|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
ANT OCCUPANCY AND ANTI-HERBIVORE DEFENSE OF Cordia alliodora, A
MATTHEW DAVID TRAGER
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
Matthew David Trager
First, I thank my graduate committee of Emilio Bruna, Heather McAuslane and
Kaoru Kitajima for their insightful questions and for sharing their invaluable expertise
throughout the research and writing process. Jack Ewel graciously allowed me to work
in the Huertos Project and also provided excellent advice on several aspects of this study.
I wish to thank La Selva Biological Station and the Organization for Tropical
Studies for access to the site and facilities. Silvino Villegas and Virgilio Alvarado
assisted with fieldwork. Jack Longino provided difficult ant identification and imparted
some of his substantial knowledge of the Cordia alliodora system. Chad Tillberg shared
data he collected on ant occupancy of C. alliodora at the site, which greatly improved
this study. Michael W. Gates (Systematic Entomology Laboratory, Agriculture Research
Service, US Department of Agriculture) identified the parasitoid wasps, specimens of
which were deposited at the US National Museum. Meghann Bernardy assisted with the
digital image analysis of herbivory. Ian Fiske and Ramon Littell assisted with the mixed
model data analysis. Funding for this study was provided by the University of Florida's
Tropical Conservation and Development program and NSF Grant DEB-0309819 awarded
to Emilio M. Bruna. The Huertos Project was funded by NSF Award LTREB 99-75235
and the Andrew W. Mellon Foundation. Collection and exportation of specimens were
conducted under Costa Rican permit DGVS-483-2004 and USDA permit 37-87383 for
plants and plant products.
Finally, I thank my parents for their constant encouragement and support.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iii
LIST OF TABLES .................................................. ......................... vi
LIST OF FIGURES ..................................................... .......... ............... .. viii
A B ST R A C T .......... ..... ...................................................................................... x
1 IN T R O D U C T IO N ............................................................................. .............. ...
2 ANT SPECIES COEXISTENCE IN Cordia alliodora, A NEOTROPICAL
M Y RM ECOPH Y TE ................ ............. .. ................. ........ .... ............... 5
Introduction ......... ................ ............................... 5
Study System ........... .. ................. ..................................................8
A nt Com m unity Com position...................................... ......................... 9
Spatial Variation in Ant Species Occupancy................................................10
Colony Founding and Expansion................. .............................................. 10
Results ........... .................... ........................... ........11
A nt Com m unity Com position...................................... ........ ............... 11
Spatial H habitat Partitioning..................................... ......................... ......... 12
Colony Founding and Expansion ............... .......................................14
Discussion .............. ........... .. ................................. .......... ........ 15
3 HERBIVORY AND ANTI-HERBIVORE DEFENSE OF Cordia alliodora:
HOW IMPORTANT IS ANT DEFENSE? ..................................... ............... 31
Introduction ........... ...................... ........................... ...................... .. 31
Methods............................... ................... .... 35
Study System .......................................................................... .35
Field Survey of Herbivory ....................... .................. ....................37
Leaf Palatability Bioassay ..................................................... ....... ...........38
R results .............. ................................................ ....................... ............ 39
Field Survey of Herbivory ....................... .................. ....................39
Leaf Palatability B ioassay ................................................... ... .... ........ 41
Discussion ........... ................................................................................... 41
4 C O N C L U SIO N S ........................ .... .... .................... .. .. ........ .... ............ 58
LIST OF REFEREN CE S ............................................................. .................... 59
B IO G R A PH IC A L SK E TCH ...................................................................... ..................67
LIST OF TABLES
2-1 Results of ANCOVA testing the effects of plant age on ant species richness, with
the number of domatia on each plant (logio-transformed) included as a covariate..24
2-2 Results of ANOVA examining the effects of plant age and the identity of the
dominant ant species on the proportion of domatia occupied. Proportional
occupancy was arcsine (square root) transformed to improve normality. Only 2-
and 5-yr-old trees primarily occupied by either Aztecapittieri or Crematogaster
carinata were included in this analysis (only one tree within these age groups,
which was dominated by Cephalotes setulifer, was excluded).............................25
2-3 Results of ANOVA testing the effects of plant age and the presence of the two
numerically dominant ant species (Aztecapittieri and Crematogaster carinata) on
the proportion of the plant occupied by the third most abundant species, Cephalotes
setulifer. Proportional occupancy was arcsine (square root) transformed to improve
normality. Only 2- and 5-yr-old trees primarily occupied by either A. pittieri or Cr.
carinata were included in this analysis (only one tree within these age groups,
which was dominated by Ce. setulifer, was excluded). ........................................26
3-1 Results of mixed model analysis testing the effects of plant age and fertilization on
the number of leaves surrounding focal domatia. ............. .................................... 47
3-2 Results of mixed model analysis testing the effects of plant age and fertilization on
the total area of leaves surrounding focal domatia.............................................48
3-3 Results of mixed model analysis testing the effects of plant age and fertilization on
the number of worker ants within the focal domatia..............................................49
3-4 Results of mixed model analysis testing the effects of plant age, fertilization, and
the number of worker ants on the proportion of leaf area missing. The number of
ants was loglo-transformed and the proportion of leaf area missing was logit-
transformed to improve the distribution for this analysis. .....................................50
3-5 Results of mixed model analysis testing the effects of fertilization and the number
of worker ants present on herbivory for 1-yr-old plants and 5-yr-old plants. The
number of ants was logl0-transformed and the proportion of leaf area missing was
logit-transformed to improve the distribution for this analysis. Although the results
are presented together, the analyses were conducted separately for the two ages...51
3-6 Results of ANOVA testing the effects of plant age, leaf age and fertilization
treatment on the leaf area consumed by one Coptocycla leprosa beetle in 24 hr,
with trial as a random block effect. ........................................ ....................... 52
LIST OF FIGURES
1-1 The inside of a Cordia alliodora domatium containing part of a Cephalotes
setulifer colony. Larvae and pupae are cylindrical and whitish. Cohabiting scale
insects are smaller, pink and attached to the wall of the domatium..........................4
2-1 Plant age significantly affected both (A) the number of domatia on a plant and (B)
the number of ant species present. Means and 95% confidence intervals are shown,
with pairwise differences (calculated with Tukeys HSD) indicated with lowercase
letters. ................................................................................2 7
2-2 Proportional occupancy of domatia in different tree microhabitats varied among
species in both (A) the 2-yr-old trees and (B) the 5-yr-old trees. Although a
number of other species were present in the 2-yr-old trees, including
Pseudomyrmexfortis, none were present in either a large number of trees or a large
number of domatia and so were not included. ......................... ............... ......28
2-3 The two most abundant species, Aztecapittieri and Crematogaster carinata,
differed in their occupation patterns and effects on other species. (A) A. pittieri
occupied relatively more domatia in trees where it was the dominant species
compared with Cr. carinata. This was true regardless of plant age. (B) The
proportional occupancy of Cephalotes setulifer was affected by the species of
dominant ant in the tree, the age of the plant and the interaction of these two factors
(T able 2-3).............................................................................................. 29
2-4 Stacked bar graph showing the expected occupation of the four most abundant ant
species in the 1-yr-old trees-based upon their occurrence in the 5-yr-old trees-
and the observed occupation frequency. All species except Aztecapittieri had
significantly different occupation patterns than expected.......................... ..... 30
3-1 The beetle Coptocycla leprosa spends its entire life cycle on Cordia alliodora. (A)
Late-instar larva with fecal shield, (B) pupa adhering to top of a leaf and (C) adult
on underside of leaf. ...................... ........ ................................. .. ..... 53
3-2 The number of worker ants present in focal domatia varied with plant age but
fertilization treatment had no effect. Boxplots show inter-quartile ranges and
expected minimum and maximum values, with values beyond the 95% CI indicated
by open circles. All of the high outliers in the 1-yr-old plants were domatia
occupied by Crematogaster carinata............................................................... 54
3-3 The number of worker ants present in focal domatia varied according to plant age
and the identity of the ant species. Boxplots show inter-quartile ranges and
expected minimum and maximum values, with values beyond the 95% CI indicated
by open circles. Pseudomyrmexfortis was not present in any domatia from 1-yr-
o ld p lan ts. ......................................................... ................ 5 5
3-4 The proportion of leaf area missing was affected at P < 0.10 by plant age and
fertilization treatment. In general 1-yr-old plants experienced more proportional
leaf damage than 5-yr-old plants, and for the 1-yr-old plants fertilization reduced
herbivore damage. Boxplots show inter-quartile ranges and expected minimum and
maximum values, with values beyond the 95% CI indicated by open circles.........56
3-5 Fertilization significantly reduced the leaf area consumed by individual Coptocycla
leprosa beetles in the leaf palatability trials. This was particularly true for young
leaves, as indicated by the marginally significant (P = 0.056) interactive effect
between fertilization and leaf age (Table 3-6) ............................................ ............ 57
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
ANT OCCUPANCY AND ANTI-HERBIVORE DEFENSE OF Cordia alliodora, A
Matthew David Trager
Chair: Emilio M. Bruna
Major Department: Interdisciplinary Ecology
I studied patterns of ant occupancy, herbivory and anti-herbivore defense in Cordia
alliodora (Boraginaceae) (Ruiz and Pavon) Oken, a common neotropical myrmecophyte.
Specifically, I investigated patterns in ant community composition in 1-, 2-, and 5-yr-old
plants and tested whether changes in ant occupancy with plant age affect the amount of
herbivory sustained by the host plant.
Although 11 ant species were present in the plants studied, four species-Azteca
pittieri Forel, Cephalotes setulifer Emery, Crematogaster carinata Mayr and
Pseudomyrmexfortis Forel-accounted for the vast majority of occupied domatia. The
relative abundance of these species varied according to plot-level environmental
variation, domatia position within plants and plant age. The two most abundant ant
species, A. pittieri and Cr. carinata, were nearly mutually exclusive in the 5-yr-old plants
and the interaction between these species affected the distribution and abundance of the
third most abundant species, Ce. setulifer. Coexistence of multiple ant species on the
same host plant in this system appears to be promoted by heterogeneity in nest site
availability, competitive interactions among species, different life history strategies and
interactions with other organisms.
The amount of herbivore damage on leaves surrounding domatia was affected by
random environmental variation, plant age, fertilization and the number of ants present.
For the 5-yr-old plants, local herbivory was not affected by fertilization but was lower as
a function of the number of workers in a domatium. However, fertilization reduced
herbivory and ant abundance had no effect for the 1-yr-old plants. A palatability trial
with a specialist herbivore, the beetle Coptocycla leprosa Boheman, showed that
fertilized plants were less palatable regardless of plant age. This suggests that C.
alliodora likely has a mixed defensive strategy in which young plants are chemically
defended or tolerant to herbivory, with the effectiveness contingent upon resource
availability, whereas older plants appear to be defended by their resident ant colonies.
Ant-plant protection mutualisms are common throughout the tropics and have
fascinated naturalists and ecologists for over a century (reviewed in Beattie 1985 and
Heil and McKey 2003). Amongst ant-plant protection mutualisms, the symbioses
between myrmecophytes, plants that produce cavities in their stems or leaves in which
ants nest, and their resident ants are considered to be the most specialized (Beattie 1985).
Although there is substantial variation in myrmecophyte-ant relationships, these
mutualisms are maintained by the reciprocal benefits afforded to both parties from
participating in the symbiosis. The ants receive housing and often food from the plant,
whereas the benefits to the plants include protection from herbivores, removal of
encroaching vegetation and, in some cases, fertilization from ant waste. At least 400
plant species in over 100 genera produce structures for housing ants, and at least that
many ant species are myrmecophyte-nesting specialists (Beattie 1985, Davidson and
McKey 1993). Myrmecophytes have served as model systems for studying diverse topics
in ecology and evolution, including plant defense strategies, conditional outcomes of
interspecific interactions, the evolution and maintenance of mutualisms, trophic cascades,
and mechanisms of species coexistence (reviewed in Bronstein 1998 and Heil and
Most myrmecophyte species are occupied by more than one ant species (Fonseca
and Ganade 1996). Ant communities of myrmecophytes often vary temporally according
to plant age or random variation (e.g., Alonso 1998, Palmer et al. 2000) and spatially due
to the interactions between habitat heterogeneity, habitat preferences and interspecific
competition (e.g., Vasconcelos and Davidson 2000, Yu et al. 2001, Palmer 2003).
Nesting space may be a limiting resource in these relationships (Fonseca 1999), and
therefore ants commonly compete to maintain occupancy of a plant, often to the point of
mutual exclusion, as the host plant grows (Janzen 1966, Davidson et al. 1989, Stanton et
al. 2002). Because ant species often differ dramatically in their ability to defend their
host plants, variation in occupancy could significantly influence plant performance
(Janzen 1975, McKey 1984, Itioka et al. 2000, Heil et al. 2001a, Bruna et al. 2004).
Addressing the factors that affect ant community composition as myrmecophytes grow
and the consequent effects on herbivore damage sustained by the host plant is a critical
component of understanding the temporal dynamics of ant-plant protection relationships
and is the focus of this study.
Cordia alliodora is a fast-growing myrmecophytic tree common in secondary
forests and fields throughout much of Central America and northern South America. The
ant associates of C. alliodora inhabit naturally hollow swellings, known as domatia,
produced by the plants at most branch nodes (Figure 1-1). In Costa Rica, more than ten
different ant species have been recorded occupying the domatia of C. alliodora, including
both specialists to that myrmecophyte and stem-nesting generalists (Wheeler 1942,
Longino 1996, Tillberg 2004). Unlike many other myrmecophytes, mature C. alliodora
trees often host colonies of multiple ant species simultaneously, making this system
particularly suitable to studies of species coexistence (Longino 1996; Tillberg 2003,
The leaves of C. alliodora are eaten by a number of insect herbivores including
lepidopteran and dipteran larvae, plant hoppers, scales and mealybugs, beetles and leaf-
cutter ants (Wheeler 1942, Flowers and Janzen 1997, Moser 2000, Tillberg 2004). The
high levels of herbivory and large number of insect herbivores on C. alliodora led
Wheeler (1942) to conclude that the ant occupants provided no substantive benefit for the
plant and, perhaps, were even parasites. However, behavioral studies, analysis of
herbivore damage and analyses using stable isotopes suggest that at least some of the ant
species attack and eat herbivorous insects (Moser 2000, Tillberg 2004).
The objectives of this study are to describe patterns in ant occupation of C.
alliodora and examine the importance of ants and other factors in the anti-herbivore
defense of 1- and 5-yr-old plants. In the second chapter I present results from a survey of
ant occupation of C. alliodora and attempt to identify potential mechanisms of
coexistence. In the third chapter I describe the effects of plant age, fertilization and ant
occupancy on the level of insect herbivory sustained by C. alliodora plants and interpret
these results in light of plant defensive strategies. In the fourth chapter I integrate the
results of these studies and interpret their results with respect to the evolution and
maintenance of the Cordia-ant relationship.
Figure 1-1. The inside of a Cordia alliodora domatium containing part of a Cephalotes
setulifer colony. Larvae and pupae are cylindrical and whitish. Cohabiting
scale insects are smaller, pink and attached to the wall of the domatium.
ANT SPECIES COEXISTENCE IN CORDIA ALLIODORA, A NEOTROPICAL
The co-occurrence of species with similar ecological requirements is ubiquitous in
natural communities, and the mechanisms promoting coexistence are a major focus of
ecological inquiry. Both the intrinsic properties of organisms (e.g., behaviors and life
history traits) and characteristics of the environment (e.g., habitat structure, disturbance
regimes, interactions with other organisms and stochastic events) allow the persistence of
species with overlapping resource needs (Tokeshi 1999, Chesson 2000, Amaresekare
2003). Because patterns of species coexistence differ across space and through time,
single factors rarely account for the observed variation in the presence and abundance of
species (Amarasekare 2003). Consequently, field studies examining the composition of
natural communities benefit greatly from the consideration of multiple, often
complementary, mechanisms of coexistence.
Ant-plant symbioses are ideal model systems in which to study the patterns of
coexistence of species with similar resource requirements (Davidson and McKey 1993,
Bronstein 1998, Heil and McKey 2003, Palmer et al. 2003). In specialized ant-plant
symbioses, myrmecophytes (i.e., ant-plants) provide nesting space and food resources for
ant inhabitants. Ant inhabitants often provide some benefit for the plant in return, such as
protection from herbivores, thereby maintaining the mutualism (Holldobler and Wilson
1990, Bronstein 1998, Heil and McKey 2003). Whereas space is rarely a limiting factor
for terrestrial ant communities (Albrecht and Gotelli 2001), ants inhabiting
myrmecophytes may face strong inter- and intra-specific competition for limited nest
sites (Janzen 1966, Davidson et al. 1989, Fonseca and Ganade 1996, Fonseca 1999,
Stanton et al. 2002). Therefore, although young myrmecophytes may experience
multiple colonization events by queens of different species or undergo a succession of ant
species as they plant grow, mature plants are usually dominated by a single ant colony
(Davidson et al. 1989, Longino 1991, Vasconcelos 1993, Young et al. 1997, Palmer et al.
2000, Feldhaar et al. 2003).
Recent studies on ant-plant relationships have provided empirical models for both
spatial (e.g., Yu et al. 2001, Palmer 2003) and temporal (e.g., Young et al. 1997, Alonso
1998) habitat partitioning among ant species that inhabit the same plant species. In some
systems, variation in host-plant characteristics, often related to underlying habitat
heterogeneity, allows fine-scale partitioning of resources among ant species (Davidson et
al. 1989, Longino 1989, Vasconcelos and Davidson 2000, Palmer 2003). In more
homogeneous conditions, coexistence may result from interspecific differences in ant life
histories and behaviors, including trade-offs between competitive dominance and
dispersal capability (Stanton et al. 2002), fecundity and dispersal ability (Cole 1983,
Vasconcelos 1993, Yu and Wilson 2001, Yu et al. 2004), or interference and exploitation
competition (Fellers 1987, Davidson 1998, Holway 1999). Positive priority effects (the
continued occupation of the species that colonizes first) may also allow ant species that
are poor competitors or even poor colonizers to occupy sites following colony
establishment despite the presence of otherwise dominant species (Longino 1989, Palmer
et al. 2002).
Characteristics of the host plants and the resident ant colonies change over time as
the plant develops and the ant colonies grow, senesce or are replaced by other species
(Vasconcelos and Casimiro 1997, Young et al. 1997, Alonso 1998, Itino and Itioka 2001,
Del Val and Dirzo 2003). The mechanisms listed above are rarely manifested at one
point in time at a single locale, but rather promote species coexistence at larger spatial
and temporal scales (Young et al. 1997; Alonso 1998; Yu et al. 2001, 2004). The early
ontogeny of myrmecophytes may be particularly important for species-sorting of ant
occupants (Davidson et al. 1989, Vasconcelos and Davidson 2000, Palmer et al. 2002).
A comprehensive understanding of ant species coexistence in ant-plant symbioses
therefore must encompass both differences in the life history strategies of the ant species
and the temporal dynamics of the symbiosis.
Cordia alliodora is a neotropical myrmecophytic tree inhabited by several ant
species, including both specialists and stem-nesting generalists. Although C. alliodora
has an extensive geographic range and is often locally abundant, little community-level
research has been conducted on the ants that inhabit this myrmecophyte (but see Longino
1996 and Tillberg 2004). Here I describe patterns of ant occupation in C. alliodora
during the course of the host plant's ontogeny from sapling to young mature tree.
Specifically, I investigated the following questions: 1) How does ant community
composition change with plant age? 2) Is there spatial variation in ant species occupancy?
3) Are there interspecific differences among ant species in colony founding and
expansion? These data are then used to identify potential mechanisms that account for
the maintenance of ant species diversity in the C. alliodora system.
Cordia alliodora (Boraginaceae) is widespread and abundant in the secondary
forests, clearings and forest edges at La Selva Biological Station (Costa Rica, Heredia
Province, 100 26' N, 830 59' W), where I conducted this study. I collected data from trees
that were planted in three replicated, monospecific blocks as part of the Huertos Project, a
long-term ecological study established at La Selva in 1991 in which C. alliodora was one
of the focal tree species. In the Huertos Project plots, single-age stands were planted in
rows at a very high initial density (2887 trees ha-1), periodically thinned, and regularly
weeded to prevent the establishment of other vegetation. Each block had three adjacent
monoculture plots of C. alliodora in 1-, 4- and 16-yr planting cycles; the results presented
here are from 1-, 2- and 5-yr-old trees sampled evenly from all three blocks. Within
plots, rows were 1.73 m apart and plants were placed at 2 m intervals within rows, such
that each individual was at the center of a hexagon 2 m from the six closest plants.
Haggar and Ewel (1995) provide further details about the site, the planting techniques
and the design of the Huertos Project. Cordia alliodora individuals were common in the
secondary forest surrounding the Huertos Project plots.
The ant associates of C. alliodora inhabit naturally hollow cauline swellings (i.e.,
domatia) that the plants produce at most branch nodes. Wheeler (1942) reported 44 ant
species from C. alliodora domatia in Panama, and more than ten ant species have been
recorded as occupants of these swollen nodes in Costa Rica (Longino 1996; Tillberg
2003, 2004). Most of these ant species are stem-nesting generalists with no particular
affinity for C. alliodora, but there are also a number of specialist ant species that are only
found in C. alliodora (Wheeler 1942, Longino 1996, Tillberg 2004). Unlike many other
myrmecophytes, individual C. alliodora trees often host colonies of multiple ant species,
making this system particularly amenable to studies on species coexistence (Longino
1996, Tillberg 2004).
Ant Community Composition
I inventoried all domatia from 18 one-year-old and 18 five-year-old C. alliodora
trees in May and June 2004. The two ages were in adjacent plots in each of the three
blocks. Additionally, Tillberg (2003) determined ant occupancy from all domatia of nine
2-yr-old trees following similar methodology in May-July 2001 and shared these data.
The 2- and 5-yr-old-trees belonged to the same cohort but individual trees were not
resampled because the domatia collection required felling the tree. Trees were sampled
evenly from three replicate blocks in the Huertos Project. The occupation status of the
small proportion of domatia encased in trunks or large branches was determined by
identifying the ants passing through the entrance or by opening the domatium in the field.
All other swollen nodes were excised from the trees and then frozen to kill ant occupants.
I then dissected the domatia in the laboratory, recorded whether they were occupied and
identified the ant inhabitants. Although dead queens were common in the domatia of 1-
yr-old trees and some empty swollen nodes showed signs of past occupation, I considered
only those domatia that contained live ants at the time of collection to be inhabited.
I tested for relationships among the number of domatia, the proportion of domatia
occupied and ant species richness using regression analysis. I compared ant species
richness among the three age classes using ANCOVA, with the number of domatia of
each plant as the covariate.
Spatial Variation in Ant Species Occupancy
To assess within-tree habitat partitioning, Tillberg (2003) classified domatia
according to their relative vertical position on the plant for 2-yr-old trees (i.e., low-, mid-
and high-level branches) and I classified domatia according to their relative age for the 5-
yr-old trees (i.e., younger domatia from terminal and subterminal branches vs. older
domatia from large branches and the bole). The domatia from higher branches of 2-yr-
old plants were similar in age and physical condition to the domatia from the terminal or
subterminal branches of the 5-yr-old plants, making these categories somewhat
comparable. Because the 1-yr-old plants did not have sufficient numbers of domatia or
the branching structure required to make such classifications, they were excluded from
I used chi-square tests to examine the frequency of species occurrence in domatia
from different parts of the trees, with expected values for the test of within-tree habitat
partitioning derived from the total proportion of domatia occupied by each species. To
examine relationships among the occupation patterns of the three most abundant species I
performed ANOVA on the proportion of domatia they occupied in 2- and 5-yr-old trees.
Colony Founding and Expansion
Although there were other mature C. alliodora individuals surrounding the
Huertos Project, I assumed that the dozens of 5-yr-old plants in the immediately adjacent
planting were the most likely source population from which foundress queens would
colonize the 1-yr-old plants. Therefore, I used chi-square tests to analyze the ant
occupation of 1-yr-old trees, with the expected values for frequency of ant occupation
derived from the proportion of domatia occupied by each species on the 5-yr-old trees.
These proportions were pooled across the three blocks.
In June 2004, I also identified the ant occupants and phase of colony development
in the distal, most recently produced, two or three domatia from a single branch
haphazardly selected from 5-yr-old trees (n = 60) that were not used in the whole-tree
analysis. I tested for interspecific differences in colonization frequency of these domatia
using chi-square tests, for which I assumed a null hypothesis that newly produced
domatia would have an equal likelihood of being empty, being colonized by the same
species that occupied the nearest node down the branch, or being colonized by a different
Although the Huertos Project replicated plant age treatments at the plot level, in
this study I treated plants as independent replicates because ant species were patchily
distributed among the plots and I was primarily concerned with variation in ant
communities at the level of individual trees or domatia within trees. Data were
transformed to satisfy the requirements of parametric tests when appropriate. Analyses
were conducted with SPSS 13.0 and R 2.2.0.
Ant Community Composition
The whole-tree inventory included 9051 domatia (n = 664 from 1-yr-old trees, n =
3430 from 2-yr-old trees and n = 4957 from 5-yr-old trees) in which we identified 11 ant
species. The four most abundant species across all plant ages, together accounting for
over 97% of occupied domatia, were Aztecapittieri (35 trees, 3113 domatia),
Crematogaster carinata (19 trees, 1677 domatia), Cephalotes setulifer (23 trees, 714
domatia), and Pseudomyrmexfortis (7 trees, 235 domatia). Others species found were
Cephalotes multispinosus, Crematogaster curvispinosa, Wasmannia auropunctata,
Pachychondyla crenata, Pheidole caltrop and unidentified species of the genera
Brachymyrmex, Pheidole and Pseudomyrmex (Tillberg 2003, pers. obs.).
When data from plants of all three ages were pooled, ant species richness was
positively related to the number of domatia present (R2 = 0.331, n = 45, P = 0.012). The
2-yr-old plants had the most domatia (Fig. la), and also hosted the highest richness of ant
species (Fig. lb), even with the number of domatia included in the model as a covariate
(Table 1). However, higher ant species richness in trees did not translate to higher
proportions of occupied domatia (R2 = 0.065, n = 45, P = 0.31). Indeed, approximately
one-third of the domatia in the 1- and 2-yr-old trees were occupied (mean = 34.2%, std.
dev. = 23.7% and mean = 31.1%, std. dev. = 20.2%, respectively), whereas nearly all of
the domatia on 5-yr-old trees (mean = 94.7%, std. dev. = 5.6%) contained ants.
Spatial Habitat Partitioning
The presence and abundance of ant species varied among the three replicate
Huertos Project blocks, among trees within blocks and among domatia within trees.
Three species A. pittieri, Cr. carinata and Ce. setulifer were present in all three
blocks. Crematogaster carinata showed a highly clumped distribution, occurring in 2-yr-
old trees in all three blocks and dominating most 1- and 5-yr-old trees in the one block
where it was most common. Aztecapittieri was present on at least some plants of all ages
in all blocks. Cephalotes setulifer was present in 2- and 5-yr-old trees from all three
plots, but was absent from 1-yr-old trees in two blocks. Pseudomyrmexfortis was present
in 5-yr-old trees in two of the three blocks but was not found on any of the 1-yr-old plants
and was very uncommon in the 2-yr-old plants included in the whole-tree occupancy
survey. Although either A. pittieri or Cr. carinata occupied the majority of domatia on
nearly all 5-yr-old trees, the smallest tree in this age class was completely inhabited by
Ce. setulifer which occupied 48 of the 49 domatia on the plant; one domatium on this tree
Ant species displayed non-random within-plant microhabitat occupancy in both
the 2- and 5-yr-old plants (x2 = 133.1, df = 4, P < 0.0001 and x2 = 76.81, df = 3, P <
0.0001, respectively). In the 2-yr-old plants, A. pittieri had a significantly higher
habitation frequency in branches from the upper stratum of C. alliodora plants (x =
32.35, df = 2, P < 0.0001), whereas Cr. carinata had a higher habitation frequency for the
lower, older branches (x2 = 90.26, df = 2, P < 0.0001). Cephalotes setulifer displayed a
non-random pattern of occupancy in the three strata (x = 10.43, df = 2, P = 0.005), but no
clear directional trend was evident (Fig. 2-2a). However, in the 5-yr-old plants, Ce.
setulifer was significantly more common in the younger terminal or subterminal domatia
than expected by chance (x2 = 56.47, df= 1, P < 0.0001). In these older plants, Cr.
carinata exhibited no difference in occupation frequency in different parts of the tree (X2
= 0, df = 1, P = 1), and both A. pittieri and P. fortis were less common than expected in
terminal or subterminal domatia (x2 = 4.64, df = 1, P = 0.033 and x2 = 15.69, df = 1, P <
0.0001, respectively; Fig. 2-2b).
The relationship among the occupation patterns of the two numerically dominant
species, A. pittieri and Cr. carinata, and the third most abundant species, Ce. setulifer,
changed substantially over the ontogeny of the symbiosis. Aztecapittieri and Cr.
carinata coexisted in many 1-yr-old plants and on eight of the nine 2-yr-old plants.
However, they were mutually exclusive, with the exception of a single foundress queen
of A. pittieri on one tree dominated by Cr. carinata, by the time trees were 5 years old.
These two dominant ant species showed markedly different occupation strategies (Fig. 2-
3a), with the proportion of occupied domatia differing according to tree age and species
(Table 2-2). When it was the dominant species, A. pittieri occupied a significantly higher
proportion of the host plants' domatia than when Cr. carinata was the dominant species.
The proportion of domatia occupied by Ce. setulifer was dependent upon whether A.
pittieri or Cr. carinata was the dominant ant species in the tree (Fig. 2-3b). Specifically,
trees dominated by Cr. carinata had more domatia occupied by Ce. setulifer regardless of
tree age, and the changes in proportional occupancy of Ce. setulifer differed among trees
dominated by the two species as the plants aged. The significant "Plant age x Ant
species" interaction term suggests that A. pittieri increasingly excluded Ce. setulifer as
the plants aged, whereas Ce. setulifer occupied an increasing proportion of domatia on
trees dominated by rC. carinata as the plants aged (Table 2-3).
Colony Founding and Expansion
Most ant species inhabiting C. alliodora were present on 1-yr-old trees only as
founding queens or very small colonies. The 2-yr-old trees were always occupied by
multiple con- and heterospecific colonies, most of them with small numbers of workers.
In these trees, domatia were colonized both through expansion of growing colonies and
by establishment of new colonies by foundress queens in unoccupied domatia (Tillberg
2003). When the plants were 5 years old and fewer unoccupied domatia were available,
the occupation of newly produced domatia appeared to occur primarily by colony
expansion. However, although almost all of the domatia of 5-yr-old trees contained ants
from established colonies, I also found some evidence of continued attempts by mated
queens to establish new colonies in most recently produced domatia of these older plants.
Chi-square analysis showed that ant occupation frequency of 1-yr-old trees was
significantly different than would be expected based on the relative abundance of the four
most abundant species in 5-yr-old trees (2 = 44.34, df= 3, P < 0.0001). Chi-squared
tests for each species showed that A. pittieri was represented proportionally in 1-yr-old
plants, but Ce. setulifer and P. fortis were underrepresented and Cr. carinata was
overrepresented (Fig. 2-4).
Colonization of apical domatia in 5-yr-old trees was not random (x2 = 78, df = 2,
P < 0.0001), and the frequency of different states of domatia habitation did not differ
among the four most common ant species (X2heterogeneity = 8.17, df = 6, P = 0.23). Rather,
new domatia produced by plant growth were most likely to be colonized by expansion of
the ant colony in the adjacent, more basal, domatium regardless of the species present (71
of 93 cases). Of the remaining apical domatia not colonized by expansion of the
neighboring ant colony, 14 (15.1%) were inhabited by workers of species nesting
elsewhere in the same plant or had been colonized by heterospecific queens and only
eight (8.6%) were not yet colonized at the time of sampling.
The parasitoid wasp Conoaxima affinis (Eurytomidae) was abundant in one of the
three blocks, where I found 13 larvae or pupae in domatia from the 1-yr-old plants with
paralyzed or dead A. pittieri queens. There was never more than one larva or pupa on a
queen. In this plot there were only 56 domatia in which I found live A. pittieri queens,
indicating an attack rate of 18.8% (13 of 69) on foundresses. This may be an
underestimate, as I frequently found domatia on 1-yr-old plants that contained only the
legs and head capsules of queens that were presumably consumed by the wasps.
The patterns of ant occupation I observed suggest that spatial heterogeneity of
nesting space among and within plants, temporal variation in habitat characteristics
related to changes in plant growth form, and interspecific competition among ant
occupants for nest space promote ant species coexistence in Cordia alliodora.
Competition, both through preemptive discovery of resources and through physical
conflict, is considered to be important for structuring many ant communities (Davidson
1980, Holldobler and Lumsden 1980, Holldobler and Wilson 1990). I did not examine
behavioral interactions among the ant species in this study, but the patterns of colony
initiation, growth and tree habitation suggest that the ant species occupying C. alliodora
engage in competitive interactions for nesting space. It appeared that the outcome of
these interactions varied according to plant age, random environmental variation among
the three blocks and differences in the life-history strategies of the most common ant
In contrast to the conclusions of Fonseca (1999), the results of the whole-tree
occupation analysis suggest that in the C. alliodora system the availability of nesting
space does not limit ant colony size when the plants are 1 or 2 years old. Indeed, more
than 60% of the domatia were unoccupied in both 1- and 2-yr-old trees. However, this
was likely the result of two very different processes. The low frequency of ant
occupation in the 1-yr-old plants was probably due to the shorter period that the domatia
had been available for colonization combined with the small number of domatia, which
may make trees less attractive for colony-founding queens. Conversely, the low
frequency of ant occupation in the 2-yr-old plants was likely due to the superabundance
of domatia and inability of young ant colonies to expand into available habitat. In fact,
the 2-yr-old plants had significantly more domatia than the 5-yr-old plants included in
this study, despite the fact that crown volume increases as the plants age (Menalled et al.
1998). It is possible that younger C. alliodora plants have denser branching that
produces relatively more domatia compared to the tiered, open crowns typical of mature
C. alliodora trees. This counterintuitive pattern of domatia production may be an artifact
of the planting design in the Huertos Project: because there was little competition for
light among young C. alliodora plants, they displayed thick, bushy growth prior to
dramatic increases in height accompanied by shedding of the lower branches as
competition for light increased (J.J. Ewel, pers. comm.).
Regardless of the generality of C. alliodora crown geometry as the plants age, in
this study the abundance of available nesting space in 2-yr-old trees may largely explain
why ant species richness was highest at this age. The changes in crown structure may
also partially explain why many of the generalist ant species found in the 2-yr-old plants
were not present in the 5-yr-old trees. If interspecific competition for nest sites limits
species coexistence as the plants age, then the presence of more unoccupied domatia in
the 2-yr-old plants may have promoted species diversity at this stage of plant
development. This could be particularly true when the ant colonies were small and
unable to wage territorial wars typical of arboreal ant communities (Holldobler and
Lumsden 1980). Tillberg (2003) showed that the lower branches of 2-yr-old plants were
inhabited by a diversity of generalist ants, which presumably did not recolonize after
these branches senesced as the plants developed the crown structure typical of mature
trees. Additionally, since many of the species found on 2-yr-old trees were generalist
live-stem nesters, they may have been only opportunistic inhabitants of the younger C.
alliodora plants and either died or moved as the plants aged and competition for nest sites
increased (Alonso 1998, Longino 1996, Palmer et al. 2000).
I found evidence of within-tree habitat partitioning among ant species in both the
2- and 5-yr-old trees. There is little basis for comparison with other ant-plant systems
because C. alliodora is unusual in housing multiple ant species in mature plants (but see
Young et al. 1997 and Palmer et al. 2003). However, many myrmecophytes can be
occupied by multiple ant species at early stages of their development and arboreal ants in
other systems have been shown to divide space based on the distribution of resources
(Cole 1983, Bluthgen et al. 2004). In the C. alliodora system, honeydew-producing
hemipteran symbionts (Pseudococcidae and Coccidae) appear to be an important food for
some of the ant species (Tillberg 2004). Because these insects rely on access to plant
vascular tissue to feed, they are likely not distributed evenly among domatia of different
ages. Therefore, variation in the ant-coccoid relationship may account for within-tree
habitat selection. Aztecapittieri was more abundant in the younger domatia in 2-yr-old
plants but more abundant in older domatia in the 5-yr-old plants. Pseudococcids and
coccids were common in A. pittieri domatia, but this species does not appear to rely
solely on honeydew for nutrition (Tillberg 2004). Crematogaster carinata showed no
preference for domatia microhabitat in the 5-yr-old trees, inhabited older branches more
frequently in 2-yr-old trees, and also only rarely tended coccids and pseudococcids inside
its domatia (Tillberg 2004). In contrast, Ce. setulifer, which disproportionately occupied
younger domatia, commonly tended hemipterans and the apparently depends primarily on
plant sources of nutrients such as the concentrated fluid excreted by the coccoids
(Tillberg 2004). Pseudomyrmexfortis was more abundant than expected in the older
domatia from large branches and trunks, which are not surrounded with the plant vascular
tissue required for coccoids to feed. The relative importance of plant or animal food
sources for P. fortis is unknown, but its occupation pattern suggests that honeydew-
producing coccoids may not be a primary source of nutrition. Work in other systems has
shown complex relationships among ants, myrmecophytes and ant-tended Hemiptera
(Gaume et al. 1998, Lapola et al. 2005), but the importance of these interactions in most
systems is unknown.
The presence of a large, polydomous colony of the habitat generalist Cr. carinata
in one of the plots apparently reduced the colonization frequency of Aztecapittieri in 1-
yr-old plants and completely prevented colonies of that otherwise dominant species from
occurring in 5-yr-old trees. The dominance of Cr. carinata in this plot could be
attributable to a number of factors, but certainly this species' mode of colonizing the 1-
yr-old trees (colony expansion from the leaf litter) allowed it to exploit the empty
domatia prior to discovery by foundress queens ofA. pittieri. Crematogaster carinata
displayed a markedly different occupation strategy than A. pittieri on older plants as well,
in which the former species left significantly more domatia uninhabited on 5-yr-old trees
where it was the numerically dominant species (Fig. 3a). It appeared that the competitive
interactions between the two dominant ants resulted in available habitat (uncolonized
domatia) that Ce. setulifer was able to occupy and then successfully defend against the
numerically dominant colonies ofA. pittieri and Cr. carinata.
Interaction with natural enemies can alter the outcome of competitive interactions,
thereby promoting coexistence (Worthen 1989, Pacala and Crawley 1992, Tokeshi 1999,
Chesson 2000). The high mortality of A. pittieri queens due to attacks by the parasitoid
wasp, C. affinis, may have mediated the success of founding queens of other species. In
his description of this genus, Brues (1922) stated that a congener wasp, C. aztecicida, was
responsible for the death of many Azteca queens colonizing Cecropia plants in Guyana.
Yu and Davidson (1997) also found that C. aztecicida accounted for a high proportion of
colony failures in two of the Azteca species in their study. Although the effects of
parasitoids on myrmecophyte ant community composition have not been tested directly,
there is evidence that wasp abundance varies across habitats and can facilitate colony
success of non-host species (Yu and Davidson 1997). In my study, A. pittieri had the
lowest proportional occupancy in the plot where C. affinis was most abundant. This was
also the plot where Cr. carinata was abundant in the 1-yr-old trees and was the only plot
where founding queens of Ce. setulifer were found in domatia of 1-yr-old plants.
Longino (1996, pers. comm.) found similar wasps killing Azteca queens in both C.
alliodora and Cecropia species. If attack by parasitoid wasps diminishes the
survivorship of A. pittieri queens to the extent that it increases the availability of nesting
habitat for other ant species, then these wasps may be an important equalizing factor in
interspecific competition (Chesson 2000).
Although I did not conduct the experiments necessary to test for interspecific
trade-offs in competition and colonization or fecundity and dispersal, the observed
patterns of ant occupation suggest that the ant species did differ in their ability to
colonize domatia and compete with other species for nest sites. The same two species, A.
pittieri and C. carinata were numerically dominant at all three plant ages included in this
study, but their proportional occupancy varied significantly between 2- and 5-yr-old
trees. Both P. fortis and C. setulifer occupied fewer domatia in 1-yr-old trees than
expected based on their relative abundance in 5-yr-old trees. This could indicate a low
resource allocation to reproductive castes in these species paired with a higher success
rate of colony founding for the few queens produced, as Cole (1983) suggested for other
species of Pseudomyrmex and Cephalotes (= Zacryptocerus) that inhabit mangrove
islands. Positive priority effects for young colonies of these subdominant species,
namely the ability to resist eviction by the larger colonies of A. pittieri or Cr. carinata
following colony establishment, may explain this phenomenon. Workers of both C.
setulifer and P. fortis are much larger than those of A. pittieri and Cr. carinata, which
could result in favorable one-on-one success in conflict for the former two species, or at
least for the aggressive P. fortis (McGlynn 2000). Both ant species also have
formidable, though very different, defenses: queens and major workers of Ce. setulifer
effectively block the entrances to their domatia with their phragmotic heads, and P. fortis
workers possess a powerful sting. Colony size alone has proven to be a strong predictor
of competitive outcome in other ant communities (Fellers 1987, Palmer 2004), but, as
shown in this and other systems, morphological and behavioral adaptations of ant species
are also important in determining the outcome of competitive interactions (Davidson
1998, Holway 1999).
In order to understand the effects of microhabitat occupation and ant competition
in the context of the mutualism, it is useful to explicitly consider the life history of the
plant that provides the resources for which the ant species presumably compete
(Bronstein 1998, Heil and McKey 2003). If one or more of the ant species provide
fitness benefits for C. alliodora individuals, then selection on plant life history traits
would favor allocation for ant-related traits (i.e., domatia production and lack of chemical
defenses against ant-tended coccoids) at the stage where acquiring a protective ant colony
is most beneficial and least costly (Brouat and McKey 2000). Because the ant species
may affect the plant differentially, plant traits should evolve to maintain the relationship
with the ant partners that provide the greatest net benefit (Stanton 2003). Although
virtually all C. alliodora plants beyond the seedling stage produce domatia at branch
junctures, most of the 1-yr-old plants in this study were still small and had only a few
domatia. However, the 2-yr-old trees had a large number of domatia-significantly more
than the 5-yr-old plants-that housed young colonies of many ant species. If rapid plant
growth during the first 2 years of development results in the production of more domatia,
then the probability of mutualistic ant species, such as A. pittieri or Cr. carinata, would
increase (Tillberg 2004). The ant community composition data showed that as the plant
cohorts aged, these two ant species were also competitively superior and dominated
nearly all the trees within 5 years of establishment. If they are indeed mutualists (see
Chapter 3 of this thesis, Tillberg 2004), then their numerical dominance over other ant
species likely provides significant fitness benefits for C. alliodora. The benefits to the
host plant afforded by mutualist ant species in turn could have affected the evolution of
allocation for domatia production and perhaps other traits that benefit these ant species in
particular (Brouat and McKey 2000).
In this study, I have described the patterns of ant occupancy in C. alliodora and
attempted to explain them through invoking the life-history characteristics of the ant
species, changes in branching structure related to the age of the host plants, the
differential abundance of honeydew-producing coccoids in different parts of the tree, the
effect of parasitoid wasps attacking one of the dominant ant species, and the interaction
of these mechanisms across space and through time. Although coexistence and
competition have been studied in ant-plant mutualisms, most research has focused on
single mechanisms that structure the relationship. However, mutualist systems often
involved multiple guilds of interacting species that likely vary in their interactions and
responses to environmental variation (Stanton 2003). As such, investigations of ant
species coexistence in myrmecophytic hosts clearly benefit from the incorporation of
multiple factors, and such approaches likely provide a better understanding of these
apparently simple systems.
Table 2-1. Results of ANCOVA testing the effects of plant age on ant species richness,
with the number of domatia on each plant (loglo-transformed) included as a
Source of variation df MS F P
Plant age 2 25.107 53.129 < 0.001
No. domatia 1 4.514 9.551 0.004
Error 41 19.375 0.473
Table 2-2. Results of ANOVA examining the effects of plant age and the identity of the
dominant ant species on the proportion of domatia occupied. Proportional
occupancy was arcsine (square root) transformed to improve normality. Only
2- and 5-yr-old trees primarily occupied by either Aztecapittieri or
Crematogaster carinata were included in this analysis (only one tree within
these age groups, which was dominated by Cephalotes setulifer, was
Source of variation
Plant age x Ant species
Table 2-3. Results of ANOVA testing the effects of plant age and the presence of the two
numerically dominant ant species (Aztecapittieri and Crematogaster
carinata) on the proportion of the plant occupied by the third most abundant
species, Cephalotes setulifer. Proportional occupancy was arcsine (square
root) transformed to improve normality. Only 2- and 5-yr-old trees primarily
occupied by either A. pittieri or Cr. carinata were included in this analysis
(only one tree within these age groups, which was dominated by Ce. setulifer,
Source of variation
Plant age x Ant species
d 100- a
1 T1 ---[-
0 I I I
1 2 5
0 a ---
0- -- ____- -- -- | -- -
1 2 5
Plant age (years)
Figure 2-1. Plant age significantly affected both (A) the number of domatia on a plant and
(B) the number of ant species present. Means and 95% confidence intervals
are shown, with pairwise differences (calculated with Tukeys HSD) indicated
with lowercase letters.
S P. fortis
* Ce. setulifer
a Cr. carinata
o A. pittiei
* Ce. setulifer
a Cr. carinata
O A. pittieri
Figure 2-2. Proportional occupancy of domatia in different tree microhabitats varied
among species in both (A) the 2-yr-old trees and (B) the 5-yr-old trees.
Although a number of other species were present in the 2-yr-old trees,
including Pseudomyrmexfortis, none were present in either a large number of
trees or a large number of domatia and so were not included.
Plant age (years)
Plant age (years)
Figure 2-3. The two most abundant species, Aztecapittieri and Crematogaster carinata,
differed in their occupation patterns and effects on other species. (A) A.
pittieri occupied relatively more domatia in trees where it was the dominant
species compared with Cr. carinata. This was true regardless of plant age.
(B) The proportional occupancy of Cephalotes setulifer was affected by the
species of dominant ant in the tree, the age of the plant and the interaction of
these two factors (Table 2-3).
[ A. pitteri
E Cr. carinata
0 A. pittieri
l Cr. carinata
o 06 IP. fortis
06l Ce. setulifer
SE Cr. canrinata
S0.4 A. pittieri
Figure 2-4. Stacked bar graph showing the expected occupation of the four most
abundant ant species in the 1-yr-old trees-based upon their occurrence in the
5-yr-old trees-and the observed occupation frequency. All species except
Aztecapittieri had significantly different occupation patterns than expected.
HERBIVORY AND ANTI-HERBIVORE DEFENSE OF CORDIA ALLIODORA: HOW
IMPORTANT IS ANT DEFENSE?
The diverse chemical and physical defenses that plants have evolved in response to
herbivory can be physiologically costly and often are produced at the expense of growth
and reproduction (Herms and Mattson 1992, Sagers and Coley 1995, Coley et al. 2005).
Although there are many theories explaining plant defensive strategies (reviewed in
Coley and Barone 1996 and Stamp 2003), in general plants should produce the type and
amount of defenses that optimize their fitness given limited resources and trade-offs in
their allocation (Coley et al. 1985, Zangerl and Rutledge 1996). Tropical forests often
have exceptionally high rates of herbivory and, therefore, selection for effective anti-
herbivore defense strategies may be particularly strong for tropical plants (Coley and
Aide 1991, Coley and Barone 1996).
Variation in anti-herbivore defenses among individuals within a plant species may
be influenced such as age, size, genotype, season, habitat, previous herbivory and plant
condition (Feeny 1970, McKey 1974, Ernest 1989, Mihaliak and Lincoln 1989, Bowers
and Stamp 1993). Additionally, nutrient availability can affect both the evolution of
defensive strategies of species and the ability of individual plants to synthesize deterrent
chemicals or continue growth despite herbivore damage (Coley et al. 1985, Nichols-
Orians 1991, Bryant et al. 1992, Folgarait and Davidson 1995, Wilkens et al. 1996,
Strauss and Agrawal 1999). Determining the net costs or benefits of chemical anti-
herbivore defense can allow the formulation of testable predictions regarding the
distribution of herbivore damage among and within plants (McKey 1974, Zangerl and
Rutledge 1996). Conversely, studying patterns of herbivory in experimental or natural
systems can elucidate the effects of plant characteristics and environmental variation on
the production and relative importance of different types of anti-herbivore defense (Coley
et al. 1985, Ernest 1989, Coley et al. 2005).
Myrmecophytes (ant-plants) are particularly interesting subjects with which to
examine plant defense strategies (Folgarait and Davidson 1994, 1995; Dyer et al. 2001,
Heil and McKey 2003). In these systems, plants provide nesting space and often food for
ant occupants, which in turn usually protect the plants from herbivory and encroaching
vegetation (reviewed in Davidson and McKey 1993, Bronstein 1998 and Heil and McKey
2003). The degree of protection afforded by the ants varies among ant-plant partnerships
(Vasconcelos and Casimiro 1997, Itioka et al. 2000, Nomura et al. 2000, Heil et al.
2001 a), and within systems according to environmental factors and individual plant and
ant characteristics (Koptur 1985, Michelangeli 2003, Faveri and Vasconcelos 2004).
Most myrmecophyte species can be occupied by multiple ant species (Fonseca and
Ganade 1996, Alonso 1998, Bruna et al. 2005), which range in their ability to defend the
plant from herbivorous insects from very efficient to completely ineffective (Janzen
1966, McKey 1984, Heil et al. 2001a, Bruna et al. 2004). Identifying this variation in
partner quality and determining the effects on the host plant are important aspects of
understanding the evolution and maintenance of ant-plant mutualisms involving guilds of
ant occupants (Stanton 2003).
Because the defensive ability of ant symbionts varies, it could benefit plant
performance to increase production of chemical or other defenses when the ants do not
prevent herbivory. Conversely, in circumstances where ants do provide effective anti-
herbivore defense, allocation for ant-related traits should increase and allocation for other
defenses should decrease. Housing and provisioning ants may bear substantial costs for
the plant, and in some ant-plant systems the plants only make these investments when the
benefits of defensive ants are realized (Heil et al. 1997, Heil et al. 2001b, Dyer et al.
2001). For example, work on Piper cenocladum, has shown that experimental removal of
defensive ants resulted in reduced investment by the host plant in food bodies for ants
and increased production of defensive chemicals. This finding supports the idea of a
trade-off between alternative defensive strategies at the level of individual plants of
myrmecophytic species, with the allocation to different defenses contingent upon both
resource availability and presence of ant defenders (Dyer et al. 2001, 2004).
In addition to ant presence, nutrient availability also affects the defensive
mechanisms of at least some ant-plants. Myrmecophytes that feed resident ants depend
on light and soil resources for both the production of food bodies and the synthesis of
defensive chemicals. Some work has shown that abundant resources may break the
apparent trade-off between these two investments in anti-herbivore defense (Dyer et al.
2004). Indeed, Folgarait and Davidson (1994) found that in Cecropia, higher light levels
resulted in increased production of both food bodies for ants and carbon-based defensive
chemicals, suggesting that defense-related traits may co-vary and correlate with plant
growth when resources are abundant. However, work on the same species of Cecropia
found that nutrient augmentation increased the production of food bodies, but
concentrations of carbon-based chemical defenses were significantly higher under low
nutrient treatments (Folgarait and Davidson 1995). For myrmecophyte species that do
not produce food bodies, nutrient availability could directly affect the rate of leaf damage
and total losses to herbivory through either the production of defensive chemicals or
variation in tolerance and resilience to leaf damage. The production of domatia, usually
associated with plant growth in general, is a key component of myrmecophyte growth
(Brouat and McKey 2000) and may be an important indirect effect of nutrient availability
in species that do not provision ant occupants.
Seedlings and very young individuals of myrmecophytic species usually are not
colonized by ants or house only very small colonies (Feldhaar et al. 2003, Dejean 2004).
This has led to the suggestion that young myrmecophytes may be more reliant upon
chemical and physical defenses than older individuals that are protected by mutualist ant
colonies. Nomura et al. (2001) found that myrmecophytic species ofMacaranga
produced more chemical defenses as saplings prior to ant occupation than as ant-occupied
adults. However, even small individuals of some myrmecophyte species may be
protected by mutualistic ants (Schupp 1986, Itino and Itioka 2001). Furthermore, young
individuals of fast-growing myrmecophytes may have high tolerance to herbivory,
allocating more resources to growth than to defense (Del Val and Dirzo 2003).
Determining the relative importance of ant defenders for preventing herbivory as
myrmecophytes grow from seedlings to mature plants is critical to understanding the
evolution and ecology of ant-plant mutualisms (Bronstein 1998, Heil and McKey 2003).
Additionally, nutrient availability is known to be important for the anti-herbivore defense
of several myrmecophytic species and should be considered in any examination of
potential trade-offs or ontogenetic shifts in plant defensive strategies. The primary
objective of this study was to investigate how the presence and identity of ant occupants,
plant age, and nutrient availability affect herbivory of Cordia alliodora, a common
neotropical myrmecophyte. Cordia alliodora is occupied by a diversity of ant species,
which are non-randomly distributed among plant ages (Chapter 2, this thesis) and likely
vary in their ability to deter herbivores, making this an ideal system to investigate the
sources of inter-plant variation in herbivory and anti-herbivore defense by ants (Stanton
Cordia alliodora (Boraginaceae) is widespread and abundant in the secondary
forests, clearings and forest edges at La Selva Biological Station (Costa Rica, Heredia
Province, 100 26' N, 830 59' W), where I conducted this research. I studied trees planted
in monospecific plots as part of the Huertos Project, a long-term ecological study
established at La Selva in 1991 in which C. alliodora was one of the focal tree species.
In each of the three replicated blocks of the Huertos Project, single-age stands of C.
alliodora were planted in rows at a very high initial density (2887 trees ha-1), periodically
thinned, and regularly weeded to prevent the establishment of other vegetation. Each
block had three adjacent plots of trees in 1-, 4- and 16-yr planting cycle; the results
presented here are from 1- and 5-yr-old trees sampled evenly from all three blocks. In
each plot, parallel rows were offset and placed 1.73 m apart, with plants placed at 2 m
intervals within rows, such that each individual was at the center of a hexagon 2 m from
the six closest plants. Haggar and Ewel (1995) provide further details about the site, the
planting techniques and the design of the Huertos Project. The plots included in this
study were divided into fertilized and unfertilized treatments. The nutrient-supplemented
plants received a liquid fertilizer of urea and NO3- (31% N volume:volume) every two
weeks, totaling 320 kg*hal*y-1 (Silver et al. 2005).
The domatia inhabited by the ant associates of C. alliodora are naturally hollow
cauline swellings that the plants produce at most branch nodes. In Costa Rica, C.
alliodora is most commonly occupied by the specialist ants Aztecapittieri and
Cephalotes setulifer, as well as several generalist live stem inhabiting ants including
species of Crematogaster, Pseudomyrmex, Cephalotes and Pheidole (Wheeler 1942,
Longino 1996, Tillberg 2004). Stable isotope and behavioral studies suggest that A.
pittieri and C. carinata, the two most abundant ant species occupying C. alliodora in the
Huertos Project plots at La Selva, patrol the plant regularly and consume insect prey
(Moser 2000, Tillberg 2004). In contrast, the isotopic profile of Cephalotes setulifer
indicates that this species subsists primarily on honeydew secreted from coccoid
hemipterans (Coccidae and Pseudococcidae) that cohabit the domatia they occupy
(Tillberg 2004). However, the relative effects of occupation by different ant species on
the amount of leaf damage sustained by their host plants are unknown.
The leaves of C. alliodora are eaten by a number of generalist and specialist insect
herbivores (Wheeler 1942, Flowers and Janzen 1997, Moser 2000, Rojas et al. 2001,
Tillberg 2004). Many of the resident ant species also tend honeydew-producing
Hemiptera inside the domatia, which could negatively affect plant performance (Moser
2000, Tillberg 2004). The most abundant and most damaging insect herbivore present on
the trees at the Huertos Project plantings of C. alliodora was the tortoise beetle
Coptocycla leprosa (Chrysomelidae: Cassidini; Fig. 3-1). This beetle spends its entire
life cycle on C. alliodora, appears to be a specialist on the genus Cordia, and can
extensively defoliate the plant when population densities are high (Wheeler 1942,
Flowers and Janzen 1997). The chronically high levels of leaf damage and the large
number of insect herbivores on C. alliodora led Wheeler (1942) to conclude that the ant
occupants provided no substantive benefit to the plant.
Field Survey of Herbivory
To determine the effects of plant and ant factors on herbivory, I collected one
subterminal domatium from a single haphazardly selected branch from 115 Cordia
alliodora plants (N = 8-10 individuals from each plant age x fertilization x block
combination). I identified the ant species present in each domatium, counted the number
of workers present, and collected all fully expanded leaves within 10 cm of the focal
domatium. I then produced digital images of these leaves with a flatbed scanner and
measured the leaf area missing with Scion Image (v. 4.02, Scion Corporation) following
the protocol described by O'Neal et al. (2002). The total area of collected leaves varied
considerably among plants, so for the analysis of herbivory I used the proportion of leaf
area missing, logit transformed to improve the distribution for parametric tests. Although
measuring leaf area missing only once provides a static view of herbivore damage and
may underestimate the impact of herbivores on longer time scales, it is nevertheless
useful for comparisons of relative herbivore damage within species (Lowman 1984,
Brown and Allen 1989, Coley and Barone 1996).
To test for the effects of plant age and fertilization on the number of leaves, total
leaf area, and the number and identity of ants in the focal domatium, I used a linear
mixed model analysis. Because plant age and fertilization were replicated at the level of
the three blocks, individual plants (treated as subplots) were nested within block x plant
age x fertilization combinations. There were no subplot-level factors or covariates in
these analyses. I also used a more complex linear mixed model procedure to test for the
effects of plot-level variation (the block x plant age x fertilization combinations) and
subplot-level variation (the number of worker ants on each plant) on the amount of leaf
y iki = u + b + a, + P k + E jk + yx ki + E jki
In this model, y ykl indicates the measure of herbivory, b is the random block effect, a, is
the fertilization effect, /f k is the plant age effect, jk is the whole-plot error, yx ,jk is the
effect of the covariate ant number within individual plants and ,kl is the subplot (plant-
level) error. In this analysis, individual plants were nested within the plots, which were
defined as plant age x fertilization x block combinations.
Because the data were not balanced, I used maximum likelihood estimation
methods to fit the models. The significance of the block effect, b i, was determined using
a likelihood ratio test and was treated as a random effect contributing to total variance in
the final model rather than a fixed effect (Pinhiero and Bates 2000). Only the main
effects of plant age and fertilization treatment were included in the final model following
a likelihood ratio tests showing that interactive effects did not improve the model fit
(Pinhiero and Bates 2000). Due to significant differences in ant occupancy and herbivory
between 1- and 5-yr-old plants, I conducted analyses similar to the model above
separately for the two ages. All mixed model analyses were conducted with the Ime
program in R 2.2.0 following the protocol of Pinhiero and Bates (2000). All P-values
were calculated with marginal (Type III) tests for significance.
Leaf Palatability Bioassay
In July 2004, I tested the palatability of Cordia alliodora leaves for adult
Coptocycla leprosa beetles with a three-factor, fully crossed laboratory trial. The factors
tested in this study were plant age (1- or 5-yr-old), leaf age (young or mature, which were
easily distinguished based on leaf color and texture) and fertilization treatment
(unfertilized or fertilized). Beetles were collected from 1-yr-old plants in the field and
then starved for approximately 24 hr in the laboratory. I then placed each of 48 randomly
selected individuals in a breathable plastic bag for 24 hr with one freshly collected C.
alliodora leaf. All leaves were collected from the same plot and all had relatively little
insect damage to control for herbivory-induced changes in palatability. I conducted three
trials with identical treatments and experimental procedures, with each trial containing
six replicates of the eight treatment combinations (N=144 beetles tested). To quantify
herbivory I measured the initial and final leaf area with a Licor 3100 area meter.
Because the young leaves lost area due to reduced turgor pressure over the course of the
experiment, I used a correction factor derived from a linear regression equation based on
the initial area to calculate their change in area due to shrinkage (y = 0.9934x 0.7378, R2
= 0.9988, P < 0.001). I used a block design ANOVA to test for main and interaction
effects of the treatments on the amount of leaf material consumed, with the trial as the
block factor to account for random temporal effects.
Field Survey of Herbivory
One to 16 leaves were present within 10 cm of focal domatia (mean SD = 6.2
3.1), and the total leaf area before herbivory ranged from 4.6 to 310.8 cm2 (mean SD =
96.9 60.9 cm2). There were no significant effects of plant age or fertilization on the
number of leaves (Table 3-1), but the total leaf area was marginally significantly larger
for the 5-yr-old plants (Table 3-2). Although the leaf area varied within both ages, the 5-
yr-old plants tended to have greater leaf area than the 1-yr-old plants (mean SD = 111.7
+ 65.2 cm2 and 82.8 53.2 cm2, respectively).
Most domatia (78.3%) contained ants. The frequency of unoccupied domatia was
approximately twice as high for the 1-yr-old plants as for the 5-yr-old plants (28.8%, and
14.3%, respectively). Aztecapittieri was the most abundant species (43.3%), followed
by Crematogaster carinata (33.3%), Cephalotes setulifer (14.4%) and Pseudomyrmex
fortis (8.9%). Significantly more worker ants were present in domatia from 5-yr-old
plants than in domatia from 1-yr-old plants, but there was no effect of fertilization on ant
number (Table 3-3, Fig. 3-2). The number of worker ants present also varied
substantially according to which species occupied the plant (Fig. 3-3). Overall, Cr.
carinata had the most workers per domatium overall and was the only species to have
more than founding queens and a few workers present in domatia from the 1-yr-old
The proportion of leaf area missing from the leaves surrounding each domatium
ranged from 0.02 to 0.45 (mean + SD = 0.12 0.09). The 1-yr-old plants tended to have
higher proportion of leaf area missing, as did the plants that were not fertilized-both of
these effects were marginally significant (P < 0.10). The number of ants, included as a
plant-level covariate in the model, significantly affected herbivory (Table 3-4). Linear
regression showed only a weak but significant negative relationship between ant number
and herbivory (R2 = 0.24, P < 0.0001), but the mixed model analysis showed that ant
presence and abundance were clearly important when plant age and fertilization were
included in the model. Although the effects of age and fertilization were not statistically
significant at the a = 0.05 level, examination of the data suggests that the 1-yr-old plants
experienced more proportional leaf damage than the 5-yr-old plants and that fertilization
reduced herbivory for the younger plants (Fig. 3-4). When the two ages were analyzed
separately, fertilization did not significantly affect the proportion of leaf area missing for
either, but the number of ants had a significant negative effect on herbivory for the 5-yr-
old plants (Table 3-5).
Leaf Palatability Bioassay
The results of the ANOVA for the leaf palatability trial showed that fertilization
significantly affected the amount of leaf area consumed by C. leprosa, whereas leaf age
had only a marginally significant effect and there was no effect of plant age (Table 3-6).
Overall, the beetles consumed more material from young leaves than from mature leaves,
and more from the leaves of unfertilized plants than from fertilized plants. However,
these main effects were mainly due to the high leaf area consumed from young,
unfertilized leaves, as indicated by the marginally significant interaction effect between
fertilization treatment and leaf age (Fig. 3-6). None of the other interactions among
factors were statistically significant.
The relationships between different types of plant defenses are often complex and
may vary over the ontogeny of the plant and according to resource availability and the
efficacy of ant defense (Folgarait and Davidson 1994, 1995, Nomura et al. 2001, Del Val
and Dirzo 2003, Dyer et al. 2004). In this study, I investigated the patterns of herbivory
on Cordia alliodora, a common myrmecophytic tree, in relation to plant age, fertilization
and ant abundance. Identifying variation in the benefits and costs associated with
different partners in mutualist guilds and determining the consequences of such variation
are major objectives of holistic examination of mutualism (Stanton 2003). The presence
of multiple ant species in C. alliodora plants, which appear to vary in their defensive
behavior (Tillberg 2004), provided an ideal system in which to examine the effects of
variation in partner quality in this system. Although I was unable to test the effects of
different ant species on herbivory directly due to nonrandom distribution of the ants in
the plants I sampled, I was able to infer that at least the two most abundant species,
Aztecapittieri and Crematogaster carinata, appeared to defend the plant from insect
herbivores. Future work on this system could include explicit tests of interspecific
differences in the amount of leaf damage allowed by the different ant occupants.
The number of worker ants had no effect on the proportional leaf damage of the 1-
yr-old plants, as might be expected given the very small number of individuals present in
domatia collected from these young plants (Table 3-5). Young individuals of many
tropical plants display a strategy of tolerance to herbivory rather than investment in anti-
herbivore defenses (Strauss and Agrawal 1999, Coley et al. 2005), and this may be the
case for C. alliodora. Del Val and Dirzo (2003) showed that leaves from young
Cecropiapeltata, another fast-growing myrmecophyte, were also more palatable to
herbivores than leaves from older plants due to reduced investment in anti-herbivore
defenses. However, despite high levels of leaf damage in the field, in controlled
environments the beetle C. leprosa consumed marginally significantly less area from
leaves taken from fertilized plants, particularly in the 1-yr-old plants (Table 3-6). This
result suggests that even the 1-yr-old C. alliodora plants also produced chemical defenses
that effectively deter at least one species of specialist herbivore. Whether these results
extend to other herbivores is unknown.
By contrast, 5-yr-old trees appeared to rely on ants for anti-herbivore defense, with
ant presence and the number of workers present in focal domatia significantly reducing
the proportion of leaf damage in the immediate vicinity (Table 3-5). Due to the non-
random distribution of the four ant species in this study, it was impossible to analyze the
effects of each on leaf damage. However, since the number of workers was significantly
negatively related to the proportional leaf damage, the variation among the ant species in
the number of workers present is suggestive of interspecific differences in plant defense
(Fig. 3-3). Aztecapittieri and Crematogaster carinata were the most abundant ant
species in this study (together accounting for 69 of the 90 occupied domatia) and had the
most workers per domatium (Figure 3-3), suggesting that they effectively reduced
herbivore damage. This result is in accordance with behavioral and stable isotopic
studies demonstrating that A. pittieri and C. carinata attack and consume insect
herbivores on C. alliodora plants (Moser 2000, Tillberg 2004). Aztecapittieri and related
Azteca species are the most abundant ants in C. alliodora throughout the range of the
plant (Wheeler 1942, Longino 1996), and the protective effect of these specialist species
may be a common result of occupation by these ants when they inhabit C. alliodora.
Conversely, C. setulifer, which is another specialist inhabitant of Cordia, was only
present in small numbers and appears to be at best a passive defender of C. alliodora
(Tillberg 2004). Because it is rarely the dominant ant species on older trees and appears
to be out-competed by A. pittieri on plants where the two species co-occur (Chapter 2,
this thesis), whether or not C. setulifer effectively defends against insect herbivores may
not affect plant performance at the individual or population level. Interestingly, C.
carinata, which is only an opportunistic occupant of C. alliodora domatia, appears to
confer benefits to plants through reduction of herbivore damage. This species is not
commonly found in C. alliodora, however, and therefore this ant species probably has
little impact on the C. alliodora-ant relationship at the population level or across the
range of the plant.
The results of the experimental leaf palatability trial with the specialist beetle
herbivore, Coptocycla leprosa, corroborated those from the survey of natural herbivory in
some ways but also differed on several important points. In the field, leaves from 1-yr-
old plants had higher proportional leaf damage than those from 5-yr-old plants, but in the
laboratory there was no difference between the two plant ages in leaf area consumed.
Although the insects responsible for the leaf damage in the field are unknown, this
difference could support the finding that ants limit C. leprosa damage in 5-yr-old trees
because it was so abundant at the site. Whereas in the laboratory trial there was less
herbivory on leaves from fertilized plants regardless of plant age, in the field the
fertilization effect was only marginally significant and appeared to have no effect at all
for the 5-yr-old plants. The contrast between the results of the leaf palatability trial and
the observed patterns of leaf damage in the field suggest that the beetles do not base their
foraging choices in nature solely on preference for the most palatable leaves. Rather,
they are most effective at attacking young plants that do not house large colonies of
Fertilization significantly reduced the leaf area consumed in the palatability trial
and reduced herbivory of 1-yr-old, but not 5-yr-old, plants in the field survey of
herbivory. Together, these findings suggest that nutrient augmentation increases the
production of defensive chemicals in C. alliodora, and that the relative importance of
these chemicals for anti-herbivore defense varies according to plant age. Even if the
chemical defenses of C. alliodora do not contain nitrogen, other work has shown that N
fertilization can increase the production of carbon-based defenses (Mihaliak and Lincoln
1985, Wilkens et al. 1996). Relatively little is known of the secondary metabolites and
anti-herbivore chemical defenses produced by C. alliodora. Chen et al. (1983) described
several triterpenoid compounds isolated from the leaves of C. alliodora that repelled
leafcutter ants in experimental trials. Gomez et al. (1999) found lower terpenoids in the
leaves of a congener, Cordia curassavica, but did not find these compounds in C.
alliodora. Additionally, a number of secondary metabolic compounds have been isolated
from the bark and wood of C. alliodora, including some with fungicidal or insecticidal
properties (Moir and Thomson 1973, Stevens et al. 1973, Manners and Jurd 1977, closet et
al. 2000, Vanisree et al. 2002). However, it is unclear what role these chemicals or
related compounds may have in plant defense against leaf herbivores. Identifying
defensive chemicals from C. alliodora leaves, assessing their effects on herbivory, and
determining how their production varies with plant age and environmental factors are
critical areas of study for fully understanding the defensive strategy of this species.
The evolution and maintenance of ant-plant mutualisms is dependent upon net
fitness benefits at the population level for both the ant and plant partners (Bronstein 1998,
Heil and McKey 2003). The primary benefit for the plants is usually protection from
herbivory, which often has negative short- and long-term fitness consequences (Marquis
1984, Ernest 1989, Doak 1992, Coley and Barone 1996). In this study, I found that the
ants reduced herbivore damage on the 5-yr-old plants but not on the 1-yr-old plants. The
increased nutrient availability marginally reduced herbivore damage overall but, at least
in the field, had no effect on herbivory of older plants (Table 3-4). However, there was
no evidence of distinct trade-offs in defensive mechanisms over plant ontogeny in this
system. Instead, it appeared that the reduction of herbivory resulting from the ant
occupants was additive, and the palatability trial suggested that whatever chemical
defenses C. alliodora produces were present in plants of both ages studied. Therefore, in
the C. alliodora-ant system, and probably other ant-plant relationships, plant growth and
the production of domatia to house mutualist ants likely represent important investments
in anti-herbivore defense for the plant over both evolutionary and ecological time scales
(Brouat and McKey 2000). This investment may not produce immediate benefits
because the ants require time to colonize the plant and produce workers. Therefore,
young plants may produce defensive chemicals if adequate resources are available or may
simply invest in rapid growth that minimizes the effects of herbivory and promotes future
Table 3-1. Results of mixed model analysis testing the effects of plant age and
fertilization on the number of leaves surrounding focal domatia.
Source of variation num. df den. df F P
Plant age 1 6 3.46 0.11
Fertilization 1 6 1.31 0.30
Table 3-2. Results of mixed model analysis testing the effects of plant age and
fertilization on the total area of leaves surrounding focal domatia.
Source of variation num. df den. df F P
Plant age 1 6 5.72 0.054
Fertilization 1 6 1.88 0.22
Table 3-3. Results of mixed model analysis testing the effects of plant age and
fertilization on the number of worker ants within the focal domatia.
Source of variation num. df den. df F P
Plant age 1 6 24.35 0.0026
Fertilization 1 6 0.17 0.70
Table 3-4. Results of mixed model analysis testing the effects of plant age, fertilization,
and the number of worker ants on the proportion of leaf area missing. The
number of ants was logio-transformed and the proportion of leaf area missing
was logit-transformed to improve the distribution for this analysis.
Source of variation
1 103 7.54
Table 3-5. Results of mixed model analysis testing the effects of fertilization and the
number of worker ants present on herbivory for 1-yr-old plants and 5-yr-old
plants. The number of ants was loglO-transformed and the proportion of leaf
area missing was logit-transformed to improve the distribution for this
analysis. Although the results are presented together, the analyses were
conducted separately for the two ages.
Plant age Source of variation num. df
Table 3-6. Results of ANOVA testing the effects of plant age, leaf age and fertilization
treatment on the leaf area consumed by one Coptocycla leprosa beetle in 24
hr, with trial as a random block effect.
Source of variation df MS F P
Plant age 1 6.25 2.03 0.16
Leafage 1 10.99 3.58 0.061
Fertilization 1 18.41 5.99 0.016
Plant age x Leaf age 1 0.35 0.12 0.74
Plant age x Fertilization 1 5.40 1.76 0.19
Leaf age x Fertilization 1 11.38 3.70 0.056
Plant age x Leaf age x
Figure 3-1. The beetle Coptocycla leprosa spends its entire life cycle on Cordia
alliodora. (A) Late-instar larva with fecal shield, (B) pupa adhering to top of a
leaf and (C) adult on underside of leaf.
50- M Fertilized
"i 40- o
c 30- -
Plant age (years)
Figure 3-2. The number of worker ants present in focal domatia varied with plant age but
fertilization treatment had no effect. Boxplots show inter-quartile ranges and
expected minimum and maximum values, with values beyond the 95% CI
indicated by open circles. All of the high outliers in the 1-yr-old plants were
domatia occupied by Crematogaster carinata.
[ 1-yr-old plants
E 5-yr-old plants
I I I I
A. pittieri C. setulifer C. carinata P. fortis
Figure 3-3. The number of worker ants present in focal domatia varied according to plant
age and the identity of the ant species. Boxplots show inter-quartile ranges
and expected minimum and maximum values, with values beyond the 95% CI
indicated by open circles. Pseudomyrmexfortis was not present in any
domatia from 1-yr-old plants.
Plant age (years)
Figure 3-4. The proportion of leaf area missing was affected at P < 0.10 by plant age and
fertilization treatment. In general 1-yr-old plants experienced more
proportional leaf damage than 5-yr-old plants, and for the 1-yr-old plants
fertilization reduced herbivore damage. Boxplots show inter-quartile ranges
and expected minimum and maximum values, with values beyond the 95% CI
indicated by open circles.
0 0 Young leaves
SE3 Mature leaves
N 8.0- o
S 6.0- 0
Figure 3-5. Fertilization significantly reduced the leaf area consumed by individual
Coptocycla leprosa beetles in the leaf palatability trials. This was particularly
true for young leaves, as indicated by the marginally significant (P = 0.056)
interactive effect between fertilization and leaf age (Table 3-6).
The overall objective of this study was to document the patterns of ant occupation
in C. alliodora and to determine the effects of this variation on the amount of herbivore
damage as the host plants aged and grew. Addressing this question first required a
detailed examination of patterns of coexistence among the ant species that inhabit C.
alliodora and discussion of the mechanisms that could account for these patterns. I then
tested the effects of ant presence and abundance, in addition to the influence of plant age
and fertilization, on the proportional damage of leaves surrounding focal domatia. I
complemented the survey of herbivory with a palatability trial using a specialist herbivore
of C. alliodora to determine ant-free preferences for leaves from plants differing in age
The two most abundant ant species I found inhabiting C. alliodora, A. pittieri and
Cr. carinata, were also present in the highest numbers in the individual domatia sampled
in the field survey of herbivory. Because the number of worker ants had a significant
negative effect on proportional leaf damage, I suggest that both of these species act as
mutualists and benefit the host plant by reducing herbivory. The other ant species found
within C. alliodora domatia were generally subdominant at the whole-tree level and had
fewer workers per domatium. However, because A. pittieri and closely related species
are ubiquitous inhabitants of C. alliodora throughout its range (Longino 1996), at the
population level it is unlikely that that less common species are important for the
maintenance of the mutualism.
LIST OF REFERENCES
Alonso, L.E. 1998. Spatial and temporal variation in the ant occupants of a facultative
ant-plant. Biotropica 30:201-213.
Amarasekare, P. 2003. Competitive coexistence in spatially structured environments: a
synthesis. Ecology Letters 6:1109-1122.
Beattie, A.J. 1985. The evolutionary ecology of ant-plant mutualisms. Cambridge
University Press, Cambridge, UK.
Bluthgen, N., N.E. Stork and K. Fiedler. 2004. Bottom-up control and co-occurrence in
complex communities: honeydew and nectar determine a rainforest ant mosaic.
Bowers, M.D. and N.E. Stamp. 1993. Effects of plant age, genotype, and herbivory on
Plantago performance and chemistry. Ecology 74:1778-1791.
Bronstein, J.L. 1998. The contribution of ant-plant protection studies to our
understanding of mutualism. Biotropica 30:150-161.
Brouat, C. and D. McKey. 2000. Origin of culinary ant-domatia and timing of their
onset in plant ontogeny: evolution of a key trait in horizontally-transmitted ant-
plant symbioses. Biological Journal of the Linnean Society 71:809-819.
Brown, B.J. and T.F.H. Allen. 1989. The importance of scale in evaluating herbivory
impacts. Oikos 54:189-194.
Brues, C.T. 1922. Conoaxima, a new genus of the hymenopteran family Eurytomidae,
with a description of its larva and pupa. Psyche 29:153-158.
Bruna, E. M., D. M. Lapola and H. L. Vasconcelos. 2004. Interspecific variation in
defensive responses of obligate ant-plant inhabitants: are there consequences for
host plants? Oecologia 238:558-565
Bruna, E.M., H.L. Vasconcelos and S. Heredia. 2005. The effect of habitat
fragmentations on communities of mutualists: Amazon ants and their host plants.
Biological Conservation 124:209-216.
Bryant, J.P., P.B. Reichardt and T.B. Clausen. 1992. Chemically mediated interactions
between woody plants and browsing mammals. Journal of Range Management
Chen, T.K., D.C. Ales, N.C. Baenziger and D.F. Wiemer. 1983. Ant-repellent
triterpenoids from Cordia alliodora. Journal of Organic Chemistry 48:3525-3531.
Chesson, P. 2000. Mechanisms of maintenance of species diversity. Annual Review of
Ecology and Systematics 31:343-366.
Cole, B.J. 1983. Assembly of mangrove and communities: colonization abilities. Journal
of Animal Ecology 52:349-355.
Coley, P.D. and T.M. Aide. 1991. Comparison of herbivory and plant defenses in
temperate and tropical broad-leaved forests. Pages 25-49 in P.W. Price, T.M.
Lewinsohn, W.W. Fernandes and W.W. Benson, eds. Plant-animal interactions:
evolutionary ecology in tropical and temperate regions. John Wiley and Sons,
Coley, P.D. and J.A. Barone. 1996. Herbivory and plant defenses in tropical forests.
Annual Review of Ecology and Systematics 27:305-335.
Coley, P.D., J.P. Bryant and F.S. Chapin, III. 1985. Resource availability and plant
antiherbivore defense. Science 230:895-899.
Coley, P.D., J. Lokvam, K. Rudolph, K. Bromberg, T.E. Sackett, L. Wright, T. Brenes-
Arguedas, D. Dvorett, S. Ring, A. Clark, C. Baptiste, R.T. Pennington and T.A.
Kursar. 2005. Divergent defensive strategies of young leaves in two species of
Inga. Ecology 86:2633-2643.
Davidson, D.W. 1980. Some consequences of diffuse competition in a desert ant
community. American Naturalist 116:92-105.
Davidson, D.W. 1998. Resource discovery versus resource domination in ants: a
functional mechanism for breaking the trade-off. Ecological Entomology 23:484-
Davidson, D.W. and D. McKey. 1993. The evolutionary ecology of symbiotic ant-plant
relationships. Journal of Hymenoptera Research 2:13-83.
Davidson, D.W., R.R. Snelling and J.T. Longino. 1989. Competition among ants for
myrmecophytes and the significance of plant trichomes. Biotropica 21:64-73.
Dejean, A., A. Quilichini, J.H.C. Delabie, J. Orivel, B. Corbara and M. Gibernau. 2004.
Influence of its associated ant species on the life history of the myrmecophyte
Cordia nodosa in French Guiana. Journal of Tropical Ecology 20:701-704.
Del Val, E. and R. Dirzo. 2003. Does ontogeny cause changes in the defensive strategies
of the myrmecophyte Cecropiapeltata? Plant Ecology 169:35-41.
Doak, D.F. 1992. Lifetime impacts of herbivory for a perennial plant. Ecology 73:2086-
Dyer, L.A., C.D. Dodson, J. Beihoffer and D.K. Letourneau. 2001. Trade-offs in
antiherbivore defenses in Piper cenocladum: ant mutualists versus plant
secondary metabolites. Journal of Chemical Ecology 27:581-592.
Dyer, L.A., D.K. Letourneau, C.D. Dodson, M.A. Tobler, J.O. Stireman III and A. Hsu.
2004. Ecological causes and consequences of variation in defensive chemistry of
a neotropical shrub. Ecology 85:2795-2803.
Ernest, K.A. 1989. Insect herbivory on a tropical understory tree: effects of leaf age and
habitat. Biotropica 21:194-199.
Faveri, S.B. and H.L. Vasconcelos. 2004. The Azteca-Cecropia association: are ants
always necessary for their host plants? Biotropica 36:641-646.
Feeny, P. 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of spring
feeding by winter moth caterpillars. Ecology 51:565-581.
Feldhaar, H., B. Fiala, R.b. Hashim and U. Maschwitz. 2003. Patterns of the
Crematogaster-Macaranga association: the ant partner makes the difference.
Insectes Sociaux 50:9-19.
Fellers, J. 1987. Interference and exploitation in a guild of woodland ants. Ecology
Flowers, R.W. and D.H. Janzen. 1997. Feeding records of Costa Rican leaf beetles
(Coleoptera: Chrysomelidae). Florida Entomologist 80:334-366.
Folgarait, P.J. and D.W. Davidson. 1994. Antiherbivore defenses of myrmecophytic
Cecropia under different light regimes. Oikos 71:305-320.
Folgarait, P.J. and D.W. Davidson. 1995. Myrmecophytic Cecropia: antiherbivore
defenses under different nutrient treatments. Oecologia 104:189-206.
Fonseca, C.R. 1999. Amazonian ant-plant interactions and the nesting space limitation
hypothesis. Journal of Tropical Ecology 15:807-825.
Fonseca, C.R. and G. Ganade. 1996. Asymmetries, compartments and null interactions in
an Amazonian ant-plant community. Journal of Animal Ecology 65:339-347.
Gaume, L., D. McKey and S. Terrin. 1998. Ant-plant-homopteran mutualism: how the
third partner affects the interaction between a plant-specialist ant and its
myrmecophytic host. Proceedings of the Royal Society of London Series B-
Biological Sciences 265:569-575.
Haggar, J.P. and J.J. Ewel. 1995. Establishment, resource acquisition, and early
productivity as determined by biomass allocation patterns of three tropical tree
species. Forest Science 41:689-708.
Heil, M., B. Fiala, U. Maschwitz and K.E. Linsenmair. 2001a. On benefits of indirect
defence: short- and long-term studies of antiherbivore protection via mutualistic
ants. Oecologia 126:395-403.
Heil, M., B. Fiala, K.E. Linsenmair, G. Zotz, P. Menke and U. Maschwitz. 1997. Food
body production in Macaranga triloba (Euphorbiaceae): a plant investment in
anti-herbivore defence via symbiotic ant partners. Journal of Ecology 85:847-861.
Heil, M., A. Hilpert, B. Fiala and K.E. Linsenmair. 2001b. Nutrient availability and
indirect (biotic) defence in a Malaysian ant-plant. Oecologia 126:404-408.
Heil, M. and D. McKey. 2003. Protective ant-plant interactions as model systems in
ecological and evolutionary research. Annual Review of Ecology, Evolution and
Herms, D.A. and W.J. Mattson. 1992. The dilemma of plants: to grow or defend.
Quarterly Review of Biology 67:283-335.
Holldobler, B. and C.J. Lumsden. 1980. Territorial strategies in ants. Science 210:732-
Holldobler, B. and E.O. Wilson. 1990. The ants. Belknap Press, Cambridge.
Holway, D.A. 1999. Competitive mechanisms underlying the displacement of native ants
by the invasive Argentine ant. Ecology 80:238-251.
Itino, T. and T. Itioka. 2001. Interspecific variation and ontogenetic change in
antiherbivore defense in myrmecophytic Macaranga species. Ecology Research
Ioset, J.R., A. Marston, M.P. Gupta and K. Hostettmann. 2000. Antifungal and larvicidal
compounds from the root bark of Cordia alliodora. Journal of Natural Products
Itioka, T., M. Nomura, Y. Inui, T. Itino and T. Inoue. 2000. Difference in intensity of ant
defense among three species ofMacaranga myrmecophytes in a Southeast Asian
dipterocarp forest. Biotropica 32:318-326.
Janzen, D.H. 1966. Coevolution of mutualism between ants and acacias in Central
America. Evolution 20:249-275.
Janzen, D.H. 1975. Pseudomyrmex nigripilosa: a parasite of a mutualism. Science
Koptur, S. 1985. Alternative defenses against herbivores in Inga (Fabaceae: Mimosideae)
over an elevational gradient. Ecology 66:1639-1650.
Lapola, D.M., E.M. Bruna, C. Granara de Willink and H.L. Vasconcelos. 2005. Ant-
tended Hemiptera in Amazonian myrmecophytes: patterns of abundance and
implications for mutualism function (Hymenoptera: Formicidae). Sociobiology
Longino, J.T. 1989. Geographic variation and community structure in an ant-plant
mutualism: Azteca and Cecropia in Costa Rica. Biotropica 21:126-132.
Longino, J.T. 1991. Azteca ants in Cecropia trees: taxonomy, colony structure, and
behavior. In: C.R. Huxley and D.F. Cutter eds. Ant-plant interactions. Oxford
University Press, Oxford. Pp 271-288.
Longino, J.T. 1996. Taxonomic characterization of some live-stem inhabiting Azteca
(Hymenoptera: Formicidae) in Costa Rica, with special reference to the ants of
Cordia (Boraginaceae) and Triplaris (Polygonaceae). Journal of Hymenoptera
Lowman, M.D. 1984. An assessment of techniques for measuring herbivory: is rainforest
defoliation more intense than we thought? Biotropica 16:264-268.
Manners, G.D. and L. Jurd. 1977. The hydroquinone terpenoids of Cordia alliodora.
Journal of the Chemical Society, Perkins Transactions I 4:405-410.
Marquis, R.J. 1984. Leaf herbivores decrease fitness of a tropical plant. Science 226:537-
McGlynn, T.M. 2000. Do Lanchester's laws of combat describe competition in ants?
Behavioral Ecology 11:686-690.
McKey, D. 1974. Adaptive patterns in alkaloid physiology. American Naturalist 108:305-
McKey, D. 1984. Interaction of the ant-plant Leonardoxa africana (Caesalpiniaceae)
with its obligate inhabitants in a rainforest in Cameroon. Biotropica 16:81-99.
Michelangeli, F.A. 2003. Ant protection against herbivory in three species of Tococa
(Melastomataceae) occupying different environments. Biotropica 35:181-188.
Mihaliak, C.A. and D.E. Lincoln. 1989. Plant biomass partitioning and chemical defense:
response to defoliation and nitrate limitation. Oecologia 80:122-126.
Moir, M. and R.H. Thomson. 1973. Naturally occurring quinines. Part XXII. Terpenoid
in Cordia spp. Journal of the Chemical Society, Perkins Transactions I 13:1352-
Moser, J. 2000. Interaktionen zwischen Herbivoren, Ameisen und Parasitoiden auf
Cordia alliodora (Boraginacea). Diplomarbeit Universitat Hannover.
Nichols-Orians, C.M. 1991. Environmentally induced differences in plant traits:
consequences for susceptibility to a leaf-cutter ant. Ecology 72:1609-1623.
Nomura, M., T. Itioka and T. Itino. 2000. Variations in abiotic defense within
myrmecophytic and non-myrmecophytic species ofMacaranga in a Bornean
dipterocarp forest. Ecological Research 15:1-11.
Nomura, M., T. Itioka and K. Murase. 2001. Non-ant antiherbivore defenses before plant-
ant colonization in Macaranga myrmecophytes. Population Ecology 43:207-212.
O'Neal, M.E., D.A. Landis and R. Isaacs. 2002. An inexpensive, accurate method for
measuring leaf area and defoliation through digital image analysis. Journal of
Economic Entomology 95:1190-1194.
Pacala, S.W. and M.J. Crawley. Herbivores and plant diversity. American Naturalist
Palmer, T.M. 2003. Spatial heterogeneity influences competition and coexistence in an
African acacia ant guild. Ecology 84:2843-2855.
Palmer, T.M. 2004. Wars of attrition: colony size determines competitive outcomes in a
guild of Acacia ants. Animal Behavior 68:993-1004.
Palmer, T.M., M.L. Stanton and T.P. Young. 2003. Competition and coexistence:
exploring mechanisms that restrict diversity within mutualist guilds. American
Palmer, T.M., T.P. Young and M.L. Stanton. 2002. Burning bridges: priority effects and
the persistence of a competitively subordinant acacia-ant in Laikipia, Kenya.
Palmer, T.M., T.P. Young, M.L. Stanton and E. Wenk. 2000. Short-term dynamics of an
acacia ant community in Laikipia, Kenya. Oecologia 123:425-435.
Pinhiero, J.C. and D.M. Bates. 2000. Mixed-effects models in S and S-PLUS. Springer-
Verlag, New York.
Rojas, L., C. Godoy, P. Hanson and L. Hilje. 2001. A survey of homopteran species
(Auchenorrhyncha) from coffee shrubs and poro and laurel trees in shaded coffee
plantations, in Turrialba, Costa Rica. Revista de Biologia Tropical 49:1057-1065.
Sagers, C.L. and P.D. Coley. 1995. Benefits and costs of defense in a neotropical shrub.
Schupp, E.W. 1986. Azteca protection of Cecropia: ant occupation benefits juvenile trees.
Stamp, N. 2003. Out of the quagmire of plant defense hypotheses. Quarterly Review of
Stanton, M.L., T.M. Palmer and T.P. Young. 2002. Competition-colonization trade-offs
in a guild of African acacia-ants. Ecological Monographs 72:347-363.
Stevens, K.L., L. Jurd and G. Manners. 1973. Alliodorin, a phenolic terpenoid from
Cordia alliodora. Tetrahedron Letters 31:2955-2958.
Strauss, S.Y. and A.A. Agrawal. 1999. The ecology and evolution of plant tolerance to
herbivory. Trends in Ecology and Evolution 14:179-185.
Tillberg, C.V. 2003. The evolution and ecology of an ant-plant symbiosis. Ph.D.
Dissertation, University of Colorado, Boulder, CO.
Tillberg, C.V. 2004. Friend or foe? A behavioral and stable isotopic investigation of an
ant-plant symbiosis. Oecologia 140:506-515.
Tokeshi, M. 1999. Species coexistence: ecological and evolutionary perspectives.
Blackwell Scientific, Oxford, UK.
Vasconcelos, H.L. 1993. Ant colonization ofMaieta guianensis seedlings, an Amazonian
ant-plant. Oecologia 95:439-443.
Vasconcelos, H.C. and A.B. Casimiro. 1997. Influence ofAzteca alfari ants on the
exploitation of Cecropia trees by a leaf-cutting ant. Biotropica 29:84-92.
Vasconcelos, H.L. and D.W. Davidson. 2000. Relationship between plant size and ant
associates in two Amazonian ant-plants. Biotropica 32:100-111.
Vanisree, M., J. Kavitha and G.V. Subbaraju. 2002. Synthesis of methyl 3-(2,4,5-
trimethoxyphenyl)-propionate, an antifungal and larvicidal constituent of Cordia
alliodora. Asian Journal of Chemistry 14:534-536.
Wheeler, W.M. 1942. Studies of Neotropical ant-plants and their ants. Bulletin of the
Museum of Comparative Zoology 90:1-262.
Wilkens, R.T., J.M. Spoerke and N.E. Stamp. 1996. Differential responses of growth and
two soluble phenolics of tomato to resource availability. Ecology 77:247-258.
Worthen, W.B. 1989. Predator-mediated coexistence in laboratory communities of
mycophagous Drosophila (Diptera: Drosophilidae). Ecological Entomology
Young, T.P., C.H. Stubblefield and L.A. Isbell. 1997. Ants on swollen-thorn acacias:
species coexistence in a simple system. Oecologia 109:98-107.
Yu, D.W. and H.B. Wilson. 2001. The competition-colonization trade-off is dead; long
live the competition-colonization trade-off. American Naturalist 49-63.
Yu, D.W., H.B. Wilson, M.E. Frederickson, W. Palomino, R. de la Colina, D.P. Edwards
and A.A. Balareso. 2004. Experimental demonstration of species coexistence
enabled by dispersal limitation. Journal of Animal Ecology 73:1102-1114.
Yu, D. W., H. B. Wilson, N. E. Pierce. 2001. An empirical model of species coexistence
in a spatially structured environment. Ecology 82: 1761-1771.
Yu, D.W. and D.W. Davidson. 1997. Experimental studies of species-specificity in
Cecropia-ant relationships. Ecological Monographs 67:273-294.
Zangerl, A.R. and C.E. Rutledge. 1996. The probability of attack and patterns of
constitutive and induced defense: a test of optimal defense theory. American
Matthew David Trager was born in Gainesville, Florida, on April 9th, 1980, to Kim
A. Trager and James C. Trager. While attending Westridge Elementary School in
Ballwin, Missouri, he won third place in his school science fair with an insect collection
comprising specimens he caught and curated. Matthew attended Grinnell College in
Grinnell, Iowa, where he earned a Bachelor of Arts degree in anthropology in 2002.
While in college he spent his summers restoring prairies for The Nature Conservancy in
Iowa, lobbying for The Wilderness Society's public policy department in Washington,
DC, and conducting research on grassland plant diversity at Kansas State University's
Konza Prairie. He also spent the fall of 2000 in Tanzania, taking classes at the University
of Dar es Salaam and conducting research in Serengeti National Park. Following his
undergraduate education, Matthew worked in the plant ecology lab at Archbold
Biological Station in south-central Florida for a year where he participated in long- and
short term ecological studies of several federally listed plant species. In the fall of 2003,
Matthew joined Dr. Emilio Bruna's lab at the University of Florida to pursue graduate
studies through the School of Natural Resources and Environment. He received his
Master of Science degree in Interdisciplinary Ecology in December, 2005.