Oecologia (2004) 140: 140?149
DOI 10.1007/s00442-004-1561-z
COMMUNITY ECOLOGY
Stacy M. Philpott . Russell Greenberg . Peter Bichier .
Ivette Perfecto
Impacts of major predators on tropical agroforest arthropods:
comparisons within and across taxa
Received: 29 October 2003 / Accepted: 19 March 2004 / Published online: 17 April 2004
# Springer-Verlag 2004
Abstract In food web studies, taxonomically unrelated
predators are often grouped into trophic levels regardless
of their relative importance on prey assemblages, multiple
predator effects, or interactions such as omnivory. Ants
and birds are important predators likely to differentially
shape arthropod assemblages, but no studies have
compared their effects on a shared prey base. In two
separate studies, we excluded birds and ants from branches
of a canopy tree (Inga micheliana) in a coffee farm in
Mexico for 2 months in the dry and wet seasons of 2002.
We investigated changes in arthropod densities with and
without predation pressure from (1) birds and (2) ant
assemblages dominated by one of two ant species (Azteca
instabilis and Camponotus senex). We first analyzed
individual effects of each predator (birds, Azteca instabilis,
and C. senex) then used a per day effect metric to compare
differences in effects across (birds vs ants) and within
predator taxa (the two ant species). Individually, birds
reduced densities of total and large arthropods and some
arthropod orders (e.g., spiders, beetles, roaches) in both
seasons. Azteca instabilis did not significantly affect
arthropods (total, small, large or specific orders). Campo-
notus senex, however, tended to remove arthropods (total,
small), especially in the dry season, and affected arthropod
densities of some orders both positively and negatively.
Predators greatly differed in their effects on Inga
arthropods (for all, small, large, and individual orders of
arthropods) both in sign (?) and magnitudes of effects.
Birds had stronger negative effects on arthropods than ants
and the two dominant ant species had stronger effects on
arthropods in different seasons. Our results show that
aggregating taxonomically related and unrelated predators
into trophic levels without prior experimental data
quantifying the sign and strengths of effects may lead to
a misrepresentation of food web interactions.
Keywords Food webs . Birds . Dominant arboreal ants .
Multi-trophic interactions . Coffee agroecosystems
Introduction
Identifying when predator taxa can be grouped into trophic
levels is critical to community ecology. In attempts to
understand food web complexity, many unrelated predator
taxa are often grouped into trophic levels or trophospecies
even when information regarding the relative importance
of different predators is unknown (e.g., Hairston et al.
1960). Yet, species do not clearly separate into homoge-
neous trophic levels in part due to omnivory, intraguild
predation, and ontogenetic or environmentally influenced
diet shifts (Polis and Strong 1996). In nature, communities
are often composed of complex webs, not of trophic
chains (Pace et al. 1999; Polis et al. 2000). While
inappropriate effects of aggregation are known for aquatic
ecosystems (Hall and Raffaelli 1991; Martinez 1993;
Abrams et al. 1996; Tavares-Cromar and Williams 1996;
Sugihara et al. 1997; Yodzis and Winemiller 1999;
Thompson and Townsend 2000; Abarca?Arenas and
Ulanowicz 2002), less is known for terrestrial ecosystems
(but see Martinez 1993; Sugihara et al. 1997). Studies on
predators that share prey could potentially provide clues
about the appropriateness of aggregating species into units
such as trophic levels or trophospecies to describe food
webs. To date, most such studies from terrestrial systems
S. M. Philpott (*)
Department of Ecology and Evolutionary Biology, University
of Michigan,
830 N. University,
Ann Arbor, MI 48109, USA
e-mail: sphilpot@umich.edu
Tel.: +1-734-7641446
Fax: +1-734-7630544
R. Greenberg . P. Bichier
Smithsonian Migratory Bird Center, National Zoo,
3001 Connecticut Ave NW,
Washington, DC 20008, USA
I. Perfecto
School of Natural Resources and the Environment, University
of Michigan,
430 E. University,
Ann Arbor, MI 48109, USA
focus on predator species within a taxonomic group and
less so on unrelated species (Simberloff and Dayan 1991;
but see Davidson et al. 1980; Jaksic and Delibes 1987).
Furthermore, many studies pool impacts of taxonomically
similar species even though this can mask considerable
variation in resource use (Davidson 1997; Murakami and
Nakano 2000).
One focus of many current investigations is to examine
the effects on food webs of multiple predators alone and in
combination to determine a basis for grouping predators in
food web models. Different predator species, even those
from one taxa or guild, may greatly differ in their effects
on prey communities (Harris 1995; Schmitz and Suttle
2001; Chalcraft and Resetarits 2003a, b) making aggre-
gation of these species in food web models challenging,
especially when individual predator species with puta-
tively similar functions create nonlinear effects when they
are combined (Sih et al. 1998; Ekl?v and Werner 2000;
Crumrine and Crowley 2003). Evidence has, however,
shown that taxonomically related predators may be
entirely substitutable in their effects on food webs and
so aggregating them in a single functional unit is
warranted (Schmitz and Sokol-Hessner 2002; Sokol-
Hessner and Schmitz 2002). Whether or not this applies
to distantly related taxa of predators remains open to
question.
This study helps to resolve that question by examining
how two unrelated predator taxa, ants and birds, affect a
terrestrial arthropod assemblage. Ants and birds are
taxonomically important predators of arthropods in natural
and agricultural settings (Way and Khoo 1992; Marquis
and Whelan 1994; Perfecto and Casti?eiras 1998; Green-
berg et al. 2000; Schmitz et al. 2000; Sanz 2001; Mols and
Visser 2002; Van Bael et al. 2003), and likely differentially
affect prey based on differences in their mobility, size, and
life history. Ants are numerous in tropical forests
(>10,000 ha?1) (Schulz and Wagner 2002; Watt et al.
2002), patrol smaller areas than birds, cooperatively forage
(using recruitment or tandem running) using chemical
signals to alert other ants (H?lldobler and Wilson 1990),
and generally cannot easily attack large or highly mobile
prey (Koptur 1984). Relative to ants, birds are scarce (10?
50 individuals ha?1 of forest or agroforest, R. Greenberg,
personal observation). But, they are highly mobile, energy
demanding, generalized, opportunistic, and vary in type
and size of arthropod prey they select (Johnson 2000). We
were interested in discerning how differences in size,
abundance, and foraging strategies between these two
predator taxa translated into effects on the arthropod
assemblage.
Here, we investigate the importance of birds and two ant
species as predators in a coffee agroecosystem. Birds and
ants are abundant in traditional farms where coffee grows
under a diverse shade canopy (Perfecto and Snelling 1995;
Greenberg et al. 1997; Moguel and Toledo 1999).
Traditional farms resemble natural forests, but relatively
low diversity therein makes them useful for exploring
functional relationships difficult to explore in more
diverse, natural systems. We experimentally removed
birds or ants from individual branches of I. micheliana
trees from the canopy of a coffee agroforest and measured
changes in arthropod assemblages. To investigate the
effects of each predator and to compare effects of
unrelated taxa (birds and ants) and taxonomically related
taxa (two dominant ants), we addressed the following
questions: (1) Do birds and/or ants affect arthropod
assemblages in canopy trees? (2) Do birds and ants differ
in their effects on arthropods? (3) Do dominant ant species
differ in their effects on arthropods?
Materials and methods
We conducted all studies at Finca Irlanda (15?11?N, 92?20?W; 900 m
elevation; 4,500 mm rain/year), a shaded coffee farm in the
Soconusco region of SW Chiapas, Mexico, during the dry and wet
seasons of 2002. The shade canopy at Finca Irlanda is diverse (>60
species), but is largely dominated by Inga spp. (Perfecto and
Vandermeer 2002). Although abundance of birds and ants likely
differ in this site, species richness of birds and ants, especially those
foraging on Inga spp. trees, is generally comparable. Based on
6 years of intensive work, we have found 59 bird species and ~60
ant species that forage on Inga spp. shade trees at this site (Dietsch
2003; S. Philpott, unpublished data).
We focused our work on dominant arboreal ants, defined as
numerically abundant and competitively superior to other ants. Such
ants are predaceous, polydomous, and tend homopterans (Leston
1973). In tropical forests and agroforests, dominant ants are spatially
arranged in mosaic patterns whereby different species form mutually
exclusive patches (Leston 1973; Room 1975; Majer 1978; Majer
and Queiroz 1993). This spatial mosaic has allowed researchers to
determine if dominant ants are associated with different homopter-
ans within ant patches (Leston 1973; Majer 1978). Feeding habits of
dominant species may be more similar to each other than to other ant
guilds (Davidson 1997). Yet with substantial heterogeneity in
resource use among dominant ants, arthropod assemblages within
ant mosaic patches may differ as well.
Many dominant ants associate with plants (including Inga spp.)
with extrafloral nectaries often resulting in lower herbivory (Bentley
1977; Horvitz and Schemske 1984; Koptur 1984; Fiala et al. 1994;
Del-Claro et al. 1996) and increased growth and reproductive output
(Janzen 1966; Fonseca 1994; Letourneau 1998; de la Fuente and
Marquis 1999). But results of ant-plant studies are not conclusive
because plant-ants may remove all herbivores presented to them,
limit some herbivores, and not others, or may not limit herbivores at
all (Schemske 1980; Koptur 1984; Barton 1986; Kelly 1986; Koptur
and Lawton 1988; Whalen and MacKey 1988; Rashbrook et al.
1992; Fiala et al. 1994; Del-Claro et al. 1996; Oliveira 1997; de la
Fuente and Marquis 1999). Thus although ants may protect Inga
spp. from herbivores, effects on ant-plant arthropod assemblages are
largely unknown (see Risch and Carroll 1982; James et al. 1999;
Gibb 2003).
We established and maintained ant exclosures on I. micheliana
trees during the dry (2 February?30 April) and wet seasons (10
May?26 July) of 2002. We located 20 trees each dominated by
either A. instabilisor C. senex ants. On each tree, we selected two
branches (4?8 leaves, 3?4 m above ground) and randomly assigned
each to a control or exclosure treatment. To exclude ants (initially
and every 2 weeks thereafter), we placed Tanglefoot (The Tangle-
foot Company, 314 Straight Avenue, S.W., Grand Rapids, Mich.,
USA) around the base of branches (over wrapped flagging tape),
clipped arboreal connections, and manually picked off ants. We also
clipped arboreal connections and wrapped flagging around control
branches. On some trees, ants crossed Tanglefoot barriers, and we
eliminated all trees with >20 ants on exclosure branches at the time
of arthropod collection. Furthermore, farm workers pruned some
shade trees accidentally cutting some experimental branches. Final
141
sample sizes were thus 6 and 14 trees for A. instabilis in the dry and
wet seasons respectively and 16 trees for C. senex in each season.
We also established bird exclosures in I. micheliana trees in the
dry (12 December 2001?12 February) and wet seasons (2 May?15
July) of 2002. Within the range of locations of ant exclosures, we
selected two areas (>800 m apart) with similar vegetation
characteristics. Within each area, we selected ten trees (30?80 m
between each), on each tree picked two branches (4?8 leaves, 3?4 m
above ground), and randomly assigned each to a control or
exclosure treatment. We eliminated birds by placing monofilament
nylon fishing nets (35?35 mm mesh) over entire branches and tying
nets to form a bag. Choice of mesh size is a trade-off between
allowing movement of large arthropods (e.g., lepidopterans) and
preventing access to small birds. Our mesh size nonetheless is
comparable to other bird-exclosure studies [e.g., Greenberg et al.
2000 (29?29 mm); Mols and Visser 2002 (25?25 mm); Van Bael et
al. 2003 (20?20 mm)]. We did not see spiders using the mesh as a
web substrate. Data analyses include 18 trees in the dry season (2
trees were lost) and 20 trees in the wet season.
Table 1 Mean (?SE) number of arthropods dg?1 collected from
Inga tree branches with (C) or without (E) bird or ant predation
during dry and wet seasons of 2002. The two ants (Azteca instabilis
and Camponotus senex) were eliminated on separate trees and data
was combined for the ?Ants? category
Birds Ants A. instabilis C. senex
C E C E C E C E
Dry season
Acarida 4.2?1.2 5.7?1.9 12.4?3.5 17.4?6.3 12.0?5.0 15.4?8.0 12.5?4.5 18.2?8.3
Aranae 8.2?1.3 24.6?4.4 12.4?3.6 14.6?3.3 27.9?10.6 11.1?2.4 6.6?1.5 15.9?4.5
Blattodea 0.4?0.2 13.2?2.8 3.7?0.9 1.3?0.4 7.8?2.6 2.8?1.0 2.2?0.5 0.7?0.3
Coleoptera 6.2?2.6 11.8?3.7 7.9?1.4 15.6?2.6 9.8?2.6 14.3?2.7 7.1?1.7 16.1?3.5
Collembola 0?0 0.8?0.3 0.5?0.3 0.3?0.2 0.3?0.3 0.3?0.3 0.6?0.3 0.4?0.2
Diptera 1.6?0.4 1.1?0.3 1.6?0.4 3.2?1.1 1.9?0.7 2.1?1.0 1.5?0.5 3.6?1.5
Hemiptera 7.1?2.3 4.3?0.8 3.4?0.8 3.6?0.9 5.3?2.1 4.1?2.0 2.8?0.7 3.4?1.0
Homoptera (non-scales) 11.8?3.4 8.5?1.5 3.9?1.0 4.7?1.3 5.7?2.5 4.3?1.7 3.2?1.1 4.8?1.8
Scales 0?0 0?0 36.1?26.5 0.2?0.1 131.6?91.1 0.4?0.3 0.2?0.2 0.1?0.1
Hymenoptera (non-ants) 6.0?3.3 6.2?2.8 4.1?1.0 4.2?0.8 2.6?1.0 4.0?1.5 4.6?1.4 4.3?0.9
Formicidae 12.3?5.5 46.0?29.6 34.0?12.6 2.6?0.9 86.2?37.3 2.0?0.8 14.5?6.3 2.8?1.2
Lepidoptera 1.8?0.6 3.3?1.1 1.6?0.5 1.3?0.4 3.6?1.7 1.1?0.8 0.9?0.3 1.4?0.5
Neuroptera 0.2?0.1 0.7?0.2 0.2?0.1 0.3?0.1 0?0 0.3?0.2 0.2?0.1 0.3?0.2
Orthoptera 0.4?0.2 0.9?0.2 0.3?0.2 0?0 0.2?0.2 0?0 0.4?0.1 0?0
Psocoptera 0?0 0?0 3.5?1.0 4.4?1.2 4.9?2.4 5.0?2.2 3.0?1.0 4.2?1.4
Thysanoptera 5.6?0.1 3.8?1.1 5.4?0.9 3.4?0.8 2.3?1.6 4.0?2.2 6.5?1.0 3.2?0.8
Total arthropods 54.1?9.6 86.3?40.6 60.9?7.7 74.5?9.2 84.2?16.2 69.0?12.8 55.2?8 76.5?11.9
Arthropods <3 mm 34.2?8.6 40.8?7.9 47.4?7.3 61.5?.2 70.9?15.1 55.4?13 38.6?7.4 63.9?11.9
Arthropods >5 mm 4.1?0.6 19.2?2.5 8.0?1.0 6.9?0.9 10.0?2.1 7.4?1.2 7.3?1.1 6.7?1.2
Wet season
Acarida 0.6?0.2 4.1?1.4 13.0?5.2 8.2?3.6 8.1?4.1 13.7?7.5 17.4?9.1 3.3?1.2
Aranae 11.4?4.4 20.5?4.0 5.1?1.0 13.6?3.5 4.3?1.7 16.5?5.9 5.8?1.2 11.0?4.1
Blattodea 3.0?1.0 34.1?11.7 4.6?1.3 3.3?1.0 5.5?2.5 4.4?2.1 3.8?1.0 2.3?0.5
Coleoptera 14.4?3.3 39.6?10.0 10.8?2.2 20.4?4.2 11.0?4.3 28.8?7.3 10.6?2.0 13.1?4.0
Collembola 0.6?0.4 0.6?0.3 0.4?0.3 0.2?0.1 0.8?0.6 1.1?0.1 0.1?0.1 0.3?0.2
Diptera 2.0?0.5 1.6?0.5 2.7?0.6 3.0?0.7 2.0?0.7 3.3?1.3 3.3?0.9 2.7?0.8
Hemiptera 2.3?0.5 3.3?0.7 1.9?0.3 2.2?0.4 2.4?0.8 2.8?0.7 1.5?0.5 1.6?0.5
Homoptera (non-scales) 4.3?0.7 7.1?5.0 5.4?1.4 5.5?0.8 7.9?2.7 6.1?1.1 3.2?0.6 5.0?1.2
Scales 0?0 0?0 94.4?64.6 0.5?0.2 202.2?135.1 0.7?0.3 0?0 0.4?0.2
Hymenoptera (non-ants) 2.3?0.5 3.3?0.8 4.7?2.0 3.3?0.7 7.4?4.3 4.3?1.2 2.3?0.6 2.3?0.6
Formicidae 7.9?2.3 17.8?4.8 55.1?22.2 4.0?1.2 88.9?45.3 3.4?1.2 24.6?9.9 4.5?2.0
Lepidoptera 6.5?1.4 6.6?1.7 12.0?2.2 17.7?4.9 10.4?2.7 21.1?10.2 15.1?3.4 14.7?2.9
Neuroptera 0.6?0.3 0.3?0.2 0.2?0.1 0.3?0.1 0.1?0.1 0.1?0.1 0.2?0.1 0.4?0.3
Orthoptera 0.4?0.1 1.1?0.4 0.4?0.2 0.1?0.1 0.4?0.3 3.3?0.2 0.4?0.3 0?0
Psocoptera 0?0 0?0 0.8?0.2 2.0?0.5 1.1?0.5 2.5?0.8 0.5?0.2 1.6?0.5
Thysanoptera 4.1?0.9 5.4?1.4 4.0?0.9 5.6?1.0 2.0?1.1 5.1?1.4 5.7?1.2 6.1?1.5
Total arthropods 52.6?8.0 128.2?19.8 69.2 8.1 85.7?11.2 68.3?12.2 109.6?18.0 70.0?11.2 64.7?12.3
Arthropods <3 mm 29.0?6.1 59.7?15.3 40.7?6.6 53.1?9.2 41.8?9.2 72.2?15.5 39.8?9.8 36.5?9.4
Arthropods >5 mm 9.6?1.1 28.0?5.3 18.1?2.6 23.5?5.2 15.3?2.7 28.8?10.6 20.5?4.2 18.8?3.0
142
Arthropod sampling for both ant and bird exclosures followed the
same protocol. To collect arthropods from Inga trees, entire
exclosure or control branches were covered with 60?90 mm plastic
bags and cut. We killed arthropods inside bags with ethyl acetate and
collected all arthropods found on leaves, branches, plastic bags, and
bird-exclosure nets. We identified all arthropods to order (and some
to family) and measured the length (mm) of each individual. We
dried and weighed all foliage collected with samples and
standardized all arthropod data as the number of individuals per
gram dry foliage (only leaves).
We first examined individual effects of each predator (birds, A.
instabilis, C. senex) on densities of total arthropods, different sized
arthropods [small (<3 mm) and large (>5 mm)], and individual
orders. We analyzed effects on total, small, and large arthropods
using separate ANOVAs with treatment (each predator alone versus
exclosure) and season as fixed factors. We examined effects on
individual arthropod orders using MANOVA with each order (>25
individuals) as dependent variables and treatment (as above) and
season as fixed factors. We followed significant MANOVAs with
ANOVAs to test for effects on particular orders. We did not include
scales or ants in arthropod totals or as dependent variables in
MANOVA (ants were included in bird MANOVA) because (1) ants
were intentionally excluded on some trees and (2) ants tend scales
(Way 1963) and we expected elevated scale densities on control
branches. We instead used separate ANOVA for scales and ants as
above. For all tests, we used square-root-transformed data to
conform to assumptions of normality.
Because our study included two separate experiments (one for
birds and one for ants) we used a metric to compare effects across
and within predator taxa (Osenberg et al. 1997; Laska and Wootton
1998; Chalcraft and Resetarits 2003b). We chose the per day effect
of predators on arthropod density such that:

r ? ln??DC ? 1?=?DE ? 1?
t (1)
where DC is the total arthropod density on control branches at the
end of the trial period, DE is the arthropod density on exclosure
branches at the end of the trial period, and t is the number of days
exclosures were maintained (Osenberg et al. 1997). The metric was
calculated on a per tree basis highlighting our paired experimental
design. We compared per day effects (?r) of birds, A. instabilis, and
C. senex on total, small, and large arthropods using ANOVAs and
compared effects on the ten most common orders using MANOVA
followed by individual ANOVAs. For each test, treatment (with
levels of birds, A. instabilis, and C. senex) was crossed with season.
Where there was a significant treatment by season interaction, we
followed with separate ANOVAs and MANOVAs for each season.
We used planned contrasts within ANOVA and MANOVA to
statistically compare effects of birds and ants (A. instabilis + C.
senex), and then to compare effects of the two ant species. We
further differentiated predator effects qualitatively by comparing
sign (positive or negative) of effects on arthropods. Because our data
included many zeros, we added one to all densities before
calculating per tree metric values. We carried out all statistical
analyses using Statistica v 6.1 for Windows.
Results
Effects of birds on arthropods
Total arthropod densities were 51% lower on control than
on exclosure branches across seasons. These reductions
were significant in the dry (by 37%; F1,34=6.46, P=0.016)
and wet seasons (59%; F1,38=17.45, P<0.001) (Table 1).
Birds reduced small (38%) and large arthropods (71%)
(Table 2), but effects on large arthropods varied with
season. Even so, birds significantly reduced large
arthropods in both the dry (79%; F1,34=50.45, P<0.001)
and wet seasons (66%; F1,38=14.18, P=0.001) (Table 1).
Birds also reduced densities of several arthropod orders
(Table 3). Although bird effects on particular orders
differed with season, there was no significant treatment by
season interaction. Birds significantly reduced roaches (by
93%), beetles (62%), orthopterans (62%), spiders (56%),
and mites (53%), and tended to reduce ants (68%) and
collembolans (57%).
Effects of ants on arthropods
On ant exclosure trees, ant densities were greatly reduced
(by 93%) on exclosure branches (F1,98=24.51,
P<0.000001) demonstrating effectiveness of Tanglefoot
treatments (Table 1). On A. instabilis trees, 97% of ants
Table 2 Effects of predators
(birds and two ant species (A.
instabilis, C. senex)) on Inga
arthropods across the wet and
dry seasons of 2002. Statistical
results show individual effects
of birds and ants on arthropod
densities compared with no-
predator treatments, and com-
parison between mean per day
effects of predators on arthro-
pods. See text for explanation of
statistics used. Bold numbers are
for significant effects, italics
show trends
Arthropods (total) Arthropods (<3 mm) Arthropods (>5 mm)
df F P df F P df F P
Individuals foliage g?1
Birds 1, 72 22.77 <0.001 1, 72 5.09 0.027 1, 72 47.61 <0.001
Season 1, 72 2.08 0.154 1, 72 0.42 0.521 1, 72 8.27 0.005
Birds ? season 1, 72 2.23 0.140 1, 72 1.66 0.202 1, 72 0.59 0.446
Ants 2, 98 1.46 0.238 2, 98 1.97 0.146 2, 98 0.06 0.941
A. instabilis 1, 98 0.03 0.869 1, 98 0.00 0.973 1, 98 0.11 0.737
C. senex 1, 98 2.78 0.099 1, 98 3.51 0.064 1, 98 0.03 0.852
Season 1, 98 0.04 0.852 1, 98 2.96 0.088 1, 98 16.54 <0.001
Ants ? season 2, 98 0.94 0.393 2, 98 0.97 0.383 2, 98 1.40 0.251
Mean per day effect (?r)
Predator 2, 84 8.036 <0.001 2, 84 0.803 0.451 2, 84 10.902 <0.001
Birds vs. ants 1, 84 14.753 <0.001 1, 84 1.579 0.212 1, 84 18.440 <0.001
A. instabilis vs. C. senex 1, 84 0.096 0.758 1, 84 0.011 0.916 1, 84 0.780 0.380
Season 1, 84 1.659 0.201 1, 84 1.432 0.235 1, 84 0.671 0.415
Predator ? Season 2, 84 3.672 0.030 2, 84 4.054 0.021 2, 84 1.002 0.372
143
were removed from no-ant branches (F1,98=48.39,
P<0.000001) (Table 1). On C. senextrees, fewer (81%),
but a still significant number of ants were eliminated
(F1,98=8.47, P=0.004) (Table 1).
Individually, A. instabilis did not reduce total, small, or
large arthropod densities (Table 2) or densities of arthro-
pod orders with the exception of scales (Table 3). Across
both seasons, scale densities increased (by 99%) on
branches with A. instabilis (Table 3). Effects of A.
instabilis on other orders differed with season, but there
was not a significant treatment by season interaction
(Table 3).
Camponotus senex tended to reduce total and small
arthropods, and these trends were slightly stronger in the
dry season (Tables 1, 2). Across both seasons, C. senex
reduced total (14%) and small (22%) arthropods and in the
dry season limited total (34%; F1,41=3.40, P=0.072) and
small (40%; F1,57=3.99, P=0.052) arthropods (Tables 1,
2). C. senex significantly affected specific arthropod orders
(Table 3). Densities of spiders (54%) and beetles (39%)
Table 3 Individual effects of
each predator on density of
particular arthropod orders
compared to no-predator con-
trols across dry and wet seasons
of 2002. Bold shows significant
effects and italics show trends
Birds A. instabilis C. senex
df F P df F P df F P
MANOVA
Treatment 14, 59 5.84 <0.001 11, 26 0.99 0.480 11, 50 2.500 0.014
Season 14, 59 7.03 <0.001 11, 26 2.69 0.019 11, 50 13.210 <0.001
Treatment ? season 14, 59 1.46 0.155 11, 26 0.99 0.477 11, 50 1.040 0.430
Individual ANOVA
Treatment
Acarida 1, 72 4.95 0.029 NA NA NA 1, 60 1.368 0.247
Aranae 1, 72 21.29 <0.001 NA NA NA 1, 60 5.844 0.019
Blattodea 1, 72 56.22 <0.001 NA NA NA 1, 60 6.023 0.017
Coleoptera 1, 72 14.47 <0.001 NA NA NA 1, 60 4.182 0.045
Collembola 1, 72 3.58 0.062 NA NA NA NA NA NA
Diptera 1, 72 2.11 0.150 NA NA NA 1, 60 0.587 0.446
Hemiptera 1, 72 0.01 0.920 NA NA NA 1, 60 0.573 0.452
Homoptera (non-scales) 1, 72 0.61 0.437 NA NA NA 1, 60 0.739 0.393
Hymenoptera (non-ants) 1, 72 0.76 0.388 NA NA NA 1, 60 0.062 0.805
Formicidae 1, 72 3.44 0.068 NA NA NA NA NA NA
Lepidoptera 1, 72 0.28 0.601 NA NA NA 1, 60 0.089 0.766
Neuroptera 1, 72 1.07 0.305 NA NA NA NA NA NA
Orthoptera 1, 72 5.61 0.021 NA NA NA NA NA NA
Psocoptera NA NA NA NA NA NA 1, 60 2.763 0.102
Thysanoptera 1, 72 0.52 0.475 NA NA NA 1, 60 2.918 0.093
Season
Acarida 1, 72 6.43 0.013 1, 36 1.16 0.288 1, 60 1.41 0.240
Aranae 1, 72 0.42 0.522 1, 36 5.46 0.025 1, 60 1.47 0.230
Blattodea 1, 72 9.63 0.003 1, 36 0.44 0.509 1, 60 7.76 0.007
Coleoptera 1, 72 23.98 <0.001 1, 36 0.77 0.385 1, 60 0.01 0.941
Collembola 1, 72 0.47 0.496 NA NA NA NA NA NA
Diptera 1, 72 0.86 0.358 1, 36 0.02 0.892 1, 60 1.36 0.248
Hemiptera 1, 72 6.01 0.017 1, 36 2.31 0.138 1, 60 7.48 0.008
Homoptera (non-scales) 1, 72 9.77 0.003 1, 36 0.73 0.399 1, 60 0.16 0.688
Hymenoptera (non-ants) 1, 72 1.60 0.211 1, 36 0.00 0.964 1, 60 4.73 0.034
Formicidae 1, 72 0.14 0.707 NA NA NA NA NA NA
Lepidoptera 1, 72 13.29 0.001 1, 36 7.10 0.011 1, 60 82.06 <0.001
Neuroptera 1, 72 0.15 0.700 NA NA NA NA NA NA
Orthoptera 1, 72 0.02 0.899 NA NA NA NA NA NA
Psocoptera NA NA NA 1, 36 9.58 0.004 1, 60 12.51 0.001
Thysanoptera 1, 72 0.04 0.848 1, 36 0.01 0.924 1, 60 0.24 0.626
ANOVA (scales)
Treatment 1, 72 2.31 0.133 1, 36 5.52 0.024 1, 60 1.429 0.237
Season 1, 72 4.985 0.029 1, 36 0.12 0.736 1, 60 0.168 0.683
Treatment ? season 1, 72 2.31 0.133 1, 36 0.13 0.716 1, 60 3.930 0.052
144
decreased whereas densities of roaches significantly
increased with C. senex presence (50%) and thysanopteran
density tended to increase (24%). Overall, C. senex did not
influence scale densities however there was a significant
treatment by season interaction (Table 3). In the wet
season, scales significantly increased on branches without
C. senex (F1,30=4.74, P=0.037).
Comparing predator effects on arthropods
In general, predators differed both across and within taxa
in the sign and magnitude of effects on arthropods (Fig. 1,
Tables 2, 4). Birds had significantly greater negative
effects than ants on total and large arthropods within and
across seasons (Fig. 1, Tables 1, 2). The signs of effect of
birds and ants differed for large arthropods. Birds and ants
also differed in their effects on individual arthropod orders
(Table 4). Negative effects of birds on spiders and beetles
Table 4 Differences in mean effect sizes (?r) of predators on Inga
arthropods measured with exclosure experiments during the dry and
wet seasons of 2002. Results show comparisons of effects of A.
instabilis, C. senex, and birds on individual arthropod orders using
MANOVA and planned contrasts within MANOVA comparing
across (birds vs ants) and within (A. instabilis vs C. senex) predator
taxa. Bold type highlights significant effects and italics show trends.
Sign shows if effects of each predator were either both positive or
negative (s) or if signs differed by predator identity (d)
Both seasons Dry season Wet season
df F P Sign df F P Sign df F P Sign
MANOVA
Predator 20, 150 2.63 <0.001 20, 56 3.081 <0.001 20, 76 1.80 0.036
Birds vs ants 10, 75 4.38 <0.001 10, 28 6.083 <0.001 10, 38 2.65 0.015
A. instabilis vs C. senex 10, 75 1.08 0.385 10, 28 1.876 0.092 10, 38 1.05 0.422
Season 10, 75 1.02 0.434
Predator ? season 20, 150 1.62 0.054
ANOVA
Birds vs ants
Acarida 1, 84 0.70 0.406 s 1, 37 0.057 0.813 s 1, 47 1.67 0.203 s
Aranae 1, 84 8.90 0.004 s 1, 37 14.416 0.001 d 1, 47 0.09 0.762 s
Blattodea 1, 84 38.44 <0.001 d 1, 37 35.134 <0.001 d 1, 47 20.30 <0.001 d
Coleoptera 1, 84 3.75 0.056 s 1, 37 0.072 0.790 s 1, 47 5.31 0.026 s
Diptera 1, 84 1.73 0.191 s 1, 37 1.479 0.232 s 1, 47 0.41 0.527 s
Hemiptera 1, 84 0.70 0.404 s 1, 37 1.258 0.269 s 1, 47 0.60 0.441 s
Homoptera (non-scales) 1, 84 0.13 0.722 s 1, 37 0.574 0.454 s 1, 47 0.13 0.723 s
Hymenoptera (non-ants) 1, 84 0.21 0.648 s 1, 37 0.001 0.977 s 1, 47 1.07 0.306 s
Lepidoptera 1, 84 0.01 0.915 s 1, 37 6.191 0.017 s 1, 47 0.45 0.505 s
Thysanoptera 1, 84 0.47 0.494 s 1, 37 0.524 0.474 s 1, 47 0.05 0.822 s
A. instabilis vs C. senex
Acarida 1, 84 2.02 0.159 s 1, 37 0.017 0.897 s 1, 47 4.59 0.037 s
Aranae 1, 84 1.66 0.202 d 1, 37 6.330 0.016 d 1, 47 1.06 0.308 s
Blattodea 1, 84 0.06 0.799 s 1, 37 0.445 0.509 s 1, 47 0.00 0.950 s
Coleoptera 1, 84 1.33 0.252 s 1, 37 0.577 0.452 d 1, 47 4.59 0.037 d
Diptera 1, 84 0.02 0.889 s 1, 37 0.650 0.425 s 1, 47 1.48 0.229 s
Hemiptera 1, 84 0.13 0.719 s 1, 37 0.171 0.681 s 1, 47 0.05 0.826 s
Homoptera (non-scales) 1, 84 0.53 0.468 s 1, 37 0.154 0.697 s 1, 47 0.47 0.495 s
Hymenoptera (non-ants) 1, 84 0.02 0.900 s 1, 37 0.053 0.819 s 1, 47 0.67 0.418 s
Lepidoptera 1, 84 0.34 0.560 d 1, 37 2.766 0.105 s 1, 47 1.31 0.259 d
Thysanoptera 1, 84 3.47 0.066 s 1, 37 2.124 0.153 s 1, 47 1.39 0.244 s
ANOVA
Scales
Predator 2, 84 8.79 <0.001 2, 37 11.66 <0.001 2, 47 3.61 0.035
Birds vs ants 1, 84 6.61 0.012 s 1, 37 10.19 0.003 s 1, 47 2.13 0.152 s
A. instabilis vs. C. senex 1, 84 14.31 <0.001 s 1, 37 19.91 <0.001 s 1, 47 5.36 0.025 s
Season 1, 84 0.10 0.748
Predator ? Season 2, 84 0.06 0.941
145
were twice as strong as those of ants, and negative bird
effects on roaches significantly differed from ants both in
magnitude (?r) and in sign of effect. Ants had
significantly greater positive effects on scales than did
birds across both seasons.
The two ant species (A. instabilis and C. senex) also
differed somewhat in their overall effects on arthropods.
Ant species did not differ in the magnitude of their effects
on total, small, or large arthropods for both seasons
(Table 2) but effect sign of ant species on large arthropods
differed. Yet, for total and small arthropods, there was a
significant predator by season interaction. During the wet
season, A. instabilis tended to have a stronger negative
effect on total arthropods than C. senex. C. senex tended to
have stronger effects on small arthropods in the dry
season, and A. instabilis tended to have greater negative
effects on small arthropods in the wet season. Further-
more, in the dry season, ant species tended to differ in
effects on individual arthropod orders, where C. senex
tended to have stronger negative effects on spiders
(Table 4). Effects of ant species differed in sign for
spiders and lepidopterans across both seasons, for spiders
and beetles in the dry season, and for beetles and
lepidopterans in the wet season. A. instabilis had
significantly greater positive effects on scales than C.
senex in both seasons.
Discussion
Birds and ants differed in their effects on total and large
arthropods indicating that these different taxa could not be
treated as an aggregate entity. Birds reduced total, small,
and large arthropod densities and reduced densities of
several arthropod orders in both dry and wet seasons. A.
instabilis ants alone did not affect arthropod densities for
any size or specific orders. In the dry season, C. senex ants
tended to reduce small arthropods, significantly reduced
spider and beetle densities, and increased roach and
thysanopteran densities. Predators (both across and within
taxa) significantly differed in both signs and magnitudes of
their effects on arthropods (total and of several orders).
The qualitative effects of ant species on small arthropods
appeared to be generally substitutable between seasons
(Fig. 1). However, this substitutability did not apply to
large arthropods.
Ants are regarded as useful biological control agents in
agricultural systems (Way and Khoo 1992; Perfecto and
Casti?eiras 1998), yet most studies examine ant effects on
particular pest species. Examinations of ant effects on
arthropod assemblages give highly variable insight into
their effects, especially as regards to limitation of large
arthropods. Risch and Carroll (1982) excluded Solenopsis
geminata from maize plants and found overall increases in
herbivore and predator abundance, but many arthropod
groups were not affected by ant removal and some ant-
tended homopterans were more abundant with ants.
Similarly, James et al. (1999) removed two species of
the Iridomyrmex rufoniger group from citrus trees and
found overall increases in beneficial and incidental
arthropods after 2 years. However, for most intermediate
sampling periods, ants did not affect total arthropods, but
did influence particular orders. Gibb (2003) removed
Iridomyrmex purpureus and found no effects on ground-
foraging arthropod assemblages. Furthermore, Koptur
(1984) found that different ants varied in their abilities
to remove caterpillars from Ingatrees. The particular ants
chosen for our study have different foraging strategies
(H?lldobler and Wilson 1990; S. Philpott, unpublished
data) and may have different diets including much
honeydew and nectar (Davidson et al. 2003). Our results
thus reinforce ideas that ant species differentially influence
particular arthropods within arthropod assemblages and
thus tend not to have substitutable effects.
Although one goal of this study was to determine how
species of one predator taxa (ants) differ, we did not
address how bird species may differ. Bird species may
strongly differ in their effects on herbivores and leaf
Fig. 1 Mean per day effects (?1 SE) of ant (A. instabilis and C.
senex) and bird predators on Inga arthropods. Graphs show
comparisons for all (total), small (<3 mm), and large arthropods
(>5 mm) for both the dry and wet seasons of 2002. Positive numbers
show net positive effects on arthropods, negative numbers show net
negative effects. Mean per day effects on arthropods were calculated
by averaging a metric {ln [(DC + 1)/(DE + 1)]/t} calculated on a per
tree basis where DC is the total arthropod density [number/dry
foliage (g)] on control branches at the end of the trial period, DEis
the arthropod density on exclosure (no-predator) branches at the end
of the trial period, and t is the number of days exclosures were
maintained
146
damage (Murakami and Nakano 2000). Thus the >100
bird species at our study site (R. Greenberg, unpublished
data) may differ in effects on arthropods, and indeed many
birds at the study site have very different foraging
strategies and likely differentially affect prey (Dietsch
2003). But because birds tend not to defend strict foraging
territories (R. Greenberg, personal observation), and form
mixed-species flocks with constantly changing composi-
tion over time (R. Greenberg, personal observation),
differential effects of bird species will be negated because
birds mix (rather than separate) in space. It is thus
impractical to look at individual bird species in this
setting.
The effects of birds and ants on arthropod assemblages
also differed from one another over time. Effects of birds
on arthropod assemblages were consistent over seasons,
whereas the sign and strength of ant species effects
differed between the dry and wet seasons. Bird effects on
total and small (but not large) arthropods were slightly
greater in the wet season?an unexpected result consider-
ing that bird abundance in the study site is higher in the
dry season (when migrants are present) (R. Greenberg,
unpublished data). Larger bird effects in the wet season
may be explained by added resource requirements (or
possible diet shifts to insectivory) for breeding tropical
resident birds (Polis 1991; Levey and Stiles 1992). In
contrast, overall ant abundance is consistent but abun-
dances of A. instabilis and C. senex may fluctuate
seasonally. In 2002, A. instabilis was more abundant in
the wet season, and C. senex was more abundant in the dry
season (S. Philpott, unpublished data) possibly explaining
changes in their relative effects. Furthermore resource use
by ants may change with ontogeny or colony reproduction
(H?lldobler and Wilson 1990). Thus seasonal differences
in ant effects may be expected.
The effects of birds and ants on arthropods differed both
by arthropod order and size. Birds affected large
arthropods and ants tended to reduce small arthropods;
expected given the size of each predator. No predator
affected densities of dipterans, hemipterans, non-scale
homopterans, or non-ant hymenopterans?all highly
mobile prey. Mobility may make capture more difficult,
or may reduce effects of exclosures if prey constantly re-
colonized branches. No predator reduced lepidopterans
(94% captured were caterpillars) but C. senex and birds
reduced spider and beetle densities. Spiders are important
predators in coffee systems (Ibarra-N??ez et al. 2001), and
many predaceous beetles (e.g., coccinellids) were cap-
tured. Omnivory by birds and ants thus may have
indirectly limited effects on lepidopterans or total
arthropods (in the case of ants), especially if other
predators compensated for the removal of birds or ants.
Birds and ants may compete for prey or interact via
intraguild predation potentially masking ant effects on
arthropod assemblages. Many studies have demonstrated
risk reduction for prey in the presence of multiple
predators because either top predators have alternative
prey (the other predator) or because of behavioral
inhibition of one or the other predator (e.g., Sih et al.
1998). Birds sometimes eat ants (Poulon and Lefebvre
1996; Strong and Sherry 2000) and in this study, birds
tended to reduce ant densities. Ants were unavailable to
birds on no-ant branches, and if ants, in coffee systems,
constitute a large part of bird diets, bird predation on other
arthropods may have significantly increased where ants
were excluded masking isolated effects of ants on
arthropods. Alternatively, aggressive ants may deter bird
feeding (Aho et al. 1999; Haemig 1992, 1996). In this
case, if ants significantly limit bird feeding when both
predators are present, bird predation also would have
increased and may have compensated for effects of ant
removal, masking large effects of one or both ant species.
At least in one case, however, combined effects of bird and
ant predators on prey were additive (Floyd 1996).
There are additional factors that may have confounded
our experimental results. First, bird and ant exclosures
were established during different times in the dry season,
perhaps resulting in some differences not attributable to
predator effects. Second, Tanglefoot may have limited
colonization by non-flying arthropods (i.e., lepidopteran
larvae, spiders, mites, or scales). Total arthropod densities
on no-bird branches were significantly higher than no-ant
branches (F1,48=4.69, P=0.035) (Table 1). In fact,
lepidopteran larvae were more abundant on no-ant
(F1,48=5.55, P=0.022) branches and mite density did not
differ (F1,48=0.09, P=0.759), but spiders tended to be more
abundant on no-bird branches (F1,48=3.67, P=0.061).
Although scales were not included in total arthropod
calculations, scale densities were higher on C. senex-
excluded branches in the wet season (even though this ant
tends scales) perhaps because Tanglefoot prevented em-
igration from treated branches. Alternatively, increased
scale density on no-ant branches may be because ants
periodically harvest homopterans especially when extra-
floral nectar resources are available (as in Inga spp.)
(Ricogray 1993; Offenberg 2001). Thus, in general,
differences in arthropod densities do not seem to be due
to experimental treatment. Third, although total richness of
birds and ants likely to forage on Inga spp. trees in our
study site is comparable (~60 species), we may have
removed a higher diversity of birds than ants from
individual trees. Ant exclosure treatments were established
on trees where A. instabilis or C. senex were visibly
abundant on quick inspection but treatments effectively
prevented all ant species from visiting branches. At the
time control branches were harvested, we found 2?3 ant
species per branch, on average, of which A. instabilis or C.
senex represented the majority of individuals. For birds, 2?
4 species may visit individual Inga trees over a 2 h period
(R. Greenberg et al, unpublished data). Based on differ-
ences in mobility between these two taxa, however,
temporal turnover in bird species visiting individual trees
may be greater than for ants. Thus, although richness of
the two taxa on Inga spp. trees is very similar, differences
in predator richness per tree may have confounded effects
of predator identity.
In conclusion, our study showed that each predator
varied in effects on arthropods. The effects of birds versus
147
ants on arthropod assemblages differed both qualitatively
and quantitatively. This also held for comparisons between
ant species. Thus, aggregating bird and ant species into
one trophospecies would not represent the effects of each
species accurately even though they overlap considerably
in arthropod prey species used. Our results thus show the
necessity of examining differential effects of predators
experimentally on the basis of effect magnitudes, rather
than on the basis of diets (e.g., Yodzis and Winemiller
1999) before aggregating either related or unrelated taxa
into trophic levels in food web models. Our study also
reiterates the general importance of ants and birds as
predators in coffee agroecosystems, even though ant
effects differed seasonally.
Acknowledgements For field help we thank J. A. Garc?a Ballinas,
G. L?pez Bautista, J. Maldonado, J. Cabrera Santos, F. Camposeco
Silvestre, B. E. Chilel, A. Gonz?lez, F. Hern?ndez G?mez, and L.
Morales. R. Burnham, E. DeMattia, T. Dietsch, P. Foster, J. Jedlicka,
M. Reiskind, O. J. Schmitz, S. Van Bael, J. Vandermeer, J. Wyatt,
and three anonymous reviewers greatly improved the manuscript. G.
Ibarra-N??ez and El Colegio de la Frontera Sur in Tapachula
provided logistical support. We thank the Peters family for
permission to work on their farm. This study was funded by NSF
no. DEB-9981526 to I.P., the Helen Olsen Brower Fellowship in
Environmental Science of the University of Michigan, and an NSF
Graduate Research Fellowship to S.P..
References
Abarca-Arenas LG, Ulanowicz RE (2002) The effects of taxonomic
aggregation on network analysis. Ecol Model 149:285?296
Abrams P, Menge B, Mittelbach G, Spiller D, Yodzis P (1996) The
role of indirect effects in food webs. In: Polis G, Winemiller K
(eds) Food webs integration of patterns and dynamics. Chap-
man and Hall, New York, pp 371?395
Aho T, Kuitunen M, Suhonen J, Jantti A, Hakkari T (1999)
Reproductive success of Eurasian treecreepers, Certhia famil-
iaris, lower in territories with wood ants. Ecology 80:998?1007
Barton AM (1986) Spatial variation in the effect of ants on an
extrafloral nectary plant. Ecology 67:495?504
Bentley BL (1977) The protective function of ants visiting the
extrafloral nectaries of Bixa orellana (Bixaceae). J Ecol 65:27?
38
Chalcraft DR, Resetarits WJ Jr (2003a) Predator identity and
ecological impacts: functional redundancy or functional
diversity. Ecology 84:2407?2418
Chalcraft DR, Resetarits WJ Jr (2003b) Mapping functional
similarity of predators on the basis of trait similarities. Am
Nat 162:390?402
Crumrine PW, Crowley PH (2003) Partitioning components of risk
reduction in a dragonfly-fish intraguild predation system.
Ecology 84:1588?1597
Davidson DW (1997) The role of resource imbalances in the
evolutionary ecology of tropical arboreal ants. Biol J Linn Soc
61:153?181
Davidson DW, Brown JH, Inouye RS (1980) Competition and the
structure of granivore communities. Bioscience 30:233?238
Davidson DW, Cook SC, Snelling R, Chua TH (2003) Explaining
the abundance of ants in lowland tropical rainforest canopies.
Science 300:969?972
de la Fuente MA, Marquis RJ (1999) The role of ant-tended
extrafloral nectaries in the protection and benefit of a
Neotropical rainforest tree. Oecologia 118:192?202
Del-Claro K, Berto V, Reu W (1996) Effect of herbivore deterrence
by ants on the fruit set of an extrafloral nectary plant, Qualea
multiflora (Vochysiaceae). J Trop Ecol 12:887?892
Dietsch TV (2003) Conservation and ecology of birds in coffee
agroecosystems of Chiapas, Mexico, PhD Thesis. In: School of
natural resources and environment. University of Michigan,
Ann Arbor
Ekl?v P, Werner EE (2000) Multiple predator effects on size-
dependent behavior and mortality of two species of anuran
larvae. Oikos 88:250?258
Fiala B, Grunsky H, Maschwitz U, Linsenmair K (1994) Diversity
of ant-plant interactions: protective efficacy in Macaranga
species with different degrees of ant association. Oecologia
97:186?192
Floyd T (1996) Top-down impacts on creosote bush herbivores in a
spatially and temporally complex environment. Ecology
77:1544?1555
Fonseca CR (1994) Herbivory and the long-lived leaves of an
Amazonian ant-tree. J Ecol 82:833?842
Gibb H (2003) Dominant meat ants affect only their specialist
predator in an epigaeic arthropod community. Oecologia
136:609?615
Greenberg R, Bichier P, Cruz Angon A, Reitsma R (1997) Bird
populations in shade and sun coffee plantations in Central
Guatemala. Conserv Biol 11:448?459
Greenberg R, Bichier P, Cruz Angon 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
Haemig PD (1992) Competition between ants and birds in a
Swedish forest. Oikos 65:479?483
Haemig PD (1996) Interference from ants alters foraging ecology of
great tits. Behav Ecol Sociobiol 38:25?29
Hairston N, Smith F, Slobodkin L (1960) Community structure,
population control and competition. Am Nat 94:421?425
Hall S, Raffaelli D (1991) Food-web patterns?lessons from a
species-rich web. J Anim Ecol 60:823?842
Harris P (1995) Are autecologically similar species also functionally
similar? A test in pond communities. Ecology 76:544?552
H?lldobler B, Wilson EO (1990) The ants. Harvard University
Press, Cambridge
Horvitz CC, Schemske DW (1984) Effects of ants and an ant tended
herbivore on seed production of a Neotropical herb. Ecology
65:1369?1378
Ibarra-N??ez G, Garcia JA, Lopez JA, Lachaud JP (2001) Prey
analysis in the diet of some ponerine ants (Hymenoptera:
Formicidae) and web-building spiders (Araneae) in coffee
plantations in Chiapas, Mexico. Sociobiology 37:723?755
Jaksic FM, Delibes M (1987) A comparative analysis of food-niche
relationships and trophic guild structure in two assemblages of
vertebrate predators differing in species richness: causes,
correlations, and consequences. Oecologia 71:461?472
James DG, Stevens MM, O?Malley KJ, Faulder RJ (1999) Ant
foraging reduces the abundance of beneficial and incidental
arthropods in citrus canopies. Biol Control 14:121?126
Janzen DH (1966) Evolution of polygynous obligate Acacia-ants in
western Mexico. Evolution 20:249?275
Johnson M (2000) Effects of shade-tree species and crop structure
on the winter arthropod and bird communities in a Jamaican
shade coffee plantation. Biotropica 32:133?145
Kelly CA (1986) Extrafloral nectaries, ants, herbivores, and
fecundity in Cassia fasciculata. Oecologia 69:600?605
Koptur S (1984) Experimental evidence for defense of Inga
(Mimosideae) saplings by ants. Ecology 65:1787?1793
Koptur S, Lawton JH (1988) Interactions between vetches bearing
extrafloral nectaries, their biotic protective agents, and
herbivores. Ecology 69:278?283
Laska MS, Wootton JT (1998) Theoretical concepts and empirical
approaches to measuring interaction strength. Ecology 92:461?
447
148
Leston D (1973) The ant mosaic?tropical tree crops and the
limiting of pests and diseases. Proc Natl Acad Sci USA 19:311?
341
Letourneau DK (1998) Ants, stem-borers, and fungal pathogens:
experimental tests of a fitness advantage in Piper ant-plants.
Ecology 79:593?603
Levey DJ, Stiles FG (1992) Evolutionary precursors of long-
distance migration: resource availability and movement patterns
in Neotropical land birds. Am Nat 140:447?476
Majer JD (1978) The maintenance of the ant mosaic in Ghana cocoa
farms. J Appl Ecol 13:123?144
Majer JD, Queiroz MVB (1993) Distribution and abundance of ants
in a Brazilian subtropical coffee plantation. Papua New Guinea
J Agric For Fish 36:29?35
Marquis RJ, Whelan CJ (1994) Insectivorous birds Increase growth
of white oak through consumption of leaf-chewing insects.
Ecology 75:2007?2014
Martinez ND (1993) Effects of resolution on food web structure.
Oikos 66:403?412
Moguel P, Toledo VM (1999) Biodiversity conservation in
traditional coffee systems of Mexico. Conserv Biol 13:11?21
Mols CMM, Visser ME (2002) Great tits can reduce caterpillar
damage in apple orchards. J Appl Ecol 39:888?899
Murakami M, Nakano S (2000) Species-specific bird functions in a
forest-canopy food web. Proc R Soc London B 267:1597?1601
Offenberg J (2001) Balancing between mutualism and exploitation:
the symbiotic interaction between Lasius ants and aphids.
Behav Ecol Sociobiol 49:304?310
Oliveira PS (1997) The ecological function of extrafloral nectaries:
herbivore deterrence by visiting ants and reproductive output in
Caryocar brasiliense(Caryocaraceae). Funct Ecol 11:323?330
Osenberg CW, Sarnelle O, Cooper SD (1997) Effect size in
ecological experiments: the application of biological models in
meta-analysis. Am Nat 150:798?812
Pace ML, Cole JJ, Carpenter SR, Kitchell JF (1999) Trophic
cascades revealed in diverse ecosystems. Trends Ecol Evol
14:483?488
Perfecto I, Casti?eiras A (1998) Deployment of the predaceous ants
and their conservation in agroecosystems. In: Barbosa P (ed)
Conservation biological control. Academic, San Diego, pp
269?289
Perfecto I, Snelling R (1995) Biodiversity and the transformation of
a tropical agroecosystem?ants in coffee plantations. Ecol Appl
5:1084?1097
Perfecto I, Vandermeer J (2002) Quality of agroecological matrix in
a tropical montane landscape: ants in coffee plantations in
southern Mexico. Conserv Biol 16:174?182
Polis GA (1991) Complex trophic interactions in deserts: an
empirical critique of food web theory. Am Nat 138:123?155
Polis GA, Strong DR (1996) Food web complexity and community
dynamics. Am Nat 147:813?846
Polis GA, Sears ALW, Huxel GR, Strong DR, Maron J (2000) When
is a trophic cascade a trophic cascade? Trends Ecol Evol
15:473?475
Poulon B, Lefebvre G (1996) Dietary relationships of migrant and
resident birds from a humid forest in central Panama. Auk
113:277?287
Rashbrook VK, Compton SG, Lawton JH (1992) Ant-herbivore
interactions?reasons for the absence of benefits to a fern with
foliar nectaries. Ecology 73:2167?2174
Ricogray V (1993) Use of plant-derived food resources by ants in
the dry tropical lowlands of coastal Veracruz, Mexico.
Biotropica 25:301?315
Risch SJ, Carroll CR (1982) Effect of a keystone predaceous ant,
Solenopsis geminata, on arthropods in a tropical agroecosys-
tem. Ecology 63:1979?1983
Room P (1975) Relative distributions of ant species in cocoa
plantations in Papua New Guinea. J Appl Ecol 12:47?62
Sanz JJ (2001) Experimentally increased insectivorous bird density
results in a reduction of caterpillar density and leaf damage to
Pyrenean oak. Ecol Res 16:387?394
Schemske DW (1980) The evolutionary significance of extrafloral
nectar production by Costus woodsonii (Zingiberaceae): an
experimental analysis of ant protection. J Ecol 68:959?967
Schmitz OJ, Sokol-Hessner L (2002) Linearity in the aggregate
effects of multiple predators in a food web. Ecol Lett 5:168?
172
Schmitz OJ, Suttle KB (2001) Effects of top predator species on
direct and indirect interactions in a food web. Ecology
82:2072?2081
Schmitz OJ, Hamback PA, Beckerman AP (2000) Trophic cascades
in terrestrial systems: a review of the effects of carnivore
removals on plants. Am Nat 155:141?153
Schulz A, Wagner T (2002) Influence of forest type and tree species
on canopy ants (Hymenoptera: Formicidae) in Budongo Forest,
Uganda. Oecologia 133:224?232
Sih A, Englund G, Wooster D (1998) Emergent impacts of multiple
predators on prey. Trends Ecol Evol 13:350?355
Simberloff D, Dayan T (1991) The guild concept and the structure
of ecological communities. Annu Rev Ecol Syst 22:115?143
Sokol-Hessner L, Schmitz OJ (2002) Aggregate effects of multiple
predator species on a shared prey. Ecology 83:2367?2372
Strong AM, Sherry TW (2000) Habitat-specific effects of food
abundance on the condition of ovenbirds wintering in Jamaica.
J Anim Ecol 69:883?895
Sugihara G, Bersier LF, Schoenly K (1997) Effects of taxonomic
and trophic aggregation on food web properties. Oecologia
112:272?284
Tavares-Cromar AF, Williams DD (1996) The importance of
temporal resolution in food web analysis: evidence from a
detritus-based stream. Ecol Monogr 66:91?113
Thompson RM, Townsend CR (2000) Is resolution the solution? the
effect of taxonomic resolution on the calculated properties of
three stream food webs. Freshwater Biol 44:413?422
Van Bael SA, Brawn JD, Robinson SK (2003) Birds defend trees
from herbivores in a Neotropical forest canopy. Proc Natl Acad
Sci USA 100:8304?8307
Watt AD, Stork NE, Bolton B (2002) The diversity and abundance
of ants in relation to forest disturbance and plantation
establishment in southern Cameroon. J Appl Ecol 39:18?30
Way MJ (1963) Mutualism between ants and honeydew-producing
Homoptera. Annu Rev Entomol 8:307?344
Way MJ, Khoo KC (1992) Role of ants in pest-management. Annu
Rev Entomol 37:479?503
Whalen MA, Mackay DA (1988) Patterns of ant and herbivore
activity on five understory Euphorbiaceous saplings in
submontane Papua New Guinea. Biotropica 20:294?300
Yodzis P, Winemiller KO (1999) In search of operational
trophospecies in a tropical aquatic food web. Oikos 87:327?340
149