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:

[email protected]

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


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


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.



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.



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.



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.



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.



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



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

Emory University - Aphids indirectly increase virulence and ... Roode et...LETTER Aphids indirectly...

Documents

Transcript of Emory University - Aphids indirectly increase virulence and ... Roode et...LETTER Aphids indirectly...