In?uence of soils and topography on Amazonian tree diversity: a
landscape-scale study
Susan G. W. Laurance, William F. Laurance, Ana Andrade, Philip M. Fearnside,
Kyle E. Harms, Alberto Vicentini & Regina C. C. Luiza?o
Abstract
Question: How do soils and topography in?uence
Amazonian tree diversity, a region with generally
nutrient-starved soils but some of the biologically
richest tree communities on Earth?
Location: Central Amazonia, near Manaus, Brazil.
Methods: We evaluated the in?uence of 14 soil and
topographic features on species diversity of rain
forest trees ( 10 cm diameter at breast height),
using data from 63 1-ha plots scattered over an area
of 400 km2.
Results: An ordination analysis identi?ed three ma-
jor edaphic gradients: (1) ?atter areas had generally
higher nutrient soils (higher clay content, carbon,
nitrogen, phosphorus, pH and exchangeable bases,
and lower aluminium saturation) than did slopes
and gullies; (2) sandier soils had lower water storage
(plant available water capacity), phosphorus and
nitrogen; and (3) soil pH varied among sites. Gra-
dient 2 was the strongest predictor of tree diversity
(species richness and Fisher?s a values), with diver-
sity increasing with higher soil fertility and water
availability. Gradient 2 was also the best predictor
of the number of rare (singleton) species, which
accounted on average for over half (56%) of all
species in each plot.
Conclusions: Although our plots invariably supported
diverse tree communities ( 225 speciesha 1), the
most species-rich sites (up to 310 species ha 1) were
least constrained by soil water and phosphorus avail-
ability. Intriguingly, the numbers of rare and common
species were not signi?cantly correlated in our plots,
and they responded differently to major soil and
topographic gradients. For unknown reasons rare
species were signi?cantly more frequent in plots with
many large trees.
Keywords: Density dependence; Distributional ecol-
ogy; Fisher?s a; Permanent plots; Rare species; Soil
chemistry; Soil texture; Soil water; Species diversity;
Species richness; Topography; Tropical trees.
Introduction
Central Amazonia sustains some of the biologi-
cally richest tree communities on Earth (Oliveira &
Mori 1999; Leigh et al. 2004) and faces escalating
pressures from forest colonization, logging and in-
frastructure expansion (Fearnside & Grac?a 2006;
Laurance & Luizao 2007). Understanding the fac-
tors that in?uence Amazonian tree diversity at
varying spatial scales is important for effective con-
servation planning and for assessing the potential
threats from imminent forest conversion on species
survival (Laurance et al. 2001; Hubbell et al. 2008).
Although a number of studies have evaluated tree
community composition and diversity at broad
geographic scales in Amazonia (Prance 1977; Gen-
try 1990; Terborgh & Andresen 1998; Oliveira &
Daly 1999; ter Steege et al. 2000; Oliveira & Nelson
2001; Pitman et al. 2002), fewer have focused on
variation at smaller landscape scales (Phillips et al.
2003; Tuomisto et al. 2003; Valencia et al. 2004).
Laurance, S.G.W. (corresponding author, susan.
laurance@jcu.edu.au) & Laurance, W.F. (bill.laurance
@jcu.edu.au): School of Marine and Tropical Biology,
James Cook University, Cairns, Qld 4870, Australia;
Smithsonian Tropical Research Institute, Apartado
0843-03092, Balboa, Anco?n, Panama.
Andrade, A. (titina@inpa.gov.br): Biological Dy-
namics of Forest Fragments Project, National
Institute for Amazonian Research (INPA) and Smith-
sonian Tropical Research Institute, C.P. 478, Manaus,
AM 69011-970, Brazil.
Fearnside, P.M. (pmfearn@inpa.gov.br): Department
of Ecology, National Institute for Amazonian Re-
search (INPA), C.P. 478, Manaus, AM 69011-970,
Brazil.
Harms, K.E. (kharms@lsu.edu): Department of Biolo-
gical Sciences, Louisiana State University, Baton
Rouge, LA 70803, USA.
Vicentini, A. (vicentini.beto@gmail.com) & Luiza?o,
R.C.C. (rccl@inpa.gov.br): Department of Ecology,
National Institute for Amazonian Research (INPA),
C.P. 478, Manaus, AM 69011-970, Brazil.
Journal of Vegetation Science 21: 96?106, 2010
DOI: 10.1111/j.1654-1103.2009.01122.x
& 2009 International Association for Vegetation Science
Working in our same central Amazonian study
area, Bohlman et al. (2008) recently assessed the in-
?uence of soils, topography and geographic distance
on tree community composition and b-diversity,
but did not consider factors affecting tree species
diversity at local scales (a-diversity). Related work
in this same study area has evaluated the in?uence
of soils on tree (Laurance et al. 1999) and liana
biomass.
Here we assess the effects of soils, topography
and tree size (stem diameter) on tree diversity and
abundance in a central Amazonian landscape span-
ning 400 km2. Like much of the Amazon, the soils
in our study area are generally heavily weathered
and nutrient-poor (Sombroek 1984, 2000), and we
hypothesize that species richness of trees will be
higher in sites that are less nutrient-starved, in line
with some current theories on the determinants of
tropical plant diversity (Gentry 1988; Givnish 1999).
We also hypothesize that plots with many large trees
(which are ecologically dominant and may reduce
the local density of other trees) will have lower tree
species richness simply because they will have fewer
trees overall.
Our analysis is based on 63 1-ha plots in which
nearly all trees [  10 cm diameter at breast height
(dbh)] have been identi?ed to species or morpho-
species level, and in which detailed data on soil
chemistry, texture and topography were collected.
Our ?ndings provide insights into how local edaphic
features in?uence tree diversity in one of the world?s
most hyper-diverse forests.
Methods
Study area
The study area is located 80km north of Man-
aus, Brazil (21300S, 601W). Today, this area is a
partially fragmented landscape spanning 1,000km2
(Lovejoy et al. 1986; Laurance et al. 2002), but the
soil and ?oristic data reported here were collected
before or during initial forest clearing, from January
1981 to January 1987. Rain forests in the study area
are evergreen and terra-?rme (not seasonally ?oo-
ded), and range from 50 to 100m elevation. The
climate is tropically hot, with total rainfall from 1900
to 3500mm. Monthly rainfall averages 4100mm
even in the dry season (June-October), but conditions
can become unusually dry during occasional El Nin?o
years. During a strong drought in 1997, for example,
dry season rainfall was less than a third of normal
(Laurance 2001).
The topography of the study area consists of
undulating plateaux dissected by many stream and
river gullies. Flat areas tend to have high clay (45-
75%) and organic carbon (0.8-3.3%) content, which
are associated with relatively high (although still
very modest) concentrations of important nutrients
such as nitrogen (N) and exchangeable bases (Laur-
ance et al. 1999; Luiza?o et al. 2004; Castilho et al.
2006). On sloping terrain, however, a ?podzoliza-
tion? process occurs over time because lateral water
movement results in the gradual destruction of clay-
rich upper soil horizons. This ultimately leads to the
creation of dendritic valley systems with increasing
sand on lower slopes and valley bottoms (Chauvel
et al. 1987; Bravard & Righi 1989).
The soils in the study area are mostly classi?ed
as xanthic ferralsols (using the FAO/UNESCO sys-
tem, or yellow latisols using the Brazilian system;
Beinroth 1975). Ferralsols are widespread in the
Amazon Basin, heavily weathered, and usually have
a low base saturation. They often are well ag-
gregated, porous and friable, with variable clay
content. Clay particles in ferralsols can form very
durable aggregations, giving the soil poor water-
holding characteristics, even with a high clay con-
tent (Richter & Babbar 1991). Xanthic ferralsols in
the Manaus area are derived from Tertiary deposits
and are typically acidic and very poor in nutrients
such as phosphorus (P), calcium (Ca) and potassium
(K) (Chauvel et al. 1987; Fearnside & Leal-Filho
2001).
Tree communities
For this study we used data from 63 square,
1-ha plots scattered over an area of 400 km2. Plots
were arrayed using a predetermined system of study
grids, irrespective of local topography or soil. With-
in each plot, all trees ( 10 cmdbh) were mapped,
marked with a numbered aluminium tag, and mea-
sured for dbh at 1.3m height or above any
buttresses. A sterile or fertile voucher specimen was
collected for nearly all trees and lodged in the
BDFFP Herbarium, Manaus, Brazil (Laurance
et al. 1998, 2006). On average, 97.6% of the trees in
each plot were identi?ed to species (or genus and
morphospecies) level (range: 94.1-99.7%). Non-
identi?ed trees were excluded from analyses. We ex-
amined relationships between tree size and diversity,
by dividing trees into two size classes: large trees
( 60 cm) and smaller trees (10-59.9 cm).
We generated ?ve parameters that either mea-
sure or potentially in?uence tree diversity in each
plot: (1) number of tree stems; (2) overall species
Amazonian tree diversity 97
richness; (3) Fisher?s a, a diversity index that is quite
insensitive to variation in sample size (Magurran
1988); (4) number of ?rare? (singleton) species, re-
presented by just one individual per plot; and (5)
number of ?common? species, having two or more
individuals per plot. We used Fisher?s a in favour of
other indices, such as rarefaction or Simpson?s in-
dex, because the latter may, by themselves, be
sensitive to varying sample size (Rosenzweig 1995).
Edaphic features
For each plot we derived 12 soil parameters
from soil surface samples (0-20 cm), using ?eld and
laboratory methods detailed in Fearnside & Leal-
Filho (2001) and brie?y summarized here. Although
we did not sample deeper soil strata, surface soils
tend to integrate the nutrient cycle in the forest and
thus represent local site characteristics, and also are
the zone where tree seedlings develop and obtain
nutrients and water (Belknap et al. 2003).
Each 1-ha plot was divided into 25 quadrats of
20m20m each. Within each plot, 9-13 quadrats
were selected for sampling, using an alternating
pattern to provide good coverage of the plot. Within
each quadrat, 15 surface samples were collected at
haphazard locations using a soil auger, then bulked
and sub-sampled. Composite samples for each
quadrat were oven-dried, cleaned by removing
stones and charcoal fragments, then passed through
20 and 2-mm sieves. In all cases, values for soil
parameters were derived separately for each quad-
rat, and then combined to yield a mean value for
each 1-ha plot.
Textural analyses were conducted to separate
samples into percentage clay (particles o0.002-mm
diameter), silt (0.002-0.05mm) and sand (0.05-
2.00mm) components, using the pipette method.
Clay and sand, the dominant soil components, were
so strongly and negatively associated (F1, 615 1049.0,
R25 94.5%, Po0.0001; linear regression) that the
clay-sand gradient could be represented by a single
variable, percentage sand content.
Plant available water capacity (PAWC), a mea-
sure of the amount of water the soil can hold in a
form extractable by plant roots, was estimated as the
difference between ?eld capacity (moisture content
retained in soil under a suction of 0.33 atmospheres)
and wilting point (moisture content retained at 15
atmospheres), using a pressure membrane appara-
tus. As is common practice, samples were dried,
sieved and re-wetted before determining available
water capacity, making the results only an index of
water available to plants in the ?eld.
A pH meter was used to measure soil pH. Total
N was determined by Kjeldahl digestion, and total
organic carbon (C) by dry combustion. Total P was
determined by digestion in HNO31, HClO4 and HF,
and reaction with ammonium molybdate. Soil
phosphate (PO4
3 ) was measured in an autoanalyser
using the molybdenum blue method. Organic
(Walkley-Black) carbon to total nitrogen (C:N)
ratios were calculated to provide an index of N
availability; if C/N415, N tends to be limiting for
plant growth.
Cation exchange capacity (CEC) was the sum of
K1, Ca21, Mg21, Al31 and H1 ions. Total ex-
changeable bases (TEB) were the sum of K1, Ca21,
Mg21 and Na1. Aluminium saturation was
((Al311H1)/CEC)100. Cation concentrations
were derived at the Brazilian Center for Nuclear
Energy and Agriculture (CENA), Piracicaba, Sa?o
Paulo, using atomic emission spectroscopy to assess
K1 and atomic absorption spectrophotometry to
determine Ca21, Mg21, Na1, Al31 and H1. Before
analysis samples were digested in HClO4, HNO3
and H2SO4, with extracts buffered to pH 7.0.
For each plot, slope was the average of the
maximum slope (determined with a clinometer) for
each of the 25 quadrats. Plot aspect (percentage of
quadrats with northern aspects, facing 1-451 or 315-
3601) was determined with a compass. Because our
study area is in the southern hemisphere, northern
aspects receive more direct insolation over the year
than do other aspects.
Data analysis
We used two strategies for data analysis. First,
Pearson correlations were used to search for asso-
ciations between the edaphic and tree diversity
variables, and among different tree diversity and
abundance variables. Where appropriate, a Bonfer-
roni-corrected a-value was employed to reduce the
likelihood of spurious correlations, using an experi-
ment-wise error rate of 0.15 to limit Type II
statistical errors (Chandler 1995). Prior to analysis,
data transformations were used as needed to im-
prove data normality and reduce outliers (i.e.
percentage slope, percentage sand and aspect data
were arcsine-square-root transformed, whereas C:N
ratios were log-transformed). None of the ?ve tree
community variables departed signi?cantly from
normality (P40.10 in all cases; Wilk-Shapiro tests),
so none were transformed.
Second, we used an ordination analysis to
identify major gradients in the edaphic data,
and then tested the effects of these gradients on tree
98 Laurance, Susan G. W. et al.
diversity using multiple linear regressions. This
approach ensures that multiple regressions do
not suffer from colinearity effects because the
ordination axes are statistically independent, and
minimizes the chances of spurious associations be-
cause only a few axes are tested. Best subsets
regressions were used to select the predictors.
Performance of the ?nal regression models was as-
sessed by comparing the standardized residuals to
the ?tted values and to each signi?cant predictor
(Crawley 1993).
We used a robust ordination method, nonmetric
multidimensional scaling (NMS), in the PC-ORD
package (McCune & Mefford 1999). All variables
were weighted equally prior to analysis with the
standardization by maximum method (Noy-Meir
et al. 1975). Randomization tests (n5 250) were
used to determine the number of ordination axes
that explained signi?cantly more variation than
expected by chance.
Results
Tree diversity and abundance
Across the 63 plots, species richness ranged
from 225 to 310 species ha 1, averaging (  SD)
261  18 species a 1. On average, 56% of the spe-
cies in each plot (range 43-65%) were classi?ed as
?rare? (singletons), with the remainder being ?com-
mon? (41 stem per plot). Notably, the numbers of
rare and common species in each plot were not sig-
ni?cantly correlated (r5  0.157, P5 0.22; Pearson
correlation; see supporting Table S1).
Tree density ranged from 521 to 731 stems per
plot, averaging 608  52 stems ha 1. Plots with
many large (  60 cmdbh) trees had lower densities
of smaller (o60 cm dbh) trees (r5  0.497,
Po0.0001), possibly because each large tree dis-
placed many smaller trees, leading to lower stem
densities where large trees were abundant (Fig. 1).
Plots with many stems tended to have somewhat
higher species richness than those with fewer stems,
although the relationship was not signi?cant
(r5 0.193, P5 0.13).
Fisher?s a-values were strongly in?uenced by
the number of rare species in each plot (r5 0.810,
Po0.0001), as expected (i.e. because if species
abundances follow a log-series, then Fisher?s a is
nearly equal to the number of singletons in the
sample; Rosenzweig 1995). Common species, how-
ever, had no signi?cant effect on Fisher?s a values
(r5 0.105, P5 0.41; all Pearson correlations).
Simple correlates of tree diversity
Pearson correlations revealed a number of sig-
ni?cant associations between tree diversity and
edaphic features, even with Bonferroni-corrected P
values (Table 1). Stem densities increased in steeper
areas (Fig. 2), which tend to have poorer soils (higher
sand content and aluminium saturation; lower C, N,
TEB and pH), possibly because such sites supported
few large, competitively dominant trees (for instance,
the density of big trees was strongly and negatively
associated with soil sand content; F1, 615 9.36,
R25 13.3%, P5 0.003; linear regression). Species
richness increased signi?cantly with soil water capa-
city (PAWC) and P. Although only weakly
associated with slope, Fisher?s a was positively cor-
related with many soil fertility variables (lower sand
content, aluminium saturation and C:N ratio, and
higher TEB, N, P and pH) as well as higher PAWC.
Notably, rare and common species had differing
associations with edaphic features (Table 1). Rare
species richness increased with higher N (lower C:N
ratios) and P availability, and also had positive but
weaker associations (Po0.067) with other fertility
variables (low aluminium saturation; high pH and
N) as well as PAWC. Common species were not
signi?cantly associated with any edaphic variable,
but were weakly and positively correlated (P5 0.04)
with PAWC. Surprisingly, there were proportion-
ally more rare species, and fewer common species, in
plots with many large trees (Fig. 3).
Ordination of edaphic gradients
Most of the 14 edaphic variables were sig-
ni?cantly intercorrelated with at least one other
Fig. 1. Relationship between the densities of large ( 60
cmdbh) and smaller (10-59.9 cmdbh) trees in central
Amazonian forest plots.
Amazonian tree diversity 99
edaphic variable (see Table S2). We therefore used
NMS ordination to extract orthogonal axes corre-
sponding to major edaphic gradients in the study
area. Three axes were selected, explaining over 92%
of the total variation (Table 2). Axis 1, which cap-
tured 56% of the variation, described a soil fertility
gradient between ?atter (high C, N and TEB) and
steeper (high sand content and aluminium satura-
tion) sites. Axis 2, capturing 25% of the variation,
described a gradient between clay-rich sites with
high PAWC and soil fertility (high P, N, C, CEC,
and TEB; low aluminium saturation) and sandy
sites with the opposite attributes. Axis 3 explained
12% of the total variation and distinguished among
sites with more acidic soils with low P, and more
basic soils with higher P.
Best subsets and multiple regressions revealed
that all of the tree community parameters were in-
?uenced by at least one major edaphic gradient
(Table 3). Tree density was positively affected by
axes 1 and 3, indicating that tree abundance was
highest in steep, sandy and low-fertility sites. Species
richness, Fisher?s a and rare species richness all
responded positively to axis 2 (Fig. 4), indicating
that all increased in clay-rich sites with higher
PAWC and soil fertility. Common species richness
was signi?cantly affected by all three axes, suggest-
ing that steepness, higher PAWC and, possibly, soil
infertility contributed to higher species numbers.
Table 1. Pearson correlations between soil or topographic features and ?ve parameters describing Amazonian tree diversity
or abundance: tree density, species richness, Fisher?s a index and numbers of ?rare? (plot-level singletons) and ?common?
(non-singletons) species. aBold values are signi?cant using a Bonferroni-corrected a-value (P  0.011). bSamples sizes:
slope, aspect, sand content and soil C, n5 63 plots; other attributes, n5 41 plots. cData arcsine-square-root transformed
prior to correlations. dData log10-transformed prior to correlations.
Attributea,b Attribute mean  SD Stem density Species richness Fisher?s a Rare species Common species
Slopec (1) 12.2  8.8 0.604 0.106 0.274  0.022 0.221
Northern aspectc (%) 20.6  16.9  0.065  0.176  0.137  0.097  0.154
Sand contentc (%) 23.1  16.8 0.483  0.138  0.424  0.157  0.011
Plant-available water capacity 7.4  1.9 0.074 0.513 0.423 0.345 0.321
Soil C (%) 1.61  0.25  0.378  0.095 0.155  0.105 0.002
C:N ratiod 9.8  1.5 0.262  0.391  0.479  0.465 0.074
Cation-exchange capacity 2.49  0.43 0.059  0.008  0.038  0.053 0.070
Aluminium saturation 92.4  1.6 0.488  0.244  0.493  0.337 0.119
Total exchangeable bases 0.196  0.055  0.406 0.176 0.394 0.245  0.090
Delta pH  0.24  0.12  0.042 0.218 0.229 0.066 0.264
Soil pH 4.16  0.25  0.516 0.206 0.441 0.313  0.145
Total N (%) 0.165  0.032  0.468 0.257 0.497 0.290  0.023
Total P (ppm) 121.3  40.6  0.308 0.425 0.527 0.429 0.040
PO4
3 (m.e./100 g dry soil) 0.030  0.006  0.285  0.092 0.106 0.011  0.172
Fig. 2. Relationship between mean slope and density of
trees in Amazonian forest plots. Fig. 3. Relationship between the density of large ( 60 cm
dbh) trees and percentage of rare (singleton) species in
Amazonian forest plots.
100 Laurance, Susan G. W. et al.
Discussion
Edaphic features and tree diversity
Central Amazonia has limited elevational, geo-
logical and climatic variability, and hence species
turnover across the landscape (b diversity) is modest
compared to other, more heterogeneous Neo-
tropical regions (Condit et al. 2002; Bolhman et al.
2008). However, local species richness (a diversity)
of central Amazonian terra-?rme forests is among
the highest recorded anywhere in the world (Oliveira
& Mori 1999). All of the 63 plots in our study area
supported very high tree diversity (  225 species
ha 1) and some were hyper-diverse, with up to
310 species ha 1.
Local edaphic factors (soils and topography)
accounted for at least some of this variability in spe-
cies diversity, consistent with earlier analyses of tree
diversity patterns across the tropics (Ashton & Hall
1992; Wright 1992, 2002; Potts et al. 2002; Leigh et
al. 2004; ter Steege et al. 2006). The most species-rich
sites in our study area appeared the least constrained
by key nutrients such as phosphorus, nitrogen and
exchangeable cations (Tables 1 and 3, Fig. 4). This
suggests that local species diversity in this region is
partly limited by soil nutrients ?especially, we be-
lieve, by phosphorus availability, which tends to be
critically limiting to plant growth in geologically old,
heavily weathered soils (Sollins 1998; Vitousek 2004;
Lambers et al. 2008; Turner 2008). Exchangeable
cations are also generally scarce in these soils (in part
because central Amazonia receives virtually no mar-
ine aerosols, which contain mobile cations such as
calcium, potassium and magnesium; Chadwick et al.
1999) and may also limit local tree diversity (Table
1). In the most nutrient-poor parts of the Amazon
(such as parts of the Rio Negro drainage, Guiana
Shield, and especially white sand soils), ?oras tend to
have lower diversity and contain specialized plant
families (e.g. Lecythidaceae, Duckeondracaeae, Ra-
pateaceae, Rhabdodendraceae, Peridiscaceae) that
tolerate highly oligotrophic conditions (Gentry 1990;
ter Steege et al. 2000, 2006). In sites where nutrient
limitation is less severe, a wider cross-section of the
regional ?ora can apparently become established,
and local tree diversity is enhanced.
The water storage capacity of soils also appears
to limit local tree diversity (Tables 1 and 3, Fig. 4;
see also Wright 1992, 2002). In terms of rainfall and
dry season intensity, the central Amazon is inter-
mediate between drier, seasonal forests of eastern
and southern Amazonia and hyper-wet forests in
western Amazonia. Drier forests in Amazonia sup-
port much lower tree diversity than do wetter,
aseasonal areas (ter Steege et al. 2006), with trees in
drier regions maintaining evergreen canopies only
by virtue of having deep root systems (Nepstad et al.
2002). The central Amazon is considered a biogeo-
graphic crossroad where distinct ?oras from drier
and wetter parts of the basin intermix (Oliveira &
Table 2. Pearson correlations between 14 Amazonian soil
and topographic variables versus three ordination axes
produced by nonmetric multidimensional scaling. aBold
values are signi?cant using a Bonferroni-corrected critical
value (P  0.011). bR2 values for correlations between
ordination distances and distances in the original n-di-
mensional space.
Variable Axis 1a Axis 2a Axis 3a
Slope 0.718  0.220  0.069
North aspect  0.269  0.381  0.358
Sand content 0.856  0.765  0.168
Plant available water capacity  0.056 0.739  0.074
Soil carbon  0.769 0.524 0.290
C:N ratio 0.106  0.404 0.250
Cation-exchange capacity  0.358 0.542 0.495
Aluminium saturation 0.783  0.471  0.066
Total exchangeable bases  0.852 0.560 0.299
Delta pH  0.020 0.163 0.194
Soil pH  0.073 0.157  0.790
Total N  0.810 0.790 0.118
Total P  0.248 0.687  0.570
Phosphate (PO4
3 ) 0.104  0.130  0.267
Variation explained (%)b 55.9 25.0 11.5
Table 3. Signi?cant predictors of tree stem density and diversity in central Amazonia, using best subsets and multiple
regressions.
Response variable Predictors Slope Multiple-regression statistics
F R2 (%) df P
Stem density Axis 1 1 19.67 50.7 2, 38 o0.0001
Axis 3 1
Species richness Axis 2 1 11.57 22.9 1, 39 0.0016
Fisher?s a Axis 2 1 12.98 35.0 1, 39 0.0009
Rare species Axis 2 1 5.97 13.3 1, 38 0.019
Common species Axis 1 1 4.10 24.9 3, 37 0.013
Axis 2 1
Axis 3 1
Amazonian tree diversity 101
Daly 1999; Oliveira & Nelson 2001). Notably, like
drier areas of the basin, the central Amazon is vul-
nerable to periodic El Nin?o droughts, which
increase tree mortality (Williamson et al. 2000;
Laurance et al. 2002) and might have a strong
structuring effect on local tree assemblages. Thus,
we believe that variation in local landform and soils
provides an underlying mechanism whereby tree
species from drier and wetter parts of the Amazon
basin can coexist. On sandy, nutrient-starved soils,
species tolerant of seasonal drought and oligo-
trophic conditions are favoured. In less severe
conditions, a broader range of species, including
drought-sensitive species from wetter parts of the
Amazon, can persist.
Local edaphic features in?uenced tree abun-
dance and size as well as tree diversity. In particular,
steeper, sandier and more nutrient-poor sites sup-
ported higher tree densities (Fig. 2). The most likely
reason for this is that steep sites contain few large
(  60 cm dbh) canopy and emergent trees (see also
Castilho et al. 2006), which competitively reduce the
abundance of smaller trees (Fig. 1). This is con-
sistent with the observation that steeper, nutrient-
poor sites in our study area have high tree densities
but low tree biomass (Laurance et al. 1999), as most
of the trees present are small. Several factors might
explain the paucity of large trees on steep, sandy
slopes: (i) scarce soil nutrients in these sites could
reduce tree growth; (ii) large trees might be prone to
uprooting in steep or sandy sites; or (iii) large trees
in sandy soils might be prone to lethal water de?cits.
Notably, Ashton & Hall (1992) also found fewer
large trees on sandy slopes in Sarawak forests, and
suggested that large trees in such sites were more
susceptible to drought-related mortality.
Diversity of rare and common species
In terms of their environmental correlates, we
encountered some surprising differences between
rare (singleton) and common (non-singleton) spe-
cies. First, within each plot, the species richness of
rare and common species was not signi?cantly cor-
related. We had expected these to be positively
associated, with both sets of species peaking in di-
versity under broadly similar environmental
conditions. The absence of such an association
highlights our limited understanding of how edaphic
and biogeochemical heterogeneity affect tropical
plant communities (John et al. 2007; Townsend et al.
2008; Turner 2008).
Second, it is intriguing that sites with more large
trees supported many locally rare species (Fig. 3).
We speculate that such a relationship might arise
indirectly ? for instance, many rare species might be
near the limit of their geographic range or environ-
mental tolerance, and thereby favour sites with less
nutrient-starved soils. Alternatively, large canopy
and emergent trees, with abundant fruit crops,
might be magnets for mobile frugivores (Kwit et al.
2004) that bring in propagules of new tree species
from afar. It would be interesting to test the gen-
erality of this pattern elsewhere, to see if rare species
tend to cluster around large trees in other tropical
forests.
Third, rare and common species responded very
differently to edaphic variables. Rare species peaked
in diversity on sites with the highest phosphorus and
nitrogen availability (low C:N ratio), whereas com-
mon species diversity showed no signi?cant
association with edaphic variables (Table 1). These
?ndings have some broad similarities to those of
Pitman et al. (2002), who found that tree stands in
western Amazonia were numerically dominated by a
relatively small number of locally common and
Fig. 4. Relationships between a major gradient in soil fer-
tility and water-storage capacity versus species richness
(above) and Fisher?s a-values (below) for Amazonian tree
communities (?water? is plant available water capacity;
?TEB? is total exchangeable bases).
102 Laurance, Susan G. W. et al.
widespread ?oligarch? species, which tended to show
only weak associations with edaphic gradients.
Our tree ?ora in central Amazonia showed a similar
oligarch structure (although ?oristically it over-
lapped little with the forests of western Amazonia,
which occur on more nutrient-rich soils; Bolhman
et al. 2008). The results of these two studies are
generally consistent with a proposal that locally
common species are often widespread edaphic gen-
eralists, whereas rare species tend to be edaphically
more specialized.
Finally, it must be emphasized that much of the
variation in tree diversity encountered was attribu-
table to differing numbers of rare (singleton) species,
which comprised from 43% to 65% of the species
richness of each plot. Such striking rarity is a con-
spicuous feature of central Amazonian forests,
possibly because of their strong nutrient limitation
(Laurance 2001) and because the vast regional spe-
cies pool in Amazonia enhances local biodiversity
via continual species colonization (Oliveira & Daly
1999). Forests such as these could be especially vul-
nerable to habitat fragmentation because their many
rare species are prone to random demographic
events (Melbourne & Hastings 2008) and environ-
mental changes in isolated fragments (Laurance et
al. 2002, 2006).
Tree diversity in central Amazonia
In terms of local tree richness, Amazonia and
northern Borneo are the two most spectacularly di-
verse regions in the world (Leigh et al. 2004). The
western Amazon, where soil fertility is much higher
than in central Amazonia because of inputs of geo-
logically young sediments from the Andes, was once
considered the biologically richest part of the Ama-
zon (Gentry 1988). It is now apparent, however, that
the zone of peak tree diversity extends well into
central Amazonia (Oliveira & Mori 1999; Laurance
2001; ter Steege et al. 2006), with much of this hyper-
diverse region overlaying heavily weathered, highly
acidic and nutrient-poor soils.
How can nutrient-poor forests sustain such high
tree diversity? First, plant species in these forests
have evolved highly ef?cient mechanisms to recycle
and scavenge scarce nutrients, especially phos-
phorus, allowing the forests to maintain relatively
high productivity (Herrera et al. 1978; Stark &
Jordan 1978). Second, as long as nutrient con-
centrations exceed a critical threshold, hyper-diverse
plant communities can evidently persist on heavily
weathered soils ? even those that, from a global
perspective, would be considered unusually nu-
trient-poor (Huston 1979; Aerts & Chapin 2000).
Under such nutrient-limited conditions, plants may
invest heavily in a diversity of defensive compounds
to reduce tissue loss from herbivores and pathogens
(Waterman 1983; Coley et al. 1985). Such diverse
chemical defences likely favour specialized over
generalized herbivores and pathogens (Waterman
1983; Coley & Barone 1996), which in turn could
drive strong density-dependent mortality that helps
to maintain local plant diversity (Janzen 1970; Con-
nell 1971;Wills et al. 1997; Harms et al. 2000; but see
Givnish 1999). Hence, in our study area, only the
most nutrient-starved soils, especially on sandy
slopes with limited water storage, have substantially
reduced tree diversity. Similar patterns might be
evident in species-rich Bornean forests, where tree
diversity peaked in soils of intermediate to low fer-
tility, rather than those that were either more fertile
or exceptionally nutrient-poor (Potts et al. 2002).
From a regional perspective, we believe a com-
bination of factors, such as the vast Amazonian
species pool that enhances local diversity via im-
migration (Terborgh 1973; Ricklefs 2004), strong
density dependence and slow plant growth rates that
collectively limit interspeci?c competition, a mixing
of drier- and wetter-adapted regional ?oras (Whit-
taker 1977; Oliveira & Daly 1999) and possible mid-
domain effects (Colwell & Lees 2000) collectively
underlie the very high diversity of central Amazo-
nian forests. On local and landscape scales, soil and
topographic features can have quite pronounced ef-
fects on local tree diversity, as well as far-reaching
in?uences on community structure (Bolhman et al.
2008) and tree abundance.
Acknowledgements. We thank John Terborgh, Egbert
Leigh, Ben Turner, Richard Condit, Phyllis Coley, Meelis
Partel and an anonymous referee for many useful com-
ments on the manuscript. Support was provided by the US
National Science Foundation, Marisla Foundation, Con-
servation, Food and Health Foundation, Blue Moon Fund,
NASA-LBA program and the Smithsonian Institution. This
is publication number 550 in the technical series of the Bio-
logical Dynamics of Forest Fragments Project.
References
Aerts, R. & Chapin, F.S. 2000. The mineral nutrition of
wild plants revisited: a re-evaluation of patterns and
processes. Advances in Ecological Research 30: 1?67.
Ashton, P.S. & Hall, P. 1992. Comparisons of structure
among mixed dipterocarp forests of north-western
Borneo. Journal of Ecology 80: 459?481.
Amazonian tree diversity 103
Beinroth, F.H. 1975. Relationships between U.S. soil
taxonomy, the Brazilian system, and FAO/UNESCO
units. In: Bornemisza, E. & Alvarado, A. (eds.) Soil
management in tropical America. pp. 97?108. North
Carolina State University, Raleigh, NC, US.
Belknap, J., Hawkes, C.V. & Firestone, M.K. 2003.
Boundaries in nature: two examples from soil.
BioScience 53: 739?749.
Bohlman, S.A., Laurance, W.F., Laurance, S.G.W.,
Nascimento, H.E.M., Fearnside, P.M. & Andrade, A.
2008. Importance of soil, topography, and geographic
distance in structuring central Amazonian tree
communities. Journal of Vegetation Science 19: 863?
874.
Bravard, S. & Righi, D. 1989. Geochemical differences in
an oxisol?spodosol toposequence of Amazonia,
Brazil. Geoderma 44: 29?42.
Castilho, C.V., Magnusson, W.E., Nazare, R., de Araujo,
O., Luizao, R.C.C., Lima, A.P. & Higuchi, N. 2006.
Variation in aboveground tree live biomass in a central
Amazon forest: effects of soil and topography. Forest
Ecology and Management 234: 85?96.
Chadwick, O.A., Derry, L.A., Vitousek, P.M., Hubert,
B.J. & Hedin, L.O. 1999. Changing sources of
nutrients during four million years of ecosystem
development. Nature 397: 491?497.
Chandler, C.R. 1995. Practical considerations in the use of
simultaneous inference for multiple tests. Animal
Behaviour 49: 524?527.
Chauvel, A., Lucas, Y. & Boulet, R. 1987. On the genesis
of the soil mantle of the region of Manaus, central
Amazonia, Brazil. Experientia 43: 234?240.
Coley, P.D. & Barone, J.A. 1996. Herbivory and plant
defenses in tropical forests. Annual Review of Ecology
and Systematics 27: 305?315.
Coley, P.D., Bryant, J.P. & Chaplin, F.S. 1985. Resource
availability and plant herbivore defense. Science 230:
895?899.
Colwell, R.K. & Lees, D.C. 2000. The mid-domain effect:
geometric constraints on the geography of species
richness. Trends in Ecology and Evolution 15: 70?76.
Condit, R., Pitman, N., Leigh, E.G., Chave, J., Terborgh,
J., Foster, R.B., Nunez V, P., Aguilar, S., Valencia, R.,
Villa, G., Muller-Landau, H.C., Losos, E. & Hubbell,
S.P. 2002. Beta-diversity in tropical forest trees.
Science 295: 666?669.
Connell, J.H. 1971. On the role of natural enemies in
preventing competitive exclusion in some marine
animals and in rain forest trees. In: Den Boer, P.J. &
Gradwell, G.R. (eds.) Dynamics of Populations. pp.
298?313. Centre for Agricultural Publishing and
Documentation, Wageningen, NL.
Crawley, M.J. 1993. GLIM for ecologists. Blackwell
Scienti?c Publications, Cambridge, UK.
Fearnside, P.M. & Grac?a, P.M.L.A. 2006. BR-319: Brazil?s
Manaus?Porto Velho Highway and the potential
impact of linking the arc of deforestation to central
Amazonia. Environmental Management 38: 705?716.
Fearnside, P.M. & Leal-Filho, N. 2001. Soil and
development in Amazonia: Lessons from the
Biological Dynamics of Forest Fragments Project. In:
Bierregaard, R.O., Gascon, C., Lovejoy, T.E. &
Mesquita, R. (eds.) Lessons from Amazonia: the
ecology and conservation of a fragmented forest. pp.
291?312. Yale University Press, New Haven, CT, US.
Gentry, A.H. 1988. Tree species richness of upper
Amazonian forests. Proceedings of the National
Academy of Sciences of the USA 85: 156?159.
Gentry, A.H. 1990. Floristic similarities and differences
between southern Central America and central
Amazonia. In: Gentry, A.H. (ed.) Four neotropical
rain forests. pp. 141?157. Yale University Press, New
Haven, CT, US.
Givnish, T.J. 1999. On the causes of gradients in tropical
tree diversity. Journal of Ecology 87: 193?210.
Harms, K., Wright, S.J., Calderon, O., Herna?ndez, A. &
Herre, E.A. 2000. Pervasive density-dependent
recruitment enhances seedling diversity in a tropical
forest. Nature 404: 493?495.
Herrera, R., Merida, T., Stark, N. & Jordan, C.F. 1978.
Direct phosphorus transfer from leaf litter to roots.
Naturwissenschaften 65: 208?209.
Hubbell, S.P., He, F., Condit, R., Borda-da-Agua, L.,
Kellner, J. & ter Steege, H. 2008. How many tree
species are there in the Amazon and how many of
them will go extinct? Proceedings of the National
Academy of Sciences USA 105: 11498?11504.
Huston, M.A. 1979. A general hypothesis of species
diversity. American Naturalist 113: 81?101.
Janzen, D.H. 1970. Herbivores and the number of tree
species in tropical forests. American Naturalist 104:
501?528.
John, R., Dalling, J.W., Harms, K.E., Yavitt, J.B.,
Stallard, R.F., Mirabello, M., Hubbell, S.P.,
Valencia, R., Navarrete, H., Vallejo, M. & Foster,
R.B. 2007. Soil nutrients in?uence spatial distributions
of tropical tree species. Proceedings of the National
Academy of Sciences USA 104: 864?869.
Kwit, C., Levey, D.J. &Greenberg, C.H. 2004. Contagious
seed dispersal beneath heterospeci?c fruiting trees and
its consequences. Oikos 107: 303?308.
Lambers, H., Raven, J.A., Shaver, G. & Smith, S.E. 2008.
Plant nutrient-acquisition strategies change with soil
age. Trends in Ecology & Evolution 23: 95?103.
Laurance, W.F. 2001. The hyper-diverse ?ora of the
central Amazon: an overview. In: Bierregaard, R.O.,
Gascon, C., Lovejoy, T.E. & Mesquita, R. (eds.)
Lessons from Amazonia: the ecology and conservation
of fragmented forest. Yale University Press, New
Haven, CT, US.
Laurance, W.F. & Luizao, R.C.C. 2007. Driving a wedge
into the Amazon. Nature 448: 409?410.
Laurance, W.F., Ferreira, L.V., Rankin-de Merona, J.M.
& Laurance, S.G. 1998. Rain forest fragmentation and
the dynamics of Amazonian tree communities.
Ecology 79: 2032?2040.
104 Laurance, Susan G. W. et al.
Laurance, W.F., Fearnside, P.M., Laurance, S.G.W.,
Delamonica, P., Lovejoy, T.E., Rankin de Merona,
J.M., Chambers, J.Q. & Gascon, C. 1999. Relationship
between soils and forest biomass: a landscape-scale
study. Forest Ecology and Management 118: 127?138.
Laurance, W.F., Cochrane, M.A., Bergen, S., Fearnside,
P.M., Delamonica, P., Barber, C., D?Angelo, S. &
Fernandes, T. 2001. The future of the Brazilian
Amazon. Science 291: 438?439.
Laurance, W.F., Lovejoy, T.E., Vasconcelos, H.L.,
Bruna, E.M., Didham, R.K., Stouffer, P.C., Gascon,
C., Bierregaard, R.O., Laurance, S.G.W. & Sampaio,
E. 2002. Ecosystem decay of Amazonian forest
fragments: a 22-year investigation. Conservation
Biology 16: 605?618.
Laurance, W.F., Nascimento, H., Laurance, S.G.,
Andrade, A., Ribeiro, J., Giraldo, J., Lovejoy, T.E.,
Condit, R., Chave, J. & D?Angelo, S. 2006. Rapid
decay of tree-community composition in Amazonian
forest fragments. Proceedings of the National Academy
of Sciences USA 103: 19010?19014.
Leigh, E.G., Davidar, P., Dick, C.W., Puyravaud, J.P.,
Terborgh, J., Ter Steege, H. & Wright, S.J. 2004. Why
do some tropical forests have so many species of trees?
Biotropica 36: 447?473.
Lovejoy, T.E., Bierregaard, R.O.J., Rylands, A.B.,Malcolm,
J.R., Quintela, C.E., Harper, L.H., Brown, K.S.J.,
Powell, A.H., Powell, G.V.N., Schubart, H.O.R. &
Hays, M.B. 1986. Edge and other effects of isolation
on Amazon forest fragments. In: Soule?, M.E. (ed.)
Conservation biology: the science of scarcity and diversity.
pp. 257?285. Sinauer Associates, Sunderland, MA, US.
Luiza?o, R.C.C., Luiza?o, F.J., Paiva, R., Monteiro, T.,
Sousa, L. & Kruijt, B. 2004. Variation of carbon and
nitrogen cycling processes along a topographical
gradient in a central Amazonian forest. Global
Change Biology 10: 592?600.
Magurran, A.E. 1988. Ecological diversity and its
measurement. Princeton University Press, Princeton,
N.J, US.
McCune, B. &Mefford, M.J. 1999. PC-ORD: multivariate
analysis of ecological data. Version 4. MjM Software
Design, Gleneden Beach, OR, USA.
Melbourne, B. & Hastings, A. 2008. Extinction risk
depends strongly on factors contributing to
stochasticity. Nature 454: 100?103.
Nepstad, D.C., Carvalho, C.R., Davidson, E.A., Jipp,
P.H., Lefebvre, P.A., Negreiros, G.H., Da Silva,
E.D., Stone, T.A., Trumbore, S.E. & Vieira, S. 2002.
The role of deep roots in the hydrological and carbon
cycles of Amazonian forests and pastures. Nature 372:
666?669.
Noy-Meir, I., Walker, D. & Williams, W.T. 1975. Data
transformations in ecological ordination. II. On the
meaning of data standardization. Journal of Ecology
63: 779?800.
Oliveira, A.A. & Daly, D.C. 1999. Geographic
distribution of tree species occurring in the region of
Manaus, Brazil: Implications for regional diversity
and conservation. Biodiversity and Conservation 8:
1245?1259.
Oliveira, A.A. & Mori, S. 1999. A central Amazonian
terra ?rme forest I. High tree species richness on poor
soils. Biodiversity and Conservation 8: 1219?1244.
Oliveira, A.A. & Nelson, B.W. 2001. Floristic relationships
of terra ?rme forests in the Brazilian Amazon. Forest
Ecology and Management 146: 169?179.
Phillips, O.L., Nunez V, P., Monteagudo, A., Cruz, A.P.,
Zans, M.-E., Sanchez, W.G., Yli-Halla, M. & Rose, S.
2003. Habitat association among Amazonian tree
species: a landscape-scale approach. Journal of
Ecology 91: 757?775.
Pitman, N., Terborgh, J., Silman, M., Nunez V, P., Neill,
D.A., Ceron, C., Palacios, W. & Aulestia, M. 2002. A
comparison of tree species diversity in two upper
Amazonian forests. Ecology 83: 3210?3224.
Potts, M.D., Ashton, P.S., Plotkin, J.B. & Kaufmann,
L.S. 2002. Habitat patterns in tropical rainforests: a
comparison of 105 plots in northwest Borneo. Ecology
83: 2782?2797.
Prance, G.T. 1977. The phytogeographic subdivisions of
Amazonia and their in?uence on the selection of
biological reserves. In: Prance, G.T. & Elias, T.S.
(eds.) Extinction is forever. pp. 195?213. New York
Botanical Garden, New York, NY, US.
Richter, B.B. & Babbar, L.I. 1991. Soil diversity in the
tropics. Advances in Ecological Research 21: 315?389.
Ricklefs, R.E. 2004. A comprehensive framework for
global patterns in biodiversity. Ecology Letters 7:
1?15.
Rosenzweig, M.L. 1995. Species diversity in space and
time. Cambridge University Press, Cambridge, UK.
Sollins, P. 1998. Factors in?uencing species composition
in tropical lowland rain forest: does soil matter?
Ecology 79: 29?30.
Sombroek, W. 1984. Soils of the Amazon region. In: Sioli,
H. (ed.) The Amazon: limnology and landscape ecology
of a mighty tropical river and its basin. pp. 521?535. Dr.
W. Junk Publishers, Dordrecht, NL.
Sombroek, W. 2000. Amazon landform and soils in
relation to biological diversity. Acta Amazonica 30:
81?100.
Stark, N. & Jordan, C.F. 1978. Nutrient retention by a
root mat of an Amazonian rain forest. Ecology 59:
434?437.
Terborgh, J. 1973. On the notion of favorableness in plant
ecology. American Naturalist 107: 481?501.
Terborgh, J. & Andresen, E. 1998. The composition of
Amazonian forests: patterns at local and regional
scales. Journal of Tropical Ecology 14: 645?664.
ter Steege, H., Sabatier, D., Castellanos, H., Van Andel,
T., Duivenvoorden, J., Oliveira, A.A., Ek, R., Lilwah,
R., Maas, P. & Mori, S. 2000. An analysis of the
?oristic composition and diversity of Amazonian
forests including those of the Guiana Shield. Journal
of Tropical Ecology 16: 801?828.
Amazonian tree diversity 105
ter Steege, H., Pitman, N., Phillips, O., Chave, J.,
Sabatier, D., Duque, A., Molion, J.-F., Prevost, M.-
F., Spichiger, R., Castellanos, H., von Hildebrand, P.
& Vasquez, R. 2006. Continental-scale patterns of
canopy tree composition and function across
Amazonia. Nature 443: 444?447.
Townsend, A.R., Asner, G.P. & Cleveland, C.C. 2008.
The biogeochemical heterogeneity of tropical forests.
Trends in Ecology and Evolution 23: 424?431.
Tuomisto, H., Ruokolainen, K., Aguilar, M. &
Sarmiento, A. 2003. Floristic patterns along a 43 km-
long transect in an Amazonian rainforest. Journal of
Ecology 91: 743?756.
Turner, B.L. 2008. Resource partitioning for soil phosphorus:
a hypothesis. Journal of Ecology 96: 698?702.
Valencia, R., Foster, R.B., Villa, G., Condit, R.,
Svenning, J.-C., Hernandez, C., Romoleroux, K.,
Losos, E., Magard, E. & Balslev, H. 2004. Tree
species distributions and local habitat variation in the
Amazon: large forest plot in eastern Ecuador. Journal
of Ecology 92: 214?229.
Vitousek, P.M. 2004. Nutrient cycling and limitation:
Hawaii as a model system. Princeton University Press,
Princeton, NJ, US.
Waterman, P.G. 1983. Distribution of secondary
metabolites in rain forest plants: toward an
understanding of cause and effect. In: Sutton, S.L.,
Whitmore, T.C. & Chadwick, A.C. (eds.) Tropical rain
forest: ecology and management. Blackwell Scienti?c,
Oxford, UK.
Whittaker, R.H. 1977. Evolution of species diversity in
land communities. In: Hecht, M.K., Steere, W.C. &
Wallace, B. (eds.) Evolutionary biology. pp. 1?67.
Plenum Publishing, New York, NY, US.
Williamson, G.B., Laurance, W.F., Oliveira, A.A.,
Delamonica, P., Gascon, C., Lovejoy, T.E. & Pohl, L.
2000. Amazonian Tree Mortality during the 1997 El
Nino Drought. Conservation Biology 14: 1538?1542.
Wills, C., Condit, R., Foster, R. & Hubbell, S.P. 1997.
Strong density-dependent and diversity-related effects
help to maintain tree species diversity in a Neotropical
forest. Proceedings of the National Academy of
Sciences USA 94: 1252?1257.
Wright, S.J. 1992. Seasonal drought, soil fertility and the
species density of tropical forest plant communities.
Trends in Ecology & Evolution 7: 260?263.
Wright, S.J. 2002. Plant diversity in tropical forests: a
review of mechanisms of species coexistence.
Oecologia 130: 1?14.
Supporting Information
Table S1. Matrix of Pearson correlations
among tree diversity and abundance variables re-
corded in 63 1-ha plots in central Amazonia.
Correlations in bold were signi?cant using a Bon-
ferroni-corrected a-value (P5 0.015).
Table S2. Matrix of Pearson correlations
among 14 soil and topographic features recorded in
Amazonian forest plots. To improve data normal-
ity, percentage slope, percentage sand and aspect
data were arcsine-square-root transformed, whereas
C:N ratios were log-transformed.
Figure S1.A dawnmist rises from the rain forest
in central Amazonia (photo by William Laurance).
In this issue, Susan and William Laurance and col-
leagues explore how nutrient-starved soils in
Amazonia sustain some of the world?s most diverse
tree communities.
Please note: Wiley-Blackwell is not responsible
for the content or functionality of any supporting
materials supplied by the authors. Any queries
(other than missing material) should be directed to
the corresponding author for the article.
Received 11 April 2009;
Accepted 24 August 2009.
Co-ordinating Editor: M. Pa?rtel.
106 Laurance, Susan G. W. et al.