This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier?s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Volume 10 ? No. 3 ? 2009 
? wsvwelsevier.cte/baae 
Bask and 
Applied Ecology 
Provided for non-commercial research and education use. 
Not for reproduction, distribution or commercial use. 
Author's personal copy
Basic and Applied Ecology 10 (2009) 246?254
Simpli?cation of a coffee foliage-dwelling beetle community under
low-shade management
Caleb E. Gordona,, Brian McGillb, Guillermo Ibarra-Nu?n?ezc,
Russell Greenbergd, Ivette Perfectoe
aBiology Department, Lake Forest College, Lake Forest, IL 60045, USA
bBiology Department, McGill University, 1205 Dr Penfield, Montreal, Que., Canada H3A 1B1
cDepartamento de Entomolog??a Tropical, El Colegio de la Frontera Sur (ECOSUR), Apartado Postal 36, Tapachula,
Chiapas, Mexico
dSmithsonian Migratory Bird Center, National Zoological Park, Washington, DC 20008, USA
eSchool of Natural Resources and Environment, University of Michigan, Dana Building, 440 Church Street, Ann Arbor,
MI 48109-1041, USA
Received 1 April 2007; accepted 28 April 2008
Abstract
Coffee agroforests may be structurally and ?oristically complex and may contain a signi?cant fraction of species
from biodiverse and threatened tropical montane forest biotas; hence, understanding the dynamics of tropical forest
biodiversity in coffee agroecosystems has emerged as a centrally important area of tropical conservation biology
research. We conducted a morphospecies analysis on foliage-dwelling beetles collected from coffee plants on four
coffee farms in southern Chiapas, Mexico, to characterize variation in the abundance, species richness, and species
composition of this mega-diverse taxon in relation to coffee cultivation system, spatio-temporal variation, and
predator removal. We constructed thirty-two cages to exclude birds and bats on four farms, each enclosing 7?10 coffee
plants and paired with an adjacent uncaged control plot, and then collected beetles from coffee foliage with D-Vac
aspirators in each plot once every 3 months for one year.
We classi?ed the 2662 beetles collected into 293 morphospecies, representing 42 families of beetles. Extrapolation
and interpolation analyses revealed a very high level of species richness, with no plateau and only a slight leveling trend
observed in our species accumulation curves. We found that low-shade systems contain equal or higher beetle
abundance, lower species richness, more highly homogenized species composition, and higher abundance of coffee
berry borer pests on coffee foliage than do high-shade systems. We observed no effect of ?ying vertebrate exclusion on
the coffee foliage beetle assemblage, but did ?nd signi?cant variation in abundance, species richness, and species
composition of coffee foliage beetles across seasons and study sites.
The increased beetle biodiversity of high-shade coffee cultivation systems has important implications both for the
preservation of native biodiversity in coffee growing regions and for the control of agricultural pests such as the coffee
berry borer.
r 2008 Gesellschaft fu?r O?kologie. Published by Elsevier GmbH. All rights reserved.
ARTICLE IN PRESS
www.elsevier.de/baae
1439-1791/$ - see front matter r 2008 Gesellschaft fu?r O?kologie. Published by Elsevier GmbH. All rights reserved.
doi:10.1016/j.baae.2008.04.004
Corresponding author. Tel.: +847 735 6051; fax: +847 735 6194.
E-mail address: gordon@lfc.edu (C.E. Gordon).
Author's personal copy
Zusammenfassung
Kaffee-Agrarwa?lder ko?nnen strukturell und ?oristisch komplex sein und ko?nnen einen signi?kanten Anteil von
Arten aus biodiversen und gefa?hrdeten tropischen montanen Waldbiotopen enthalten. Deshalb hat sich das
Versta?ndnis der Dynamik der tropischen Waldbiodiversita?t in Kaffee-Agraro?kosystemen als ein zentrales Gebiet der
Forschung in der tropischen Naturschutzbiologie entwickelt. Wir fu?hrten eine Morphospezies-Untersuchung an
laubbewohnenden Ka?fern durch, die auf Kaffeep?anzen in vier Kaffeefarmen im su?dlichen Chiapas, Mexiko,
gesammelt wurden, um die Variation in der Abundanz, im Artenreichtum und in der Artenzusammensetzung dieses
megadiversen Taxons in Bezug zu setzen zum Kaffee-Anbausystem, zur raumzeitlichen Variation und zur Entfernung
der Pra?datoren. Wir konstruierten 32 Ka??ge um Vo?gel und Flederma?use auf vier Farmen auszuschlie?en, von denen
jeder 7-10 Kaffeep?anzen enthielt, und bildeten Paare mit naheliegenden, nicht eingeschlossenen Kontroll?a?chen. Wir
sammelten dann in jeder Fla?che u?ber ein Jahr lang einmal in drei Monaten die Ka?fer mit einem D-Vac-Saugapparat
von den Kaffeebla?ttern. Wir klassi?zierten die 2662 gesammelten Ka?fer in 293 Morphospecies, die 42 Ka?ferfamilien
repra?sentierten. Extrapolations- und Intrapolationsanalysen lie?en einen sehr hohen Grad des Artenreichtums
erkennen, und die Artenakkumulationskurven verliefen ohne Plateau und nur mit einer leicht abfallenden Steigung.
Wir fanden, dass Systeme mit wenig Schatten eine a?hnliche oder ho?here Ka?ferabundanz, einen geringeren
Artenreichtum, eine viel sta?rker homogene Artenzusammensetzung und eine ho?here Abundanz von Scha?dlingen,
die sich in Kaffeebohnen vermehren, aufweisen als Systeme mit viel Schatten. Wir fanden keine Auswirkung des
Ausschlusses von ?iegenden Vertebraten auf die Ka?ferzusammensetzung auf den Kaffeebla?ttern. Wir fanden jedoch
eine signi?kante Vera?nderung in der Abundanz, im Artenreichtum und in der Artenzusammensetzung der Ka?fer auf
den Kaffeebla?ttern mit der Jahreszeit und in den Untersuchungsgebieten. Die erho?hte Ka?ferdiversita?t der schattigen
Kaffeep?anzungen hat wichtige Implikationen sowohl fu?r die Erhaltung der vorhandenen Biodiversita?t in
Kaffeeanbauregionen, als auch fu?r die Kontrolle von landwirtschaftlichen Scha?dlingen, wie dem Kaffeebohnenbohrer.
r 2008 Gesellschaft fu?r O?kologie. Published by Elsevier GmbH. All rights reserved.
Keywords: Coleoptera; Biodiversity conservation; Diversity estimators; Morphospecies; Top-down effects; b diversity; Predator
exclusion; Conservation agroecology
Introduction
Coffee agroecosystems are tropical agroforests managed
to produce an agricultural commodity. Understanding
how tropical forest biotas respond to coffee management
regimes represents a vitally important knowledge frontier
in conservation biology for several reasons: (1) the tropical
forested regions in which coffee is grown are high in bio-
diversity and endemism (Mittermeier, Meyers, Thomsen,
de Fonseca, & Olivieri, 1998; Moguel & Toledo, 1999),
(2) coffee agroecosystems are of great areal and
economic signi?cance worldwide in tropical regions
(FAO, 2007), (3) coffee agroecosystems may contain
complex forest-like vegetation structure and harbour
signi?cant biodiversity (reviewed in Donald, 2004;
Perfecto & Armbrecht, 2003; Perfecto, Armbrecht,
Philpott, Soto-Pinto, & Dietsch, in press; Somarriba,
Harvey, Samper, Anthony, Gonza?lez et al., 2004).
Of particular interest is the response of biodiversity to
different shade management strategies. Variation in the
cultivation techniques used by coffee growers has
created tremendous variation in the structure and
?oristic diversity of the shade stratum, or canopy layer,
of coffee agroecosystems. Despite the complexity of this
variation, studies of biodiversity in coffee to date allow a
meaningful distinction to be made between high-shade
systems (e.g. ??rustic coffee?? of Moguel & Toledo,
1999, ??diverse shade?? of Perfecto, Vandermeer, Lo?pez
Bautista, Ibarra Nun?ez, Greenberg et al., 2004, ??bajo monte
coffee?? of Gordon, Manson, Sundberg, & Cruz-Ango?n,
2007) and low-shade systems (e.g. ??sun coffee,?? ??shaded
monoculture,?? of Moguel & Toledo, 1999, ??monodo-
minant shade?? of Perfecto et al., 2004). High-shade
systems typically harbor more species than do low-shade
systems for a wide variety of taxa (e.g. birds: Gordon
et al., 2007; butter?ies: Mas & Dietsch, 2004; spiders:
Pinkus Rendo?n, Leo?n Corte?s, & Ibarra Nun?e?z,
2006; but see Klein, Steffan-Dewenter, Buchori, &
Tscharntke, 2002 for a counter example with trap-
nesting bees and wasps). This difference has been
attributed to various features of high-shade systems,
including their tall canopies (Gordon et al., 2007),
shade tree species richness (Philpott, Perfecto, &
Vandermeer, 2006), abundant and diverse epiphytes
(Hietz, 2005), and dense shade (Perfecto, Armbrecht,
Philpott, Soto-Pinto, & Dietsch, in press) in relation to
low-shade systems.
This study makes two principal contributions to our
understanding of biodiversity in high- vs. low-shade
coffee agroecosystems. First, we present an extensive
and taxonomically broad study of species richness and
composition patterns in a beetle assemblage from a
ARTICLE IN PRESS
C.E. Gordon et al. / Basic and Applied Ecology 10 (2009) 246?254 247
Author's personal copy
coffee agroecosystem. Most prior studies of beetle
faunas in coffee have examined total abundance
patterns from foliage-collected beetles (Philpott,
Greenberg, Bichier, & Perfecto, 2004), have sampled
from a few shade trees (Perfecto, Hansen, Vandermeer,
& Cart??n, 1997), or have included only dung and/or
carrion beetles (e.g. Arellano, Favila, & Huerta, 2005;
Pineda, Moreno, Escobar, & Halffter, 2005). Using
beetle morphospecies analysis with pyrethrum knock-
down samples from coffee plants in western Ecuador,
Richter, Klein, Tscharntke, and Tylianakis (2007) found
high beetle species richness near coffee plantation edges,
but no difference between traditionally managed, and
abandoned coffee plantations. The species richness of
foliage-dwelling beetles in tropical forest communities is
well-known, and high enough to have prompted order
of magnitude increases in the estimates of the total
species diversity of the Earth (Erwin, 1982). Perhaps
related to this extreme species richness, foliage-dwelling
beetles represent a wide diversity of trophic guilds and
ecological roles, including many different types
of predators, herbivores, and fungivores (Arnett &
Thomas, 2001; Arnett, Thomas, Skelley, & Frank,
2002). Species-level data in this megadiverse group
therefore provides an extremely information-rich ecolo-
gical ?ngerprint, with potential to make signi?cant
contributions to our understanding of the dynamics of
tropical forest biodiversity in coffee agroecosystems.
Second, we deepen our understanding of the ecologi-
cal processes that underlie biodiversity patterns in coffee
agroecosystems by analyzing beetle community varia-
tion over time, space, cultivation technique, and in
response to predator removal. Previous studies have
suggested that birds and/or bats can signi?cantly
depress the abundance of arthropods on the foliage of
Inga shade trees in coffee plantations (Philpott et al.,
2004), or on the foliage of the coffee bushes, themselves
(Greenberg, Bichier, Cruz-Ango?n, MacVean, & Perez
et al., 2000; Perfecto et al., 2004). Jedlicka, Greenberg,
Perfecto, Philpott, and Dietsch (2006) found that ?ying
vertebrates depressed arthropod abundance in the
canopy of Inga shade trees but not in the coffee bushes
of the understory. Herein, we add to this emerging
picture of top-down effects in coffee agroecosystems by
analyzing the effects of ?ying vertebrate exclusion on the
abundance, species richness, and species composition of
beetles on coffee foliage.
Methods
Study sites
This study was conducted on four coffee farms in the
Soconusco region of the state of Chiapas in southern
Mexico: Belen high-shade (151150N, 991220W); Belen low-
shade (151150N, 921190W); Irlanda (151110N, 921200W);
and Hamburgo (151100N, 921190W). These four farms
collectively represent two pairs of adjacent farms. The
pairs are separated by approximately 10km, and each pair
contains one farm under high-shade cultivation and one
farm under low-shade cultivation. No coffee berry borer
control techniques were being practiced on any of the
farms during the period of this study. Detailed character-
izations of the vegetation structure, ?oristics, and manage-
ment of these four farms can be found in Mas and Dietsch
(2003), Philpott et al. (2006), Philpott, Perfecto, and
Vandermeer (in press) Belen high- and low-shade farms
are located within the municipality of Huixtla, are
described as traditional and commercial polycultures
(Moguel & Toledo, 1999), respectively, by Philpott et al.
(in press), and are hereafter referred to as Huixtla high-
shade and Huixtla low-shade. The other two farms,
Irlanda (high-shade) and Hamburgo (low-shade), are
located in the municipality of Tapachula, are described as
commercial polyculture and shade monoculture (Moguel
& Toledo, 1999), respectively, by Philpott et al. (in press),
and are hereafter referred to as Tapachula high-shade and
Tapachula low-shade. All farms are located between 950
and 1150m elevation above sea level and receive ca.
4500mm of rain per year (data provided by B. Peters of
Irlanda farm). Although these four farms can generally be
classi?ed as high-shade and low-shade, they are all
managed in different ways, and represent a gradient of
management intensi?cation based on diversity, density
and height of shade trees and percent shade cover with
Huixtla high-shade 4Tapachula high-shade 4Huixtla
low-shade 4Tapachula low-shade (Mas & Dietsch 2003;
Philpott et al., 2006).
Flying vertebrate exclusion
Large (ca. 10m 5m 3m) bird/bat exclosures
were established in each of the four farms: 6 in Huixtla
high-shade, 6 in Huixtla low-shade, 10 in Tapachula
high-shade and 10 in Tapachula low-shade, for a total
of 32 exclosures. Exclosures were constructed of
transparent mono?lamentous nylon (5 cm mesh) ?shing
net and established in November of 2000. Each
exclosure enclosed at least ten coffee plants, with the
exception of three exclosures that enclosed seven plants,
one that enclosed eight plants, and one that enclosed
nine plants. The same numbers of control plants were
selected from a parallel row of coffee approximately
2?3m from the paired exclosure, resulting in a total
of 616 coffee plants sampled, half inside the enclosures
and half outside.
Beetle sampling and classi?cation
Arthropods were sampled using a D-vac, a reversed
leaf blower modi?ed with a ?ne mesh that allowed for
ARTICLE IN PRESS
C.E. Gordon et al. / Basic and Applied Ecology 10 (2009) 246?254248
Author's personal copy
the collection of very small arthropods. Two coffee
branches were randomly selected for arthropod sam-
pling from each of the coffee bushes in each plot.
Samples were taken a few days after the installment
of the exclosures (November 2000), and at 3 months
(February 2001), 6 months (May 2001), and 9 months
(August 2001) after the establishment of the
exclosures (see Appendix A: Graph 2). Arthropod
collection was always performed during the ?rst
three hours of daylight, and not during rain or
excessive wind. All arthropods were sorted to order.
Beetles (Order: Coleoptera) were placed into separate
vials containing 70% alcohol for each plot date
sample.
Beetle morphospecies analysis was conducted
by Gordon, who examined all beetle specimens and
?rst identi?ed them to family using Arnett and
Thomas (2001), Arnett et al. (2002), Borror, Triplehorn,
and Johnson (1989), following the taxonomy of Arnett
and Thomas (2001) and Arnett et al. (2002). Morpho-
logical features were then used to classify all specimens
in each family into morphospecies. To provide limited
ground-truthing of the morphospecies classi?cation,
specimens from ?ve families representing 38 of the
293 total morphospecies (13%) were examined and
identi?ed by taxonomic specialists in their respective
groups.
Statistical methods
One primary goal was to analyze species richness
(a-diversity) according to the various treatment factors.
To perform comparative analysis between different
factors, we pooled all data within a factor and then
used rarefaction analysis, in which the richness is
estimated for some Nomax (N1,N2), and then com-
pared across treatment factors (Heck, Van Belle, &
Simberloff, 1975; Hurlbert, 1971; Sanders, 1968;
Simberloff, 1972).
The second main goal was to explore b-diversity,
the variation in species composition across space or
time (Legendre & Legendre, 1998). To do this, we used
the Morisita?Horn Index, and the Jaccard Index (see
Appendix A: Table 1 for analyses using additional
diversity indices). The Morisita?Horn Index takes
abundance information into account, whereas the
Jaccard Index uses only presence/absence information
(Magurran, 2004). These measures of community
similarity were all converted to dissimilarity measures
(i.e. distance) ? (1?similarity). These distances were
then used to create a dendrogram using nearest
neighbor-joining on group averages (UPGMA) as
implemented in MATLAB?s LINKAGE function. All
statistical analyses were performed in MATLAB v
2006b (Mathworks Incorporated, 2006).
ARTICLE IN PRESS
Fig. 1. Rarefaction comparisons. Each graph plots expected species richness (S) on the y-axis vs. sample size (N) on the x-axis. The
dotted enclosing shapes represent 95% con?dence intervals. (A) High-shade vs. low-shade management. (B) Flying vertebrate
exclosure cage vs. uncaged control. (C) Site effects: the curves with fewer than 500 individuals are for the two Huixtla sites, while the
curves with much greater than 500 individuals represent the Tapachula sites. The dashed lines are for low-shade and the solid lines
are for high-shade. (D) Seasonal effects.
C.E. Gordon et al. / Basic and Applied Ecology 10 (2009) 246?254 249
Author's personal copy
Results
Accuracy of morphospecies classi?cation
Our ability to ground-truth the morphospecies
classi?cation was severely limited both by the limited
availability of expertise for many beetle taxa, and also
by the presence of undescribed species in all taxa
examined by taxonomic specialists. Among the 38
morphospecies from ?ve arbitrarily selected beetle
families sent out for review by taxonomic experts, the
morphospecies analysis contained two instances of
inaccurate splitting, one instance of inaccurate lumping,
and one instance of probable inaccurate splitting. This
suggests that the correspondence between morphospe-
cies and actual species was generally high.
a-Diversity (species richness)
In a sample with 2662 individual beetles, we identi?ed
293 morphospecies from 42 families. The rarefaction
curves depicted in Fig. 1 are still rising rapidly even at
the largest sample sizes, indicating that the assemblage
of beetles dwelling on coffee plants is extraordinarily
rich.
Effects of treatments on a-diversity
The rarefaction curves indicate that:
1. High-shade sites have signi?cantly greater richness
than low-shade sites (Fig. 1A).
2. One area (Huixtla) has signi?cantly greater richness
than the other area (Tapachula) (Fig. 1C).
3. There is signi?cant variation in richness between
months, with February and May (dry season and
beginning of wet season, respectively) signi?cantly
less species rich than August and November (wet
season and end of wet season, respectively) (Fig. 1D).
4. There is virtually no difference in richness between
plots with ?ying vertebrate exclusion cages and
control plots (Fig. 1B).
b-Diversity (species composition)
Species composition patterns reveal a homogenizing
effect of low-shade management. The high value of the
Morisita?Horn index obtained for the comparison
between the two low-shade sites stands out from all of
the other pairwise comparisons between sites (Fig. 2A,
and see Appendix A: Table 1). The patterns depicted in
Fig. 2 suggest that moving to a low-shade management
ARTICLE IN PRESS
0 0.2 0.4 0.6 0.8 1
Huixtla low shade
Tapachula low shade
Tapachula high shade
Huixtla high shade
0 1
Tapachula low shade
Tapachula high shade
Huixtla low shade
Huixtla high shade
0.40.2 0.6 0.8
Fig. 2. Homogenization effect of production. The dendrograms illustrate the relationships between the two geographic areas and
the two management types, using the abundance-based Morisita?Horn index (A) and the incidence-based Jaccard index (B). The
x-axis indicates branch length.
C.E. Gordon et al. / Basic and Applied Ecology 10 (2009) 246?254250
Author's personal copy
style moves the two sites away from the strong
differences in composition found in the high-shade
management to a highly homogenized community
structure (i.e. both low-shade sites are similar). This
pattern holds true for all indices of similarity that
use abundance information but is less distinct for
indices that use only presence/absence information (see
Appendix A: Table 1).
Species composition patterns also reveal high
b-diversity over both space and time in this system.
The abundance-based Morisita?Horn Index re?ects
signi?cant changes in species composition between
months and between the two geographically separated
areas (see Appendix A: Table 1). The between-month
and between-area values for this index stand in contrast
to the higher values of compositional similarity between
?ying vertebrate exclusion and control plots, indicating
relatively weak effects of predator removal on beetle
community composition.
Abundance
Comparisons of the numbers of individual beetles
collected per sample reveal high beetle abundance
in May relative to the other 3 months of sampling
(Fig. 3C), and at the Tapachula low-shade site relative
to the other three sites (Fig. 3A). Pairwise comparisons
between May and each of the other months, and
between Tapachula low-shade and each of the other
sites were statistically signi?cant (two-tailed t-test not
assuming equal variance across samples, po0.05), and
were the only statistically signi?cant differences between
months or between sites in this system. Fig. 3B contains
a suggestion of slightly higher beetle abundance in the
uncaged control plots relative to the ?ying vertebrate
exclusion plots, but this difference was not statistically
signi?cant (two-tailed t-test not assuming equal variance
across samples, p ? 0.19).
Discussion
Our data suggest that low-shade cultivation simpli?es
the understory foliage-dwelling beetle community of
coffee agroecosystems. This simplifying effect can be
seen in two distinct ways. First, low-shade sites
demonstrate lower a-diversity (reduced species richness)
relative to high-shade sites (Fig. 1). Fig. 1C reveals that
while each high-shade site is signi?cantly richer than its
ARTICLE IN PRESS
20
15
10
5
0
Huixtla high
shade
Huixtla low
shade
Tapachula low
shade
Caged UncagedTapachula high
shade
20
15
10
5
0
25
20
15
10
5
0
November February May August
N
um
be
r o
f b
ee
tle
s p
er
 sa
m
pl
e
Fig. 3. Beetle abundance patterns. Mean (7standard error) number of beetles per sample is shown for (A) the four study sites
(shading represents high-shade management). The abundance of beetles recorded at the Huixtla high-shade site may not be directly
comparable to that of the other three sites because of a slight variation in the amount of foliage sampled for arthropods at this site;
(B) plots in which ?ying vertebrates had been experimentally excluded (caged) vs. control plots (uncaged); and (C) the four different
months in which beetles were sampled.
C.E. Gordon et al. / Basic and Applied Ecology 10 (2009) 246?254 251
Author's personal copy
low-shade neighbor, the two geographically separate
areas differ in species richness such that the low-shade
site in the Huixtla area appears roughly equivalent in
species richness to the high-shade site in the Tapachula
area. The most likely explanation for this result is that
both cultivation type and geographic variation exert
signi?cant in?uences on species richness in this system.
Second, low-shade cultivation homogenizes the spe-
cies composition of beetle communities (Fig. 2, see
Appendix A: Table 1). In effect, low-shade coffee
cultivation reduces geographic b-diversity in this system.
It is important to note that this homogenizing effect is
most strongly seen in the Morisita?Horn index, which
takes species? abundances into account. This pattern
appears to be driven more by similar relative abundance
pro?les of particular species in these communities than
by shared sets of species that are exclusive to one
cultivation type or another. The species that best
illustrates this is a species of major agricultural
signi?cance, the coffee berry borer, Hypothenemus
hampei (Ferr.). We sampled 72 and 129 individuals of
this species in the two low-shade sites, and 16 and 7
individuals in the two high-shade sites, respectively,
suggesting that high-shade cultivation may confer some
degree of control of this pest species, consistent with the
??insurance hypothesis?? (Perfecto et al., 2004; Yachi &
Loreau, 1999).
The strong effects of cultivation type evident in our
data set are particularly notable given that differences in
cultivation technique mostly impact the shade, or
canopy stratum, while our samples were collected in
the coffee understory. We expect that our sample
contained mostly species that feed or forage directly
on coffee, or on the herbaceous vegetation that grows in
between coffee bushes. However, it is possible that an
in?ux of beetles from a highly diverse pool of canopy-
restricted species into the understory may account for
the higher levels of a- and b-diversity we observed in
high-shade coffee farms. This pattern could also have
been produced by a high degree of interconnectedness
and faunal exchange between the canopy and the
understory, or by a microclimatic effect of increased
shade on the beetle fauna of the understory, itself.
A ?nal possibility is a landscape cultivation interac-
tion effect, wherein high-shade cultivation permits an
in?ux of beetles from a species-rich pool in the
surrounding landscape.
Our analysis suggests that ?ying vertebrates exert a
negligible effect upon beetles in the coffee stratum in this
system. This is suggested by the lack of a signi?cant
effect of ?ying vertebrate exclusion on the abundance
(Fig. 3B), species richness (Fig. 1B), or species composi-
tion (see Appendix A: Table 1) of beetles. This result
contrasts strongly with the results of Greenberg et al.
(2000), who demonstrated signi?cant depression of large
(45mm in length) arthropod abundance on coffee
plants by ?ying vertebrates in low-shade coffee planta-
tions in the Polochic Valley in Guatemala. Beetles
smaller than 5mm in length were not analyzed by
Greenberg et al. (2000) but dominate our samples,
representing 78% of the species and 90% of the
individuals in our data set. However, large beetles
(X5mm in length) show the same lack of ?ying
vertebrate exclusion effects in our study (total abun-
dances of 102 and 110 large beetles in post-cage samples
from caged and control plots, respectively, two-tailed
t-test not assuming equal variance p ? 0.69). Of the
293 beetle morphospecies in our data set, only six
occurred in signi?cantly higher abundance in either
experimental treatment at the po0.05 level (two-tailed
t-tests not assuming equal variance), fewer than would
be expected at random if there were no ?ying vertebrate
exclusion effects. This rules out the possibility of any
signi?cant effect of ?ying vertebrates on any particular
taxon or guild of beetles. We also found no evidence of
signi?cant seasonal variation in the effects of ?ying
vertebrates on beetles. We therefore conclude that the
direct or indirect trophic links between ?ying vertebrates
and foliage-dwelling beetles in the coffee understory in
this system are weak.
Our results, combined with those of several other
studies in our system, suggest that the effects of ?ying
vertebrates on arthropods are much stronger in the
canopy stratum of coffee plantations than in the coffee-
dominated understory. At one of our study sites,
Philpott et al. (2004) demonstrated that ?ying verte-
brates depress the abundance of arthropods, including
beetles, on branches of Inga shade trees. Jedlicka et al.
(2006) combined ?ying vertebrate exclosures in Inga
shade trees with a subset of our exclosures in coffee at
one of our study sites to show that ?ying vertebrates
signi?cantly depress the abundance of both large
(45mm) and small (o5mm) arthropods in Inga foliage
of the canopy, but not in the coffee foliage of the
understory. However, Perfecto et al. (2004) demon-
strated increased removal rates of lepidopteran larvae,
presumably by ?ying vertebrates, from coffee foliage
outside of ?ying vertebrate exclosures in one of our
high-shade sites.
The variation in species richness and composition
across months reveals signi?cant but complex seasonal
variation in this system. The climate in this region is
characterized by distinct wet-dry seasonality (Cardoso,
1979). Arthropod communities in seasonal tropical
forests are known to undergo abrupt and dramatic
increases in abundance and activity in response to the
onset of the wet season (Janzen, 1983). Our May
samples were collected at the end of April and beginning
of May, roughly 1 week after the onset of signi?cant wet
season rains in 2001 (see Appendix A: Graph 2). These
samples show a distinct abundance spike relative to the
other months of sampling (Fig. 3C), yet the species
ARTICLE IN PRESS
C.E. Gordon et al. / Basic and Applied Ecology 10 (2009) 246?254252
Author's personal copy
richness and species composition comparisons do not
reveal a simple wet?dry season grouping of the monthly
samples. The species accumulation curve of the
February sample, taken during the heart of the dry
season, is very similar to that of the May sample taken
in the early wet season (Fig. 1D). Both of these months
appear less species rich than do August and November
in the middle, and toward the end of the wet season,
respectively (Fig. 1D). The pairwise comparisons of
species compositional similarity reveal roughly equiva-
lent levels of species turnover across all 3-month
intervals, hence, there is no clear separation of dry and
wet season beetle faunas re?ected in our data.
The seasonality of rainfall in the year of our study was
typical of the long-term average pattern in the region
(see Appendix A: Graph 1).
Acknowledgments
We thank the owners of Finca Irlanda, Belen and
Hamburgo for allowing us to conduct studies in their
farms. Gustavo Lopez Bautista, Alvaro Ballinas,
Juan Carlos Lo?pez de Leo?n, Aurelio Mendizabal and
Peter Bichier helped with insect collection and sorting.
Rachel Gratis and Kimberly Lawser helped with the
morphospecies analysis. The following beetle taxono-
mists examined and identi?ed beetle specimens from
selected families: Yves Bousquet (Monotomidae), Gene
Hall (Corylophidae, Ptiliidae), Michael C. Thomas
(Silvanidae), Robert Gordon (Coccinellidae). The In-
stituto de Ecolog??a, A. C. in Xalapa, Veracruz, Mexico,
and Lake Forest College, Lake Forest, Illinois provided
lab space and equipment for the beetle morphospecies
analysis. Financial support was provided by NSF grant
DEB 9981526 awarded to IP, RG, and GIN, NSF
International Research Fellowship Award INT-0076201
awarded to CEG, and by Lake Forest College.
Appendix A. Supporting Information
Supplementary data associated with this article can be
found in the online version at doi:10.1016/j.baae.
2008.04.004.
References
Arellano, L., Favila, M. E., & Huerta, C. (2005). Diversity of
dung and carrion beetles in a disturbed Mexican tropical
montane cloud forest and on shaded coffee plantations.
Biodiversity and Conservation, 14, 601?615.
Arnett, R. H., Jr., & Thomas, M. C. (2001). American beetles,
Vol. 1. Boca Raton: CRC Press.
Arnett, R. H., Jr., Thomas, M. C., Skelley, P. E., & Frank, J. H.
(2002). American beetles, Vol. 2. Boca Raton: CRC Press.
Borror, D. J., Triplehorn, C. A., & Johnson, N. F. (1989). An
introduction to the study of insects (6th ed.). Orlando:
Harcourt, Brace & Company.
Cardoso, M. D. (1979). El clima de chiapas y tabasco. Mexico
City: Universidad Nacional Autonoma de Me?xico.
Donald, P. F. (2004). Biodiversity impacts of some agricultural
commodity production systems. Conservation Biology, 18,
17?37.
Erwin, T. L. (1982). Tropical forests: Their richness in
Coleoptera and other arthropod species. The Coleopterists?
Bulletin, 36, 74?75.
Food and Agriculture Organization of the United Nations
(FAO). (2007). FAOstat agricultural database, /http://
faostat.fao.org/S accessed 20 January 2007.
Gordon, C., Manson, R., Sundberg, J., & Cruz-Ango?n, A.
(2007). Biodiversity, pro?tability, and vegetation structure
in a Mexican coffee agroecosystem. Agriculture, Ecosys-
tems, and Environment, 118, 256?266.
Greenberg, R., Bichier, P., Cruz-Ango?n, A., MacVean, C.,
Perez, R., & Cano, E. (2000). The impact of avian
insectivory on arthropods and leaf damage in some
Guatemalan coffee plantations. Ecology, 81, 1750?1755.
Heck, K. L. J., Van Belle, G., & Simberloff, D. (1975). Explicit
calculation of the rarefaction diversity measurement and the
determination of suf?cient sample size. Ecology, 56, 1459?1461.
Hietz, P. (2005). Conservation of vascular epiphyte diversity in
Mexican coffee plantations. Conservation Biology, 19,
391?399.
Hurlbert, S. H. (1971). The nonconcept of species diversity:
A critque and alternative parameters. Ecology, 52, 577?586.
Janzen, D. H. (1983). Insects. In D. H. Janzen (Ed.), Costa
rican natural history (pp. 619?779). Chicago: The University
of Chicago Press.
Jedlicka, J. A., Greenberg, R., Perfecto, I., Philpott, S. M., &
Dietsch, T. V. (2006). Seasonal shift in the foraging niche of
a tropical avian resident: Resource competition at work?
Journal of Tropical Ecology, 22, 1?11.
Klein, A. M., Steffan-Dewenter, I., Buchori, D., & Tscharntke,
T. (2002). Effects of land-use intensity in tropical agrofor-
estry systems on coffee ?ower-visiting and trap-nesting bees
and wasps. Conservation Biology, 16, 1003?1014.
Legendre, P., & Legendre, L. (1998). Numerical Ecology
(2nd ed.). Amsterdam: Elsevier.
Magurran, A. E. (2004). Measuring biological diversity (2nd
ed.). Oxford: Blackwell.
Mas, A. H., & Dietsch, T. V. (2003). An index of management
intensity for coffee agroecosystems to evaluate butter?y
species richness. Ecological Applications, 13, 1491?1501.
Mas, A. H., & Dietsch, T. V. (2004). Linking shade coffee
certi?cation programs to biodiversity conservation: Butter-
?ies and birds in Chiapas, Mexico. Ecological Applications,
14, 642?654.
Mathworks, Incorporated. (2006). MATLAB v2006b. Natick.
Mittermeier, R., Meyers, N., Thomsen, J. B., de Fonseca,
G. A. B., & Olivieri, S. (1998). Biodiversity hotspots
and major tropical wilderness areas: Approaches to
setting conservation priorities. Conservation Biology, 12,
516?520.
ARTICLE IN PRESS
C.E. Gordon et al. / Basic and Applied Ecology 10 (2009) 246?254 253
Author's personal copy
Moguel, P., & Toledo, V. M. (1999). Biodiversity conservation
in traditional coffee systems of Mexico. Conservation
Biology, 13, 11?21.
Perfecto, I., & Armbrecht, I. (2003). The coffee agroecosystem
in the Neotropics: Combining ecological and economic
goals. In J. Vandermeer (Ed.), Tropical agroecosystems
(pp. 159?194). Boca Raton: CRC Press.
Perfecto, I., Armbrecht, I., Philpott, S. M., Soto-Pinto, L., &
Dietsch, T. V. (in press). Shade coffee and the stability of
forest margins in Northern Latin America. In:
T. Tscharntke, M. Zeller, & C. Leuschner (Eds.), The
stability of tropical rainforest margins: Linking ecological,
economic and social constraints. Berlin: Springer.
Perfecto, I., Hansen, P., Vandermeer, J., & Cart??n, V. (1997).
Arthropod biodiversity loss and the transformation of a
tropical agro-ecosystem. Biodiversity and Conservation, 6,
935?945.
Perfecto, I., Vandermeer, J., Lo?pez Bautista, G., Ibarra Nun?ez,
G., Greenberg, R., Bichier, P., et al. (2004). Greater
predation in shaded coffee farms: The role of resident
neotropical birds. Ecology, 85, 2677?2681.
Philpott, S., Greenberg, R., Bichier, P., & Perfecto, I. (2004).
Impacts of major predators on tropical agroforest arthro-
pods: Comparisons within and across taxa. Oecologia, 140,
140?149.
Philpott, S. M., Perfecto, I., & Vandermeer, J. (2006). Effects
of management intensity and season on arboreal ant
diversity and abundance in coffee agroecosystems. Biodi-
versity and Conservation, 15, 139?155.
Philpott, S. M., Perfecto, I., & Vandermeer, J. (in press). Effects
of predatory ants on lower trophic levels across a gradient of
coffee management complexity. Journal of Animal Ecology.
Pineda, E., Moreno, C., Escobar, F., & Halffter, G. (2005).
Frog, bat, and dung beetle diversity in the cloud forest and
coffee agroecosystems of Veracruz, Mexico. Conservation
Biology, 19, 400?410.
Pinkus Rendo?n, M. A., Leo?n Corte?s, J. L., & Ibarra Nun?e?z, G.
(2006). Spider diversity in a tropical habitat gradient in
Chiapas, Mexico. Diversity and Distributions, 12, 61?69.
Richter, A., Klein, A. M., Tscharntke, T., & Tylianakis, J. M.
(2007). Abandonment of coffee agroforests increases insect
abundance and diversity. Agroforestry Systems, 69,
175?182.
Sanders, H. L. (1968). Marine benthic diversity: A compara-
tive study. American Naturalist, 102, 243?282.
Simberloff, D. S. (1972). Properties of the rarefaction diversity
measurement. American Naturalist, 106, 414?418.
Somarriba, E., Harvey, C. A., Samper, M., Anthony, F.,
Gonza?lez, J., Staver, C., et al. (2004). Biodiversity
conservation in neotropical coffee (Coffea arabica) planta-
tions. In G. Schroth, G. A. B. da Fonseca, C. A. Harvey, C.
Gascon, H. L. Vasconcelos, & A. M. N. Izac (Eds.),
Agroforestry and biodiversity conservation in tropical land-
scapes (pp. 198?226). Washington, DC: Island Press.
Yachi, S., & Loreau, M. (1999). Biodiversity and ecosystem
productivity in a ?uctuating environment: The insurance
hypothesis. Proceedings of the National Academy of
Sciences, USA, 96, 1463?1468.
ARTICLE IN PRESS
C.E. Gordon et al. / Basic and Applied Ecology 10 (2009) 246?254254