterogeneity of the distribution of A. sciophi-
lum. The palm can be found on a wide variety of soil types:
latosols (lateritic soils), the summits of table mountains with
an underlying lateritic base, podzols, soils on granite, gneiss
and schist. All these soil types are encountered along the
coast, where the species is widespread (Boyé et al., 1979;
Charles-Dominique et al., 2001b). Moreover, where the spe-
cies is present, it can be found in valleys, near poorly drained
areas, as well as on the well-drained slopes and ridges of hills,
and it is abundant (on the order of 500 individuals per hec-
tare). About 1300 m to the south of the Nouragues station,
the edge of the coastal population crosses the middle of a pla-
teau, then runs diagonally across a valley and up the other
side onto another plateau.
DISCUSSION
We have shown that A. sciophilum, unlike most other conge-
neric species, is a slow-growing palm, with a maturation age
of over 170 years (over 55 years with 95% confidence). This
slow growth rate is consistent with C14 dating of large trees
(Chambers et al., 1998) and direct diameter growth measure-
ments (Worbes, 1989; Condit et al., 1999). If trees reach
maturity above 10 cm d.b.h., with an average growth rate of
c. 1 mm /year, their typical maturation age compares with that
of A. sciophilum.
This species is dispersed only by scatter-hoarding rodents
over short distances. Dispersal by water is highly improbable
as the fresh seed sinks in water. Our seed-dispersal experi-
ments and supporting calculations suggest an extreme dispersal
Fig. 5 Simulated pictures of a ‘virtual forest’ using the
individual-based model troll. A. sciophilum seeds
were introduced at the left extremity. The size of the
simulation is 2000 × 400 m, and the density of A. scio-
p
hilum stems is recorded in quadrats of size 25 × 25 m
(white: no individual, black: more than 500 indi-
viduals per hectare). After 500 years, we observe a
progression of about 500 m and a heterogeneous pat-
tern of distribution, close to the situation presently ob-
served in the field.
246 P. Charles-Dominique et al.
© 2003 Blackwell Publishing Ltd, Global Ecology & Biogeography, 12, 237– 248
distance of 125 m in the 55 years estimated for individuals
to reach reproductive maturity. Therefore, the upper-bound
speed of a front of A. sciophilum palms is 125/55 = 2.3 m/
year. Thus, A. sciophilum can be considered as a poor colo-
nist (Clark et al., 1999). This holds only if we can positively
exclude long-distance dispersal events, i.e. seed movements
well beyond 125 m. They would enhance the invasion speed
considerably, and this has been suggested as a probable mech-
anism to explain the rapid advance of vegetation in temperate
areas after the last glacial (e.g. Clark et al., 1999). However,
we have found no direct evidence of any long-distance trans-
port for A. sciophilum, and the present study suggests that
long-distance seed dispersal is unlikely to be related to sec-
ondary dispersal by scatter-hoarding rodents.
The spatial distribution of A. sciophilum is highly aggre-
gated. Several hypotheses could explain this clumping, and
we shall discuss four of these: (1) the current range of A. sci-
ophilum could be limited by edaphic constraints; (2) it could
be related to human settlements; (3) it could be highly sensi-
tive to present-day climatic fluctuations; (4) finally, we could
witness a slow, continuous, invasion stage, lasting since the
last major climatic disturbance.
The regional survey performed all around French Guiana
showed consistently that A. sciophilum is a habitat generalist.
Hence, hypothesis (1) can be rejected. Anthropogenic influence
(hypothesis 2) should not be dismissed (Bush et al., 2000),
since A. sciophilum has been recorded as a useful palm in
Venezuela (Galeano, 1992) and in Suriname (Wessels Boer,
1965). However, several lines of evidence suggest that this
hypothesis is incorrect. A. sciophilum’s only recorded use is
in the production of oil, which results in the destruction of
the seeds. Several other species of the same genus are much
more common and convenient for oil production (e.g. A. vul-
gare, A. paramaca). Hence, we also reject this hypothesis.
Temperatures are relatively uniform throughout French
Guiana, but rainfall does vary. However, current rainfall
distribution cannot explain the sharp edge observed near the
Nouragues, or that in the central and southern landscapes of
French Guiana (the spatial distribution of the species does not
overlap with the isohyet lines). Therefore, we also reject the
third hypothesis.
While current climatic restrictions cannot explain A. sci-
ophilum’s present spatial distribution, former environmental
stresses have probably contributed to it. This idea was the
starting point of the refuge hypothesis (Haffer, 1969; Prance,
1973; Granville, 1982). Although this hypothesis has been
criticized strongly for the Neotropics (Nelson et al., 1990; Bush,
1994; Colinvaux et al., 2000), it is generally acknowledged
that recurrent climatic events may have modified the species
composition of neotropical forests during the Holocene, while
maintaining a closed canopy (Bush et al., 2000; Colinvaux et al.,
2000, 2001). It is therefore reasonable to suggest that the
observed heterogeneous distribution of the species could be
related to past disturbances from which the species is recover-
ing at a slow pace. In our study of the population front, we
found that the individuals located on the population edge are
as fit as those in the bulk of the population. Moreover, the gre-
ater proportion of young individuals found there suggests that
we are observing a true population advance within the forest.
Charcoal analyses undertaken at the Nouragues research
station (Tardy, 1998) showed that areas that are covered
nowadays with a mature wet tropical forest were burnt sev-
eral times during the Holocene (Charles-Dominique et al.,
1998). These analyses indicate a less diverse tree community
between 12 000 and 6000 years bp in comparison to the
period 5000 years bp to the present day. In addition, taxa
such as Tabebuia and Swartzia, now conspicuous in the
semideciduous forest of Venezuela, were abundant in French
Guiana between 12 000 and 6000 years bp. It is likely that
such fires modified the spatial distribution of many slowly
dispersing species, such as Eperua spp. or Vouacapoua amer-
icana (Caesalpiniaceae). In addition, local blowdowns might
have played an important role in these forests (Nelson et al.,
1994). However, as juveniles of A. sciophilum possess a well-
protected subterranean apex, short fire episodes probably did
not affect this population greatly. In a 25-ha block of forest
that was clear-cut in 1976 at Piste de Saint Elie (ARBOCEL
plot), a mean of 210 individuals/ha A. sciophilum had
resprouted in 1994, some with leaves over 6 m high (P.
Chareyre, unpublished report). In addition, the rapid reg-
eneration of this palm in abandoned slash-and-burn areas
(personal observation, P.C.D.) suggests that the absence of A.
sciophilum from certain areas of French Guiana is probably
related more to disturbances that lasted long enough to
impede the regeneration of this species.
The reconstitution of the forest in response to long-term
regional climatic change and for species such as this palm
probably from a few scattered locales, has probably occurred
at a variable pace depending on the species colonization
ability, thus leading to a non-linear increase in diversity over
time. Such ‘within-forest invasions’ might hold the key to
variations in plant diversity in tropical forest ecosystems (ter
Steege & Hammond, 2001). Complementary research should
include genetic work on this species, as well as on other spe-
cies (Caron et al., 2000; Dutech et al., 2000).
ACKNOWLEDGMENTS
We thank warmly Mireille Charles-Dominique, Sylvie
Jouard, Christine Poixblanc, Desmo Bétian and Wémo
Bétian for their help during this study. We are indebted to
Frans Bongers, Anya Cockle, Pierre-Michel Forget, Helene
Muller-Landau, Ran Nathan, Jean-Christophe Pintaud
and Joe Wright for insightful comments at various stages
of this research, to Daniel Sabatier and Jean Olivier for
sharing with us observations of predation by caterpillars,
Spatial patterns of Astrocaryum sciophilum 247
© 2003 Blackwell Publishing Ltd, Global Ecology & Biogeography, 12, 237–248
and to Luc Legal and Joël Minet for identifying them.
This work is a publication of the ECOFIT research pro-
gramme (ECOsystèmes Forestiers Inter-Tropicaux, CNRS/
GDR 489). The Nouragues Research Station is funded by
the CNRS (UPS 656). J.C. was supported by a postdoctoral
fellowship from the Andrew W. Mellon Foundation and
the David and Lucille Packard Foundation (grant 99–8307).
REFERENCES
Ashton, P.S. (1969) Speciation among tropical forest trees: some
deductions in light of recent evidences. Biological Journal of the
Linnean Society London, 1, 155–196.
Bongers, F., Charles-Dominique, P., Forget, P.-M. & Théry, M., eds
(2001) Nouragues: dynamics and plant–animal interactions in a
neotropical rainforest. Kluwer Academic Publishers, Dordrecht.
Boyé, M., Cabaussel, G. & Perrot, Y. (1979) Atlas des départe-
ments Français d’outre-mer. La Guyane. CNRS-ORSTOM, Paris.
Bush, M.B. (1994) Amazonian speciation: a necessarily complex
model. Journal of Biogeography, 21, 5–17.
Bush, M.B., Miller, M.C., De Oliveira, P.E. & Colinvaux, P.A. (2000)
Two histories of environmental change and human disturbance in
eastern lowland Amazonia. Holocene, 10, 543–553.
Caron, H., Dumas, S., Marque, G., Messier, C., Bandou, E.,
Petit, R.J. & Kremer, A. (2000) Spatial and temporal distribution
of chloroplast DNA polymorphism in a tropical tree species.
Molecular Ecology, 9, 1089–1098.
Chambers, J.Q., Higuchi, N. & Schimel, J.P. (1998) Ancient trees in
Amazonia. Nature, 391, 135–136.
Charles-Dominique, P., Blanc, P., Larpin, D., Ledru, M.P., Riéra, B.,
Rosique, T., Sarthou, C., Servant, M. & Tardy, C. (2001b) Palaeo-
climates and their consequences on forest composition. Nouragues:
dynamics and plant–animal interactions in a neotropical rainforest
(ed. by F. Bongers, P. Charles-Dominique, P.-M. Forget and M. Théry),
pp. 79–88. Kluwer Academic Publishers, Dordrecht.
Charles-Dominique, P., Blanc, P., Larpin, D., Ledru, M.P., Riéra, B.,
Sarthou, C., Servant, M. & Tardy, C. (1998) Forest perturbations
and biodiversity during the last ten thousand years in French
Guiana. Acta Oecologica, 1, 295–302.
Charles-Dominique, P., Chave, J., Vezzoli, C., Dubois, M.-A. &
Riéra, B. (2001a) Growth strategy of the understorey palm
Astrocaryum sciophilum in the rainforest of French Guiana. Life
forms and dynamics in tropical forests (ed. by G. Gottsberger and
S. Liede), pp. 153–163. Dissertationes Botanicae 346, Berlin/
Stuttgart.
Chave, J. (1999) Study of structural, successional and spatial patterns
in tropical rain forests using TROLL, a spatially explicit forest
model. Ecological Modelling, 124, 233–254.
Chave, J. (2001) Spatial patterns and persistence of woody plant
species in ecological communities. American Naturalist, 157, 51–65.
Clark, J.S., Silman, M., Kern, R., Maklin, E. & HilleRisLambers, J.
(1999) Seed dispersal near and far: patterns across temperate and
tropical forests. Ecology, 80, 1475–1494.
Colinvaux, P.A., de Oliveira, P.E. & Bush, M.B. (2000) Amazonian
and neotropical plant communities on glacial time-scales: the failure
of the aridity and refuge hypotheses. Quaternary Science Reviews,
19, 141–169.
Colinvaux, P.A., Iron, G., Rasanen, M.E., Bush, M.B. & De Mello, J.
(2001) A paradigm to be discarded: geological and paleoecological
data falsify the Haffer & Prance refuge hypothesis of Amazonian
speciation. Amazoniana, 16, 609–646.
Condit, R., Ashton, P.S., Manokaran, N., LaFrankie, J.V., Hubbell, S.P.
& Foster, R.B. (1999) Dynamics of the forest communities at Pasoh
and Barro Colorado: comparing two 50-ha plots. Philosophical
Transactions of the Royal Society of London B, 354, 1739–1748.
Dutech, C., Maggia, L. & Joly, H.I. (2000) Chloroplast diversity in
Vouacapoua americana (Caesalpiniaceae), a neotropical forest
tree. Molecular Ecology, 9, 1427–1432.
Forget, P.-M. (1991) Scatterhoarding of Astrocaryum paramaca by
Proechymis in French Guiana: comparison with Myoprocta exilis.
Tropical Ecology, 94, 255– 261.
Forget, P.-M. (1997) Effect of microhabitat on seed fate and
seedling performance in two rodent-dispersed tree species in rain
forest in French Guiana. Journal of Ecology, 85, 693–703.
Fragoso, J.M.V. (1997) Tapir-generated seed shadows: scale dependent
patchiness in the Amazon rain forest. Journal of Ecology, 85, 519–529.
Galeano, G. (1991) Las Palmas de la Region de Araracuara. Studies
on the Colombian Amazon 01 (ed. by J.G. Saldarriaga and T. van
der Hammen). Tropenbos-Colombia, Bogota, 180pp.
Gartland, J.S., Newbery, D.McC., Thomas, D.W. & Waterman, P.G.
(1986) The influence of topography and soil phosphorus on the
vegetation of Korup Forest reserve, Cameroun. Vegetatio, 65, 131–148.
Gentry, A.H. & Dodson, C. (1987) Diversity and phytogeography of
neotropical vascular epiphytes. Annals of the Missouri Botanical
Garden, 74, 205–233.
de Granville, J.-J. (1989) La distribucion de las palmas en la Guyana
francesa. Acta Amazônica, 19, 115–138.
Grubb, P.J. (1977) The maintenance of species richness in plant com-
munities: the importance of the regeneration niche. Biological
Reviews, 53, 107–145.
Haffer, J. (1969) Speciation in amzonian forest birds. Science, 165,
131–137.
Harms, K., Condit, R., Hubbell, S.P. & Foster, R.B. (2001) Habitat
associations of trees and shrubs in a 50-ha neotropical forest plot.
Journal of Ecology, 89, 947–959.
Henderson, H., Galeano, G. & Bernal, R. (1995) Field guide to the
palms of the Americas. Princeton University Press, Princeton.
Henry, O. (1999) Frugivory and the importance of seeds in the
diet of the orange-rumped agouti (Dasyprocta leporina) in French
Guiana. Journal of Tropical Ecology, 15, 291–300.
Hubbell, S.P. (1979) Tree dispersion, abundance, and diversity in a
tropical dry forest. Science, 203, 1299–1309.
Hubbell, S.P. (2001) The unified neutral theory of biodiversity and
biogeography. Monographs in population biology. Princeton
University Press, Princeton.
Hubbell, S.P. & Foster, R.B. (1986) Biology, chance and history and
the structure of tropical rain forest tree communities. Community
Ecology (ed. by J. Diamond and T.J. Case), pp. 314–329. Harper
& Row, New York.
Jansen, P.A. & Forget, P.M. (2001) Scatter hoarding and tree regeneration.
Nouragues: dynamics and plant-animal interactions in a neotropical
rainforest (ed. by F. Bongers, P. Charles-Dominique, P.M. Forget &
M. Théry), pp. 275–288. Kluwer Academic Publisher, Dordrecht.
Jansen, P.A., Bartholomeus, M., Bongers, F., Elzinga, J.A., Den Ouden, J.
& Van Wieren, S.E. (2002) The role of seed size. Dispersal by
248 P. Charles-Dominique et al.
© 2003 Blackwell Publishing Ltd, Global Ecology & Biogeography, 12, 237– 248
a scatter-hoarding rodent. Seed dispersal and frugivory: ecology,
evolution and conservation (ed. by D.J. Levey, W.R. Silva and
M. Galetti), pp. 209–225. CABI Publishing, Wallingford.
Julliot, C. (1997) Impact of seed dispersal of red howler monkeys
Alouatta seniculus on the seedling population in the understorey of
tropical rain forest. Journal of Ecology, 85, 431–440.
Kahn, F. (2000) Two Amazonian palm species revalidated: Astrocaryum
farinosum and A. Sociale. Palms, 45, 29–36.
Kahn, F. & de Granville, J.-J. (1992) Palms in forest ecosystems of
Amazonia. Ecological Studies 95. Springer-Verlag, New York.
Kahn, F. & Millán, B. (1992) Astrocaryum (Palmae) in Amazonia, a
preliminary treatment. Bulletin de l’Institut Francais d’Etudes
Andine, 21, 459–531.
Kiltie, R.A. (1981) Distribution of palm fruits on a rain forest floor: why
white-lipped peccaries forage near objects. Biotropica, 13, 234–236.
Nelson, B.W., Ferreira, C.A.C., da Silva, M.F. & Kawasaki, M.L.
(1990) Endemism centres, refugia and botanical collection density
in Brazilian Amazonia. Nature, 345, 714–716.
Nelson, B.W., Karpos, V., Adams, J.B., Oliveira, W.J., Braun, O.P.G.
& do Amaral, I.L. (1994) Forest disturbance by large blowdowns
in the Brazilian Amazon. Ecology, 75, 1994.
Pintaud, J-C., Second, G. & Kahn, F. (in press). Species relationships
and species concepts in Amazonian Astrocaryum (Palmae): an evalu-
ation based on cladistic and phenetic analyses of morphological and
DNA (AFLP) characters. American Journal of Botany, in press.
Pitman, N.C.A., Terborgh, J., Silman, M.R. & Nuñez, V.P. (1999)
Tree species distribution in an upper Amazonian forest. Ecology,
80, 2651–2661.
Poncy, O., Riéra, B., Larpin, D., Belbenoit, P., Jullien, M., Hoff, M.
& Charles-Dominique, P. (1998) The permanent field research
station ‘Les Nouragues’ in the tropical forest of French Guiana:
current projects and preliminary results on tree diversity, structure and
dynamics. Measuring and Monitoring Tropical Forest Diversity:
a Network of Biodiversity Plots. Man and the biosphere series
(ed. by F. Dallmeier and J. Comiskey), pp. 398 – 414. UNESCO
and Parthenon Publishing Group, Carnforth, Lancashire.
Portnoy, S. & Willson, M.F. (1993) Seed dispersal curves: behavior of
the tail of the distribution. Evolutionary Biology, 7, 25–44.
Poulsen, A.D. & Balslev, H. (1991) Abundance and cover of ground
herbs in an Amazonian rain forest. Journal of Vegetation Science,
2, 315–322.
Prance, G.T. (1973) Phytogeographic support for the theory of Pleis-
tocene forest refuges in the Amazon basin, based on evidence from
distribution patterns in Caryocaraceae, Chrysobalanaceae, Dicha-
petalaceae and Lecythidaceae. Acta Amazônica, 3, 5–28.
Riéra, B. & Alexandre, D. (1988) Surface des chablis et temps de renou-
vellement en forêt dense tropicale. Acta Oecologica, 9, 211–220.
Sabatier, D. & Prévost, M.F. (1987) Une forêt a cacaoyers sauvages
sur le haut-camopi, en Guyane Française. ORSTOM, Paris.
Sist, P. (1989a) Peuplement et phénologie des palmiers en forêt
guyanaise (Piste de Saint-Elie). Revue d’Ecologie (Terre et Vie), 44,
113–151.
Sist, P. (1989b) Demography of Astrocaryum sciophilum, an under-
storey palm of French Guiana. Principes, 33, 142–151.
ter Steege, H. & Hammond, D.S. (2001) Character convergence,
diversity, and disturbance in tropical rain forest in Guyana. Eco-
logy, 82, 3197–3212.
Svenning, J.-C. (1999) Microhabitat specialization in a species-rich palm
community in Amazonian Ecuador. Journal of Ecology, 87, 55–65.
Tardy, C. (1998) Paléoincendies naturels, feux anthtropiques et
environnements forestiers de Guyane Française, du Tardiglaciaire à
l’Holocène récent, Approches Chronologique et Anthracologique.
Thèse de Doctorat, Université de Montpellier II.
Tomlinson, P.B. (1990) The structural biology of palms. Clarendon
Press, Oxford.
Tuomisto, H. (1998) What satellite imagery and large-scale field
studies can tell about biodiversity patterns in Amazonian forests.
Annals of the Missouri Botanical Garden, 85, 48–62.
Turchin, P. (1998) Quantitative analysis of movement. Sinauer,
Sunderland, MA.
Van der Steege, J.G. (1983) Bladproductie En Bladfytomassa Van
Een Aantal Palmsoorten Van Het Tropisch Regenbos Van Suriname.
CELOS rapporten no. 138. Universiteit van Suriname, Paramaribo.
Van Roosmalen, M.G.M. (1985) Fruits of the Guianan flora. Insti-
tute of Systematic Botany, Utrecht University, Drukkerij Veenman
B.V., Wageningen.
Wessels Boer, J.G. (1965) The indigenous palms of Suriname. E.J.
Brill, Leiden, the Netherlands.
Willson, M.F. (1993) Dispersal mode, seed shadows, and coloniza-
tion patterns. Vegetatio, 107/108, 261–280.
Worbes, M. (1989) Growth rings, increment and age of trees in inun-
dation forests, savannas and a mountain forest in the neotropics.
IAWA (International Association of Wood Anatomists) Bulletin,
10, 109–122.
Zar, J.H. (1996) Biostatistical Aanalysis, 3rd edn. Prentice Hall,
Englewood Cliffs.
BIOSKETCHES
Pierre Charles-Dominique is a senior researcher at the
CNRS Brunoy. He has worked in Africa and French Guiana
on a broad array of topics in tropical ecology: plant–animals
interactions, seed dispersal by vertebrates and palaeoecology
of the tropics.
Jérôme Chave is a theoretical ecologist at the CNRS
Toulouse. He works on the mechanisms of co-existence of
tropical tree species, and on forest growth modelling.
Marc Dubois is a senior researcher at the CEA Saclay. He is
the director of the ECOFIT programme, and works on
numerous questions ranging from forest modelling to the
dynamics of diseases.
Jean-Jacques de Granville is a botanist at IRD Cayenne and
a specialist in palm systematics. He is the Senior Curator of
the Cayenne Herbarium.
Bernard Riera is a researcher at the CNRS Brunoy. He works
in French Guiana on sylvigenesis, gaps dynamics and
regeneration.
Cécile Vezzoli is a graduate student in animal ecology
working on seed dispersal by neotropical rodents.
