Inter- and intraspecific comparisons of antiherbivore defenses in three species of rainforest understory shrubs

Article abstract of DOI:10.1007/s10886-008-9432-4

Plants defend themselves against herbivores and pathogens with a suite of morphological, phenological, biochemical, and biotic defenses, each of which is presumably costly. The best studied are allocation costs that involve trade-offs in investment of resources to defense versus other plant functions. Decreases in growth or reproductive effort are the costs most often associated with antiherbivore defenses, but trade-offs among different defenses may also occur within a single plant species. We examined trade-offs among defenses in closely related tropical rain forest shrubs (Piper cenocladum, P. imperiale, and P. melanocladum) that possess different combinations of three types of defense: ant mutualists, secondary compounds, and leaf toughness. We also examined the effectiveness of different defenses and suites of defenses against the most abundant generalist and specialist Piper herbivores. For all species examined, leaf toughness was the most effective defense, with the toughest species, P. melanocladum, receiving the lowest incidence of total herbivory, and the least tough species, P. imperiale, receiving the highest incidence. Although variation in toughness within each species was substantial, there were no intraspecific relationships between toughness and herbivory. In other Piper studies, chemical and biotic defenses had strong intraspecific negative correlations with herbivory. A wide variety of defensive mechanisms was quantified in the three Piper species studied, ranging from low concentrations of chemical defenses in P. imperiale to a complex suite of defenses in P. cenocladum that includes ant mutualists, secondary metabolites, and moderate toughness. Ecological costs were evident for the array of defensive mechanisms within these Piper species, and the differences in defensive strategies among species may represent evolutionary trade-offs between costly defenses.

Full text of DOI:10.1007/s10886-008-9432-4

J Chem Ecol (2008) 34:558–574
DOI 10.1007/s10886-008-9432-4
Inter- and Intraspecific Comparisons of Antiherbivore
Defenses in Three Species of Rainforest Understory Shrubs
R. M. Fincher & L. A. Dyer & C. D. Dodson &
J. L. Richards & M. A. Tobler & J. Searcy & J. E. Mather &
A. J. Reid & J. S. Rolig & W. Pidcock
Received: 2 July 2007 /Revised: 24 November 2007 /Accepted: 9 December 2007 /Published online: 4 March 2008
#
Springer Science + Business Media, LLC 2008
Abstract Plants defend themselves against herbivores and
pathogens with a suite of morphological, phenological,
biochemical, and biotic defenses, each of which is
presumably costly. The best studied are allocation costs
that involve trade-offs in investment of resources to defense
versus other plant functions. Decreases in growth or
reproductive effort are the costs most often associated with
antiherbivore defenses, but trade-offs among different
defenses may also occur within a single plant species. We
examined trade-offs among defenses in closely related
tropical rain forest shrubs (Piper cenocladum, P. imperiale,
and P. melanocladum) that possess different combinations
of three types of defense: ant mutualists, secondary
compounds, and leaf toughness. We also examined the
effectiveness of different defenses and suites of defenses
against the most abundant generalist and specialist Piper
herbivores. For all species examined, leaf toughness was
the most effective defense, with the toughest species, P.
melanocladum, receiving the lowest incidence of total
herbivory, and the least tough species, P. imperiale,
receiving the highest incidence. Although variation in
toughness within each species was substantial, there were
no intraspecific relationships between toughness and
herbivory. In other Piper studies, chemical and biotic
defenses had strong intraspecific negative correlations with
herbivory. A wide variety of defensive mechanisms was
quantified in the three Piper species studied, ranging from
low concentrations of chemical defenses in P. imperiale to a
complex suite of defenses in P. cenocladum that includes
ant mutualists, secondary metabolites, and moderate tough-
ness. Ecological costs were evident for the array of
defensive mechanisms within these Piper species, and the
differences in defensive strategies among species may
represent evolutionary trade-offs between costly defenses.
.
.
Keywords Piper cenocladum Piper imperiale
.
.
.
Piper melanocladum Piperaceae Amides
.
.
Secondary metabolites Chemical defense
.
.
Plant-herbivore interactions Costa Rica Tropical rain forest
Introduction
In response to strong selective forces exerted by a broad
array of enemies, which include insect herbivores, verte-
brate herbivores, and pathogens (Coley et al. 1985), plants
employ a variety of morphological, biochemical, and biotic
defenses (Maxwell et al. 1972; Johnson 1975; Coley 1983;
Kursar and Coley 1992; Harborne 2001). Differences in
plant palatability that result from these defenses contribute
to substantial variation in the extent of herbivore damage to
plants (Coley 1983). However, each kind of defense is
potentially associated with allocation costs (Levin 1976;
Herms and Mattson 1992; Gershenzon 1994; Bergelson and
Purrington 1996; Koricheva 2002; Strauss et al. 2002).
Many studies have addressed intraspecific trade-offs
between defense and other plant metabolic functions
(Zangerl et al. 1997). Because most plants produce several
:
:
R. M. Fincher L. A. Dyer (*) M. A. Tobler
Department of Ecology and Evolutionary Biology,
Tulane University,
400 Boggs,
New Orleans, LA 70118, USA
e-mail: ldyer@tulane.edu
:
:
:
:
C. D. Dodson J. L. Richards J. Searcy J. E. Mather
:
:
A. J. Reid J. S. Rolig W. Pidcock
Department of Physical and Environmental Sciences,
Mesa State College,
1100 North Ave.,
Grand Junction, CO 81501, USA

J Chem Ecol (2008) 34:558–574
559
different defenses (reviewed in Romeo et al. 1996),
intraspecific trade-offs may occur between defenses (e.g.,
between chemical and biotic defenses or between two
different chemical defenses; Mattson et al. 1988; Steward
and Keeler 1988; but see Koricheva et al. 2004). In this
case, rather than reducing plant allocation of resources to
reproduction or growth, plants may express the costs of
heightened defense production by diverting resources from
one defense to another. These trade-offs among different
defenses are referred to as “ecological costs” (Heil 2002)
since reducing a given defense only incurs costs under the
appropriate ecological conditions.
Evolutionary trade-offs among defenses may also be
evident between closely related species because sympatric
plant species that differ in the type or quantity of their
defenses may be at a selective advantage when defending
themselves against specialist herbivores that have evolved
resistance (Cates and Rhoades 1977). Producing multiple
defenses may be selectively advantageous either because of
synergistic action, where the effect of a combination of
defenses exceeds the sum of the effect of all individual
defenses (e.g., Berenbaum and Neal 1985; Berenbaum et al.
1991; Calcagno et al. 2002; Dyer et al. 2003), or because
multiple defenses may protect a plant against a diverse suite
of enemies (Hay et al. 1994; Lindroth and Hwang 1996;
Nelson and Kursar 1999; Dyer et al. 2003; Koricheva et al.
2004).
Ant–plant associations ranging from obligate mutualisms
to loose facultative relationships are common (Heil and
McKey 2003) and have long been used as model systems in
the study of plant defense (e.g., Janzen 1966; Rehr et al.
1973; Dyer et al. 2001). Allocation of resources to
production of domatia, food bodies, nectar, or other ant
rewards in these systems presumably incurs some costs
(Folgarait and Davidson 1995; Dyer et al. 2001; Heil et al.
2002). Leaf toughness and production of secondary
metabolites are two widely distributed defenses in tropical
plants (Coley and Barone 1996). Several studies have
shown that toughness is an effective defense (Feeny 1970;
Coley 1983; Sagers and Coley 1995; Coley and Barone
1996), but relatively few studies include toughness in
examinations of ecological costs. Plant secondary metabo-
lites are often effective in reducing herbivory (reviewed by
Harborne 2001) but can be metabolically costly to produce
(Gershenzon 1994; Zangerl et al. 1997; Dyer et al. 2001;
reviewed by Bergelson and Purrington 1996, and Strauss et
al. 2002), potentially necessitating trade-offs in resource
allocation. For example, allocation of limited plant resour-
ces may necessitate trade-offs between allocation of carbon
to ant rewards, construction of structural carbohydrates to
enhance toughness, and/or chemical defense. Similarly,
nitrogen may be allocated to ant rewards, nitrogen-based
chemical defenses, or the enzymes and physiological
machinery necessary to produce tougher leaves or higher
levels of chemical defense. Plant species or individuals that
invest in ant mutualisms often show reduced chemical
defense when compared to closely related species or
individual plants within a species that lack ant defenders
(Rehr et al. 1973; Dyer et al. 2001; Heil et al. 2002).
By using field experiments and surveys, we examined
relative efficacies of and trade-offs among three ecologi-
cally important tropical plant defenses: ant mutualists,
secondary metabolites, and leaf toughness. Three closely
related species in the genus Piper [P. imperiale (Miq.) C.
DC., P. cenocladum C. DC., and P. melanocladum C. DC.)
served as a model system. The phytochemistry of the
diverse tropical genus Piper is variable and well docu-
mented (reviewed by Parmar et al. 1997; Dyer et al. 2004a).
All species investigated to date (more than 112) produce
mixtures of secondary metabolites, and the compounds
discovered include alkaloids/amides/imides, lignans, neo-
lignans, terpenes, propenylphenols, steroids, kavapyrones,
chalcones, flavones, flavanones, and piperolides (Dyer et al.
2004a). The three Piper species utilized are closely related
(as part of the sect. Macrostachys clade, Tepe et al. 2004),
and share many species of generalist and specialist
herbivores (Marquis 1991; L. Dyer, personal observations).
They display a unique combination of ant mutualism,
chemical defense, and leaf toughness. Piper cenocladum is
defended by ant mutualists and nitrogen-containing chem-
icals, P. imperiale has a facultative relationship with ants,
but its chemistry has not previously been investigated, and
P. melanocladum has no relationship with ants, and its
chemistry was previously unknown. The relative leaf
toughness of all three has not previously been measured.
We tested for inter- and intra-specific trade-offs between
multiple putative plant defenses and their effects on
herbivory in order to address these two general questions:
How is intra- and interspecific variation in plant defense
associated with herbivory by specialists and generalists?
Are ecological costs of different types of defense evident
within and among these three Piper species?
We predicted that generalist herbivores would respond
more to chemical defenses, while small, specialists would
be deterred by ant defenders or leaf toughness (Dyer et al.
2004b). We expected to find trade-offs in allocation of
resources to toughness, ant associations, and chemical
defense within and among the three Piper species.
Materials and Methods
Study System Piper cenocladum is a tall (to 5 m) understory
plant with large, long-lived leaves (Letourneau and Dyer
1998) found in the lowland wet forest of Costa Rica
(Burger 1971). This species has defenses in the form of

560
J Chem Ecol (2008) 34:558–574
amide/imide secondary metabolites (Dodson et al. 2000)
and resident obligate ant mutualists (Risch and Rickson
1981; Risch 1982). Piper cenocladum leaves contain two
imides and one amide [Appendix, structures 1–3: piplartine
(1), 4′-desmethylpiplartine (2), and cenocladamide (3);
Dodson et al. 2000] in relatively high concentrations of
up to 3.8% dry weight (Dyer et al. 2004b). Piplartine is
cytotoxic in vitro (Duh et al. 1990), and all three
compounds in combination act synergistically to directly
and indirectly affect the fitness and feeding preferences of
generalist and specialist herbivores from a wide range of
taxa (Dyer et al. 2003). Pheidole bicornis Forel (Hyme-
noptera: Formicidae: Myrmicinae) ant colonies are housed
inside of P. cenocladum sheathing leaf bases (petioles), and
the plants produce amino acid and lipid rich opalescent
food bodies on the adaxial sides of the petioles (Risch and
Rickson 1981). These food bodies comprise the majority of
the ants’ diet (Fischer et al. 2002). In return, the ants
remove insect eggs, vines, and small particles from the
surface of the leaves, and kill small lepidopteran larvae and
stem boring weevils (Risch et al. 1977; Letourneau 1983;
Letourneau and Dyer 1998). When plants are not inhabited
by ants, the production of these food bodies declines, and
amide production increases threefold or greater (Dodson et
al. 2000; Dyer et al. 2001).
(Dyer et al. 2004c), individuals that we collected were
separated by at least 5 m. On each plant, we chose the first
fully mature leaf from the top of the plant for our
measurements to standardize leaf age. We measured leaf
toughness by using a modified penetrometer described by
Feeny (1970). This device measures leaf toughness as the
force, measured in grams of sand, needed to punch a 5-mm
steel rod through a taut leaf. Immediately after removing
the leaf from the plant, we took three toughness measure-
ments, two on one side of the midrib and one on the other
side. These measurements were made near the leaf tip, thus
avoiding all major veins. The mean of these three toughness
values was used in all analyses.
We used a translucent grid to measure the percent area
removed from each leaf by each of the following major
Piper herbivores: Atta cephalotes (leaf-cutter ants; Hyme-
noptera: Formicidae: Myrmicinae), katydids (Orthoptera:
Tettigoniidae), beetles (Coleoptera: Chrysomelidae and
Curculionidae), Quadrus cerealis caterpillars (Lepidoptera:
Hesperiidae), and at least two species of Eois caterpillars
(Lepidoptera: Geometridae), which were grouped together.
Each type of herbivore damage is easily discernible based
upon the pattern and shape of damage. Total percent
damage was calculated for each leaf as the sum of damage
by all herbivore types.
Piper imperiale is a large shrub or small tree to 6 m tall
found in moist, shaded forest areas (Burger 1971). This
species is characterized by large leaves and tubercles on
stems and leaves but in many ways is morphologically and
ecologically similar to P. cenocladum (Burger 1971).
Several species of ants facultatively inhabit the sheathing
leaf bases, but the plants produce no food bodies inside of
the domatia. Whether the ants benefit the plants in any way
is unknown.
Piper melanocladum is a small understory shrub to
1.6 m tall, with thick, glabrous, shiny lanceolate leaves
(Burger 1971). It has small, partially open leaf bases, and
has no demonstrated relationship with ants. We describe the
first characterization of the nitrogen-based secondary
metabolites of P. imperiale and P. melanocladum.
Imide Isolation and Quantification Imides were isolated
from P. imperiale and P. melanocladum, and their structures
were determined with the methods for P. cenocladum
described by Dodson et al. (2000). All imide structures
were confirmed by synthesis (Appendix). A randomly
selected subset of P. imperiale (N=30), P. melanocladum
(N=44), and P. cenocladum (N=34) leaves were air-dried
and analyzed to determine secondary chemical content.
Each leaf was extracted ×2 with ethanol overnight. The
resultant extract was resuspended in a 3:1 water/ethanol
mixture and exhaustively extracted with chloroform in a
separatory funnel. Combined extracts were dried, resus-
pended in chloroform, and analyzed by gas chromatogra-
phy. Standards for analysis were synthesized at Mesa State
College (Appendix). For a more detailed description of
extraction and GC methods, see Dodson et al. (2000) and
Dyer et al. (2001).
Field Survey This study took place in two lowland tropical
wet forests at Tirimbina Rainforest Center and Estacion
Biologica La Selva, Heredia Province, Costa Rica. La Selva
is located at 10°25′N 84°05′W at circa 100 m elevation on
the Caribbean slope. Tirimbina is located nearby at circa
210 m elevation, 10°24′4″ N, 84°6′29″ W.
Atta cephalotes Bioassays To determine the effectiveness of
the newly identified P. imides as defenses, we performed a
feeding choice experiment with Atta cephalotes colonies.
We modeled our experiment on those performed by
Folgarait et al. (1996). We presented foraging A. cephalotes
colonies with a selection of leaf fragments coated with
different compounds. Treatments applied to experimental
leaf fragments included piplaroxide (5), the newly discov-
ered compounds (6) and (4), and a control that contained
We sampled P. imperiale (N=81), P. cenocladum (N=
84), and P. melanocladum (N=94) plants along trails at
Tirimbina and La Selva. Because P. imperiale and P.
cenocladum reproduce vegetatively through fragmentation

J Chem Ecol (2008) 34:558–574
561
only methanol (all imides were in methanol solution).
Treatments were applied to Hyeronima alchorneoides
Allemao (Euphorbiaceae) leaves (a species shown to be
palatable to A. cephalotes; Folgarait et al. 1996) and allowed
to dry. Randomly distributed A. cephalotes colonies were
simultaneously presented with five fragments of each of the
four treatments (for a total of 20 leaf fragments per trial) in a
cafeteria-style display. These trials continued for 15 min or
until all 5 fragments of one of the treatments had been
removed by the ants. It was repeated with 22 different
colonies (N=22).
normality for parametric statistics due to the low frequency
of occurrence of herbivory, we used two methods of
analyses to test for differences among Piper species in
herbivory by different herbivores. First, to examine quali-
tatively host choice by herbivores, we categorized leaves as
experiencing herbivory or escaping from herbivory, and we
performed chi-square tests for contingency tables to test for
differences among plant species in the frequency of escape
from herbivory. We then quantitatively examined the extent
of feeding by herbivores once feeding was initiated (i.e.,
excluding all plants lacking herbivory) by performing an
analysis of covariance (ANCOVA) with plant species as the
independent variable, herbivore damage as the dependent
variable, and toughness as a covariate, followed by Tukey’s
multiple range tests. For this analysis, we removed all zero
herbivory values from the data set and log transformed
damage and toughness values to meet assumptions of
residual normality and homogeneity of variance. The
covariate and the interactive term were insignificant in all
tests. In combination, these two analyses compare both the
frequency of herbivory and the amount of herbivore
damage among plant species. Both of these analyses were
repeated for all damage types (total herbivory, A. cepha-
lotes, tettigoniid, beetle, Eois, and Q. cerealis). To test for
interspecific differences in toughness and concentration of
amides/imides, we performed an analysis of variance
(ANOVA) followed by Tukey’s multiple range test.
We modified this A. cephalotes feeding trial to use P.
imperiale leaf ethanol extracts, which would contain most
secondary metabolites and other ethanol-soluble com-
pounds produced by the plant, and to force the ants to cut
fragments from a large leaf section, so that their mouthparts
would be exposed to any secondary metabolites during the
cutting process. We applied an extract of P. imperiale to
leaves of Hieronyma alchorneoides and compared the
amount of leaf material removed from leaves treated with
extract to control leaves. Fresh and dry mass measurements
from 144 strips of H. alchorneoides were used to create a
conversion factor for estimation of the dry mass of fresh
leaf strips used in feeding choice trials (dry mass=0.2512
(fresh mass) +0.0054, R2=0.83). Leaves were cut into
paired strips (two from each leaf) weighing 0.25 g fresh
weight, avoiding all major veins, and the leaf area for each
strip was measured with a LICOR leaf area meter (leaf area
mean SE=12.27 0.15 cm2). We cut a pair of treatment
and control leaf strips from the same leaf, applied P.
imperiale extract at high or low concentrations of 0.0095%
leaf dry weight compound 4 (Appendix) and 0.0051% leaf
dry weight compound 5 (Appendix) (N=42), or 0.0024%
leaf dry weight compound 4 (Appendix) and 0.0013% leaf
dry weight compound 5 (Appendix) (N=36; dissolved in
0.24 and 0.06 ml ethanol, respectively) to the treatment leaf
strip and a corresponding amount of ethanol to the control
leaf strip and allowed the extract to dry. These concen-
trations are an order of magnitude lower than those found in
P. imperiale leaves, and so should measure the lower limits
of ant abilities to detect and respond to secondary
metabolites. Leaves were placed in pairs (one control leaf
and one leaf treated with P. imperiale extract) in an actively
foraging column of leaf-cutter ants (modeled after the
methods of Folgarait et al. 1996). Each trial was allowed to
continue for 2 hr or until the ants had removed all of one
leaf strip. Leaf strips were large enough that ants cut small
(approximately 1 cm2) sections to carry away. After each
trial, final leaf weight and leaf area were measured.
We used logistic regression to examine intraspecific
relationships among chemical defense (with chemical
defense content used as a predictor variable), toughness,
and herbivory by different herbivores. One test was
performed for each type of herbivore (with the exception
of Q. cerealis and A. cephalotes, which were excluded
because of rarity) and for total percent herbivory. Damage
by each type of herbivore was classified as either present or
absent. We tested for evidence of intraspecific trade-offs
among different defenses by examining correlations be-
tween average leaf toughness and amide content for each
Piper species.
Due to intrinsic difficulties in analyzing cafeteria-style
feeding choice data (Lockwood 1998, and works cited
therein) and nonnormality of the data, we analyzed the first
A. cephalotes bioassay feeding preference data by using
log-linear models (Floyd 2001). Each leaf fragment
treatment was used as a variable in specified models. We
used the maximum likelihood method for parameter
estimation of linear models and Chi-square statistics for
hypothesis testing (Folgarait et al. 1996). Because we were
interested in testing specific hypotheses, we used nonhier-
archical models to test the significance of leaf treatments as
predictors of A. cephalotes preference. For the second A.
cephalotes bioassay feeding preference experiment that
Statistical Analyses Because we were not able to transform
the percent herbivory data to meet requirements of

562
J Chem Ecol (2008) 34:558–574
used leaf extracts, two paired t tests were used to test for
differences in the leaf area and mass removed from control
and P. imperiale leaf extract treated leaf strips, for each
concentration of extract.
Results
Secondary Metabolites Piper imperiale contains two
imides (Appendix, structures 4, 5). The first imide,
compound 4, is an analog of piplartine, a compound we
previously found in P. cenocladum. Compound 4 (5′-
desmethoxydihydropiplartine) has not been isolated previ-
ously from a species of Piper, and is accompanied by its
epoxide deviative, compound 5, which has been isolated
previously from P. tuberculatum by Capron and Wiemer
(1996) and is named piplaroxide. Total imide content
ranged from 0.0044% to 0.029% dry weight. P. imperiale
did not have alkaloids but contained at least five different
sesquiterpenes (Appendix, Experimental).
We isolated compounds 4 and 5 (piplaroxide; Appendix) as
well as the 4′-desmethyl analog of piplaroxide from P.
melanocladum. The latter compound is also new to the
genus Piper and has structure 6 (Appendix). These imides
are present at high levels (ranging from 0.016% to 0.40%
dry weight) in the leaves. No other defensive compounds were
detected in P. melanocladum (Appendix, Experimental).
Piper cenocladum total amide/imide content ranged from
0.17% to 1.068% leaf dry weight. This range of concen-
trations is consistent with ranges reported in other studies
with this species where synergy (Dyer et al. 2003) and
trade-offs (Dodson et al. 2000) have been demonstrated. No
other defensive compounds were detected in P. cenocladum
(Appendix, Experimental).
Fig. 1 Differences in: a mean leaf toughness ( 1 SE) and b mean
amide/imide content ( 1 SE) among three species of Piper shrubs (P.
cenocladum: N=84, P. imperiale: N=84, P. melanocladum: N=94).
Data were analyzed using analysis of variance followed by Tukey’s
multiple range test; different letters above Piper species indicate
significant differences
Interspecific Differences The three Piper species were
different in their leaf toughness (F2, 256=168.2, P<0.001;
Fig. 1a). Piper melanocladum is the toughest species with a
penetrometer value of 559.3 13.9 g, nearly twice as tough
as the least tough P. imperiale (296.6 6.7 g). The
toughness of P. cenocladum was intermediate (389.9
13.9 g). The concentration of amides/imides in P. cenocla-
dum was higher than that in P. melanocladum, which in
turn was higher than that in P. imperiale (F2, 100=107.18,
P<0.001; Fig. 1b).
escaping total herbivory and beetle herbivory was lowest
for P. imperiale, while the number of individuals escaping
Eois herbivory was lowest for P. cenocladum. Piper
melanocladum individuals escaped all three types of
herbivory more frequently than the other two Piper species.
Damage by Atta cephalotes and Q. cerealis was rare (N=4
for each herbivore), indicating that these herbivores avoid
all three Piper species. Tettigoniids showed no preference
for one Piper species over another (X2=4.0, df=2, P=0.1).
Once a plant is attacked by an herbivore, the extent of
damage differs among Piper species and among herbivores.
When individuals with no herbivore damage were omitted
from the analysis, beetle and total herbivore damage were
significantly different between Piper species (Table 1,
Fig. 2b). Herbivores fed more on Piper imperiale than on
P. cenocladum or P. melanocladum (Fig. 2b). Beetles
When leaves were categorized as either experiencing or
escaping herbivory, Piper species had significant differ-
ences in frequency of leaves escaping damage from beetles,
Eois, and all herbivores combined (Fig. 2a; beetle: X2=
17.3, df=2, P<0.001; Eois: X2=13.1, df=2, P=0.001; total:
X2=21.4, df=2, P<0.001). The number of individuals

J Chem Ecol (2008) 34:558–574
563
(R=−0.63, P=0.05). No other correlations between herbiv-
ory and secondary metabolites were found. We found no
intraspecific relationships between toughness and herbivory
(Table 2), but intraspecific variance was low compared to
interspecific variance (Fig. 1a). We also found no evidence
for trade-offs between toughness and secondary metabolites
in the three species (P. cenocladum N=34, R=−0.22, P>
0.05; P. imperiale N=30, R=0.033, P>0.05; P. melano-
claddum N=44, R=0.10, P>0.05).
Atta cephalotes Feeding Trial In P. imperiale and P.
melanocladum, compound 4, piplaroxide (5), and com-
pound 6 are clearly deterrent to A. cephalotes (Table 3,
Fig. 3). The loglinear model revealed significant associa-
tions in removal of different leaf disks based on this
deterrence. The (compound 4×compound 6) and (com-
pound 4×piplaroxide) interactions are significant since the
frequencies of removal of both disk types are low in each
case. In addition, associations between removal of control
and treatment disks reflect the fact that the likelihood of the
control disk being taken while the treatment disk was not
taken was high, and the likelihood of the control disk being
left while the treatment disk was taken was low. Whole leaf
extract of P. imperiale was also deterrent to A. cephalotes
(Fig. 4). Ants removed less leaf area and mass of leaves
treated with leaf extracts in high (T=−2.38, df=41, P=0.02;
T=−2.86, df=41, P=0.007) and low (T=−2.60, df=35,
P=.01; T=−2.10, df=35, P=0.04) concentrations.
Fig. 2 Naturally occurring interspecific variation in herbivory for
three closely related species of tropical shrubs, Piper spp. a
Percentage of individuals completely escaping herbivory. Chi-square
tests for contingency tables were utilized to test for differences
between plant species in the frequency of escape from herbivory.
Beetle, Eois and total herbivory differed significantly between Piper
species (P. cenocladum: N=84, P. imperiale: N=84, P. melanocladum:
N=94). A value of 100% indicates all individuals of that species
escaped herbivory from a particular herbivore taxon. Total herbivory
includes escape from all categories of herbivory, thus, it is always has
a lower percentage than other categories. b Percent herbivory ( 1 SE)
on individual plants that did not escape herbivores. Analysis of
Covariance (with toughness as a covariate) was utilized to test for
differences in herbivory between species. P. cenocladum: N=72, P.
imperiale: N=67, P. melanocladum: N=56)
Discussion
We found evidence for ecological costs of the different
defense mechanisms in the three Piper species studied.
Secondary metabolites and their concentrations differed
among species, along with toughness and the strength of
the association of the plants with ants. Piper melanocla-
dum, which has no ant association and low levels of
chemical defense, has tough leaves. Piper cenocladum has
the strongest association with ants and the highest levels of
chemical defenses, but the leaves of this species are not
likely to use toughness as a deterrent. Piper imperiale has
tender leaves, low levels of defenses, and a loose
association with ants. In addition to these contrasting suites
of defenses, it is possible that there are trade-offs between
the traits measured here and chemical defenses that are
unknown or were not quantified, such as the sesquiterpenes
that are present in P. imperiale. A comparison of toughness
for the non-ant plant, P. melanocladum, with the ant-plants,
P. cenocladum (obligate) and P. imperiale (facultative),
provides support for the hypothesis that ant plants invest
less in other defenses (Fig. 1a; Dyer et al. 2001; Heil et al.
showed a significant preference for P. imperiale over P.
cenocladum but did not differentiate between P. melano-
cladum and the other two Piper species (Table 1). Piper
damage by Atta cephalotes and Q. cerealis was excluded
from the analysis due to small sample size. Levels of
damage by Eois and tettigoniids were not significantly
different among Piper species (Table 1).
Intraspecific Relationships Among Secondary Metabolites,
Toughness, and Herbivory Secondary metabolite content in
P. melanocladum was negatively correlated with levels of
tettigoniid herbivory on leaves that experienced herbivory

564
J Chem Ecol (2008) 34:558–574
Table 1 Qualitative differences in frequency of herbivory
Damage Type
Mean SE (N)
F (df)
Plant Species (Total N)
P. imperiale (81)
P. cenocladum (84)
P. melanocladum (94)
Total percent damage
9.6 1.21 (72)
6.1 1.1% (52)
6.0 1.3 (35)
6.1 1.9 (17)
(0)
4.5 1.1 (67)
1.7 0.4% (43)
2.8 0.8 (31)
3.0 0.6 (25)
22.9 20.8 (3)
2.4 (1)
5.0 1.4 (56)
2.6 0.6% (31)
5.9 2.7 (27)
4.2 1.7 (8)
1.0 (1)
9.6** (2)
10.7** (2)
3.0* (2)
1.4 (2)
–
Beetle percent damage
Tettigoniid percent damage
Eois percent damage
A. cephalotes percent damage
Q. cerealis percent damage
26.9 23.9 (2)
6.6 (1)
–
a Analysis of covariance
b Only herbivores experiencing herbivory were included in analyses, thus N varies by damage type and different types of herbivory do not add up
to total percent damage.
*P<.06
**P<.0001
2002). On the other hand, the hypothesis that ant plants
invest less than non-ant plants in chemical defenses, such as
amides/imides, is only supported when comparing concen-
trations of these compounds in P. imperiale and P.
melanocladum (Fig. 1b).
dum vs. the other two species, although P. cenocladum is of
intermediate toughness. Eois feeding may be deterred by
both toughness (in P. melanocladum) and secondary
metabolite content (in P. imperiale). Alternatively, as
relatively specialized feeders, Eois species, some of which
sequester imides/amides (Dyer et al. 2003), may be adapted
to the defensive compounds in P. cenocladum. Tettigoniids,
which are relatively generalized feeders, do not appear to
have a host preference possibly because they feed on the
young, expanding leaves of these species, which may have
low toughness and secondary metabolite content (Kursar
and Coley 1992).
Intraspecific variation in leaf toughness is not associated
with changes in total herbivory, suggesting that feeding
preferences within a single plant species may be motivated
by other factors, such as nutrient content or secondary
metabolism. Indeed, individual Piper species with lower
levels of chemical defenses in their tissues may be preferred
by insect enemies, such as leaf-cutter ants. We have shown
here that piplaroxide (5) and compounds 4 and 6 (found in
P. imperiale and P. melanocladum) are strongly deterrent to
The effectiveness of different defensive mechanisms in
these three species varies widely and differs according to
the taxon of the attacking herbivore. The strong negative
relationship between herbivory and toughness across the
three species supports the findings of previous studies that
demonstrate that toughness is an important defense (Coley
1983; Sagers and Coley 1995; Coley and Barone 1996).
When choosing among multiple host species, generalist and
oligophagous herbivores may avoid extremely tough
species. In an examination of different types of anti-
herbivore defenses and other plant characteristics across
46 tropical tree species, Coley (1983) found that leaf
toughness was the plant characteristic most correlated with
reduced herbivory. However, all patterns of herbivory
cannot be accounted for with differences in leaf toughness.
Specialist Eois caterpillar damage is higher on P. cenocla-
Table 2 Logistic regression: toughness as predictor of herbivore damage
Species
Total Damage
Beetle Damage
Tettigoniid Damage
Likelihood
Ratio χ2, Standardized
Estimate, Estimate χ2
Eois Damage
Likelihood
Ratio χ2, Standardized
Estimate, Estimate χ2
Likelihood
Likelihood
Ratio χ2, Standardized
Estimate, Estimate χ2
Ratio χ2, Standardized
Estimate, Estimate χ2
Piper cenocladum (N=
84)
0.40, −0.54, 0.31
0.78, −0.73, 0.63
1.86, −1.13, 1.33
0.19, 0.42, 0.18
Piper imperiale (N=81)
Piper melanocladum
(N=94)
2.80, 1.67, 2.71
2.11, 1.44, 2.04
0.42, 0.65, 0.42
2.86, 1.72, 2.75
0.28, 0.53, 0.28
0.30, 0.56, 0.30
0.29, 0.53, 0.28
0.15, 0.38, 0.15
a Data analyzed with logistic regression; degrees of freedom=1.
b P>0.05 for all likelihood ratios, indicating that models are a good fit.
c P>0.05 for all standardized estimates, indicating no association between toughness and herbivore damage.

J Chem Ecol (2008) 34:558–574
565
Table 3 Effects of Piper imides on Atta cephalotes feeding preference
Source
Degrees of Freedom
Chi-Square
P Value
Standardized Estimate
Compound 5×Compound 6
Compound 5×Compound 4
Compound 6
Control×Compound 6
Control×Compound 6
Control
Control×Compound 5
Compound 6a
Compound 5a
1
1
1
1
1
1
1
1
1
4
30.58
23.29
7.67
4.41
4.38
3.79
0.49
0.91
0.06
5.13
<0.0001
<0.0001
0.0281
0.0499
0.0250
0.0486
0.5792
0.3396
0.8001
0.2738
5.5303
4.8243
2.7695
2.1002
2.0924
−1.9470
0.6967
0.9549
−0.2230
Likelihood ratio
Data analyzed utilizing logit models
a Nonsignificant variables included in former models.
Compound 4=1-[(2E)-3-(3′,4′-dimethoxyphenyl)prop-2-enoyl]-5,6-dihydropridin-2(1H)-one; Compound 5=piplaroxide; Compound 6=3-[(2E)-
3-(4-hydroxy-3-methylphenyl)prop-2-enoyl]-7-oxa-3-azabicyclo[4.1.0]heptan-2-one
A. cephalotes (see also Capron and Wiemer 1996). In
previous studies, we demonstrated also that the three
imides/amides found in P. cenocladum deter feeding by A.
cephalotes and act synergistically against a variety of
herbivores (Dyer et al. 2003). Clearly, the imide/amide
nitrogen-based defenses are effective deterrents to leaf-
cutter ants, and the presence of these compounds is a likely
explanation for why these Piper species are avoided by A.
cephalotes.
The two sister species, P. cenocladum and P. imperiale
exhibit different patterns of defense, tolerance, and trade-
offs among defenses, despite the fact that these two species
are morphologically similar, unresolved in molecular
phylogenies (Tepe et al. 2004), and share almost identical
herbivore fauna. In P. imperiale, herbivory is common, as
the species appears to be investing few resources in defense
and relying on tolerance. Dyer et al. (2004c) found that
asexual reproductive success of P. imperiale was not
affected by herbivory, which is consistent with a tolerance
hypothesis. Piper imperiale fragments placed on the surface
of lowland rainforest soils to simulate asexual reproduction
through natural fragmentation appeared tolerant of high
Fig. 3 The frequency of experimental leaf fragment removal by
leafcutting ants, Atta cephalotes (Hymenoptera). Experimental fragments
were treated with imides from two species of tropical shrubs, P. imperiale
and P. melanocladum, and control fragments were treated only with
solvent. A loglinear model uncovered significant associations between
removal of pairs of leaf fragments with different solutions applied (e.g.,
when high numbers of control fragments were removed, low numbers of
piplaroxide fragments were removed). Based on significant associations
in this model, all defensive imides were deterrent to A. cephalotes. Imide
4=1-[(2E)-3-(3′,4′-dimethoxyphenyl)prop-2-enoyl]-5,6-dihydropridin-2
(1H)-one; Imide 5=piplaroxide; Imide 6=3-[(2E)-3-(4-hydroxy-3-
methylphenyl)prop-2-enoyl]-7-oxa-3-azabicyclo[4.1.0]heptan-2-one
Fig. 4 Mean leaf area ( 1 SE) removed from leaves of the tropical
shrub, P. imperiale, by foraging leaf-cutter (Atta cephalotes; Hyme-
noptera) workers. Large leaf fragments were treated with P. imperiale
extracts at high (T=−2.38, d.f.=41, P=0.02) and low (T=−2.60, d.f.=
35, P=0.01) concentrations and ants were presented with a control leaf
and a treatment leaf during a 2 h trial

566
J Chem Ecol (2008) 34:558–574
Acknowledgement This manuscript was improved by comments
from G. Gentry, R. Forkner, M. Heil, and 2 anonymous reviewers.
Thanks to M. Araya, Earthwatch Institute volunteers, G. Vega, and H.
Garcia for help in the field. This research was supported by funding
from Sigma Xi, Earthwatch Institute, The Garden Club of America,
OTS, NSF DEB 0508552 and DEB 0344250.
levels of herbivory present on fragments before planting. In
contrast, P. cenocladum exhibits high allocation of resour-
ces to defense, which may reflect low tolerance of
herbivory in this species. Piper cenocladum relies primarily
on vegetative fragmentation for reproduction (Greig 1993),
and Dyer et al. (2004c) showed that P. cenocladum asexual
reproductive success is reduced by natural herbivory on and
artificial damage of the leaves of fragments. Prevention of
herbivory by chemical and ant defense is associated with
increased asexual reproductive fitness. One might predict
that such high investment in defense would lead to trade-
offs among different defenses. In P. cenocladum, a trade-off
between secondary metabolite content and ant defense is
indicated by a threefold increase in amide concentrations
when ants are excluded from P. cenocladum plants (Dodson
et al. 2000). In the current study, this increase in secondary
metabolite content was not correlated with Eois herbivory,
which is consistent with other studies that demonstrate that
amides/imides do not protect P. cenocladum plants from
feeding by specialized herbivores (Dyer et al. 2004b).
There is evidence that the Ph. bicornis ant mutualists
effectively protect plants against specialized herbivores,
but they have little effect on large, generalist herbivores
such as katydids or other caterpillars, which may be
deterred by increased amide/imide concentrations (Dyer
and Letourneau 1999; Dyer et al. 2001, 2004b). Although
we found that P. cenocladum exhibits moderate investment
in leaf toughness, we were unable to detect any correla-
tions between toughness and herbivory in this species.
Nevertheless, investment in leaf toughness could make it
difficult for rare herbivores such as hesperiids (Lepidoptera)
and Atta (Hymenoptera) from adding this plant species as a
consistent food resource.
In summary, we have demonstrated a wide variety of
defensive mechanisms among three closely related species
of tropical shrubs. It is clear that redundancy in defense
provides the plant with protection against a variety of
herbivores. Previous work demonstrated intraspecific trade-
offs between chemical and ant-mediated defense in P.
cenocladum, but our interspecific results here were consis-
tent with both the trade-off hypothesis as well as the
alternative hypothesis that there are no trade-offs between
indirect biotic and direct chemical defenses in plants (Heil
et al. 2002). We were also unable to measure trade-offs
between defense and other functions, but tough leaves may
come at the expense of lower rates of growth or
reproduction. Studies of chemical redundancy, ecological
costs of antiherbivore defense, and interspecific variation in
defensive attributes should consider the complex interaction
between forces that select for optimum levels of defense
against a variety of different herbivores and trade-offs
between defenses, growth, and reproduction that allow
plants to minimize the cost of defense.
Appendix
Structures 1–3. Defensive metabolites isolated from Piper
cenocladum
O
CH3O
N
CH3O
O
CH3O
1
O
O
CH3O
HO
N
CH3O
2
O
CH3O
HO
N
O
CH3O
3

J Chem Ecol (2008) 34:558–574
567
Structure 4–6. Defensive metabolites isolated from Piper
melanocladum
performed on Aldrich silica gel (200–400 mesh, 60 Å).
Tetrahydrofuran (THF) was redistilled from sodium/benzo-
phenone. Dichloromethane (CH2Cl2) was redistilled from
CaH2. All other reagents and solvents were used as
received unless otherwise noted.
O
CH3O
N
1-(Methoxymethyl)piperidin-2-one (13) A 500-ml round
bottom flask was flushed under N2 while δ-valerolactam
was melted on the oven. δ-Valerolactam (15.147 g,
153 mmol) and distilled THF (160 ml) were added to the
flask, and the mixture was cooled on an ice bath for 10 min.
After cooling, 2.25 M n-BuLi (50 ml) was added to the
mixture slowly via syringe. CH3OCH2Cl (11.8 g,
157 mmol) was added to the flask, and the mixture stirred
at room temperature for 2 h. The reaction mixture was
transferred to a separatory funnel with hexane (50 ml), and
washed with water (2×100 ml) and brine (1×50 ml). The
combined aqueous layers were extracted with CH2Cl2 (5×
50 ml). The combined organic layers were washed with
brine (1×60 ml), dried over anhydrous magnesium sulfate,
and filtered. The solvent was removed under reduced
pressure to yield a yellow liquid (15.2 g), which was
purified via vacuum distillation (92–94°C @ 4 mm Hg)
using a Vigeroux fractionating column to yield 13 as a
CH3O
O
4
O
O
CH3O
CH3O
N
O
5
1
colorless liquid [10.3 g, 49% yield; H NMR (CDCl3): δ
4.77 (2H, s), δ 3.32 (2H, t, J=5 Hz), δ 3.26 (3H, s), δ 2.39
(2H, t, J=5 Hz), δ 1.78 (4H, m). 13C NMR (CDCl3): δ
171.1, δ 77.1, δ 56.0, δ 46.2, δ 32.4, δ 23.1, δ 21.3].
O
O
CH3O
HO
N
1-(Methoxymethyl)-3-(phenylthio)piperidin-2-one (14) A
250-ml three-necked round bottom was oven dried, fitted
with an addition funnel and a septum, and flushed under
N2. Distilled THF (30 ml) and diisopropylamine (8 ml)
were added to the flask and the solution stirred on ice for
10 min. 2.25 M n-BuLi (24 ml) was added through the
septum via syringe and the mixture stirred on ice for
10 min. The mixture was cooled to −78°C in a dry ice/
isopropanol bath. Compound 13 (4 g, 27.9 mmol) was
dissolved in distilled THF (15 ml) and added to the reaction
through the dropping funnel over 10 min. The mixture
stirred at −78°C for an additional 45 min. A mixture of
phenyldisulfide (6.01 g) and HMPA (4.8 ml) in distilled
THF (15 ml) was added to the reaction mixture over
20 min. The reaction stirred at −78°C for an additional
40 min and was then allowed to warm to room temperature.
The reaction mixture was transferred to a separatory funnel
with water (60 ml) and extracted with diethyl ether (3×
60 ml). The combined ether layers were washed with 3 M
NaOH (1×40 ml), water (1×40 ml), 3 M HCl (1×40 ml),
water (1×40 ml), and brine (1×40 ml). The solution was
dried over magnesium sulfate, filtered, and evaporated to
O
6
Synthesis of 4, 5, and 6 (Schemes 1, 2, 3, 4, 5, 6 and 7):
The syntheses of 1-[(2E)-3-(3’′,4’′-dimethoxyphenyl)prop-
2-enoyl]-5,6-dihydropridin-2(1H)-one (4), 3-[(2E)-3-(3,4-
dimethylphenyl)prop-2-enoyl]-7-oxa-3-azabicyclo-[4.1.0]
heptan-2-one (piplaroxide, 5) and 3-[(2E)-3-(4-hydroxy-3-
methylphenyl)prop-2-enoyl]-7-oxa-3-azabicyclo[4.1.0]-
heptan-2-one (6) were accomplished via similar convergent
approaches. These involved the synthesis of an appropriate
functionalized pyridinone piece and its coupling with an
appropriately activated acid derivative.
Experimental
General Information All H NMR and 13C NMR spectra
1
1
yield 14 as an orange oil [6.25 g, 89% yield; H NMR
were recorded at 300 MHz. All flash chromatography was
(CDCl3): δ 7.54 (2H, m), δ 7.29 (3H, m), δ 4.83 (2H, q, J=

568
J Chem Ecol (2008) 34:558–574
O
O
O
R1
R1
N
N
R2
R2
O
O
O
R1
Cl
HN
R2
O
O
R1
OH
HN
R2
Scheme 1 Convergent retrosynthesis of “Piper amides” isolated from P. melanocladum
O
10 Hz), δ 3.89 (1H, t, J=5.5 Hz), δ 3.39 (2H, m), δ 3.30
(3H, s), δ 2.06 (3H, m), δ 1.7 (1H, m). 13C NMR (CDCl3):
δ 169.8, δ 134.4, δ 132.7, δ 129.1, δ 127.7, δ 77.5, δ 56.2,
δ 49.0, δ 45.9, δ 28.4, δ 20.4].
3-(Phenylthio)piperidin-2-one (15) Compound 14 (10 g)
was placed in a 1-l round bottom flask with 95% ethanol
(270 ml) and conc. HCl (54 ml). The reaction mixture was
heated at reflux for 6 h. The volume was reduced to 100 ml
O
O
O
SPh
HN
O
N
O
N
1. n-BuLi
1. LDA
2. PhSSPh
2. CH3OCH2Cl
12
13
14
O
O
SPh
HN
HN
EtOH
1. m-CPBA
2.
∆
HCl (aq)
15
16
Scheme 2 The dihydropyridinone moiety was prepared from δ-
valerolactam (12). The lactam nitrogen was protected (13) by treating
the lithium salt of the lactam with chloromethyl methyl ether
(MOMCl). The enolate of compound 13 was generated by treating
the protected lactam with two equivalents of LDA in THF at −78°C
and sulfenylated in the α-position with phenyl disulfide (PhSSPh).
After deprotection of the sulfenylated lactam (14), the sulfide was
oxidized to the sulfoxide and subjected to thermal elimination to
produce 5,6-dihydropyridin-2(1H)-one (16)

J Chem Ecol (2008) 34:558–574
569
O
O
CH3O
CH3O
CH3O
oxalyl chloride
OH
Cl
DMF
CH2Cl2
CH3O
7
17
Scheme 3 The hydrocinnamoyl derivatives were prepared from
appropriate carboxylic acid precursors. Consequently, commercially
available 3,4-dimethoxyhydrocinnamic acid (7) was converted to 3,4-
dimethoxyhydrocinnamoyl chloride (17) by treating the acid with
oxalyl chloride in dichloromethane using catalytic amounts of dry
DMF
under reduced pressure, and the mixture was extracted with
CH2Cl2 (2×250 ml). The acid layer was neutralized with
6 M NaOH and extracted with CH2Cl2 (2×250 ml). Each
CH2Cl2 layer was washed with saturated NaHCO3 (2×
80 ml) and brine (1×80 ml). The combined organic layers
were dried over sodium sulfate, filtered, and evaporated to
yield a brown residue (7.4 g). The residue was recrystal-
lized from ethyl acetate/hexane to yield 15 as an off-white
δ 7.31 (3H, m), δ 6.05 (1H, broad s), δ 3.83 (1H, t,
J=6 Hz), δ 3.32 (2H, t, J=4 Hz), δ 2.03 (3H, m), δ 1.74
(1H, m)]. 13C NMR (CDCl3): δ 170.4, δ 134.1, δ 132.9, δ
129.1, δ 127.8, δ 48.4, δ 42.5, δ 28.3, δ 20.2.
5,6-Dihydropyridin-2(1H)-one (16) Compound 15 (3.0 g)
was weighed into a 250-ml round bottom flask. CH2Cl2
(100 ml) and saturated NaHCO3 (20 ml) were added to the
flask, which was then cooled on an ice bath for 10 min.
1
solid [5.3 g, 64% yield; H NMR (CDCl3): δ 7.54 (2H, m),
O
O
CH3O
CH3O
HCO2- +NH4
OH
OH
10% Pd/C
CH3OH
HO
HO
18
19
O
CH3O
OTBDMS
TBDMSCl
imidazole
DMF
TBDMSO
20
O
CH3O
Cl
oxalyl chloride
DMF
TBDMSO
CH2Cl2
21
Scheme 4 In the case of the 4-hydroxy-3-methoxy derivative, the
appropriate acid (19) was first obtained via catalytic transfer
hydrogenation of commercially available ferulic acid (18) with
ammonium formate using 10% Pd/C in methanol. The free phenolic
hydroxyl group was then protected as the TBDMS ether (20) using t-
butyldimethylsilyl chloride and imidazole in dry DMF. Protection was
necessary to avoid polymerization during and after the formation of
the acyl halide. While this step also introduced a TBDMS group at the
carboxyl oxygen, this group was directly converted to the acyl
chloride (21) by treating the silyl ester with oxalyl chloride in
dichloromethane using catalytic amounts of DMF

570
J Chem Ecol (2008) 34:558–574
O
O
O
Li+ -N
CH3O
RO
CH3O
RO
Cl
N
THF
O
R = CH3
(
17
)
R = CH3
4
( )
TBDMS (21
)
TBDMS (22
)
Scheme 5 In all cases, the next step involved the coupling of compound 16 with the appropriate acyl halide. Thus, the lithium salt of compound
16 was generated by treatment with KHMDS in dry THF and condensed with ether 17 or 21 to yield 4 or 22, respectively
Seventy-seven percent 3-chloroperbenzoic acid (3.1 g) was
added in five portions over 10 min. The mixture was
vigorously stirred on an ice bath 1 h. The reaction mixture
was transferred to a separatory funnel with CH2Cl2
(200 ml) and washed with saturated NaHCO3 (2×30 ml).
The solution was dried over sodium sulfate, filtered, and
evaporated to yield a brown residue (3.2 g). Toluene (50 ml)
was added to the brown residue and the solution was heated
at reflux for 1 h. The solvent was removed under reduced
pressure, and the brown oil was purified via flash chroma-
tography (EtOAc) to yield compound 16 [1.25 g, 90% yield;
1H NMR (CDCl3): δ 6.64 (1H, m) δ 6.45 (1H, broad s),
δ 5.89 (1H, m), δ 3.42 (2H, m), δ 2.34 (2H, m)]. 13C NMR
(CDCl3): δ 166.6, δ 141.9, δ 129.2, δ 39.7, δ 23.9.
3-(4-Hydroxy-3-methylphenyl)propanoic acid (19) A mix-
ture of 18 (2.0 g, 10 mmol), ammonium formate (2.0 g,
32 mmol), and 10% Pd/C (0.1 g) in methanol (50 ml) was
stirred under N2 at RT for 24 h. Celite filter aid (0.5 g) was
added, and the mixture was vacuum filtered. The volatiles
were removed under reduced pressure. The residue was
transferred to a separatory funnel with the aid of 1 M HCl
(50 ml) and dichloromethane (50 ml). The organic layer
was isolated, and the aqueous solution was extracted with
more dichloromethane (1×50 ml). The combined organic
layers were washed with brine, dried over anhydrous
MgSO4, and filtered. The volatiles were removed under
reduced pressure to yield 19 as a pure white solid [1.9 g,
95%; 1H NMR (CDCl3): δ 11.17 (1H, br s), δ 6.84 (1H, d),
δ 6.71 (2H, m), δ 5.54 (1H, br s), δ 3.87 (3H,s), δ 2.89 (2H,
t), δ 2.65 (2H, d). 13C NMR (CDCl3): δ 179.11, δ 146.50, δ
144.14, δ 132.14, δ 120.91, δ 114.47, δ 110.97, δ 55.94, δ
36.07, δ 30.41].
3-(3,4-Dimethoxyphenyl)propanoyl chloride (17) A 100-ml
round-bottomed flask was flushed with N2 and charged
with 7 (2.1 g, 10 mmol), freshly distilled dichloromethane
(40 ml), and dry DMF (four drops). Oxalyl chloride
(1.75 ml, 20 mmol) was added slowly via pipet. The flask
was fitted with a drying tube (CaCl2), and the mixture was
stirred at RT for 22 h. The volatiles were removed under
reduced pressure to yield 17 as a pale yellow oil sufficiently
tert-Butyldimethylsilyl 3-(4-t-butyldimethylsilyloxy-3-
methyoxyphenyl)propanoate (20) A mixture of 19 (2.7 g,
14 mmol), TBDMSCl (4.52 g, 30 mmol) and imidazole
(4.1 g, 60 mmol) in dry DMF (12 ml) was stirred under N2
at RT for 5 days. The mixture was poured into a separatory
funnel with diethyl ether (75 ml) and hexanes (25 ml). The
mixture was washed with water (1×50 ml), saturated
NaHCO3 (1×50 ml), and water (1×50 ml), and brine (1×
50 ml). The organic solution was dried over anhydrous
MgSO4 and filtered. Solvents were removed under reduced
1
pure for use [2.3 g, 100%; H NMR (CDCl3): δ 6.82–6.69
(3H, m, ArH), δ 3.87 (3H, s, -OCH3), δ 3.86 (3H, s,
-OCH3), δ 3.19 (2H, t, J=7.4 Hz, -CH2-), δ 2.96 (2H, t, J=
7.4 Hz, -CH2-). 13C NMR (CDCl3): δ 173.22, δ 149.09, δ
147.94, δ 131.22, δ 120.30, δ 111.61, δ 111.43, δ 55.99, δ
55.94, δ 48.92, δ 30.75].
O
O
CH3O
CH3O
N
N
NaOCl
Et2O / DMF
RO
RO
O
O
O
R = CH3
( )
4
R = CH3
( )
5
TBDMS (22
)
TBDMS (23
)
Scheme 6 The corresponding epoxides (5 and 23) were prepared via epoxidation of the respective alkenes by treatment with NaOCl in a mixture
of diethyl ether and DMF

J Chem Ecol (2008) 34:558–574
571
O
O
O
O
CH3O
CH3O
HO
N
N
NH4F
CH3OH
TBDMSO
O
O
(
23
)
6
( )
Scheme 7 Finally, compound 6 was prepared after deprotection of 23 with methanolic NH4F
pressure to yield a colorless oil. The remaining volatiles
[470 mg, 40% yield; 1H NMR (CDCl3): δ 6.86 (1H, d of t),
δ 6.78 (3H, s), δ 5.97 (1H, d of t), δ 3.96 (2H, t), δ 3.86
(3H, s), δ 3.84 (3H, s), δ 3.24 (2H, t), δ 2.93 (2H, t), δ 2.37
(2H, m); 13C NMR (CDCl3): δ 175.7 (s), δ 165.5 (s), δ
149.0 (s), δ 147.5 (s), δ 145.3 (d), δ 133.8 (s), δ 126.0 (d),
δ 120.4 (d), δ 111.9 (d), δ 111.2 (d), δ 56.0 (q), δ 55.9 (q), δ
41.2 (t), δ 41.1 (t), δ 30.8 (t), δ 24.7 (t). LREIMS m/z (rel.
int.): 289 (37), 192 (51), 164 (100), 151 (50). HRMS:
Found m/z 289.1315, calculated for C16H19O4N 289.1314.
Anal. Calcd. for C16H19NO4: C, 66.42; H, 6.62; N, 4.84.
Found: C, 66.52; H, 6.62; N, 4.65.
were removed under high vacuum to yield 20 as a pure,
1
pale yellow oil [5.77 g, 99%; H NMR (CDCl3): δ 6.74
(1H, d), δ 6.68 (1 H, d), δ 6.63 (1H, d of d), δ 3.78 (3H, s),
δ 2.85 (2H, t), δ 2.62 (2H, t), δ 0.98 (9 H, s), δ 0.90 (9H, s),
δ 0.23 (6H, s), δ 0.13 (6H, s). 13C NMR (CDCl3): δ 173.57,
δ 150.81, δ 143.36, δ 134.22, δ 120.84, δ 120.35, δ 112.38,
δ 55.53, δ 37.78, δ 30.88, δ 25.82, δ 25.63, δ 18.52, δ
17.67, δ −4.58, δ −4.74].
3-(4-t-butyldimethylsilyloxy-3-methyoxyphenyl)propanoyl
chloride (21) A 250-ml round-bottomed flask was flushed
with N2 and charged with 20 (5.768 g, 13.6 mmol), freshly
distilled dichloromethane (25 ml), and dry DMF (five
drops). Oxalyl chloride (1.80 ml, 20.6 mmol) was added
slowly via pipet. The flask was fitted with a drying tube
(CaCl2), and the mixture was stirred at RT for 20 h. The
volatiles were removed under reduced pressure to yield 21
as a pale yellow oil sufficiently pure for further use [2.3 g,
3-[3-(3,4-dimethoxyphenyl)propanoyl]-7-oxa-3-azabicyclo
[4.1.0]heptan-2-one (5) A 250-ml round-bottomed flask
was charged with 4 (3.47 g, 12 mmol), DMF (100 ml) and
Et2O (100 ml). The mixture was cooled to 5°C on an ice
bath. Cold, 6% NaOCl(aq) (Clorox, 20 ml, 24 mmol) was
added, and the resultant mixture was vigorously stirred for
10 min. The reaction mixture was then transferred to a
separatory funnel with 5% Na2S2O3 (100 ml) and Et2O
(50 ml). The layers were mixed and separated. The aqueous
layer was further extracted with Et2O (2×100 ml). The
combined organic layers were then washed with brine (1×
100 ml) and dried over anhydrous MgSO4. The volatiles
were evaporated under reduced pressure to yield 5 as a pale
solid that was sufficiently pure for further use (1.23 g,
34%). Further purification can be accomplished via
recrystallization from 95% EtOH [1H NMR (CDCl3): δ
6.77 (3H, br s), δ 4.32 (3H, dddd), δ 3.87 (3H, s), δ 3.85
(3H, s), δ 3.68 (1H, dd), δ 3.56 (1H, d), δ 3.21 (2H, m), δ
3.15 (1H, dd), δ 2.90 (2H, t), δ 2.41 (1H, dm), δ 1.98 (1H,
dddd). 13C NMR (CDCl3): δ 174.8 (s), δ 169.7 (s), δ 148.8
(s), δ 147.4 (s), δ 133.5 (s), δ 120.4 (d), δ 111.9 (d), δ 111.2
(d), δ 56.0 (q), δ 55.9 (q), δ 53.6 (d), δ 52.5 (d), δ 41.5 (t),
δ 35.8 (t), δ 30.6 (t), δ 24.0 (t). LREIMS m/z (rel. int.): 305
(24), 192 (35), 164 (100), 151 (62). Anal. Calcd. for
C16H19NO5: C, 62.94; H, 6.27; N, 4.59. Found: C, 63.05;
H, 6.08; N, 4.54.
1
100%; H NMR (CDCl3): δ 6.77 (1H, d), δ 6.66 (1H,d), δ
6.62 (1H,d), δ 3.79 (3H, s), δ 3.18 (2H, t), δ 2.93 (2H, t), δ
0.98 (9H, s), δ 0.13 (6H, s). 13C NMR (CDCl3): δ 173.25,
δ 151.03, δ 143.91, δ 132.09, δ 121.09, δ 120.45, δ 112.38,
δ 55.58, δ 48.95, δ 30.84, δ 25.80, δ 18.52, δ −4.56].
1-[3-(3,4-Dimethoxyphenyl)propanoyl]-5,6-dihydropyridin-
2(1H)-one (4) A 100-ml three-necked, round-bottomed
flask was oven dried and flushed with N2 while cooling.
Freshly distilled THF (15.6 ml) and a solution of KHMDS
(8.2 ml, 0.5 M in toluene) were added to the flask and
cooled on ice for 10 min. Compound 16 (0.4 g, 4 mmol)
was dissolved in THF (4.5 ml) and added dropwise to the
reaction mixture, which was then stirred on ice for 10 min.
Compound 17 (1.0 g, 4.4 mmol) was added to the reaction
mixture, which was then stirred on ice for 10 min. The
reaction mixture was transferred to a separatory funnel with
water (20 ml), ethyl acetate (30 ml), and hexane (10 ml).
The organic layer was isolated and washed with saturated
NaHCO3 (2×20 ml), 0.1 M HCl (1×20 ml), and brine (1×
20 ml). The organic solution was dried over sodium sulfate,
filtered, and the solvent was evaporated to yield an orange
oil (800 mg). The oil was purified via flash chromatography
(4:1 EtOAc/hexane) to yield compound 4 as a white solid
1-[3-(4-t-Butyldimethylsilyloxy-3-methyoxyphenyl)-
propanoyl]-5,6-dihydropyridin-2(1H)-one (22) A 250-ml
three neck, oven dried, round bottom was cooled while
flushing with N2. KHMDS (0.5 M in toluene, 28.4 ml) was

572
J Chem Ecol (2008) 34:558–574
1
added and the flask cooled on an ice bath for 10 min.
Compound 16 (1.38 g, 14 mmol) was dissolved in THF
(30 ml) and added dropwise to the reaction mixture which
then stirred for 10 min. Compound 21 (4.5 g, 14 mmol) was
added to the reaction which then stirred for 10 min. The
reaction mixture was transferred to a separatory funnel with
water (50 ml), ethyl acetate (75 ml), and hexanes (15 ml).
The water layer was removed and the organic layer was
washed with saturated NaHCO3 (2×50 ml), 0.1M HCl (1×
50 ml), and brine (1×50 ml). The solution was dried over
sodium sulfate, filtered, and evaporated to yield a brown oil
(5.2 g). The oil was purified via flash chromatography (1:1
EtOAc/hexanes) to yield compound 22 [2.2 g, 42% yield;
1H NMR (CDCl3): δ 6.86 (1H, d of t), δ 6.70 (3H, m), δ
5.97 (1H, d of t), δ 3.95 (2H, t), δ 3.77 (3H, s), δ 3.21 (2H,
t), δ 2.90 (2H, t), δ 2.36 (2H, m), δ 0.97 (9H, s), δ 0.12
(6H, s). 13C NMR (CDCl3): δ 175.81, δ 165.44, δ 150.73, δ
145.26, δ 143.26, δ 134.64, δ 125.97, δ 120.72, δ 120.66,
δ 112.68, δ 55.55, δ 41.14, δ 41.08, δ 30.94, δ 25.82, δ
24.72, δ 18.51, δ 14.28, δ -4.55].
[0.41 g, 48%; H NMR (CDCl3): δ 6.83–6.70 (3H,m), δ
5.47 (1H,s), δ 4.35 (1H, ddt), δ 3.87 (3 H, m), δ 3.68 (1 H,
br t), δ 3.56 (1H, d), δ 3.21 (2H, t), δ 3.21 (1H, m), δ 2.90
(2 H, t), δ 2.40 (1H, m), δ 1.98 (1H, ddd). 13C NMR
(CDCl3): δ 175.1 (s), δ 169.9 (s), δ 146.6 (s), δ 144.2 (s), δ
133.0 (s), δ 121.2 (d), δ 114.3 (d), δ 111.2 (d), δ 56.0 (q), δ
53.5 (d), δ 52.4 (d), δ 41.5 (t), δ 35.7 (t), δ 30.6 (t), δ 23.9
(t). LREIMS m/z (rel. int.): 291 (13), 178 (23), 150 (100),
137 (50). ES(+)MS m/z 292.1 [M+H], 314.1 [M+Na]. Anal.
Calcd. for C15H17NO5: C, 61.85; H, 5.88; N, 4.81. Found:
C, 62.16; H, 6.11; N, 4.81.
Other potential compounds All three species were screened
for basic alkaloids using a standard acid/base partitioning
technique followed by TLC with visualization using the
alkaloid specific spray reagent, iodoplatinic acid, and found
to be negative. Each species was examined by GC-MS
using the same conditions described above for imide/amide
quantitation. Piper cenocladum and P. melanocladum were
found to be devoid of compounds other than the imides/
amides described (other peaks in the chromatograms had
areas at least 1,000 times less than the imides/amides).
Piper imperiale was found to contain a series of seven
peaks with areas similar to or greater than the imides
described. The mass spectra of all of these compounds are
consistent with sesquiterpenes or mono-oxygenated sesqui-
terpenes. Five peaks had mass spectra that were excellent
matches to the library spectra of copaene, cyclopropylazu-
lene, (-) spathulenol, ledrol, and caryophyllene oxide. The
peak for caryophyllene oxide was the largest in the
chromatogram. Lack of standards for these compounds
precluded confirmation of their identities and quantifica-
tion. Other classes of natural products were not actively
pursued.
3-[3-(4-t-Butyldimethylsilyloxy-3-methoxyphenyl)-
propanoyl]-7-oxa-3-azabicyclo[4.1.0]heptan-2-one (23) A
250-ml round-bottomed flask was charged with 22 (1.34 g,
3.44 mmol), DMF (36 ml) and Et2O (36 ml). The mixture
was cooled to 5°C on an ice bath. Cold, 6% NaOCl(aq)
(Clorox, 20 ml, 17.2 mmol) was added, and the resultant
mixture was vigorously stirred for 1 h. The reaction mixture
was then transferred to a separatory funnel with 5%
Na2S2O3 (180 ml) and Et2O (90 ml). The layers were
mixed and separated. The aqueous layer was further
extracted with Et2O (2×90 ml). The combined organic
layers were then washed with brine (1×180 ml) and dried
over anhydrous MgSO4. The volatiles were evaporated
under reduced pressure to yield 23 as a pale orange oil that
was sufficiently pure for further use (1.27 g, 91%). Further
purification can be accomplished via flash chromatography
1
on silica [1:1 EtOAc/hexanes; H NMR (CDCl3): δ 6.75–
6.64 (3H, m), δ 4.32 (1H, dddd), 3.78 (3H, s), 3.68 (1H,
dd), 3.55 (1H, d), 3.18 (2H, m), 3.13 (1H, dd), 2.34 (1H,
dm), δ 1.98 (1H, dddd), 0.98 (9H, s), 0.13 (6H, s)].
References
BERENBAUM, M., and NEAL, J. J. 1985. Synergism between myristicin
and xanthotoxin, a naturally co-occurring plant toxicant. J.
Chem. Ecol. 11:1349–1349.
3-[3-(4-hydroxy-3-methoxyphenyl)propanoyl]-7-oxa-3-aza-
bicyclo[4.1.0]heptan-2-one (6) A mixture of 23 (1.27 g,
3.12 mmol), ammonium fluoride (0.579 g, 15.6 mmol), and
methanol (75 ml) was heated at reflux for 5 min. The
reaction mixture was transferred to a separatory funnel with
the aid of water (90 ml). The aqueous layer was extracted
with CH2Cl2 (2×90 ml). The combined organic extracts
were washed with brine (1×90 ml) and dried over
anhydrous MgSO4. The volatiles were evaporated under
reduced pressure to yield a tan solid. This was recrystallized
from EtOAc-hexanes to yield 6 as an off-white solid
BERENBAUM, M. R., NITAO, J. K., and ZANGERL, A. R. 1991.
Adaptive significance of furanocoumarin diversity in Pastinaca
sativa (Apicaceae). J. Chem. Ecol. 17:207–207.
BERGELSON, J., and PURRINGTON, C. B. 1996. Surveying patterns in
the cost of resistance in plants. Am. Nat. 148:536–558.
BURGER, W. 1971. Flora costaricensis. Fieldiana Bot. 35:1–1.
CALCAGNO, M. P., COLL, J., LLORIA, J., FAINI, F., and ALONSO-
AMELOT, M. E. 2002. Evaluation of synergism in the feeding
deterrence of some furanocoumarins on Spodoptera littoralis. J.
Chem. Ecol. 28:175–175.
CAPRON, M. A., and WIEMER, D. F. 1996. Piplaroxide an ant-repellent
piperidine epoxide from Piper tuburculatum. J. Nat. Prod.
59:794–794.

J Chem Ecol (2008) 34:558–574
573
CATES, R. G., and RHOADES, D. 1977. Patterns in the production of
antiherbivore chemical defenses in plant communities. Biochem.
Syst. Ecol. 5:185–185.
COLEY, P. D. 1983. Herbivory and defensive characteristics of tree
species in a lowland tropical forest. Ecol. Monogr. 34:209–234.
COLEY, P. D., and BARONE, J. A. 1996. Herbivory and plant defenses
in tropical forests. Annu. Rev. Ecol. Syst. 27:305–305.
COLEY, P. D., BRYANT, J. P., and CHAPIN, F. S. III 1985. Resource
availability and plant anti-herbivore defense. Science 230:895–
899.
HARBORNE, J. B. 2001. Twenty-five years of chemical ecology. Nat.
Prod. Rep. 18:361–379.
HAY, M. E., KAPPEL, Q. E., and FENICAL, W. 1994. Synergisms in
plant defenses against herbivores-interactions of chemistry,
calcification, and plant-quality. Ecology 75:1714–1726.
HEIL, M. 2002. Ecological costs of induced resistance. Curr. Opin.
Plant Biol. 5:345–350.
HEIL, M., and MCKEY, D. 2003. Protective ant-plant interactions as
model systems in ecological and evolutionary research. Annu.
Rev. Ecol. Syst. 34:425–453.
DODSON, C. D., DYER, L. A., SEARCY, J., WRIGHT, Z., and
LETOURNEAU, D. K. 2000. Cenocladamide, a dihydropyridone
alkaloid from Piper cenocladum. Phytochemistry 53:51–51.
DUH, C. Y., WU, Y. C., and WANG, S. K. 1990. Cytotoxic pyridine
alkaloids from Piper aborescens. Phytochemistry 29:2689–2689.
DYER, L. A., and LETOURNEAU, D. K. 1999. Trophic cascades in a
complex terrestrial community. Proc. Nat. Acad. Sci. USA
96:5072–5072.
HEIL, M., DELSINNE, T., HILPERT, A., SCHURKENS, S., ANDARY, C.,
LINSENMAIR, K. E., SOUSA, M. S., and McKEY, D. 2002.
Reduced chemical defence in ant-plants? A critical re- evaluation
of a widely accepted hypothesis. Oikos 99:457–468.
HERMS, D. A., and MATTSON, W. J. 1992. The dilemma of plants: to
grow or defend. Q. Rev. Biol. 67:283–283.
JANZEN, D. H. 1966. Coevolution of mutualism between ants and
Acacias in Central America. Evolution 20:249–249.
DYER, L. A., DODSON, C. D., BEIHOFFER, J., and LETOURNEAU, D. K.
2001. Trade-offs in antiherbivore defenses in Piper cenocladum:
Ant mutualists versus plant secondary metabolites. J. Chem.
Ecol. 27:581–581.
DYER, L. A., DODSON, C. D., STIREMAN, J. O., TOBLER, M. A.,
SMILANICH, A. M., FINCHER, R. M., and LETOURNEAU, D. K.
2003. Synergistic effects of three Piper amides on generalist and
specialist herbivores. J. Chem. Ecol. 29:2499–2499.
DYER, L. A., DODSON, C. D., and RICHARDS, J. 2004a. Isolation,
synthesis, and evolutionary ecology of Piper amides, pp. 117–
139, in L. A. Dyer, and A. N. Palmer (eds.). A Model Genus for
Studies of Evolution, Chemical Ecology, and Trophic Interac-
tionsKluwer Academic Publishers, Boston.
DYER, L. A., DODSON, C. D., LETOURNEAU, D. K., TOBLER, M. A.,
HSU, A., and STIREMAN, J. O. I. 2004b. Ecological causes and
consequences of variation in defensive chemistry of a neotropical
shrub. Ecology 85:2795–2803.
JOHNSON, H. B. 1975. Plant pubescence: An ecological perspective.
Bot. Rev. 41:233–258.
KORICHEVA, J. 2002. Meta-analysis of sources of variation in fitness
costs of plant antiherbivore defenses. Ecology 83:176–176.
KORICHEVA, J., NYKANEN, H., and GIANOLI, E. 2004. Meta-analysis
of trade-offs among plant antiherbivore defenses: are plants
jacks-of-all-trades, masters of all?. Am. Nat. 163:E64–E75.
KURSAR, T. A., and COLEY, P. D. 1992. Delayed greening in tropical
leaves: An antiherbivore defense?. Biotropica 24:256–256.
LETOURNEAU, D. K. 1983. Passive aggression: An alternative
hypothesis for the Piper-Pheidole association. Oecologia
60:122–122.
LETOURNEAU, D. K., and DYER, L. A. 1998. Experimental manipu-
lations in lowland tropical forest demonstrate top-down cascades
through four trophic levels. Ecology 79:1678–1678.
LEVIN, D. A. 1976. The chemical defenses of plants to pathogens and
herbivores. Annu. Rev. Ecol. Syst. 7:121–121.
DYER, L. A., GENTRY, G., and TOBLER, M. A. 2004c. Fitness
consequences of herbivory: Impacts on asexual reproduction of
tropical rain forest understory plants. Biotropica 36:68–68.
FEENY, P. 1970. Seasonal changes in oak leaf tannins and nutrients as
a cause of spring feeding by winter moth caterpillars. Ecology
51:565–565.
FISCHER, R. C., RICHTER, A., WANEK, W., and MAYER, V. 2002.
Plants feed ants: food bodies of myrmecophytic Piper and their
significance for the interaction with Pheidole bicornis ants.
Oecologia 133:186–186.
FLOYD, T. 2001. Causal categorical data analysis, pp. 197–216, in S.
M. Scheiner, and J. Gurevitch (eds.). Design and Analysis of
Ecological ExperimentsOxford University Press, Oxford.
FOLGARAIT, P. J., and DAVIDSON, D. W. 1995. Myrmecophytic
Cecropia: Antiherbivore defences under different nutrient treat-
ments. Oecologia 104:189–189.
FOLGARAIT, P. J., DYER, L. A., MARQUIS, R. J., and BRAKER, H. E.
1996. Leaf-cutting ant (Atta cephalotes) preferences for five
native tropical plantation tree species growing under different
light conditions. Entomol. Exp. App. 80:521–521.
LINDROTH, R. L., and HWANG, S. Y. 1996. Diversity, redundancy, and
multiplicity in chemical defense systems of aspen, pp. 25–56, in
J. T. Romeo, J. A. Saunders, and P. Barbosa (eds.). Phytochemical
Diversity and Redundancy in Ecological InteractionsPlenum
Press, New York.
LOCKWOOD, J. R. 1998. On the statistical analysis of multiple-choice
feeding preference experiments. Oecologia 116:475–475.
MARQUIS, R. J. 1991. Herbivore fauna of Piper (Piperaceae) in a Costa
Rican wet forest: diversity, specificity, and impact, pp. 179–208, in
P. W. Price, T. M. Lewinsohn, G. W. Fernandes, and W. W.
Benson (eds.). Plant-Animal Interactions: Evolutionary Ecology in
Tropical and Temperate RegionsJohn Wiley & Sons, New York.
MATTSON, W. J., LAWRENCE, R. K., HAACK, R. A., HERMS, D. A., and
CHARLES, P. J. 1988. Defensive strategies of woody plants against
different insect-feeding guilds in relation to plant ecological
strategies and intimacy of association with insects, pp. 3–38, in W.
J. Mattson, J. Levieux, and C. Bernard-Dagan (eds.). Mechanisms
of Woody Plant Defenses Against InsectsSpringer, New York.
MAXWELL, F. G., JENKINS, J. N., and PARROTT, W. L. 1972.
Resistance of plants to insects. Adv. Agron. 24:187–265.
NELSON, A. C., and KURSAR, T. A. 1999. Interactions among plant
defense compounds: a method for analysis. Chemoecology 9:81–92.
PARMAR, V. S., JAIN, S. C., BISCT, K. S., JAIN, R., TANEJA, P., JHA, A.,
TYAGI, O. D., PRASAD, A. K., WENGEL, J., OLSEN, C. E., and
BOLL, P. M. 1997. Phytochemistry of the genus Piper. Phyto-
chemistry 46:597–597.
GERSHENZON, J. 1994. The cost of plant chemical defense against
herbivory: A biochemical perspective, pp. 105–173, in E.
Bernays (ed.). Insect-plant interactions, Vol VCRC Press, Boca
Raton, IL.
GREIG, N. 1993. Regeneration mode in Neotropical Piper: Habitat and
species comparisons. Ecology 74:2125–2135.

574
J Chem Ecol (2008) 34:558–574
REHR, S. S., FEENY, P. P., and JANZEN, D. H. 1973. Chemical
defence in Central American non-ant-acacias. J. An. Ecol.
42:405–405.
STEWARD, J. L., and KEELER, K. H. 1988. Are there trade-offs among
antiherbivore defenses in Ipomoea (Convolvulaceae)?. Oikos
53:79–79.
RISCH, S. 1982. How Pheidole ants help Piper plants. Brenesia 19/
20:545–548.
RISCH, S. J., and RICKSON, F. R. 1981. Mutualism in which ants must
be present before plants produce food bodies. Nature 291:149–
150.
RISCH, S., MCCLURE, M., VANDERMEER, J., and WALTZ, S. 1977.
Multualism between three species of tropical Piper (Piperaceae)
and their ant inhabitants. Am. Midl. Nat. 98:433–433.
ROMEO, J. T., SAUNDERS, J. A., and BARBOSA, P. 1996. Phytochem-
ical Diversity and Redundancy in Ecological Interactions.
Plenum Press, New York.
STRAUSS, S. Y., RUDGERS, J. A., LAU, J. A., and IRWIN, R. E. 2002.
Direct and ecological costs of resistance to herbivory. Trends
Ecol. Evol. 17:278–278.
TEPE, E., VINCENT, M., and WATSON, L. 2004. Phylogenetic patterns,
evolutionary trends, and the origin of ant-plant associations in
Piper section Macrostachys: Burger's hypothesis revisited, pp.
156–178, in L. A. Dyer, and A. D. N. Palmer (eds.). Piper: A
Model Genus for Studies of Phytochemistry, Ecology, and
EvolutionKluwer Academic Publishing, New York.
ZANGERL, A. R., ARNTZ, A. M., and BERENBAUM, M. R. 1997.
Physiological price of an induced chemical defense: photosyn-
thesis, respiration, biosynthesis and growth. Oecologia 109:433–
433.
SAGERS, C. L., and COLEY, P. D. 1995. Benefits and costs of defense
in a neotropical shrub. Ecology 76:1835–1835.