Ecological Applications, 19(8), 2009, pp. 2142–2156
? 2009 by the Ecological Society of America
Alpha and beta diversity of plants and animals
along a tropical land-use gradient
MICHAEL KESSLER,1,2,10STEFAN ABRAHAMCZYK,1,2,3MERIJN BOS,3,4DAMAYANTI BUCHORI,5DADANG DWI PUTRA,6
S. ROBBERT GRADSTEIN,1PATRICK HO¨HN,3JU¨RGEN KLUGE,1FRIEDERIKE OREND,1RAMADHANIEL PITOPANG,7
SHAHABUDDIN SALEH,7CHRISTIAN H. SCHULZE,8SIMONE G. SPORN,1INGOLF STEFFAN-DEWENTER,3,9
SRI S. TJITROSOEDIRDJO,5AND TEJA TSCHARNTKE3
1Albrecht-von-Haller-Institute of Plant Sciences, University of Go ¨ttingen, 37073 Go ¨ttingen, Germany
2Systematic Botany, University of Zu ¨rich, Zollikerstrasse 107, CH-8008 Zu ¨rich, Switzerland
3Agroecology, University of Go¨ttingen, 37073 Go¨ttingen, Germany
4Louis Bolk Institute, Hoofdstraat 24, 3972 LA Driebergen, The Netherlands
5Faculty of Biology, Bogor Agricultural University, Jalan Padjajaran, 16144 Bogor, Indonesia
6Celebes Bird Club, c/o Balai Penelitian dan Pengembangan Zoologi, Puslitbang Biologi—LIPI, Jl. Raya Bogor Jakarta Km 46,
16911 Gibinong, Indonesia
7Faculty of Agriculture, Tadulako University, Palu, Indonesia
8Department of Population Ecology, Faculty of Life Sciences, University of Vienna, Althanstraße 14, A-1090 Vienna, Austria
9Department of Animal Ecology I, University of Bayreuth, Universita ¨tsstrasse 30, 95440 Bayreuth, Germany
insurmountable task for ecologists. Therefore, researchers frequently sample selected taxa that
they believe reflect general biodiversity patterns. Usually, these studies focus on the
congruence of a diversity (the number of species found per sampling unit) between taxa
rather than on b diversity (turnover of species assemblages between sampling units). Such
approaches ignore the potential role of habitat heterogeneity that, depending on the
taxonomic group considered, can greatly enhance b diversity at local and landscape scales. We
compared a and b diversity of four plant groups (trees, lianas, terrestrial herbs, epiphytic
liverworts) and eight animal groups (birds, butterflies, lower canopy ants, lower canopy
beetles, dung beetles, bees, wasps, and the parasitoids of the latter two) at 15 sites in Sulawesi,
Indonesia, that represented natural rain forest and three types of cacao agroforests differing in
management intensity. In total, we recorded 863 species. Patterns of species richness per study
site varied strongly between taxonomic groups. Only 13–17% of the variance in species
richness of one taxonomic group could be predicted from the species richness of another, and
on average 12–18% of the variance of b diversity of a given group was predicted by that in
other groups, although some taxon pairs had higher values (up to 76% for wasps and their
parasitoids). The degree of congruence of patterns of a diversity was not influenced by
sampling completeness, whereas the indicator value for b diversity improved when using a
similarity index that accounts for incomplete sampling. The indication potential of a diversity
for b diversity and vice versa was limited within taxa (7–20%) and virtually nil between them
(0–4%). We conclude that different taxa can have largely independent patterns of a diversity
and that patterns of b diversity can be more congruent. Thus, conservation plans on a
landscape scale need to put more emphasis on the high heterogeneity of agroforests and the
overarching role of b diversity shaping overall diversity patterns.
Assessing the overall biological diversity of tropical rain forests is a seemingly
species richness; Sulawesi, Indonesia; tropical rain forest.
agroforests; biodiversity indication; community similarity; indicator species; Indonesia;
The immense diversity of tropical plant and animal
communities faces ever-increasing risk of extinction
(e.g., Jetz et al. 2007), but our inadequate taxonomic
knowledge of tropical taxa continues to limit the scope
and extent of biodiversity assessments. Accordingly,
researchers frequently employ selected surrogate taxa
believed to reflect overall levels of biodiversity (Pearson
1994, Prendergast 1997, Pharo et al. 1999, Rodrigues
and Brooks 2007). Such surrogate taxa can be flagship
species (chosen for their charisma), focal taxa (individ-
ual species of particular conservation concern), keystone
species (with high ecological impact), umbrella species
(requiring large areas of habitat, thereby providing
space for other taxa), or indicator taxa (with the same
habitat requirements as the species or communities that
they indicate) (e.g., Paine 1969, Wilcox 1984, Landres et
al. 1988, Mittermeier 1988, Lambeck 1997, Simberloff
Manuscript received 11 June 2008; revised 5 March 2009;
accepted 6 March 2009. Corresponding Editor: H. Hillebrand.
1998, Leader-Williams and Dublin 2000, Bond 2001).
Recent reviews however, concluded that it is not yet
possible to decide when and where surrogate approaches
are effective, how surrogate species should best be
selected, and how to assess the chances of successful
surrogacy (Favreau et al. 2006, Wolters et al. 2006,
Rodrigues and Brooks 2007).
For selecting indicator taxa, the basic approach is to
compare the degree to which patterns of diversity and
community composition coincide or differ between taxa
(Landres et al. 1988, Lawton et al. 1998, Barlow et al.
2007). The congruence of such patterns can be assessed
at various spatial scales and levels of abstraction
(Whittaker 1960, 1972, Hubbell 2001). The classical
distinction is between a diversity derived from species
richness at a given site or spatial level and b diversity
derived from changes in species composition from one
site or spatial level to another (Whittaker 1960, 1972,
Cody 1986, Crist et al. 2003). Most of the diversity at
larger spatial scales is made up of b diversity (Crist et al.
2003, Crist and Veech 2006, Gabriel et al. 2006).
Moreover, a and b diversity may reveal contrasting
spatiotemporal patterns (Tylianakis et al. 2005), and the
extent to which a diversity can predict b diversity can
differ between taxonomic groups. Hence, comparing a
and b diversity at local and landscape scales is an
important yet little-understood area of basic and applied
Although the basic conceptual distinction between a
and b diversity is well established in the ecological
literature, the analytical problems that arise from this
distinction are complex and not yet fully explored,
especially in the context of biodiversity conservation.
First, levels of diversity can be measured at different
spatial scales, which may require partitioning of the
diversity (Whittaker 1960, 1972, Crist et al. 2003, Crist
and Veech 2006). Second, a fundamental complication
arises from the fact that a diversity is assessed based on
sampling units whereas b diversity is calculated based on
differences between sampling units (e.g., Legendre et al.
2005, Tuomisto and Ruokolainen 2006, Jost 2007).
Accordingly, the number and quality of data points
differ greatly. For example, in a study of 10 sites, 10 data
points can be obtained for a diversity and 45 data points
(pairwise similarity values) for b diversity. At the level of
a diversity, species identities are irrelevant during data
analysis, whereas at the level of b diversity it is the
difference between the identities that is taken into
account. At the level of a diversity, regression or
correlation analyses as well as canonical analyses may
be adequate (Magurran 2004). In contrast, at the level of
b diversity, methods for comparing matrices have to be
applied (e.g., the Mantel test) (Legendre and Legendre
1998). Thus, variances analyzed at these different levels
of abstraction are not comparable and there is no simple
relationship between them (Legendre et al. 2005),
although numerous previous studies have confused them
(Tuomisto and Ruokolainen 2006).
Within taxonomic groups, the conclusions drawn
from patterns in a and b diversity can be similar
(Clough et al. 2007), but may also differ greatly
(Tylianakis et al. 2005) because two sites with equal
species richness can share between all and none of their
species. Local species richness of mobile taxonomic
groups may approach regional species richness (Oliver et
al. 1998), whereas assemblages of less mobile species are
expected to differ between sites, leading to an increase in
species turnover and, consequently, in regional diversity.
Ultimately, the assumption that patterns for a and b
diversity change in a similar way, which is underlying
many biodiversity assessments, has been rarely tested
and may not hold true (Tylianakis et al. 2005).
In the tropics, the congruence of diversity patterns
between different taxa has mostly been studied across
large geographical regions (e.g., Beccaloni and Gaston
1995, Carroll and Pearson 1998, Oliver et al. 1998,
Myers et al. 2000, Moore et al. 2002, Duque et al. 2005,
Tushabe et al. 2006, Larsen et al. 2007, McKnight et al.
2007). Only four studies have compared small-scale
changes in taxonomically diverse groups along gradients
of land use within tropical landscapes (Lawton et al.
1998, Schulze et al. 2004, Barlow et al. 2007, No ¨ ske et al.
2008; see Plate 1). This paucity of studies is largely due
to the difficulty of sampling and identifying the
enormous biodiversity of tropical forests. As in most
assessments of indicator taxa (Wolters et al. 2006), two
of these studies focused only on the congruence of a
diversity (Lawton et al. 1998, Schulze et al. 2004).
Additionally, Barlow et al. (2007) and No ¨ ske et al.
(2008) also assessed congruencies between patterns of b
diversity of various taxa along land-use gradients and
found higher congruence of b diversity than of a
diversity between taxa. However, their analyses were
based on observed data for a diversity and a similarity
matrix for b diversity and are therefore statistically
incomparable. Furthermore, none of these studies
compared a diversity with b diversity within taxa.
In the present study, we linked a and b diversity of
four plant groups (trees, lianas, terrestrial herbs,
epiphytic liverworts) and eight animal groups (birds,
butterflies, lower canopy ants, lower canopy beetles,
dung beetles, bees, wasps, and the parasitoids of the
latter two) at 15 sites in natural rain forest and in three
types of cacao agroforests differing in management in
Sulawesi, Indonesia. For a direct comparison of a and b
diversity, we analyzed a diversity not only with linear
regressions, but also by comparing differences in species
numbers, a level we called Da (Appendix A), an additive
analogue to the factorial similarity in species composi-
tion used in analyses of b diversity. Furthermore,
because sampling is typically incomplete in species-rich
tropical communities (Lawton et al. 1998), we compared
observed species richness and similarity as well as
estimated species richness and similarity indices (Colwell
and Coddington 1994).
December 20092143TROPICAL a AND b DIVERSITY
We focused our study on plants and insects as
potential biological indicator taxa because they repre-
sent ;80% of all described species (Herrera and Pellmyr
2002) and determine important ecosystem processes
(Loreau et al. 2001, 2003, Kremen 2005). In particular,
because trees are the main structural elements of forests
and represent crucial food resources for many verte-
brates and insects (Daniels et al. 1992, Davis and Sutton
1998, Fermon et al. 2000, Greenberg et al. 2000, Willott
et al. 2000, Green et al. 2005), they are commonly used
to determine overall forest biodiversity (e.g., Oliver et al.
1998, Williams-Linera et al. 2005). Birds were included
because they are the best-known major group of
organisms and are much-used biodiversity indicators
(Garson et al. 2002, Schulze et al. 2004, Jetz et al. 2007).
By comparing the predictive values of a, Da, and b
diversity of floral and faunal groups, we provide basis
data for the use of indicator taxa in the design of policies
that aim at biodiversity conservation in tropical
Study area and site selection
The study took place in an area of ;10 km2in and
around the village of Toro in the Kulawi Valley, Central
Sulawesi, Indonesia (183002400S, 1208201100E, 800–900 m
above sea level; Fig. 1). Toro is located at the western
border of the 231000-ha Lore Lindu National Park,
;100 km south of Palu, the capital city of Central
Sulawesi. The region has an annual temperature of 24.08
6 0.168C (mean 6 SE) and a monthly rainfall of 143.7
6 22.74 mm. There are no clear seasonal precipitation
fluctuations. The natural vegetation of the National
Park around the village is submontane rain forest.
The agricultural landscape in the region is highly
heterogeneous, consisting of a patchy mosaic of pasture,
hedges, and cacao-dominated agroforests, which is
typical for the region. Cacao production in the region
increased strongly in the 1990s when large areas of
coffee agroforest were converted into cacao agroforests
(Steffan-Dewenter et al. 2007). Cacao agroforests in the
trees; AD, agroforest with diverse planted shade trees; AF, agroforest with few planted shade tree species) around Toro Village in
the Kulawi Valley, Central Sulawesi, Indonesia. The inset shows the location of the island of Sulawesi and of the study region
Location of the study sites in mature forest (MF) and different agroforestry systems (AN, agroforest with natural shade
MICHAEL KESSLER ET AL.2144
Vol. 19, No. 8
Toro village are owned and managed by small-scale
farmers. Shade tree management in the region is
dynamic and farmers tend to remove shade trees in
mature agroforestry systems to increase cacao produc-
tion (Steffan-Dewenter et al. 2007).
We defined a priori four habitat types with a gradient
of shade tree diversity (Gradstein et al. 2007). (1)
Mature forest sites (MF sites) were selected close to the
village, but within the national park, at least 300 m away
from forest sites where selective logging occurred, and
representative of the submontane forest in the area
(Kessler et al. 2005). In the selected sites minor rattan
extraction occurred. (2) We chose cacao agroforests with
diverse, natural shade trees (AN sites), retained after
thinning of the previous forest cover and underplanted
with cacao trees and few fruit trees. These sites had a
long history of cultivation (approximately 20–40 years,
converted from coffee to cacao agroforest approximate-
ly 10 years ago). These agroforests still had high
numbers of native shade trees, including some endemic
species. (3) We selected cacao agroforests with shade tree
stands dominated by various species of planted shade
trees (AD sites). These sites had a shorter history of
cultivation between 15 and 20 years (cacao cultivation
since approximately 10 years ago, sometimes converted
from coffee agroforests), and the majority of the former
forest canopy trees were replaced by various planted
fruit and timber trees that provided the owners with
non-market products. Among these trees were some
native species, including a few endemics. (4) Cacao
agroforests with a low diversity of planted shade trees
(AF sites) had a history of cultivation after a clearcut
12–30 years ago (converted 4–12 years ago from coffee
plantations, rice, or cornfields). Management of these
agroforests was aimed at maximum cacao productivity.
Shade was provided predominantly by the nonindige-
nous leguminous trees Gliricidia sepium (Jacq.) Walp.
and Erythrina subumbrans Merr., which are nitrogen-
fixing. Few native timber or fruit tree species were also
grown, none of which were endemic.
We selected four replicates of each of the four habitat
types, except for AN, where one plot had to be excluded
from the analysis because the local owner cut many of
the shade trees before our sampling was completed.
Agroforest sites were selected based on the age of the
cacao trees, which was at all sites between 4 and 17
years. At the time of this study, agroforestry was non-
intensive in each site, with little use of fertilizers and
pesticides. Farmers regularly pruned trees and weeded
the plantations (two to four times per year).
The minimum distance between study sites was 300 m
and the maximum distance was ;5 km. All sites were
between 850 m and 1100 m above sea level. The
agroforests did not have sharp borders with other
habitat types, but gradually changed into other forms
of land use and at the landscape scale formed a
continuous band along the forest margin. Boundaries
between agroforests were based on ownership rather
than on any physical boundaries. We marked core areas
of 50350 m2in the middle of each site, whose land use
and shade-tree composition was as constant as possible.
Sites belonging to the different habitat types were
geographically interspersed so that none of the individ-
ual habitat types were spatially clustered. Geographical
distance between sites was calculated as the linear
distance between the study plots based on GPS readings
and was log-transformed prior to analysis.
Species collection and determination
Trees.—Each plot was subdivided into 25 subplots of
10 3 10 m2. Within each subplot all trees with diameter
at breast height (dbh; measured at 1.3 m above the
ground surface) ?10 cm were mapped and individually
numbered with aluminum tags, their dbh was measured,
and their trunk height and total height were estimated.
Specimens of all recognizable morphospecies of trees
were collected in sets of at least seven duplicates.
Identification of the plant specimens was done by R.
Pitopang, partly in collaboration with Dr. P. J. A.
Keßler (Leiden, Holland), at Herbarium Celebense,
CEB (Universitas Tadulako, Palu) and Herbarium
Bogoriense, BO (Bogor). Vouchers were deposited in
the herbaria BIOT, BO, CEB, GOET, and L. Specimens
that could not reliably be named were grouped into
Lianas and herbs.—In each study plot of 50350 m210
subplots of 232 m2each were randomly placed. Within
these, all herb and liana species were inventoried,
collected, and determined as detailed for the trees.
Epiphytic liverworts from lower canopy trees.—Two
trees with a height up to 8 m, a dbh ranging between 20
and 60 cm, and comparable bark texture were selected in
each study plot. In the agroforest sites these require-
ments for tree structure were fulfilled by cacao trees and
in natural forest sites by trees in the understory. Each
tree was divided into zone 1 (tree base up to the first
ramification), zone 2 (inner crown), and zone 3 (outer
crown), according to modified Johansson zones for
small trees (Johansson 1974). Within subplots of 200
cm2, liverworts were sampled from each cardinal
direction in all three zones. The identification was
carried out by M. Burghardt, S. R. Gradstein, and
S. G. Sporn (Go ¨ ttingen, Germany). Vouchers were
deposited in CEB, GOET, and L. Specimens that could
not reliably be named were sorted to morphospecies.
Birds.—In February and March 2007, each plot was
visited on two mornings from 05:30 to 10:30. Birds were
recorded visually and acoustically and by systematic
tape recordings (Parker 1991, Abrahamczyk et al. 2008).
For every species we recorded the number of individuals
present simultaneously in the plot. During the second
visit, only additional records (new species or more
individuals of the same species) were considered. Species
flying only above the canopy such as swifts (Apodidae)
and raptors were excluded from the analysis. For
taxonomy we followed Coates and Bishop (1997).
December 20092145TROPICAL a AND b DIVERSITY
Butterflies.—Butterflies were captured alive in traps
baited with rotten mashed bananas. A detailed descrip-
tion of the trap design can be found in Daily and Ehrlich
(1995). Traps were suspended from tree branches with
strings ;1.5 m above the ground. To prevent ants from
entering the traps, branches touching the traps were
removed and the string was prepared with sticky glue.
At each location four traps were set up on the corners of
each study plot. Trapping was conducted in March
2007, with nine days of trapping per study site. The
majority of specimens could be identified in the field
and, therefore, trapped specimens were released imme-
diately afterwards. To avoid pseudoreplicates all but-
terflies were marked with a number on their forewing.
Butterflies were identified according to Aoki et al.
(1982), D’Abrera (1985), Tsukada et al. (1985), Tsukada
(1991), and Vane-Wright and Fermon (2003). Butterfly
communities are known to vary seasonally (Barlow et al.
2007), but in our study area seasonality appears to affect
only the abundance of individual species and not species
richness and composition (C. H. Schulze, unpublished
Ants and beetles from lower canopy trees.—Within
each study plot, four trees were selected, which were of
similar age and size. These were cacao trees in the
agroforests (n ¼ 48; height, 3.4 6 0.56 m) and small,
shade-dwelling lower canopy trees (n¼15; height, 6.3 6
1.90 m) with canopy sizes similar to those of the selected
cacao trees at natural forest sites. At one forest site, ants
and beetles from only three trees could be sampled due
to a technical problem. In order to characterize the
forest insect fauna as completely as possible, we also
sampled insects on a diverse set of trees in the forest
understory. The 15 trees in the forest sites were identified
by R. Pitopang and belonged to 14 species of 10 families.
Only on one occasion were two sampled trees in the
same forest site of the same family. None of the forest
trees were recorded flowering or fruiting at the time of
sampling. At the time of the survey, cacao in the region
was between a main flowering and a harvesting period,
although minor flowering and fruiting occurred
throughout the year. The lower canopy-dwelling ant
and beetle fauna were sampled using canopy knock-
down fogging, which is an effective and widely used
technique for collecting arthropods from tree crowns
(Perfecto et al. 1997, Lawton et al. 1998). With a
SwingFog TF35, a fog of 1% pyrethroid insecticide
(Permethrin; Mitra Envitech, Bogor, Indonesia) was
blown horizontally into the target canopy to avoid
collecting insects from higher canopy layers. Killed
arthropods were collected from a 4-m2sheet of white
canvas placed directly under each tree. We randomly
selected one site per day and sampled all four trees
between 08:00 and 09:00 at the time of day of lowest
wind speed and rainfall probability between 17 Decem-
ber 2003 and 1 January 2004. The collected beetles and
ants were sorted into units based on external morphol-
ogy (morphospecies). Ant sorting was carried out by A.
Rizali (IPB Bogor, Indonesia), based on literature
(Bolton 1994) and reliable digital resources (available
online).11Beetles were identified and sorted by C. Bayer,
B. Bu ¨ che, and A. Rizali. Where necessary, beetle
morphospecies were sorted based on genitalia prepara-
tions. All morphospecies were photographed and posted
on the Internet through which more than 50 specialists
internationally contributed with identifications based on
the photographs and continued further taxonomic work.
Dung beetles.—Dung beetles were collected using
baited pitfall traps as described in Shahabuddin et al.
(2005). Traps were baited with ;20 g of fresh cattle (Bos
taurus) dung. The dung was wrapped in a small square
of textile and fixed with a string at the top of the trap,
which was covered by a metal roof as protection against
sun and rain. Five traps were set up along an 80-m
transect and exposed for two days (Shahabuddin 2007).
After removal from the traps, specimens were preserved
in Scheerpelz solution (75% ethanol, 5% acetic acid, 20%
water) as recommended by F. T. Krell (unpublished
manuscript). Dung beetles were sampled once per month
between March and August 2005. Identification of
specimens was done in close collaboration with B.
Bu ¨ che (Berlin, Germany) using available keys (e.g.,
Balthasar 1963) and the reference collection of the
Bogor Zoological Museum of LIPI (Bogor, Indonesia).
Unidentified species were sorted to morphospecies.
Bees, wasps, and their parasitoids.—Trap nests offer
standardized nesting sites for aboveground-nesting bees
and wasps and can therefore be used to experimentally
study these insects. They were constructed from PVC
tubes with a length of 28 cm and a diameter of 14 cm.
Internodes of the reed Saccharum spontaneum (Poaceae)
with varying diameter (3–25 mm) and a length of 20 cm
were inserted into these tubes to provide nesting sites
(following Tscharntke et al. 1998). Twelve trap nests
(four in each stratum) were installed from October 2004
until September 2005 in three different heights from
understory (U) and intermediate tree height (I) to the
canopy (C), where we placed the trap nests with a
crossbow and a line. Trap nests were checked every
month and bee and wasp larvae were reared for later
identification. Understory was defined as below the
cacao tree canopy, and trap nests were placed 1.5 m
above ground. Intermediate height trap nests were
placed above the cacao tree canopy and below the
shade tree canopy (4 m above the ground in high-
intensity plots and 7 m in primary forest, depending on
canopy structure). Due to technical constraints we
placed the canopy trap nests in the lower part of the
shade tree canopy. Here, trap nest heights varied
between forest habitat type due to different canopy
heights, with higher nests in primary forests and low-
intensity agroforestry systems (primary forest, 19.13 6
0.438 m; low-intensity AF, 20.89 6 0.746 m, n ¼ 16,
11hhttp://www.antweb.orgi and hhttp://www.antbase.neti
MICHAEL KESSLER ET AL.2146
Vol. 19, No. 8
respectively) and lower nests in medium- and high-
intensity agroforestry systems (medium-intensity AF,
16.36 6 0.619 m; high-intensity AF, 15.29 6 0.844 m, n
¼16, respectively). Sticky glue was applied every month
to the edge of the PVC tube to deter ants from
colonizing the trap nests. Individuals from the four trap
nests per plot and stratum and the whole year were
pooled for analysis.
a diversity.—For the analyses of aobsdiversity we used
the number of species recorded in each plot (Appendix
A). Because observed species richness values in field
studies are typically an underestimate of the actual
number of species occurring at a site (Colwell and
Coddington 1994), sampling completeness and estimat-
ed species richness (aestdiversity) were also calculated
using the Chao2 richness estimator (Chao 1987) with
study sites as sample units, using the program EstimateS
8 Windows (Colwell 2006). The Chao2 estimator is
recommended by Walther and Moore (2005) and is
analogous to the estimator used for b diversity. To
assess the potential impact of sampling completeness, we
correlated sampling completeness with the correlation
Da diversity.—In order to more directly compare a
and b diversity, we calculated differences in species
richness between plots using Mantel analyses. As for a
diversity, this was done for the observed and the
estimated richness values. The correlations of sampling
completeness with the correlation values that involved
two taxonomic groups were done using the lower of the
two values of sampling completeness.
b diversity.—Similarities in the species composition of
site pairs were quantified with the quantitative Sørensen
similarity index (also known as Bray-Curtis index),
which takes into account species abundances (Magurran
2004). To correct for incomplete sampling, we further
used the similarity index of Chao et al. (2005), which is
based on the above index but includes an estimation of
incompleteness. As for a diversity, this was done for the
observed and the estimated richness values, as well as
correlating sampling completeness with the Mantel
Correlation analyses.—The correlation of diversity
values between taxonomic groups was calculated using
Spearman correlations (a diversity) and Mantel analyses
(Da and b diversity). Mantel analyses are correlation
tests between matrices consisting of pairwise similarities
or dissimilarities (Legendre and Legendre 1998). Prob-
abilities are assigned by repeatedly randomizing the
arrangement of similarity matrices, each time recalcu-
lating correlation coefficients and comparing the ob-
served correlation value to the randomly generated
ones. All Mantel analyses were conducted with PCOrd
4.5 (McCune and Mefford 1999), applying 9999
randomizations. Mantel analyses were also used to
assess (1) the relationship between turnover in species
composition (b diversity) and geographical distance
between sites and (2) the relationship between Da
diversity and b diversity. Correspondences of the three
measures of diversity (a, Da, and b diversity) within
study groups were assessed using Spearman correla-
tions. When averaging R2values, two different ap-
proaches were used because many original R values had
negative signs that were lost when squaring them. First,
we did not consider negative values. Second, for those
R2values based on negative R values, we maintained the
signs. This was done because negative signs only can be
considered to have high indication value (R2values) if
the negative richness relationship between taxa is known
a priori. Usually, however, the implicit assumtion of
biodiversity indication is that the diversity patterns of
the taxa are positively correlated. In this case, R2values
resulting from squaring negative R values would be
Species richness per site: a and Da diversity
In total, we recorded 863 species, with total species
richness per taxonomic group ranging from nine cavity-
nesting bee species to 198 canopy beetle species (Table
1). Estimated sampling completeness ranged from 29%
for canopy beetles to 89% for cavity-nesting bees (Table
Patterns of species richness per sample site using the
observed species numbers (aobsdiversity) varied between
taxonomic groups along the land-use gradient (Fig. 2).
The aobsdiversity of trees, lianas, liverworts, and dung
beetles was highest in either mature forest stands or
agroforest stands with natural shade trees and declined
with decreasing shade-tree diversity, whereas that of
herbs and canopy beetles showed the opposite pattern.
With highest richness recorded in agroforests with
diverse natural or planted shade trees, aobsdiversity of
butterflies, bees, wasps, and their parasitoids showed a
hump-shaped pattern. Birds had an inversely hump-
(Sobs), total number of species in the study region estimated
through the Chao2 estimator (Sest), sampling completeness
(SC; %Sobsof Sest) for the study region in the Kulawi Valley,
Central Sulawesi, Indonesia, and range of SC for the
individual plots, for the 12 study groups.
Number of observed species in the 15 study sites
SC study region
December 20092147TROPICAL a AND b DIVERSITY
shaped pattern, and ants were the only group with no
Linear correlation coefficients of aobsdiversity per site
between the different taxa varied enormously, ranging
from?0.92 to 0.77 (Figs. 2 and 3, Appendix B). Only 15
of the 66 pairwise comparisons were statistically
significant and 10 of these significant correlations were
negative. The same analysis based on richness values
corrected for incomplete sampling with the Chao
estimator for species richness (aestdiversity) resulted in
similar overall results, with the correlation values of aobs
and aestwithin study groups ranging from R ¼ 0.64 for
butterflies to 0.98 for parasitoids (mean for all groups, R
¼ 0.90; Table 2).
A comparison of Daobsdiversity among taxa using
Mantel correlation analyses gave similar results to those
obtained based on aobsdiversity (correlation between
aobsand Da, R ¼ 0.99, P , 0.001), although R values
were lower and the number of significant relationships
increased to 30 (15 negative; Fig. 3, Appendix B).
Similar results were obtained comparing Daestand aest
For none of the a diversity parameters were correla-
tion coefficients between taxa significantly correlated
with sampling completeness (aobs, R ¼ 0.09, P ¼ 0.63;
Daobs, R¼0.08, P ¼0.69; aest, R¼0.03, P¼0.86; Daest,
R¼0.04, P¼0.76), indicating that differential sampling
completeness did not directionally bias the analyses.
Regional species richness: b diversity
Mantel tests of correlations of bobsdiversity between
taxonomic groups recovered 37 significant relationships,
all of which were positive (Fig. 3, Appendix B).
Repeating the same analysis with values corrected for
incomplete sampling (bestdiversity) led to similar results
(Table 2), although the correlation values were some-
what higher (Fig. 3, Appendix B). With the exception of
liverworts, ants, and bees, which mostly showed low,
nonsignificant correlation values of assemblage similar-
ity with the other groups, turnover in the species
composition of most study groups was significantly
Geographical distance between the study plots was
significantly correlated with turnover in species compo-
sition only for birds, wasps, and parasitoids, in terms of
bobsand bestdiversity (Appendix B).
a and Da vs. b diversity
Within the study groups, patterns of a and Da
diversity were highly correlated, both for observed and
natural shade trees; AD, agroforest with diverse planted shade trees; AF, agroforest with few planted shade tree species). The box
plots show medians (horizontal line in box), 50% (the box limits), and 95% of all values (brackets). Estimated values of a diversity
(aest) show roughly similar patterns and are therefore not shown.
Species richness (aobs) of 12 study groups in mature forest (MF) and three land-use systems (AN, agroforest with
MICHAEL KESSLER ET AL.2148
Vol. 19, No. 8
estimator-corrected values, with values ranging from R
¼ 0.62 to R ¼ 1.00 and averaging R ¼ 0.96 (Table 2).
Similarly, bobsand bestdiversity were highly correlated
(R¼0.94). In contrast, all comparisons of patterns of a
and b diversity resulted in low correlation values,
ranging from R ¼ ?0.35 to R ¼ 0.56 and averaging R
¼ 0.13 (Table 2; Appendix B).
Taxonomic groups differed greatly in their response to
the change from natural forests to shaded cacao agro-
forests and the subsequent loss of shade-tree diversity.
Species richness of trees, lianas, liverworts, and dung
beetles declined with increasing land-use intensity,
whereas species richness of the other taxa either did not
respond at all (ants), had various hump-shaped re-
sponses (birds, butterflies, bees, wasps, and their
parasitoids), or even increased (herbs and canopy
beetles). This variability of a diversity patterns resulted
in low congruence and predictability between the
taxonomic groups (Figs. 1–3). The patterns of species
richness were roughly similar when analyzed with the
different measures (aobs, Daobs, aest, and Daest), which
suggests that the method of analysis and sampling
incompleteness did not strongly influence results (Fig.
3, Table 2).
One of the main questions of our study was the degree
to which patterns of species richness of a given
taxonomic group can be predicted by those of another.
In our comparisons, only between 13% and 17% of the
variance of species richness of one taxonomic group
could be predicted by the species richness of another
(Table 3). These values are in the range of those found
by Lawton et al. (1998) for eight animal groups along a
land-use gradient in Cameroon, but less than the 48%
found by Schulze et al. (2004) for five groups in
Indonesia and the 21–52% found by Barlow et al.
(2007) for 15 groups in Brazil. These numerical
differences likely reflect differences in the studied taxa
as well as the extent of the land-use gradient. The
(the difference of species numbers between plots), and b diversity (the difference of species composition between plots) calculated
with the observed species numbers (obs) and the numbers corrected for sampling incompleteness (est) of each of the 12 study groups
relative to the other 11 groups. The box plots show medians (vertical line in box), 50% (the box limits), and 95% of all values
(brackets), outliers (circles), and far outliers (triangles).
Box plot representation of the correlation (R) values of a diversity (the number of species per study plot), Da diversity
December 20092149TROPICAL a AND b DIVERSITY
habitats studied by Schulze et al. (2004) ranged from
natural forests to annual crop fields devoid of trees, and
those of Barlow et al. (2007) ranged from natural forests
to Eucalyptus plantations. Our study focused on a
comparatively limited range of natural forests and
agroforestry systems. However, our R2values do not
reflect the negative signs of many of the individual
correlation (R) values. If these signs are maintained,
then our average R2values range from ?0.01 to 0.02
(Table 3). Thus, unless the nature of the diversity
relationship between two given study groups was known
a priori, the predictive value of patterns of species
richness between study groups in our study was
essentially zero. Thus, the differences between compar-
isons of taxonomic groups along tropical land-use
gradients likely reflect differences in the study taxa as
well as the extent of the land-use gradient.
The independent patterns that we recorded for the a
diversity of a wide range of taxonomic groups are
comparable with those reported by studies that com-
pared taxa along natural ecotones. For example, along
an elevational gradient in Costa Rica, richness of palms
is highest at 100 m (Chazdon 1987), of trees at 300 m
(Lieberman et al. 1996), of moths and vascular
epiphytes at 1000 m (Cardelu ´ s et al. 2006, Brehm et al.
2007), and of ferns at 1800 m (Kluge et al. 2006). This
difficulty of predicting patterns of species richness by
indicator taxa has as well been documented for forests
on different soil types in Amazonia (Duque et al. 2005,
Tuomisto and Ruokolainen 2005) and for various taxa
in Europe and North America (e.g., Prendergast 1997,
Su et al. 2004, Wolters et al. 2006, Billeter et al. 2008).
Despite the fact that trees are the main structural
elements of forests, tree species richness was in our study
only positively correlated with the species richness of
birds, dung beetles, and lianas. This result is in support
of the idea that large groups of species depend on food
resources that are not directly related to tree species
richness, as has been suggested for canopy beetles
(Wagner 2001), social bees (Klein et al. 2003), and
canopy ants (Philpott and Foster 2005). For example,
canopy beetle assemblages in the study area were
dominated by species associated with dead wood and
related fungi (M. M. Bos and B. Bu ¨ che, unpublished
data), while communities of canopy ants on cacao were
mostly affected by microclimatic changes that were
largely independent of changes in species richness of
shade trees (Bos et al. 2007).
The observation of decreased species richness in five
of seven studied insect groups in mature forest could be
related to pronounced vertical stratification reported for
tropical forest insects (e.g., Rodgers and Kitching 1998,
DeVries et al. 1999, Schulze et al. 2001, Fermon et al.
2005, Diwakar and Balakrishnan 2007). While in some
taxa species richness is more pronounced in one
vegetation layer, others contribute equally to lower
strata as well as the upper canopy (e.g., DeVries et al.
1997, Schulze et al. 2001, Stork and Grimbacher 2006).
Natural and anthropogenic forest disturbance can cause
a breakdown of vertical stratification as documented for
butterflies in selectively logged forest (Dumbrell and Hill
2005, Fermon et al. 2005), at tree-fall gaps (Hill et al.
2001), and forest edges (DeVries et al. 1997). Conse-
quently, lower vegetation strata at agroforestry sites,
particularly sites at the forest edge, may be characterized
by a rich mixture of canopy and understory species,
while lower vegetation strata of forest interior sites are
characterized predominantly by relatively few forest
Congruence of b diversity patterns between study
groups was higher than between a diversity patterns
(Figs. 2 and 3, Table 3). Unlike the results from the a
diversity comparisons, the only negative correlations
were found for liverworts. On average, 12–18% of the
variance in b diversity could be predicted by that in
another taxonomic group, although certain pairs of taxa
TABLE 2.Linear correlations (R values) between different diversity values within the 12 study groups.
Note: Abbreviations are: aobs, observed alpha diversity, i.e., the counted species number per plot; aest, estimated alpha diversity,
i.e., the estimated total species number per plot; Daobs, the difference between the aobsvalues of two plots; Daest, the difference
between the aestvalues of two plots; bobs, observed beta diversity, i.e., the observed similarity in species composition between two
plots; best, estimated beta diversity, i.e., the estimated similarity in species composition between two plots.
MICHAEL KESSLER ET AL.2150
Vol. 19, No. 8
had much higher values, as exemplified by wasps and
their parasitoids (76%) and wasps and canopy beetles
The low correlation coefficients between b diversity of
bees and other study groups (even with their parasitoids)
probably relate to the comparably even distribution of
the nine aboveground nesting bee species in the study
area, with most species present in all habitat types. In
contrast, the other two ‘‘independent’’ study taxa,
namely liverworts and ants, showed marked turnover
between plots and habitats. Their low correlations to
other study groups are presumably based on taxon-
specific ecological requirements. Intriguingly, in Britain
hot spots of liverwort diversity corresponded to cold
spots of birds and butterflies (Prendergast et al. 1993). In
the other taxonomic groups, variability in patterns of b
diversity did not seem to relate to general taxonomic
(e.g., plants vs. animals) or ecological (e.g., relative to
trophic level, mobility) differences, but were rather
group specific. For example, the extremely high turnover
in beetle species relates to the fact that the majority of
the recorded species aggregated on resources that were
little related to habitat type, such as deadwood and fungi
(M. M. Bos and B. Bu ¨ che, unpublished data).
a vs. b diversity
In accordance with previous studies (e.g., Su et al.
2004, Tuomisto and Ruokolainen 2005, Barlow et al.
2007, McKnight et al. 2007, No ¨ ske et al. 2008) our
results show that among the 12 investigated groups, the
use of indicator taxa is most valuable when taking into
account the patterns in b diversity rather than a
diversity. The higher congruence of patterns of b
diversity than of a diversity is readily explained
biologically (Su et al. 2004, McKnight et al. 2007,
Rodrigues and Brooks 2007). Along an ecological
gradient (land use, elevation, climate, soil fertility,
etc.), b diversity of all taxa will tend to shift more
strongly the more divergent the ecological conditions
are. In contrast, species richness can be similar even
under strikingly different conditions (Fig. 4).
Within the studied taxonomic groups, the congruence
of patterns of a and b diversity was fairly low, with R2
values averaging 0.07–0.20 (Table 3). Across the
different taxa, only 1–4% of the variance of b diversity
of a given group could be predicted by the patterns of a
diversity of another group, and vice versa. If the sign of
negative correlations is maintained in calculating these
values, the values are further reduced to 0–1%. The
main conclusion to be drawn from this is that indication
of patterns of b diversity through patterns of a diversity
of another taxon is practically impossible within our
study system and difficult using the same taxon. These
results are not surprising considering the low correlation
values within each of the diversity levels and the fact
measures of diversity.
Mean determination coefficients (R2values) of a given study group relative to the other 11 study groups, for six
Notes: Because the directions of the relationships (negative, positive) are lost when R values are squared, R2values were
calculated both not maintaining the original signs (?) and maintaining them (þ). Abbreviations are: aobs, observed alpha diversity,
i.e., the counted species number per plot; aest, estimated alpha diversity, i.e., the estimated total species number per plot; Daobs, the
difference between the aobsvalues of two plots; Daest, the difference between the aestvalues of two plots; bobs, observed beta
diversity, i.e., the observed similarity in species composition between two plots; best, estimated beta diversity, i.e., the estimated
similarity in species composition between two plots.
December 20092151TROPICAL a AND b DIVERSITY
that concordances across levels of diversity are neces-
sarily less tight (Tylianakis et al. 2005, Clough et al.
Implications for biodiversity sampling
Our study shows that indication of diversity patterns
of a given taxon by another taxon remains a difficult
task (Prendergast 1997, Favreau et al. 2006, Wolters et
al. 2006, Billeter et al. 2008). This has previously been
shown for patterns of a diversity both at the local level
(Lawton et al. 1998, Schulze et al. 2004, Favreau et al.
2006, Wolters et al. 2006) and on regional to continental
scales (Beccaloni and Gaston 1995, Carroll and Pearson
1998, Myers et al. 2000, Moore et al. 2002, Tushabe et
al. 2006, Billeter et al. 2008). We also found that
different approaches to calculating a diversity (species
numbers vs. differences between species numbers;
observed vs. estimated values) resulted in roughly
similar values. In particular, the sampling completeness
of the different study groups did not appear to influence
the observed patterns of a diversity directionally or in a
Our study further showed that b diversity is more
consistent across the study taxa, and although overall
levels of congruence were rather low (see also Tuomisto
and Ruokolainen 2005, No ¨ ske et al. 2008), negative
correlation values were very rare and indication is
possible, if on a low level. Furthermore, in contrast to a
diversity, the correction for sampling incompleteness
increased the indicative value of b diversity, showing
that this approach should be preferred when comparing
b diversity of incomplete samples.
Implications for biodiversity conservation
in tropical landscapes
Because the potential conservation value of an area
depends more on which species occur there, rather than
how many species (Gaston 1996), the reasonably good
turnover of species assemblages between sampling units, also known as the Bray-Curtis similarity) between different taxa. (A) The
species richness of trees and herbs are closely inversely correlated due to the high abundance of herbs in open plantations with few
shade trees. This high species richness of herbs in turn correlates positively with the species richness of butterflies (C), many of
which use the herbs as food plants. Species numbers of trees and butterflies are not significantly correlated (B). In contrast to a
diversity, b diversity is positively correlated between trees and herbs (D) as well as between herbs and butterflies (F), and slightly,
but not significantly so, between trees and butterflies (E). Thus, changes in land use affect b diversity of different taxa in roughly the
same way, whereas trends in a diversity can be completely unrelated or even opposite. Trend lines in nonsignificant relationships are
shown by dotted lines for clarity (B and E).
* P , 0.05; ** P , 0.01; *** P , 0.001.
Examples of the relationships of a diversity (the number of species found per sampling unit) and b diversity (the
MICHAEL KESSLER ET AL.2152
Vol. 19, No. 8
indication of b diversity, in contrast to a diversity, is of
considerable practical interest (McKnight et al. 2007).
Due to contrasting patterns, indication of b diversity
through a diversity is also not possible, suggesting that
diversity assessments based exclusively on patterns of a
diversity miss the important b diversity component.
Perhaps most importantly, our study shows that, at
least at the scale of our study, trees, despite being the
main structural components of forests and agroforests
and providing resources to many other organisms, are
not better suited as indicators than other taxa. Whether
the relatively high correlation values obtained for both
a and b diversity patterns for some taxon pairs such as
trees and birds (a diversity) or wasps and their
parasitoids (b diversity) represent general relationships
that can be used for biodiversity indication remains to
be confirmed by further studies. Our results thus
support those from other multi-taxa studies of the
consequences of land-use change and suggest that the
impact of management changes on the diversity and
composition of a given taxon in tropical agroforests
cannot be predicted reliably from other taxa, although
changes in species composition (b diversity) appear to
be more consistent than those of species richness (a
This study was funded by the German Research Foundation
(DFG), grant SFB-552 STORMA (Stability of Tropical Rain-
forest Margins). We thank Pak Mann, Arifin, Daniel Stieten-
roth, Adam Malik, Wolfram Lorenz, Surya Tarigan, and all
plantation owners for their help in this work. We also thank
Mirkka Jones and Hanna Tuomisto for comments on the
Abrahamczyk, S., M. Kessler, D. D. Putra, M. Waltert, and T.
Tscharntke. 2008. The value of differently managed cacao
plantations for forest bird conservation in Sulawesi, Indo-
nesia. Bird Conservation International 18:349–362.
Aoki, T., S. Yamaguchi, and Y. Uemura. 1982. Satyridae,
Libytheidae. Pages 1–500 in E. Tsukada, editor. Butterflies of
the South East Asian Islands 3. Plapac, Tokyo, Japan.
Balthasar, V. 1963. Monographie der Scarabaeidae und
Aphodiidae der palaearktischen und orientalischen Region,
Band 1. Verlag der Tschechoslowakischen Akademie der
Wissenschaften, Prague, Czechoslovakia.
Barlow, J., et al. 2007. Quantifying the biodiversity value of
tropical primary, secondary, and plantation forests. Proceed-
ings of the National Academy of Sciences (USA) 104:18555–
Beccaloni, G. W., and K. J. Gaston. 1995. Predicting the species
richness of neotropical forest butterflies: Ithomiinae (Lepi-
doptera: Nymphalidae) as indicators. Biological Conserva-
equally high species numbers for many other plant or animal groups in natural forests. Instead, the good light conditions on the
forest floor in plantations favor the growth of herbs and associated fauna. Photo credits: primary forest, S. R. Gradstein; cacao
plantation, Daniele Cicizza.
Although primary forests (left) have many more tree species than cacao plantations (right), this does not translate into
December 20092153 TROPICAL a AND b DIVERSITY
Billeter, R., et al. 2008. Indicators for biodiversity in
agricultural landscapes: a pan-European study. Journal of
Applied Ecology 45:141–150.
Bolton, B. 1994. Identification guide to the ant genera of the
world. Harvard University Press, Cambridge, Massachusetts,
Bond, W. 2001. Keystone species: Hunting the snark? Science
Bos, M. M., I. Steffan-Dewenter, and T. Tscharntke. 2007. The
contribution of cacao agroforests to the conservation of
lower canopy ant and beetle diversity in Indonesia. Biodi-
versity and Conservation 16:2429–2444.
Brehm, G., R. K. Colwell, and J. Kluge. 2007. The role of
environment and mid-domain effect on moth species richness
along a tropical elevational gradient. Global Ecology and
Cardelu ´ s, C. L., R. K. Colwell, and J. E. Watkins. 2006.
Vascular epiphyte distribution patterns: explaining the mid-
elevation richness peak. Journal of Ecology 94:144–156.
Carroll, S. S., and D. L. Pearson. 1998. Spatial modeling of
butterfly species diversity using tiger beetles as a bioindicator
taxon. Ecological Applications 8:531–543.
Chao, A. 1987. Estimating the population size for capture–
recapture data with unequal catchability. Biometrics 43:783–
Chao, A., R. L. Chazdon, R. K. Colwell, and T.-J. Shen. 2005.
A new statistical approach for assessing similarity of species
composition with incidence and abundance data. Ecology
Chazdon, R. L. 1987. The palm flora of Braulio Carrillo
National Park, Costa Rica. Brenesia 28:107–116.
Clough, Y., A. Holzschuh, D. Gabriel, T. Purtauf, D. Kleijn, A.
Kruess, I. Steffan-Dewenter, and T. Tscharntke. 2007. Alpha
and beta diversity of arthropods and plants in organically
and conventionally managed wheat fields. Journal of Applied
Coates, B. J., and K. D. Bishop. 1997. A guide to the birds of
Wallacea. Dove, Alderley, Queensland, Australia.
Cody, M. L. 1986. Diversity, rarity, and conservation in
Meditarranean-climate regions. Pages 123–152 in M. Soule ´ ,
editor. Conservation biology: the science of scarcity and
diversity. Sinauer, Sunderland, Massachusetts, USA.
Colwell, R. K. 2006. EstimateS: Statistical estimation of species
richness and shared species from samples. Version 8. hhttp://
Colwell, R. K., and J. A. Coddington. 1994. Estimating
terrestrial biodiversity through extrapolation. Philosophical
Transactions of the Royal Society B 345:101–118.
Crist, T. O., and J. A. Veech. 2006. Additive partinioning of
rarefaction curves and species–area relationships: unifying
alpha, beta, and gamma diversity with sample size and
habitat area. Ecology Letters 9:923–932.
Crist, T. O., J. A. Veech, J. C. Gering, and K. S. Summerville.
2003. Partioning species diversity across landscapes and
regions: a hierarchical analysis of a, b, c diversity. American
D’Abrera, B. 1985. Nymphalidae, Satyridae and Amathusiidae.
Pages 245–534 in B. D’Abrera, editor. Butterflies of the
Oriental Region II. Hill House, Ferny Creek, Victoria,
Daily, G. C., and P. R. Ehrlich. 1995. Preservation of
biodiversity in small rainforest patches: rapid evaluations
using butterfly trapping. Biodiversity and Conservation 4:35–
Daniels, R. J. R., N. V. Joshi, and M. Gadgil. 1992. On the
relationship between bird and woody plant species diversity
in the Uttara Kannada District of South India. Proceedings
of the National Academy of Sciences (USA) 89:5311–5315.
Davis, A. J., and S. L. Sutton. 1998. The effects of rainforest
canopy loss on arboreal dung beetles in Borneo: implications
for the measurement of biodiversity in derived tropical
ecosystems. Diversity and Distributions 4:167–173.
DeVries, P. J., D. Murray, and R. Lande. 1997. Species
diversity in vertical, horizontal, and temporal dimensions of a
fruit-feeding butterfly community in an Ecuadorian rain-
forest. Biological Journal of the Linnean Society 62:343–364.
DeVries, P. J., T. R. Walla, and H. F. Greeney. 1999. Species
diversity in spatial and temporal dimensions of fruit-feeding
butterflies from two Ecuadorian rainforests. Biological
Journal of the Linnean Society 68:333–353.
Diwakar, S., and R. Balakrishnan. 2007. Vertical stratification
in an acoustically communicating ensiferan assemblage of a
tropical evergreen forest in southern India. Journal of
Tropical Ecology 23:479–486.
Dumbrell, A. J., and J. K. Hill. 2005. Impacts of selective
logging on canopy and ground assemblages of tropical forest
butterflies: implications for sampling. Biological Conserva-
Duque, A. J., J. F. Duivenvoorden, J. Cavalier, M. Sa ´ nchez, C.
Polanı´a, and A. Leo ´ n. 2005. Ferns and Melastomataceae as
indicators of vascular plant composition in rain forests of
Colombian Amazonia. Plant Ecology 178:1–13.
Favreau, J. M., C. A. Drew, G. R. Hess, M. J. Rubino, F. H.
Koch, and K. A. Eschelbach. 2006. Recommendations for
assessing the effectiveness of surrogate species approach.
Biodiversity and Conservation 15:3949–3969.
Fermon, H., M. Waltert, T. B. Larsen, U. Dall’Asta, and M.
Mu ¨ hlenberg. 2000. Effects of forest management on diversity
and abundance of nymphalid butterflies in southeastern Cˆ ote
d’Ivoire. Journal of Insect Conservation 4:173–189.
Fermon, H., M. Waltert, R. I. Vane-Wright, and M. Mu ¨ hlen-
berg. 2005. Forest use and vertical stratification in fruit-
feeding butterflies of Sulawesi, Indonesia: impacts for
conservation. Biodiversity and Conservation 14:333–350.
Gabriel, D., I. Roschewitz, T. Tscharntke, and C. Thies. 2006.
Beta diversity at different spatial scales: plant communities in
organic and conventional agriculture. Ecological Applica-
Garson, J., A. Aggarwal, and S. Sarkar. 2002. Birds as
surrogates for biodiversity: an analysis of a data set from
southern Que ´ bec. Journal of Biosciences 27:347–360.
Gaston, K. J. 1996. Biodiversity: a biology of numbers and
difference. Blackwell, Oxford, UK.
Gradstein, S. R., M. Kessler, and R. Pitopang. 2007. Tree
species diversity relative to human land uses in tropical rain
forest margins in Central Sulawesi. Pages 321–334 in T.
Tscharntke, C. Leuschner, M. Zeller, E. Guhardja, and A.
Bidin, editors. The stability of tropical rainforest margins:
linking ecological, economic and social constraints of land
use and conservation. Springer, Berlin, Germany.
Green, R. E., J. C. Stephen, J. P. W. Scharlemann, and A.
Balmford. 2005. Farming and the fate of wild nature. Science
Greenberg, R., P. Bichier, and A. C. Angon. 2000. The
conservation values for birds of cacao plantations with
diverse planted shade in Tabasco, Mexico. Animal Conser-
Herrera, C. M., and O. Pellmyr. 2002. Plant–animal interac-
tions: an evolutionary approach. Blackwell, Oxford, UK.
Hill, J. K., K. C. Hamer, J. Tangah, and M. Dawood. 2001.
Ecology of tropical butterflies in rainforest gaps. Oecologia
Hubbell, S. P. 2001. The unified neutral theory of biodiversity
and biogeography. Monographs in Population Biology 32.
Princeton University Press, Princeton, New Jersey, USA.
Jetz, W., D. S. Wilcove, and A. P. Dobson. 2007. Projected
impacts of climate and land-use change on the global
diversity of birds. PLoS Biology 5:1211–1219.
Johansson, D. R. 1974. Ecology of vascular epiphytes in West-
African rain forest. Acta Phytogeographica Suecica 59:1–136.
MICHAEL KESSLER ET AL.2154
Vol. 19, No. 8
Jost, L. 2007. Partitioning diversity into independent alpha and
beta components. Ecology 88:2427–2439.
Kessler, M., P. J. A. Keßler, S. R. Gradstein, K. Bach, M.
Schmull, and R. Pitopang. 2005. Tree diversity in primary
forest and different land use systems in Central Sulawesi,
Indonesia. Biodiversity and Conservation 14:547–560.
Klein, A.-M., I. Steffan-Dewenter, and T. Tscharntke. 2003.
Pollination of Coffea canephora in relation to local and
regional agroforestry management. Journal of Applied
Kluge, J., M. Kessler, and R. R. Dunn. 2006. What drives
elevational patterns of diversity? A test of geometric
constraints, climate and species pool effects for pteridophytes
on an elevational gradient in Costa Rica. Global Ecology and
Kremen, C. 2005. Managing ecosystem services: What do we
need to know about their ecology? Ecology Letters 8:468–
Lambeck, R. J. 1997. Focal species: a multi-species umbrella for
nature conservation. Conservation Biology 11:849–856.
Landres, P. B., J. Verner, and J. W. Thomas. 1988. Ecological
uses of vertebrate indicator species, a critique. Conservation
Larsen, F. W., J. Bladt, and C. Rahbek. 2007. Improving the
performance of indicator groups for the identification of
important areas for species conservation. Conservation
Lawton, J. H., et al. 1998. Biodiversity inventories, indicator
taxa and effects of habitat modification in tropical forest.
Leader-Williams, N., and H. T. Dublin. 2000. Charismatic
megafauna as ‘‘flagship species.’’ Pages 53–81 in A. Entwistle
and N. Dunstone, editors. Priorities for the conservation of
mammalian diversity: Has the panda had its day? Cambridge
University Press, Cambridge, UK.
Legendre, P., D. Borcard, and P. R. Peres-Neto. 2005.
Analyzing beta diversity: partitioning the spatial variation
of community composition data. Ecological Monographs 75:
Legendre, P., and L. Legendre. 1998. Numerical ecology.
Second English edition. Developments in environmental
modelling 20. Elsevier, Amsterdam, The Netherlands.
Lieberman, D., M. Lieberman, R. Peralta, and G. S. Hart-
shorn. 1996. Tropical forest structure and composition on a
large-scale altitudinal gradient in Costa Rica. Journal of
Loreau, M., N. Mouquet, and A. Gonzalez. 2003. Biodiversity
as spatial insurance in heterogeneous landscapes. Proceedings
of the National Academy of Sciences (USA) 100:12765–
Loreau, M., S. Naeem, P. Inchausti, J. Bengtsson, J. P. Grime,
A. Hector, D. U. Hooper, M. A. Huston, D. Raffaelli, B.
Schmid, D. Tilman, and D. A. Wardle. 2001. Biodiversity
and ecosystem functioning: current knowledge and future
challenges. Science 294:804–808.
Magurran, A. E. 2004. Measuring biological diversity. Black-
well, Oxford, UK.
McCune, B., and M. J. Mefford. 1999. PC-ORD: multivariate
analysis of ecological data. Version 4.5. MjM Software
Design, Gleneden Beach, Oregon, USA.
McKnight, M. W., P. S. White, R. I. McDonald, J. F.
Lamoreux, W. Sechrest, R. S. Ridgely, and S. N. Stuart.
2007. Putting beta-diversity on the map: broad-scale congru-
ences and coincidences in the extremes. PLoS Biology 5:
Mittermeier, R. A. 1988. Primate diversity and the tropical
forest: case studies from Brazil and Madagascar and the
importance of the megadiversity countries. Pages 145–154 in
E. O. Wilson, editor. Biodiversity. National Academy Press,
Washington, D.C., USA.
Moore, J. L., A. Balmford, T. Brooks, N. D. Burgess, L. A.
Hansen, C. Rahbek, and P. H. Williams. 2002. The
performance of sub-Saharan African vertebrates as indicator
groups for conservation priority setting. Conservation
Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. da
Fonseca, and J. Kent. 2000. Biodiversity hotspots for
conservation priorities. Nature 403:853–858.
No ¨ ske, N. M., N. Hilt, F. A. Werner, G. Brehm, K. Fiedler,
and S. R. Gradstein. 2008. Disturbance effects on diversity of
epiphytes and moths in a montane forest in Ecuador. Basic
and Applied Ecology 9:4–12.
Oliver, I., A. J. Beattle, and A. York. 1998. Spatial fidelity of
plant, vertebrate, and invertebrate assemblages in multiple-
use forest in Eastern Australia. Conservation Biology 12:
Paine, R. T. 1969. A note on trophic complexity and
community stability. American Naturalist 103:91–93.
Parker, T. A. 1991. On the use of tape recorders in avifaunal
surveys. Auk 108:443–444.
Pearson, D. L. 1994. Selecting indicator taxa for the
quantitative assessment of biodiversity. Philosophical Trans-
actions of the Royal Society B 345:75–79.
Perfecto, I., J. Vandermeer, P. Hanson, and V. Cartı´n. 1997.
Arthropod biodiversity loss and the transformation of a
tropical agro-ecosystem. Biodiversity and Conservation 6:
Pharo, E. J., A. J. Beatti, and D. Binns. 1999. Vascular plant
diversity as a surrogate for bryophyte and lichen diversity.
Conservation Biology 13:282–292.
Philpott, S. M., and P. F. Foster. 2005. Nest-site limitation in
coffee agroecosystems: artificial nests maintain diversity of
arboreal ants. Ecological Applications 15:1478–1485.
Prendergast, J. R. 1997. Species richness covariance in higher
taxa: empirical tests of the biodiversity indicator concept.
Prendergast, J. R., R. M. Quinn, J. H. Lawton, B. C. Eversham,
and D. W. Gibbons. 1993. Rare species, the coincidence of
diversity hotspots and conservation strategies. Nature 365:
Rodgers, D. J., and R. L. Kitching. 1998. Vertical stratification
of rainforest collembolan (Collembola: Insecta) assemblages:
description of ecological patterns and hypotheses concerning
their generation. Ecography 21:392–400.
Rodrigues, A. S. L., and T. M. Brooks. 2007. Shortcuts for
biodiversity conservation planning: the effectiveness of
surrogates. Annual Review of Ecology, Evolution, and
Schulze, C. H., K. E. Linsenmair, and K. Fiedler. 2001.
Understorey versus canopy: patterns of vertical stratification
and diversity among Lepidoptera in a Bornean rain forest.
Plant Ecology 153:133–152.
Schulze, C., M. Waltert, P. J. A. Keßler, R. Pitopang,
Shahabuddin, D. Veddeler, M. Mu ¨ hlenberg, S. R. Gradstein,
C. Leuschner, I. Steffan-Dewenter, and T. Tscharntke. 2004.
Biodiversity indicator taxa of tropical land-use systems:
comparing plants, birds, and insects. Ecological Applications
Shahabuddin, G. 2007. Effect of land use on dung beetles
(Coleoptera: Scarabaeidae) diversity and dung decomposi-
tion in Central Sulawesi, Indonesia. Dissertation. Bogor
Agricultural University, Bogor, Indonesia.
Shahabuddin, G., C. H. Schulze, and T. Tscharntke. 2005.
Changes of dung beetle communities from rainforests
towards agroforestry systems and annual cultures. Biodiver-
sity and Conservation 14:863–877.
Simberloff, D. 1998. Flagships, umbrellas, and keystones: Is
single-species management passe in the landscape era?
Biological Conservation 83:247–257.
December 2009 2155TROPICAL a AND b DIVERSITY
Steffan-Dewenter, I., et al. 2007. Tradeoffs between income,
biodiversity, and ecosystem function during rainforest
conversion and agroforestry intensification. Proceedings of
the National Academy of Sciences (USA) 104:4973–4978.
Stork, N. E., and P. S. Grimbacher. 2006. Beetle assemblages
from an Australian tropical rainforest show that the canopy
and the ground strata contribute equally to biodiversity.
Proceedings of the Royal Society B 273:1969–1975.
Su, J. C., D. M. Debinski, M. E. Jakubauskas, and K.
Kindscher. 2004. Beyond species richness: community
similarity as a measure of cross-taxon congruence for
coarse-filter conservation. Conservation Biology 18:167–173.
Tscharntke, T., A. Gathmann, and I. Steffan-Dewenter. 1998.
Bioindication using trap-nesting bees and wasps and their
natural enemies: community structure and interactions.
Journal of Applied Ecology 35:708–719.
Tsukada, E. 1991. Nymphalidae (2). Pages 1–576 in E.
Tsukada, editor. Butterflies of the South East Asian Islands
5. Plapac, Tokyo, Japan.
Tsukada, E., Y. Nishiyama, and M. Kaneko. 1985. Nympha-
lidae (1). Pages 1–558 in E. Tsukada, editor. Butterflies of the
South East Asian Islands 4. Plapac, Tokyo, Japan.
Tuomisto, H., and K. Ruokolainen. 2005. Environmental
heterogeneity and the diversity of pteridophytes and Mela-
stomataceae in western Amazonia. Biologist Skrifter 55:37–
Tuomisto, H., and K. Ruokolainen. 2006. Analyzing or
explaining beta diversity? Understanding the targets of
different methods of analysis. Ecology 87:2697–2708.
Tushabe, H., J. Kalema, A. Byaruhanga, J. Asasira, P.
Ssegawa, A. Balmford, T. R. B. Davenport, and J. Fjeldsa ˚ .
2006. A nationwide assessment of the biodiversity value of
Uganda’s important bird areas network. Conservation
Tylianakis, J. M., A.-M. Klein, and T. Tscharntke. 2005.
Spatiotemporal variation in the diversity of Hymenoptera
across a tropical habitat gradient. Ecology 86:3296–3302.
Vane-Wright, R. I., and H. Fermon. 2003. Taxonomy and
identification of Lohora Moore (Lepidoptera: Satyrinae), the
Sulawesi bush browns. Invertebrate Systematics 17:129–141.
Wagner, M. 2001. Seasonal changes in the canopy arthropod
fauna in Rinorea beniensis in Budongo Forest, Uganda. Plant
Walther, B., and J. L. Moore. 2005. The concepts of bias,
precision and accuracy, and their use in testing the
performance of species richness estimators, with a literature
review of estimator performance. Ecography 28:815–829.
Whittaker, R. H. 1960. Vegetation of the Siskiyou Mountains,
Oregon and California. Ecological Monographs 30:279–338.
Whittaker, R. H. 1972. Evolution and measurement of species
diversity. Taxon 21:213–251.
Wilcox, B. A. 1984. In situ conservation of genetic resources:
determinants of minimum area requirements. Smithsonian
Institution Press, Washington, D.C., USA.
Williams-Linera, G., M. Palacios-Rios, and R. Herna ´ ndez-
Go ´ mez. 2005. Fern richness, tree species surrogacy, and
fragment complementarity in a Mexican tropical montane
cloud forest. Biodiversity and Conservation 14:119–133.
Willott, S. J., D. C. Lim, S. G. Compton, and S. L. Sutton.
2000. Effects of selective logging on the butterflies of a
Bornean rainforest. Conservation Biology 14:1055–1065.
Wolters, V., J. Bengtsson, and A. S. Zaitsev. 2006. Relationship
among species richness of different taxa. Ecology 87:1886–
An example of the calculation of the six different diversity parameters used in the study (Ecological Archives A019-090-A1).
Linear and Mantel correlations between study groups (Ecological Archives A019-090-A2).
MICHAEL KESSLER ET AL.2156
Vol. 19, No. 8