Emory University - Aphids indirectly increase virulence and ... Roode et...LETTER Aphids indirectly...
Transcript of Emory University - Aphids indirectly increase virulence and ... Roode et...LETTER Aphids indirectly...
L E T T E RAphids indirectly increase virulence and transmission potential
of a monarch butterfly parasite by reducing defensive chemistry
of a shared food plant
Jacobus C. de Roode1*, Rachel M.
Rarick1, Andrew J. Mongue1, Nicole
M. Gerardo1 and Mark D. Hunter2
1Department of Biology, Emory
University, 1510 Clifton Road,
Atlanta, GA 30322, USA2Department of Ecology and
Evolutionary Biology, University of
Michigan, 1141 Natural Sciences
Building, 830 North University,
Ann Arbor, MI 48109–1048, USA
*Correspondence: E-mail:
AbstractParasites and hosts live in communities consisting of many interacting species, but few studies have examined
how communities affect parasite virulence and transmission. We studied a food web consisting of two species
of milkweed, two milkweed herbivores (monarch butterfly and oleander aphid) and a monarch butterfly-
specific parasite. We found that the presence of aphids increased the virulence and transmission potential of the
monarch butterfly�s parasite on one milkweed species. These increases were associated with aphid-induced
decreases in the defensive chemicals of milkweed plants. Our experiment suggests that aphids can indirectly
increase the virulence and transmission potential of monarch butterfly parasites, probably by altering the
chemical composition of a shared food plant. These results indicate that species that are far removed from
host–parasite interactions can alter such interactions through cascading indirect effects in the food web.
As such, indirect effects within ecological communities may drive the dynamics and evolution of parasites.
KeywordsAphis nerii, Asclepias, cardenolide, Danaus plexippus, disease ecology, host–parasite interaction, Ophryocystis
elektroscirrha, trait-mediated, tri-trophic, virulence evolution.
Ecology Letters (2011) 14: 453–461
INTRODUCTION
Host–parasite models have shown that host density as well as parasite
virulence (i.e. parasite-induced host mortality) and transmission play
key roles in the spread of parasites in host populations (Anderson &
May 1991; Hudson et al. 2002). A major limitation of these models is
that they generally represent host and parasite species in isolation, and
assume that transmission and virulence are under the full control of
host and parasite (reviewed in Restif & Koella 2003). However, hosts
and parasites are members of larger communities of interacting
species (Lafferty et al. 2006), which may impact host density, parasite
transmission and virulence, and thereby influence parasite dynamics
and disease outbreaks.
Studies in community ecology have long shown that species can
affect each other directly and indirectly. Indirect effects occur when
one species affects the performance of a second species through its
effect on a third. They can not only occur due to changes in species
density (density-mediated indirect effects), but may also be mediated
by traits that are involved in the interaction between species (trait-
mediated indirect effects: Wootton 1994; Abrams 1995; Werner &
Peacor 2003). For example, one herbivore may induce a phenotypic
increase in defence traits in its food plant, which may subsequently
affect a second herbivore species (Van Zandt & Agrawal 2004).
Like predators and competitors, parasites are well known to affect
the interactions between other species (MacNeil et al. 2003; Packer
et al. 2003; Mouritsen & Poulin 2005; Hatcher et al. 2006). There is
also increasing evidence that different parasite species can affect each
other�s virulence and transmission when co-infecting the same host
(Graham 2008; Jolles et al. 2008; Telfer et al. 2010). However, little is
known about the indirect effects of other free-living species on host–
parasite interactions. A number of studies have addressed density-
mediated indirect effects, and shown that lower species diversity can
result in increased absolute or relative abundance of host species,
resulting in greater disease prevalence and severity (e.g. Mitchell et al.
2002; Keesing et al. 2006; Johnson et al. 2008; Borer et al. 2009). Many
of these studies have dealt with generalist parasites, where the relative
abundance of more and less competent host species will affect the
overall disease prevalence and severity in a population. However,
studies on the effects of community structure on host-specific
parasites are still rare. Moreover, few studies have investigated the role
of trait-mediated indirect effects (Raffel et al. 2008, 2010) on parasite
virulence and transmission, two traits that crucially affect the spread of
parasites in populations and the severity of disease outbreaks
(Anderson & May 1991; Hudson et al. 2002).
Here, we studied how community composition may affect the
virulence and transmission of a host-specific parasite through indirect
effects. We used a naturally occurring food web consisting of four
species and three trophic levels: milkweed plants (Asclepias curassavica
or Asclepias incarnata), two milkweed herbivores (the monarch butterfly
Danaus plexippus and the oleander aphid Aphis nerii) and a monarch
butterfly-specific protozoan parasite (Ophryocystis elektroscirrha) (Fig. 1).
This food web provides a suitable system to study the indirect effects
on virulence and transmission for several reasons.
First, virulence and transmission potential of the parasite
O. elektroscirrha (McLaughlin & Myers 1970) are relatively straightfor-
ward to measure. This parasite forms dormant spores on the outside
of adult monarch butterflies. Spores are transmitted to eggs and
milkweed during oviposition, after which hatching larvae become
infected by ingesting them. The parasite then replicates during larval
and pupal stages and forms spores on the newly developed adult
butterfly. Spores replicate no further on the adult butterfly, and higher
sporeloads result in greater transmission to larvae (De Roode et al.
2008b, 2009). Thus, the sporeload on an adult butterfly provides a
measure of the full transmission potential of the parasite. Besides
Ecology Letters, (2011) 14: 453–461 doi: 10.1111/j.1461-0248.2011.01604.x
� 2011 Blackwell Publishing Ltd/CNRS
providing greater transmission, higher sporeloads also result in greater
virulence, with heavily infected monarchs having reduced eclosion and
mating success and reduced adult lifespan (De Roode et al. 2008b).
Second, monarch butterfly caterpillars are specialists on milkweed
plants and their close relatives (Ackery & Vane-Wright 1984).
Milkweeds are toxic to most herbivores as they produce secondary
chemicals called cardenolides (Malcolm & Zalucki 1996; Agrawal &
Fishbein 2006). Monarchs suffer little from these chemicals, and in
fact sequester them to protect themselves against predators (Brower &
Fink 1985). Recent studies have also suggested that these chemicals
can reduce the virulence and transmission potential of the monarch�sprotozoan parasite O. elektroscirrha (De Roode et al. 2008a; Lefevre
et al. 2010). More specifically, these studies found that parasites
produced lower sporeloads and caused lower virulence on monarchs
reared on a milkweed species with high cardenolide concentrations
(tropical milkweed: A. curassavica) than on a species with low
concentrations (swamp milkweed: A. incarnata).
Third, monarch caterpillars regularly share their milkweed food
plants with co-occurring oleander aphids, Aphis nerii (e.g. Koch et al.
2005), especially late in the season when aphids colonize many plants
within milkweed patches and reach high numbers on individual plants
(Helms et al. 2004; Helms & Hunter 2005). Like monarch butterflies,
oleander aphids are specialist herbivores on milkweed and sequester
cardenolides as a defence against predation (Malcolm 1991). Impor-
tantly, studies have suggested that high-cardenolide milkweed species
can reduce their cardenolide concentrations in response to aphid
herbivory (Martel & Malcolm 2004; Zehnder & Hunter 2007b). It is
thought that such reductions may both save resources and limit aphid
population growth, as lower levels of cardenolides should increase the
susceptibility of aphids to predators (Malcolm & Zalucki 1996; Martel
& Malcolm 2004; Zehnder & Hunter 2007b).
Overall, the studies described above suggest that aphids are likely to
indirectly increase the virulence and transmission potential of
monarch butterfly parasites on high-cardenolide milkweeds. This is
because aphids can reduce the cardenolide concentrations of high-
cardenolide milkweeds and because lower cardenolide levels correlate
with higher parasite virulence and transmission potential. Here, we
tested this hypothesis by rearing monarch caterpillars on A. curassavica
(high cardenolide) and A. incarnata (low cardenolide) plants with or
without aphids, and by measuring parasite virulence, parasite
transmission potential and cardenolide concentrations. We predicted
that aphids would reduce the cardenolide concentrations of the high-
cardenolide A. curassavica and thereby indirectly increase parasite
virulence and transmission potential in monarchs reared on this
species. In contrast, we predicted no such effects on the low-
cardenolide species A. incarnata.
MATERIALS AND METHODS
Monarch, aphid, plant and parasite sources
Monarchs used in this experiment were the lab-reared grand-offspring
of butterflies collected in Miami, FL, in March 2008. Monarchs from
three full-sib family lines were allocated randomly to experimental
treatments. Aphids were the clonal offspring of a single aphid, also
collected in Miami, FL, in March 2008. Aphids had been maintained on
milkweed by transferring two adult and four mid-instar aphids to fresh
plants every 2 weeks; aphids and plants were maintained in a climate-
controlled chamber at 22 �C, 50% RH and a 16L : 8D light cycle.
The parasite (denoted C1F3-P2-1) was a clone derived from a monarch
butterfly collected in Miami, FL, in 2004. Milkweed seeds were obtained
from Butterfly Encounters, CA. Seeds were germinated in seedling mix
and repotted into Fafard 3B (Conrad Fafard, Inc., Agawam, MA, USA)
soil 2 weeks post-germination. They were maintained in a climate-
controlled greenhouse until used in the experiment.
Experimental design
The experiment was fully factorial, with milkweed species (A. curassavica
vs. A. incarnata), aphid treatment (with or without aphids) and parasite
treatment (infected vs. uninfected) as fixed factors. Each milkweed
species by aphid treatment by parasite treatment combination initially
had between 10 and 30 replicate larvae, as shown in Table 1. Few
larvae died during the experiment, and final numbers on which
statistical analyses were based are also shown in Table 1.
Figure 1 A food web consisting of milkweed (Asclepias curassavica
or Asclepias incarnata), aphids (Aphis nerii), monarch butterfly
caterpillars (Danaus plexippus) and a parasite of the monarch
butterfly (Ophryocystis elektroscirrha). Solid arrows show the direction
of energy flow. The dashed arrows depict the indirect effects of
aphids on the interaction between the monarch butterfly and its
parasite. Results suggest that aphids have a negative effect on the
defensive chemistry of A. curassavica (indicated with the �)� sign),
which subsequently increases the virulence and transmission
potential of the monarch butterfly�s parasite (indicated with a �+�sign). However, as aphids do not alter the chemistry of A. incarnata,
aphids are not expected to indirectly affect the interaction
between the monarch butterfly and its parasite when monarchs
feed on this species (hence the absence of dashed lines connecting
aphids with A. incarnata and the interaction between the monarch
butterfly and its parasite).
454 J. C. de Roode et al. Letter
� 2011 Blackwell Publishing Ltd/CNRS
On day 0 of the experiment, we randomly assigned A. curassavica
and A. incarnata plants to aphid and parasite treatments and placed
them into individual 1.34-L plastic tubes with meshed lids. We added
two adult and four mid-instar aphids to plants in the aphid treatments,
and maintained all plants in a climate-controlled chamber held at 26 �C,
50% RH and a 16L : 8D light cycle. Aphids were counted on days 3, 6,
10 and 13. On day 13, 2-day-old monarch caterpillars were placed
individually in 10-cm petri dishes that contained a moist filter paper and
a leaf disc obtained from the individual plant that was assigned to each
caterpillar. Day 13 was chosen to allow aphid populations to grow to
typical field densities (Helms et al. 2004). Caterpillars in the infected
treatments received 10 parasite spores on their leaf discs, whereas
control caterpillars did not. Upon complete consumption of their leaf
disc (within 24 h), caterpillars were transferred to their assigned plants
in 1.34-L plastic tubes. Thus, caterpillars completed their development
on individual live plants, and in the case of aphid treatment, aphids
remained present on these plants. Plants were watered as needed, and
extra food was provided to caterpillars that ran out of milkweed before
reaching the pupal stage. In such cases, we took care to provide
caterpillars with milkweed of the appropriate species that had or did not
have aphids, depending on the treatment.
Seven days after pupation (an average of 2 days before eclosion),
pupae were transferred to a separate room maintained at the same
conditions as the climate-controlled chamber. This was done to avoid
the potential contamination of the larval rearing chamber with spores
derived from eclosing infected adult monarchs. Once monarchs
eclosed from their pupal cases, they were transferred to individual
glassine envelopes and held in a chamber at 12 �C without food and
water. We then recorded their time of death and calculated a measure
of monarch adult lifespan – referred to from hereon as lifespan index
– as the difference (in days) between the day of eclosion and the day
of death. Lifespan index provides a combined index of adult monarch
lifespan and starvation resistance. We have shown previously that the
lifespan index responds to parasite infection and increasing parasite
numbers in a similar way as lifespan under more natural conditions
(De Roode et al. 2009). Although a more natural way of measuring
lifespan might be preferable, allowing monarchs to fly in cages for the
duration of their life results in the loss of > 90% of parasite spores
(De Roode et al. 2009), resulting in highly inaccurate estimates of
parasite transmission potential.
After monarchs died, we removed their wings and vortexed their
bodies in 5 mL H2O for 5 min to shake off parasite spores. Spores
were subsequently counted using a hemocytometer to provide a
measure of parasite transmission potential (sporeload).
Cardenolide identification and quantification
On days 0 and 13, we removed a small leaf from a subset of plants
(see Table 1) for chemical analysis. Cardenolide concentrations were
estimated using established methods (Malcolm & Zalucki 1996;
De Roode et al. 2008a). Leaves were collected in glassine envelopes
and freeze-dried. They were weighed and pulverized using a bead mill,
then extracted in 100% methanol. Extracts were analysed using
reverse phase high-performance liquid chromatography (HPLC).
Peaks were detected by diode array at 218 nm, and absorbance
spectra were recorded from 200 to 300 nm with digitoxin as an
internal standard. Peaks with symmetrical absorption maxima between
217 and 222 nm were recorded as cardenolides (Malcolm & Zalucki
1996). We detected a total of 16 different cardenolide peaks in
A. curassavica and four in A. incarnata. Concentrations of each peak
were estimated relative to the internal standard, and total cardenolide
concentration was calculated as the sum of all individual cardenolide
peaks.
Statistical analysis
We used analyses of variance (ANOVAs) to determine the effects of
milkweed species and aphid treatment on monarch adult lifespan
index and parasite sporeload. In these analyses, we first fitted full
models with infection treatment, milkweed species, aphid treatment
and the interaction between these factors. Parasite sporeload was
log10-transformed to ensure homogeneity of variance and normality of
errors.
We also used ANOVAs to determine the effects of day (0 vs. 13) and
aphid treatment on overall cardenolide concentrations and the
concentrations of individual cardenolides in milkweed plants. These
analyses were carried out separately for A. curassavica and A. incarnata.
Where necessary, the concentrations of individual cardenolides were
square-root-transformed to ensure homogeneity of variance and
normality of errors. We also used principal components analysis to
describe the composition of cardenolides in A. curassavica plants (this
was not possible for A. incarnata, as most plants contained only two
types of cardenolides). As dictated by the results (below), we analysed
certain cardenolide data for day 13 (when caterpillars were infected)
separately to confirm that cardenolides differed between plants with
and without aphids on this day. We then used ANOVA to investigate the
effect of day and aphid treatment on all principal components that
explained > 10% of the variance of the cardenolide composition.
We also used linear models to analyse the relationship between
Table 1 Experimental treatments and numbers of replicate larvae used for each treatment combination
Milkweed species Aphid treatment Parasite treatment Initial number Number used for analysis Plants used for cardenolides
Asclepias incarnata With aphids Infected 30 26 15
Asclepias incarnata With aphids Uninfected 10 9 –
Asclepias incarnata Without aphids Infected 30 29 15
Asclepias incarnata Without aphids Uninfected 10 9 –
Asclepias curassavica With aphids Infected 30 27 15
Asclepias curassavica With aphids Uninfected 10 9 –
Asclepias curassavica Without aphids Infected 20 19 15
Asclepias curassavica Without aphids Uninfected 10 10 –
Total 150 138 60
Numbers of larvae used for the analysis were lower than initial numbers due to some monarch mortality.
Letter Community affects virulence and transmission 455
� 2011 Blackwell Publishing Ltd/CNRS
cardenolides (overall concentration, pca1 and individual concentra-
tions) and parasite sporeload; these analyses were restricted to
cardenolides measured on day 13 when plants were used to infect
caterpillars. The analyses were also restricted to cardenolides that
differed significantly on day 13 due to the presence of aphids, so as to
reduce the inflation of Type I error rates. Thus, this analysis was carried
out only for cardenolides RT584 and RT650 (see Results section).
All ANOVAs and linear models were carried out in R 2.7.0
(R Foundation for Statistical Computing, Vienna, Austria). Signifi-
cance (P < 0.05) was determined by comparing the explanatory power
of models that included or did not include an explanatory variable
using the ANOVA function in R (Crawley 2002). Principal components
analysis was carried out in SAS (SAS Institute Inc., Cary, NC, USA).
Aphid population growth was analysed using a linear mixed effects
model, in which milkweed species and day were fixed effects, and
milkweed individual was a random effect. The model was fitted using
maximum likelihood. To determine whether milkweed species
significantly affected aphid reproduction, models with and without
the interaction between milkweed species and day were compared
using the ANOVA procedure in SPLUS 7.0 (Crawley 2002); the degrees of
freedom and P-value for this comparison are reported. Aphid
numbers were log-transformed prior to the analysis.
RESULTS
Effects of aphid-free milkweed species on virulence and
transmission of monarch butterfly parasites
As expected on the basis of previous studies (De Roode et al. 2008a;
Lefevre et al. 2010), we found that, in the absence of aphids, parasites
produced much lower sporeloads (2.37 times lower) on monarchs
reared on the high-cardenolide A. curassavica than on monarchs reared
on the low-cardenolide A. incarnata (Fig. 2a, white bars). Conse-
quently, infected adult monarchs lived much longer (an additional
4.8 days, which is 65% longer) when reared on A. curassavica than on
A. incarnata (Fig. 2b, white bars). As expected, chemical analyses also
confirmed that the A. curassavica plants used in this study had much
higher cardenolide concentrations than A. incarnata plants (Fig. 3).
Effects of aphid-infested milkweed species on virulence and
transmission of monarch butterfly parasites
In contrast to results on aphid-free plants, in the presence of aphids,
parasites produced similar sporeloads on monarchs reared on the
high-cardenolide A. curassavica and on the low-cardenolide A. incarnata
(Fig. 2a, grey bars), and infected adult monarchs had a similar lifespan
index when reared on either milkweed species (Fig. 2b, grey bars).
These results suggest that the presence of aphids increased parasite
sporeload and reduced monarch adult lifespan index on A. curassavica
to similar levels as on A. incarnata (Fig. 2). ANOVAs restricted to
monarchs reared on A. curassavica indeed confirmed that monarchs
reared on A. curassavica carried higher parasite sporeloads
(F1,44 = 4.46; P = 0.040) and lived shorter lives (F1,44 = 10.8;
P = 0.002) when reared on plants with aphids than on plants without
aphids.
To confirm that the effect of aphids on monarch lifespan index was
mediated through effects on parasite sporeload, we included sporeload
as a covariate in the analysis of monarch lifespan index. This showed
that sporeload was a strong predictor of monarch adult lifespan index
(F1,95 = 76.1, P < 0.0001), but that the interaction between milkweed
species and aphid presence (see Fig. 2 legend) was no longer
significant (F1,95 = 3.15, P = 0.079). These results suggest that aphids
can increase the sporeload of O. elektroscirrha on monarchs reared on
the high-cardenolide A. curassavica and that this increased sporeload
consequently results in a reduced lifespan index in the monarch
butterfly host.
Effects of aphids on milkweed chemistry
Aphid populations grew rapidly on both milkweed species, but grew
less rapidly on A. curassavica than on A. incarnata plants (Fig. 4; mixed-
effects model: d.f. = 1, P = 0.0008), possibly due to the higher levels
of cardenolides in A. curassavica (Agrawal 2004; Zehnder & Hunter
2007a). As expected (Martel & Malcolm 2004; Zehnder & Hunter
2007b), the rapid population growth of aphids affected the cardeno-
lide concentrations in the high-cardenolide A. curassavica, but not in
the low-cardenolide A. incarnata, probably because the latter species
has little room to reduce its cardenolide concentrations below
constitutive levels in response to aphid herbivory.
Aphid effects on the chemistry of the high-cardenolide A. curassavica
Although plants were assigned randomly to treatment groups,
A. curassavica plants that were assigned to receive aphids started off
with lower overall cardenolide concentrations on day 0, and this
difference was maintained throughout to day 13 (Fig. 3a,b;
F1,52 = 12.7, P = 0.0008). However, these differences in the overall
cardenolide concentrations did not affect our results, because the
4.4
4.6
4.8
5.0
5.2
5.4
5.6
5.8
6.0
Lo
g10
(sp
ore
load
)
0
4
8
12
Infe
cted
mo
nar
ch a
du
ltlif
esp
an in
dex
(d
ay)
Without aphidsWith aphids
Food plant species Food plant speciesA. curassavica A. incarnata A. curassavica A. incarnata
(a) (b)
Figure 2 Effects of milkweed species and the presence of aphids
on parasite sporeload and virulence. (a) In the absence of aphids,
parasite sporeload was lower on monarchs reared on the high-
cardenolide Asclepias curassavica than on monarchs reared on the
low-cardenolide Asclepias incarnata; when aphids were present, this
milkweed effect disappeared (milkweed species · aphid treatment
interaction: F1,96 = 16.91, P = 0.045). (b) Consequently, infected
adult monarch butterflies lived longer when reared on A. curassavica
than A. incarnata in the absence of aphids, but not in the presence
of aphids (milkweed species · aphid treatment interaction:
F1,97 = 6.93, P = 0.0098). Bars show mean + 95% confidence
intervals.
456 J. C. de Roode et al. Letter
� 2011 Blackwell Publishing Ltd/CNRS
overall cardenolide concentrations had no effect on parasite sporeload
(Fig. 5A; F1,25 = 0.0028, P = 0.96).
Aphids significantly changed the overall composition of cardeno-
lides in A. curassavica over time. Principle components analysis was
used to describe this composition, and together, principle component
axes 1–3 (pca1–3) explained more than 60% of the variance in
cardenolides (Supporting Information). Of these, the first axis (pca1)
showed a significant interaction between day and aphid treatment:
pca1 was lower in plants with than without aphids on day 0, but higher
in plants with than without aphids on day 13 (day · aphid treatment
interaction: F1,51 = 11.27, P = 0.001). Restricting the analysis to day
13 only confirmed that pca1 scores were higher on that day:
F1,28 = 5.20, P = 0.030). Interactions between day and aphid treat-
ment were not significant for pca2 and pca3 (F1,51 = 0.13, P = 0.72
and F1,51 = 0.16, P = 0.69). The finding that aphids significantly
increased pca1 between days 0 and 13 demonstrates that they changed
the overall composition of cardenolides in the high-cardenolide
A. curassavica.
0.0
(a) (b)
(c) (d)
2.0
4.0
6.0
0.0
0.4
0.8
1.2Without aphidsWith aphids
0.0
2.0
4.0
6.0
0.0
0.4
0.8
1.2
0.0
0.1
0.2
0.3
0.4
0.0
0.1
0.2
0.3
0.0
0.1
0.2
0.3
0.4
0.0
0.1
0.2
0.3
Co
nce
ntr
atio
n (
mg
g–1
)C
on
cen
trat
ion
(m
g g
–1)
A. curassavica, day 0
Tota
l
Cardenolide
A. curassavica, day 13
Tota
l
Tota
l
Cardenolide Cardenolide
A. incarnata, day 0 A. incarnata, day 13
RT0
65R
T092
RT1
15R
T162
RT1
75R
T240
RT2
56R
T280
RT3
20R
T355
RT4
52R
T551
RT5
65R
T584
RT6
13R
T650
RT6
50
RT6
13
RT5
84
RT5
65
RT5
51
RT4
52
RT2
80
RT2
56
RT2
40
RT1
75
RT1
62
RT1
15
RT0
92
RT0
65
Tota
l
Cardenolide
RT3
20
RT4
52R
T551
RT5
65R
T584
RT6
13R
T650
RT0
65R
T092
RT1
15R
T162
RT1
75R
T240
RT2
56R
T280
RT3
20
RT3
55
RT3
55R
T452
RT5
51R
T565
RT5
84R
T613
RT6
50
RT0
65R
T092
RT1
15R
T162
RT1
75R
T240
RT2
56R
T280
RT3
20R
T355
RT4
52R
T551
RT5
65R
T584
RT6
13R
T650
* *
*
*
Figure 3 Cardenolide concentrations in the high-cardenolide Asclepias curassavica and the low-cardenolide Asclepias incarnata. Cardenolides were measured on day 0 (a and c for
A. curassavica and A. incarnata respectively) and day 13 (b and d for A. curassavica and A. incarnata respectively) in plants that received aphids or not on day 0. Asterisks in (b)
highlight four cardenolides for which there was a significant interaction between aphid presence and day. The initial differences in concentrations of RT162 and RT256 on day
0 had disappeared by day 13 (day · aphid treatment interaction: F1,51 = 4.45, P = 0.04 and F1,51 = 11.4, P = 0.001 for RT162 and RT256 respectively), whereas RT584 and
RT650 increased to higher levels on day 13 in plants without aphids than in plants with aphids (see text for details). Cardenolides are named based on their relative retention
time (RT) under reverse-phase HPLC; increasing RTs represent decreasing polarity. Bars show mean + 95% confidence intervals. Note that y-axis scales differ between panels
to facilitate interpretation of bar heights.
0 3 6 9 120
100
200
300
400
Ap
hid
nu
mb
er
A. curassavicaA. incarnata
Day
Figure 4 Aphid population growth. Population growth of Aphis nerii grown on the
high-cardenolide Asclepias curassavica and the low-cardenolide Asclepias incarnata.
Each plant received six aphids on day 0. Data points show mean + 95% confidence
intervals.
Letter Community affects virulence and transmission 457
� 2011 Blackwell Publishing Ltd/CNRS
When analysing individual cardenolides, we found that aphids
appeared to mitigate an ontogenetic increase of the concentration of
cardenolide RT584 (Fig. 3a,b; day · aphid treatment interaction:
F1,51 = 3.90, P = 0.054), while significantly mitigating an ontogenetic
increase of cardenolide RT650 between days 0 and 13 (Fig. 3a,b;
day · aphid treatment interaction: F1,51 = 7.87, P = 0.007). As a
result, A. curassavica plants with aphids had lower concentrations of
these two chemicals than did plants without aphids on day 13, when
caterpillars were infected with parasites (F1,28 = 7.92, P = 0.009 and
F1,28 = 8.76, P = 0.006 for RT584 and RT650 respectively).
Aphid effects on the chemistry of the low-cardenolide A. incarnata
In the low-cardenolide A. incarnata, total cardenolide concentrations
and concentrations of RT065 and RT355 decreased between day 0
and 13 (F1,56 = 29.5, P < 0.0001; F1,55 = 22.5, P < 0.0001; and
F1,56 = 39.6, P < 0.0001 for total, RT065 and RT355 respectively)
but aphids did not affect these decreases (day · aphid treatment:
F1,54 = 0.005, P = 0.94; F1,54 = 2.38, P = 0.13; and F1,54 = 1.30,
P = 0.26 for total, RT065 and RT355 respectively). Thus, aphids had
no effect on the cardenolide composition of the low-cardenolide
A. incarnata.
Aphid-mediated indirect effects on virulence and transmission of
monarch butterfly parasites
Aphid-mediated indirect effects on the high-cardenolide A. curassavica
As shown above, the presence of aphids increased the virulence and
transmission potential of parasites in monarchs reared on the high-
cardenolide A. curassavica. In addition, aphids changed the overall
composition of cardenolides and mitigated an ontogenetic increase of
two cardenolides in A. curassavica. Here, we tested whether these
effects on milkweed chemistry can account for the observed changes
in parasite virulence and transmission potential.
In terms of overall cardenolide composition, we found that higher
values of pca1 resulted in higher parasite sporeloads (Fig. 5b;
F1,25 = 8.75, P = 0.007), suggesting that the higher sporeloads in
monarchs infected on A. curassavica plants with aphids could be
explained by their different cardenolide compositions on day 13.
Similarly, regression analysis showed that higher concentrations of
the individual chemicals RT584 and RT650 were associated with
lower parasite sporeloads (Fig. 5c,d; F1,25 = 5.41, P = 0.028 and
F1,25 = 4.85, P = 0.037 for RT584 and RT650 respectively). Because
the P-values of these regression analyses were relatively high and
because the sporeload values were used in both analyses, we carried
out a partial Bonferroni correction to ensure that the significance of
these tests was not the result of carrying out multiple tests. We used
the web-based statistical program SISA (http://www.quantitatives-
kills.com/sisa/calculations/bonfer.htm), which accounts for variable
dependence while correcting for inflated Type I error rates (Garcıa
2004). Given a Pearson correlation coefficient of 0.657 (P < 0.0001)
between RT584 and RT650 concentrations on day 13, the adjusted afor each of the two regression analyses is 0.039; hence, we conclude
that the negative associations between sporeload and the concentra-
tions of RT584 (P = 0.028) and RT650 (P = 0.037) are indeed
significant. In these regression analyses, aphid presence had no
further effect on sporeload when cardenolide concentration was
included in the model as a covariate (F1,24 = 0.15, P = 0.70 and
F1,24 = 0.11, P = 0.74 for analyses with RT584 and RT650
respectively). These results suggest that the lower concentrations of
RT584 and RT650 cardenolides in A. curassavica plants with aphids
may explain the higher parasite sporeloads that we observed
(Fig. 2a).
Overall, these results suggest that aphids indirectly increase the
virulence and transmission potential of monarch butterfly parasites
by decreasing the defensive chemistry of the high-cardenolide
A. curassavica.
0 3 6 94.4
4.9
5.4
5.9
6.4
0.0 0.4 0.8 1.2 1.64.4
4.9
5.4
5.9
6.4
–4 –3 –2 –1 0 1
0.0 0.1 0.2 0.3 0.4 0.5
Concentration (mg g–1)
Concentration (mg g–1) Concentration (mg g–1)
pca1
Log 1
0 (sp
orel
oad)
Log 1
0 (sp
orel
oad)
Total cardenolide concentration Total cardenolides: pca
RT584 RT650
(a) (b)
(c) (d)
Figure 5 Relationships between cardenolide concentration and
parasite sporeload in the high-cardenolide Asclepias curassavica. The
overall concentration of cardenolides was not associated with
parasite sporeload (a), but the overall composition of cardenolides
was (b). Furthermore, higher concentrations of the individual
cardenolides RT584 (c) and RT650 (d) were associated with lower
parasite sporeloads. Cardenolide concentrations shown are those
measured on day 13, when monarch caterpillars were infected.
Lines show least-squares regression lines. Note that cardenolide
data were obtained from plants for a subset of experimental
animals only (Table 1), so that numbers of animals in these graphs
are lower than those used in Fig. 2. Note that since we found
no effects of aphids on the cardenolide concentrations in the
low-cardenolide A. incarnata, no data are shown for this species.
(s) Without aphids; (d) With aphids.
458 J. C. de Roode et al. Letter
� 2011 Blackwell Publishing Ltd/CNRS
Aphid-mediated indirect effects on the low-cardenolide A. incarnata
As shown above, aphids had no effects on the parasite sporeload of
monarchs reared on the low-cardenolide A. incarnata, and also did not
affect the defensive chemistry of this species. These results suggest
that aphids did not indirectly affect parasite virulence and transmission
potential in monarchs reared on A. incarnata.
Effects of aphids and milkweed plants on uninfected monarchs
Overall, uninfected monarchs lived much longer than infected
monarchs (compare Fig. S1, Fig. 2b; F1,128 = 427, P < 0.0001).
There was also a significant three-way interaction between infection
status, milkweed species and aphid treatment (F1,128 = 10.46,
P = 0.0015). This interaction indicates that the lifespan index of
uninfected monarchs was more strongly decreased by aphids on the
low-cardenolide A. incarnata, whereas the lifespan index of infected
monarchs was only decreased by aphids on the high-cardenolide
A. curassavica (compare Fig. S1, Fig. 2b).
When restricting the analysis to uninfected monarchs only, we
found that aphids reduced the adult lifespan index of monarchs that
were reared on both milkweed species (Fig. 2b; F1,32 = 12.8,
P = 0.001). Additionally, uninfected monarchs lived longer when
reared on the low-cardenolide A. incarnata than on the high-
cardenolide A. curassavica (Fig. 2b; F1,32 = 12.9, P = 0.001). The
interaction between aphid treatment and milkweed species was nearly
significant (F1,33 = 4.15, P = 0.0502), suggesting that aphids more
strongly reduced the adult lifespan index of uninfected monarchs
reared on A. incarnata than on A. curassavica.
DISCUSSION
We have shown that the ecological community within which hosts and
parasites interact can have important effects on parasite virulence and
transmission potential. In particular, our results suggest that oleander
aphids can reduce the levels of some secondary compounds in
A. curassavica that appear to inhibit a specific parasite of monarch
butterflies. In doing so, aphids can remove the medicinal effects of
these milkweeds and increase the virulence and transmission potential
(sporeload) of monarch parasites. Importantly, our results also show
that the indirect effects on virulence and transmission potential caused
by a fourth species (oleander aphid) depend on the identity of the
third species: specifically, aphids have indirect effects when infected
monarchs develop on the high-cardenolide A. curassavica. In contrast,
aphids have no effect on the virulence and transmission potential of
parasites when infected monarchs develop on the low-cardenolide
A. incarnata, probably because this species has little room to reduce its
cardenolide concentrations below constitutive levels in response to
aphid herbivory.
Our study confirms that species outside the direct interaction
between host and parasite can have indirect effects on host–parasite
traits. Most studies to date have shown that third species can have
density-mediated indirect effects on host–parasite interactions, by
changing the absolute or relative abundance of host species (Mitchell
et al. 2002; Keesing et al. 2006; Johnson et al. 2008; Borer et al. 2009).
In contrast, studies of trait-mediated indirect effects on parasite-host
dynamics are still relatively rare (Raffel et al. 2008, 2010). Where they
have been carried out, such studies have confirmed that indirect
effects may shape virulence and transmission parameters. For
example, feeding by gypsy moth larvae on oak foliage causes
increases in foliar phenolics that then help to protect larvae from
viral infection (Hunter & Schultz 1993). Additionally, studies have
shown that predatory midges can enhance bacterial parasite
transmission by tearing apart bacteria-filled waterfleas (Caceres et al.
2009), that predators can suppress the immunity of a herbivorous
beetle to its parasites (Ramirez & Snyder 2009) and that co-infecting
parasite species can affect each other�s transmission through
modulating host immunity (Graham 2008; Jolles et al. 2008). Our
study supports these findings, but also goes further by illustrating
that a fourth species even further removed from the direct host–
parasite interaction in the food web may affect a host–parasite
interaction by altering the traits – in this case the defensive
chemistry – of a third species.
Our results have important implications for understanding host–
parasite dynamics. Theoretical models have indicated that virulence
and transmission are two key parameters that affect the spread of
parasites in populations and the severity of disease outbreaks
(Anderson & May 1991; Hudson et al. 2002). Moreover, a trade-off
between virulence and transmission is generally expected to select for
parasites with intermediate levels of host exploitation, at which level
the costs of exploitation (parasite-induced host mortality) are
balanced by the benefits (between-host transmission) (Anderson &
May 1982; Van Baalen & Sabelis 1995; De Roode et al. 2008b; Alizon
et al. 2009). These models generally assume that virulence and
transmission are under the full control of host and parasite (reviewed
in Restif & Koella 2003) and that other species in the food web do
not affect these traits. However, our results suggest that other species
may have strong impacts on virulence and transmission. In our study,
the plants on which monarchs feed as larvae affect these traits, as do
the aphids that share this food source with monarchs. How indirect
effects on virulence and transmission affect parasite dynamics would
likely depend on the strength of these effects as well as the
environmental heterogeneity in species composition. Theoretical
models will be required to formally predict the ecological and
evolutionary consequences of such indirect effects (e.g. Choo et al.
2003).
The parasites of herbivorous insects are often affected by the plants
upon which insects feed (Cory & Hoover 2006). Such plant effects
may occur in two ways: either the plant chemicals directly interfere
with the parasite or the plant affects the general health and immunity
of the insect. The results obtained here suggest that milkweed
chemicals directly interfere with the monarch�s parasite and that the
indirect effect of aphids is mediated through these chemicals. We did
find that aphids reduced the adult lifespan index of uninfected
monarchs reared on both milkweed species (Fig. S1), suggesting that
monarchs that co-occur with aphids are in worse overall health. This
may be due to increased competition for resources, or aphid-induced
reductions in food quality, such as reduced levels of nitrogen.
However, this reduction in lifespan index did not appear to be related
to disease susceptibility because the lifespan index of infected
monarchs was only reduced by aphids when they were reared on
A. curassavica. Moreover, our results show that the presence of aphids
on the high-cardenolide A. curassavica changed its cardenolide
composition (Fig. 3) and that these changes were associated with
changes in parasite sporeload and adult lifespan index of infected
monarchs (Fig. 5). These results also suggest that any factor that alters
the chemical composition of milkweeds may have indirect effects on
the transmission potential and virulence of the monarch butterfly�sparasite.
Letter Community affects virulence and transmission 459
� 2011 Blackwell Publishing Ltd/CNRS
As outlined in the introduction, previous studies have found that
the presence of oleander aphids can result in reductions in cardenolide
concentrations of high-cardenolide milkweeds (Martel & Malcolm
2004; Zehnder & Hunter 2007b). Here, we did not find aphid-induced
reductions in cardenolide concentrations nor did we find a change in
overall cardenolide levels. Instead, we found that the presence
of aphids prevented a developmental increase in the concentration of
two particular non-polar cardenolides in the high-cardenolide species
A. curassavica (Fig. 3). This finding supports the view that overall
cardenolide concentrations of milkweeds may not be fully informative
of their ecological effects and that it is important to analyse individual
compounds (Zehnder & Hunter 2007b). This was also suggested by
Fordyce & Malcolm (2000) who found that overall cardenolide
concentrations in milkweed stems with weevil eggs were lower than
those without, but contained more non-polar forms, which are
thought to be more toxic.
In conclusion, our results show that species in food webs may
indirectly alter the virulence and transmission parameters of parasites,
and that these effects may be mediated through the traits of yet other
species. Our study of trait-mediated indirect effects complements others
that have illustrated density-mediated indirect effects of community
structure on host–parasite interactions. Studying virulence and
transmission in a community context will not only increase our
understanding of the ecology and evolution of parasites, but will also
help in predicting the effects of human-altered communities on the
disease risks of humans and wild species alike (Keesing et al. 2006, 2010).
ACKNOWLEDGEMENTS
We thank A. Chiang and N. Lowe for help with the experiment, and
the Emory IHoP journal club as well as four anonymous referees,
M. Holyoak and K. Lafferty for helpful and constructive comments
on previous versions of this manuscript. Funding was provided by
Emory University (to JCdR) and National Science Foundation awards
DEB-0814340 to MDH and DEB-1019746 to JCdR and MDH.
REFERENCES
Abrams, P.A. (1995). Implications of dynamically variable traits for identifying,
classifying, and measuring direct and indirect effects in ecological communities.
Am. Nat., 146, 112–134.
Ackery, P.R. & Vane-Wright, R.I. (1984). Milkweed Butterflies: Their Cladistics and
Biology. Cornell University Press, Ithaca, NY.
Agrawal, A.A. (2004). Plant defense and density dependence in the population
growth of herbivores. Am. Nat., 164, 113–120.
Agrawal, A.A. & Fishbein, M. (2006). Plant defense syndromes. Ecology, 87, S132–
S149.
Alizon, S., Hurford, A., Mideo, N. & Van Baalen, M. (2009). Virulence evolution
and the trade-off hypothesis: history, current state of affairs and the future.
J. Evol. Biol., 22, 245–259.
Anderson, R.M. & May, R.M. (1982). Coevolution of hosts and parasites. Parasi-
tology, 85, 411–426.
Anderson, R.M. & May, R.M. (1991). Infectious Diseases of Humans – Dynamics and
Control. Oxford University Press, Oxford.
Borer, E.T., Mitchell, C.E., Power, A.G. & Seabloom, E.W. (2009). Consumers
indirectly increase infection risk in grassland food webs. Proc. Natl Acad. Sci.
U.S.A., 106, 503–506.
Brower, L.P. & Fink, L.S. (1985). A natural toxic defense system – cardenolides in
butterflies versus birds. Ann. N Y Acad. Sci., 443, 171–188.
Caceres, C.E., Knight, C.J. & Hall, S.R. (2009). Predator-spreaders: predation can
enhance parasite success in a planktonic host-parasite system. Ecology, 90, 2850–
2858.
Choo, K., Williams, P.D. & Day, T. (2003). Host mortality, predation and the
evolution of parasite virulence. Ecol. Lett., 6, 310–315.
Cory, J.S. & Hoover, K. (2006). Plant-mediated effects in insect-pathogen inter-
actions. Trends Ecol. Evol., 21, 278–286.
Crawley, M.J. (2002). Statistical Computing: An Introduction to Data Analysis Using S-Plus.
John Wiley & Sons, Chichester.
De Roode, J.C., Pedersen, A.B., Hunter, M.D. & Altizer, S. (2008a). Host plant
species affects virulence in monarch butterfly parasites. J. Anim. Ecol., 77,
120–126.
De Roode, J.C., Yates, A.J. & Altizer, S. (2008b). Virulence-transmission trade-offs
and population divergence in virulence in a naturally occurring butterfly parasite.
Proc. Natl Acad. Sci. U.S.A., 105, 7489–7494.
De Roode, J.C., Chi, J., Rarick, R.M. & Altizer, S. (2009). Strength in numbers: high
parasite burdens increase transmission of a protozoan parasite of monarch
butterflies (Danaus plexippus). Oecologia, 161, 67–75.
Fordyce, J.A. & Malcolm, S.B. (2000). Specialist weevil, Rhyssomatus lineaticollis, does
not spatially avoid cardenolide defenses of common milkweed by ovipositing
into pith tissue. J. Chem. Ecol., 26, 2857–2874.
Garcıa, L.V. (2004). Escaping the Bonferroni iron claw in ecological studies. Oikos,
105, 657–663.
Graham, A.L. (2008). Ecological rules governing helminth-microparasite coinfec-
tion. Proc. Natl Acad. Sci. U.S.A., 105, 566–570.
Hatcher, M.J., Dick, J.T.A. & Dunn, A.M. (2006). How parasites affect interactions
between competitors and predators. Ecol. Lett., 9, 1253–1271.
Helms, S.E. & Hunter, M.D. (2005). Variation in plant quality and the population
dynamics of herbivores: there is nothing average about aphids. Oecologia, 145,
197–204.
Helms, S.E., Connelly, S.J. & Hunter, M.D. (2004). Effects of variation among plant
species on the interaction between an herbivore and its parasitoid. Ecol. Entomol.,
29, 44–51.
Hudson, P.J., Rizzoli, A., Grenfell, B.T., Heesterbeek, H. & Dobson, A.P. (2002).
The Ecology of Wildlife Diseases. Oxford University Press, Oxford.
Hunter, M.D. & Schultz, J.C. (1993). Induced plant defenses breached? Phyto-
chemical induction protects an herbivore from disease. Oecologia, 94, 195–
203.
Johnson, P.T.J., Hartson, R.B., Larson, D.J. & Sutherland, D.R. (2008). Diversity
and disease: community structure drives parasite transmission and host fitness.
Ecol. Lett., 11, 1017–1026.
Jolles, A.E., Ezenwa, V.O., Etienne, R.S., Turner, W.C. & Olff, H. (2008). Inter-
actions between macroparasites and microparasites drive infection patterns in
free-ranging African buffalo. Ecology, 89, 2239–2250.
Keesing, F., Holt, R.D. & Ostfeld, R.S. (2006). Effects of species diversity on
disease risk. Ecol. Lett., 9, 485–498.
Keesing, F., Belden, L.K., Daszak, P., Dobson, A., Harvell, C.D., Holt, R.D. et al.
(2010). Impacts of biodiversity on the emergence and transmission of infectious
diseases. Nature, 468, 647–652.
Koch, R.L., Venette, R.C. & Hutchison, W.D. (2005). Influence of alternate prey on
predation of monarch butterfly (Lepidoptera: Nymphalidae) larvae by the mul-
ticolored Asian lady beetle (Coleoptera: Coccinellidae). Environ. Entomol., 34,
410–416.
Lafferty, K.D., Dobson, A.P. & Kuris, A.M. (2006). Parasites dominate food web
links. Proc. Natl Acad. Sci. U.S.A., 103, 11211–11216.
Lefevre, T., Oliver, L., Hunter, M.D. & De Roode, J.C. (2010). Evidence for trans-
generational medication in nature. Ecol. Lett., 13, 1485–1493.
MacNeil, C., Dick, J.T.A., Hatcher, M.J., Terry, R.S., Smith, J.E. & Dunn, A.M.
(2003). Parasite-mediated predation between native and invasive amphipods.
Proc. R. Soc. Lond., B, Biol. Sci., 270, 1309–1314.
Malcolm, S.B. (1991). Cardenolide-mediated interactions between plants and her-
bivores. In: Herbivores: Their Interactions With Secondary Plant Metabolites (eds
Rosenthal, G.A. & Berenbaum, M.R.). Academic Press, San Diego, pp. 251–
296.
Malcolm, S.B. & Zalucki, M.P. (1996). Milkweed latex and cardenolide induction
may resolve the lethal plant defence paradox. Entomol. Exp. Appl., 80, 193–
196.
Martel, J.W. & Malcolm, S.B. (2004). Density-dependent reduction and induction
of milkweed cardenolides by a sucking insect herbivore. J. Chem. Ecol., 30, 545–
561.
460 J. C. de Roode et al. Letter
� 2011 Blackwell Publishing Ltd/CNRS
McLaughlin, R.E. & Myers, J. (1970). Ophryocystis elektroscirrha sp. n., a neogregarine
pathogen of monarch butterfly Danaus plexippus (L.) and the Florida queen
butterfly D. gilippus berenice Cramer. J. Protozool., 17, 300–305.
Mitchell, C.E., Tilman, D. & Groth, J.V. (2002). Effects of grassland plant species
diversity, abundance, and composition on foliar fungal disease. Ecology, 83, 1713–
1726.
Mouritsen, K.N. & Poulin, R. (2005). Parasites boosts biodiversity and changes animal
community structure by trait-mediated indirect effects. Oikos, 108, 344–350.
Packer, C., Holt, R.D., Hudson, P.J., Lafferty, K.D. & Dobson, A.P. (2003).
Keeping the herds healthy and alert: implications of predator control for
infectious disease. Ecol. Lett., 6, 797–802.
Raffel, T.R., Martin, L.B. & Rohr, J.R. (2008). Parasites as predators: unifying
natural enemy ecology. Trends Ecol. Evol., 23, 610–618.
Raffel, T.R., Hoverman, J.T., Halstead, N.T., Michel, P.J. & Rohr, J.R. (2010).
Parasitism in a community context: trait-mediated interactions with competition
and predation. Ecology, 91, 1900–1907.
Ramirez, R.A. & Snyder, W.E. (2009). Scared sick? Predator-pathogen facilitation
enhances exploitation of a shared resource. Ecology, 90, 2832–2839.
Restif, O. & Koella, J.C. (2003). Shared control of epidemiological traits in a
coevolutionary model of host-parasite interactions. Am. Nat., 161, 827–836.
Telfer, S., Lambin, X., Birtles, R., Beldomenico, P., Burthe, S., Paterson, S. et al.
(2010). Species interactions in a parasite community drive infection risk in a
wildlife population. Science, 330, 243–246.
Van Baalen, M. & Sabelis, M.W. (1995). The dynamics of multiple infection and the
evolution of virulence. Am. Nat., 146, 881–910.
Van Zandt, P.A. & Agrawal, A.A. (2004). Community-wide impacts of herbivore-
induced plant responses in milkweed (Asclepias syriaca). Ecology, 85, 2616–2629.
Werner, E.E. & Peacor, S.D. (2003). A review of trait-mediated indirect interactions
in ecological communities. Ecology, 84, 1083–1100.
Wootton, J.T. (1994). The nature and consequences of indirect effects in ecological
communities. Annu. Rev. Ecol. Syst., 25, 443–466.
Zehnder, C.B. & Hunter, M.D. (2007a). A comparison of maternal effects and
current environment on vital rates of Aphis nerii, the milkweed-oleander aphid.
Ecol. Entomol., 32, 172–180.
Zehnder, C.B. & Hunter, M.D. (2007b). Interspecific variation within the genus
Asclepias in response to herbivory by a phloem-feeding insect herbivore. J. Chem.
Ecol., 33, 2044–2053.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online
version of this article:
Figure S1 Effects of milkweed species and aphids on lifespan index of
uninfected monarchs.
Table S1 Eigenvalues and per cent of variance explained by principal
component axes 1–16 of the principal components analysis of
cardenolide concentrations of Asclepias curassavica plants.
Table S2 Correlations between each factor (cardenolide) and principal
component axis of the principal components analysis of cardenolide
concentrations of Asclepias curassavica plants.
As a service to our authors and readers, this journal provides
supporting information supplied by the authors. Such materials are
peer-reviewed and may be re-organized for online delivery, but are not
copy edited or typeset. Technical support issues arising from
supporting information (other than missing files) should be addressed
to the authors.
Editor, Kevin Lafferty
Manuscript received 16 September 2010
First decision made 20 October 2010
Second decision made 7 December 2010
Manuscript accepted 31 January 2011
Letter Community affects virulence and transmission 461
� 2011 Blackwell Publishing Ltd/CNRS
Supplementary information Belonging to: De Roode, J.C., Rarick, R.M., Mongue, A.J., Gerardo, N.M. & Hunter, M.D. (2011) Aphids indirectly increase virulence and transmission potential of a monarch butterfly parasite by reducing defensive chemistry of a shared host plant.
0
5
10
15
20
25
30
35
Uni
nfec
ted
mon
arch
adu
lt
life
span
inde
x (d
ays)
without aphidswith aphids
Food plant speciesA. incarnataA. curassavica
Figure S1. Effects of milkweed species and aphids on lifespan index of uninfected monarchs. Aphids reduced the lifespan index of monarchs reared on both species, and monarchs lived longer when reared on the low-cardenolide A. incarnata than on the high-cardenolide A. curassavica. Bars show mean + 95% confidence intervals.
Table S1. Eigenvalues and percent of variance explained by principal component axes 1-16 of the principal components analysis of cardenolide concentrations of A. curassavica plants.
Principal component axis Eigenvalue
Percent of variance
Cumulative percent of variance
1 5.00129 31.3 31.3 2 2.90128 18.1 49.4 3 2.06907 12.9 62.3 4 1.38963 8.7 71 5 1.04545 6.5 77.5 6 0.82272 5.1 82.7 7 0.65755 4.1 86.8 8 0.65327 4.1 90.9 9 0.36407 2.3 93.2
10 0.28618 1.8 94.9 11 0.26438 1.7 96.6 12 0.19267 1.2 97.8 13 0.13849 0.9 98.7 14 0.09725 0.6 99.3 15 0.07177 0.4 99.7 16 0.04492 0.3 100
Table S2. Correlations between each factor (cardenolide) and principal component axis of the principal components analysis of cardenolide concentrations of A. curassavica plants. Factor 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 RT065 0.1623 -0.1042 0.3337 -0.1016 0.4325 0.1457 0.252 -0.6933 0.1165 0.2304 0.0427 0.0962 0.0173 0.0201 -0.1008 -0.0075
RT092 0.1188 -0.4001 0.2901 -0.2227 -0.2819 0.0823 -0.1223 0.0198 -0.2004 0.1775 -0.561 -0.1363 0.1503 0.1979 0.104 0.336
RT115 -0.2132 -0.1143 -0.4006 -0.1054 0.4111 -0.2837 0.2718 0.1021 -0.3241 0.2215 -0.1293 0.299 0.1026 0.2554 0.3155 0.0239
RT162 0.2908 -0.3569 0.0531 0.102 0.0328 0.1807 -0.181 0.0294 -0.0285 -0.4027 0.1542 0.6095 -0.1862 -0.0835 0.2905 0.1526
RT175 -0.1579 -0.2961 0.0084 0.1772 0.4936 0.2779 -0.5091 0.2421 0.043 0.2792 0.205 -0.2564 0.1389 -0.0823 0.0153 0.0582
RT240 -0.01 -0.1936 0.4004 -0.4724 -0.0503 -0.2318 0.1867 0.3978 -0.012 0.266 0.4058 0.0679 -0.2568 -0.1233 -0.0679 -0.0606
RT256 0.3557 -0.1654 -0.2628 0.1097 0.069 -0.0112 -0.0652 0.0388 -0.2658 0.2332 -0.329 0.0311 -0.4351 -0.2845 -0.3773 -0.3279
RT280 -0.3666 -0.155 0.1615 0.1748 0.0338 -0.0248 0.1917 0.0286 0.3989 -0.0019 -0.378 -0.0769 -0.2753 -0.3555 0.4406 -0.2033
RT320 0.2194 -0.1634 -0.1776 0.1268 -0.0142 0.5469 0.6373 0.3202 0.1184 -0.0012 0.069 -0.1669 0.0761 0.103 -0.041 0.0695
RT355 -0.1554 -0.2657 0.3278 0.4469 0.11 -0.3066 0.1189 0.1842 0.078 -0.2126 -0.1205 0.1672 0.1678 0.288 -0.4738 -0.1209
RT452 0.1454 -0.4631 -0.2005 -0.0636 -0.2499 -0.1948 0.0125 -0.1819 0.0509 -0.0196 0.1899 -0.1006 0.5241 -0.2059 0.1231 -0.4529
RT551 -0.3049 -0.2858 -0.0585 -0.2565 0.089 0.135 0.0083 -0.1841 -0.2866 -0.5 0.0863 -0.3699 -0.3399 0.2546 -0.0504 -0.1917
RT565 0.3316 -0.1626 -0.1613 0.2346 -0.0578 -0.385 -0.055 -0.1006 0.3256 0.1468 0.1804 -0.2946 -0.359 0.4186 0.1976 0.1668
RT584 -0.2942 -0.2835 -0.3339 -0.0501 -0.0881 -0.1456 0.1039 -0.172 0.1438 -0.0288 0.0668 0.0252 -0.0399 -0.3703 -0.3535 0.5995
RT613 -0.229 -0.0389 0.1682 0.5152 -0.3333 0.079 0.1266 -0.193 -0.5274 0.2806 0.2914 -0.0102 -0.1075 -0.0631 0.1484 0.0482
RT650 -0.3289 -0.102 -0.193 -0.1131 -0.3262 0.3198 -0.1714 -0.0897 0.3146 0.3232 0.0195 0.3793 -0.1241 0.3766 -0.1544 -0.2377